ENSCs engraft, form neo-ganglia, and project neuronal fibers within the aganglionic colons of Ednrb-KO mice. ENSCs were isolated from the GI tract of 2- to 3-week-old Wnt1-tdT mice, in which all neural crest–derived cells, including enteric neurons and glia, express tdTomato (tdT) (25) (Figure 1A). ENSCs were expanded in culture as enteric neurospheres (Figure 1, A and B) that contain p75+ enteric neural crest–derived cells (Figure 1C) and Hu-expressing neurons (Figure 1D). Neurospheres were transplanted via needle injection into the distal aganglionic colons of 7- to 10-day-old Ednrb-KO mice using an anorectal approach (Figure 1, A and E). Two weeks after surgery, transplanted tdT+ ENSCs were identified in recipient colons (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.179755DS1) by costaining with the pan-neuronal marker, Tuj1 (Figure 1F, magnified view of the dotted box in F is shown in H; Supplemental Figure 1B). Transplanted cells formed clusters that resembled “neo-ganglia” and consisted of Tuj1+ neuronal cell bodies and fibers (Figure 1, H–J, arrows). ENSCs also projected neural fibers along the extrinsic-derived fibers that are a feature of the aganglionic colon (Figure 1G, arrows). Furthermore, neo-ganglia contained differentiated neurons that express nNOS (Figure 1, K–N, arrows), the enzyme required for producing nitric oxide (NO), which is involved in neurogenic relaxation of GI smooth muscle. These findings suggest that ENSCs can engraft; differentiate into enteric neurons, including enteric neuronal subtypes; and project fibers within the aganglionic gut environment in vivo, features that are necessary for normal ENS formation.

Transplant of Wnt1-tdT ENSCs to Ednrb-KO mice. Schematic of experimental overview (A), including isolation of ENSCs from the gastrointestinal tract of Wnt1-tdT mice, their expansion as neurospheres (B), and subsequent transplantation into the aganglionic distal colons of Ednrb-KO mice via anorectal needle injection (A and E). Enteric neurospheres contain p75+ neural crest cells (C) and Hu+ neurons (D). Transplanted cells were observed 2 weeks following surgery (F), projecting fibers along host-derived Tuj1+ extrinsic nerves (G, arrows) and forming neo-ganglia (H–J, arrows) that contain donor-derived nNOS immunoreactive neurons (K–N, arrows). Scale bars: 50 μm (C, D, and K–N), 100 μm (F–J), and 200 μm (B).

ENSC transplantation restores motor responses in the aganglionic smooth muscle. We next performed organ bath studies to determine the effect of ENSC transplantation on aganglionic smooth muscle activity. Aganglionic colonic smooth muscle excised from Ednrb-KO mice exhibited spontaneous, rhythmic, high-amplitude contractions at baseline (Figure 2A, red tracing). This activity was not observed in the ganglionated colon from normal, Ednrb WT mice (Figure 2A, green tracing). Importantly, the spontaneous contractions seen in Ednrb-KO colon were significantly reduced following cell transplantation (Figure 2A, blue tracing), as summarized graphically in Figure 2B. Following application of electrical field stimulation (EFS), an immediate contractile response was elicited in colonic smooth muscle from WT mice (Figure 2A, green tracing), whereas no response was observed in Ednrb-KO colon (Figure 2A, red tracing). However, after ENSC transplantation, the contractile response to EFS was restored in Ednrb-KO mice (Figure 2A, blue tracing). To test whether the contractile response was neurally mediated, the voltage-gated sodium channel blocker, tetrodotoxin (TTX) was added to the bath. In the presence of TTX, the contractile response was blocked in Ednrb WT as well as ENSC transplanted colon, confirming that the response was mediated by the transplanted enteric neurons. A quantitative analysis of these responses is provided in Figure 2C.

EFS demonstrates functional recovery of smooth muscle contractility in Ednrb-KO mice after cell transplantation. Representative traces of smooth muscle contractions during the spontaneous, after EFS, and under NANC conditions (A). Quantifications of spontaneous muscle contractility (B), EFS-induced contractility (C), and under NANC conditions (D) are shown. The amplitude of EFS contractions reflects maximal contractility as an absolute change from baseline and is markedly reduced in the presence of TTX (C). Effects of ACh (E) and KCl (F) on muscle activity. All the values represent the mean of 2–4 animals for each group, repeated 2–3 times. Data are shown as the mean ± SEM. Statistical significance was determined by the 1-way ANOVA with a post hoc Tukey’s test; *P < 0.05, ***P < 0.001, and ###P < 0.001 are statistically significant. EFS, electrical field stimulation; TTX, tetrodotoxin.

To determine if ENSC transplantation also restores the nitrergic inhibitory (relaxation) response in the aganglionic colon, EFS was performed in the presence of adrenergic and cholinergic antagonists, referred to as nonadrenergic, noncholinergic (NANC) conditions, to reveal the effects of NO, the main inhibitory neurotransmitter in the gut (26). In Ednrb WT colon, EFS elicited a relaxation response (Figure 2A, green tracing) but did not do so in Ednrb-KO colon (Figure 2A, red tracing). After cell transplantation, this inhibitory response was restored (Figure 2, A and D). Organ bath studies were also performed using aganglionic colon from sham-operated (vehicle injected) Ednrb-KO mice, and no responses were observed (Supplemental Figure 3). To determine the integrity and sensitivity of cholinergic receptors expressed by colonic smooth muscle preparations, 100 μM acetylcholine (ACh) was added to the organ bath and the force contraction was measured. Smooth muscles obtained from all 3 groups demonstrated an ability to contract in response to ACh, with no significant differences observed (Figure 2E). Addition of KCl also confirmed intact muscle contractile responses with no significant differences in responses among the 3 groups (Figure 2F). These findings indicate that transplanted ENSC-derived neurons restore functional contractile and relaxation responses in the recipient aganglionic colon.

Optogenetics confirms functional neuromuscular connectivity between transplanted cell–derived neurons and aganglionic colonic smooth muscle. We next tested if neuromuscular responses could be demonstrated in transplanted gut by selectively activating only the transplant-derived neurons. ENSCs were isolated from Baf53b:Cre;R26-Channelrhodopsin-2 tdTomato (Baf53b-ChR2tdT) mice in which all neurons express the light-sensitive ion channel, channelrhodopsin-2 (ChR2), and thus can be activated by blue light stimulation (BLS). Immunohistochemical characterization of colonic muscle preparations dissected from 3-week-old Baf53b-ChR2tdT mice demonstrated complete overlap between Baf53b-tdT and Hu-immunoreactive enteric neurons (Figure 3, A–D), confirming that all gut-derived cells used for transplant are capable of responding to light stimulation. Following isolation and expansion of ENSCs from Baf53b-ChR2tdT gut tissues, we injected cells, as neurospheres, into the aganglionic colons of Ednrb-KO mice via the anorectal approach as described above. Ten days later, tdT+ ENSCs had formed neo-ganglia (Figure 3, E–G, arrows) containing Tuj1+ neurons with extensive neuronal processes (Figure 3, E–G, open arrows). Tuj1 staining also revealed hypertrophic nerve bundles within the aganglionic recipient colon (Figure 3G, arrowheads). Organ bath studies showed that aganglionic colon obtained from Ednrb-KO mice (Figure 3H, red tracing) or sham-operated Ednrb-KO mice (Supplemental Figure 3) exhibited spontaneous myogenic activity, as seen in Figure 2A but no contractile response to BLS. In contrast, BLS induced a robust contractile response following transplantation of ChR2+ neurons (Figure 3H, blue tracing). Quantitative comparisons demonstrated a significant increase in the amplitude of BLS-evoked muscle contractions in transplanted Ednrb-KO colons (Figure 3I). This effect was abolished by addition of TTX, confirming that the responses were mediated by transplant-derived neurons (Figure 3, H and I).

Optogenetics demonstrates neuromuscular connectivity between ENSCs and recipient aganglionic colon. Immunohistochemical evaluation of ENS in the Baf53b-ChR2tdT mice confirmed that Hu+ enteric neurons express ChR2tdT (A–D, arrows). Two weeks after surgery, transplanted cells were visualized (E). High-power images show that transplanted cells form neuronal cell clusters (F and G, arrows) with projecting fibers (F and G, open arrows), and hypertrophic nerve bundles (F and G, arrowheads) within the aganglionic colon. Traces depict spontaneous contractions and smooth muscle responses to BLS (H). While Ednrb-KO and WT colon show no response to BLS, transplantation of ChR2-expressing ENSCs leads to robust smooth muscle contraction (I), which is significantly reduced by the addition of TTX (I). Scale bars: 50 μm (B–D), 100 μm (A), 200 μm (F and G), and 500 μm (E). All the values represent the mean of 2–4 animals for each group, repeated 2–3 times. Data are shown as the mean ± SEM. Statistical significance was determined by the 1-way ANOVA with a post hoc Tukey’s test. **P < 0.01 and ***P < 0.001 are statistically significant. BLS, blue light stimulation; ChR2, channelrhodopsin-2; TTX, tetrodotoxin.

Isolation, expansion, and characterization of ENSCs from Plp1GFP;Baf53b-tdT mice. To allow more thorough cell characterization and determination of the fate of transplant-derived enteric neurons and glia, we generated a novel dual reporter transgenic mouse line (22) by crossing Plp1GFP (glial reporter) mice (27) with Baf53b;R26-tdT (neuronal reporter) mice (28). Immunohistochemical characterization of the longitudinal muscle-myenteric plexus (LMMP) obtained from 6-week-old Plp1GFP;Baf53b-tdT mice showed that Baf53b-tdT+ enteric neurons were immunoreactive for the pan-neuronal marker Hu (Figure 4, A–D, arrows), and Plp1GFP cells colocalized with GFAP-expressing enteric glial cells (Figure 4, E–H, arrows). ENSCs were isolated from these mice and expanded as neurospheres (Figure 4, I–P), which contained both Baf53b-tdT–expressing neurons and Plp1GFP-expressing glia. Immunostaining demonstrated colabeling of enteric neurons with Baf53b-tdT and Tuj1 (Figure 4, I–K) and of enteric glia with Plp1GFP and S100β (Figure 4, L–N). Dissociated neurospheres were plated on a fibronectin-coated surface, where they gave rise to neurons (Figure 4, Q–S, Tuj1, arrows) and glial cells (Figure 4, T–V, arrows), confirming the presence of neuroglial progenitors within the neurospheres.

Isolation, expansion, and differentiation of ENSCs from Plp1GFP;Baf53b-tdT mice. Plp1GFP;Baf53b-tdT mice, in which Baf53b/Hu+ neurons express tdT (A–D, arrows) and PLP1/GFAP+ glial cells express GFP (E–H, arrows) were used to isolate ENSCs and generate enteric neurospheres (I–P), which express markers for neurons (Tuj1;J) and glia (S100β;M). Upon dissociation and culturing on fibronectin, neurospheres give rise to neurons (Q–S, Tuj1, arrows) and glial cells (T–V, GFAP, arrows). Scale bars: 50 μm (B–D, I–K, and L–N), and 100 μm (A, E–H, and O–V).

Neurospheres are enriched in enteric glia/progenitors. To determine ENSC fate, including generation of neural subtypes, prior to and following transplant to aganglionic colon in vivo, we isolated ENSCs from Plp1GFP;Baf53b-tdT mice and performed multiple injections of neurospheres into the midcolon of 2- to 3-week-old Ednrb-KO mice via laparotomy (Figure 5A). Recipient colons were examined 2 weeks following surgery. Transplanted ENSCs engrafted, migrated, and formed neo-ganglia (Figure 5, B–E, and Supplemental Figure 1, C and D) that contained Hu+ neuronal cell bodies with extensive fiber projections within the aganglionic gut environment. These observations are similar to our findings with Wnt1-tdT cell transplants (Figure 1) but with increased cell coverage due to performing multiple injections (4.9 ± 1.4 mm2 following single anorectal injection, n = 4, vs. 12.3 ± 4.5 mm2 after multiple injections via laparotomy, n = 3; Supplemental Figure 1E). We also observed that neural fibers projected extensively from the transplant sites and notably penetrated the muscle layers to reach the submucosal layer, similar to normal ENS (Supplemental Video 1). Further characterization of neo-ganglia formed by transplanted ENSCs demonstrated that both nNOS+ (Figure 5, F–I) and calretinin+ (Figure 5, J–M) enteric neuronal subtypes were present at 2 weeks after transplantation.

ENSCs transplanted into Ednrb-KO mice via laparotomy formed neo-ganglia that contain enteric neuron subtypes. The experimental design involves isolation of ENSCs from Plp1GFP;Baf53b-tdT mice, expansion as enteric neurospheres, and transplantation into the midcolon of recipient HSCR mice by multiple injections via laparotomy (A). Two weeks following surgery, transplanted cells are present in the aganglionic recipient colon (B). Many cell clusters contain neurons (C–E, arrows), and extensive fiber projections are seen (C and D, arrowheads). Transplanted ENSC-derived neo-ganglia contain nNOS-immunoreactive (F–I, arrows) and calretinin-immunoreactive (J–M, arrows) neurons with fibers (G, H, K, and L, arrowheads). Cell compositions in “Neurospheres in vitro” and “Transplant-derived neo-ganglia” were compared with those in the enteric ganglia of small or large bowel of 1- to 2-month-old WT mice (N). Statistical significance was determined by Fishers’ exact test. Scale bars: 25 μm (F–M) and 500 μm (B–E).

To determine the extent of neurogenesis and the fate of transplanted cells, quantitative evaluation of ENS composition of the neo-ganglia was performed and the results were compared with the cell types present within neurospheres in vitro prior to transplantation (Figure 5N and Supplemental Figure 2, A–A′′′) and to the endogenous enteric ganglia in 1- to 2-month-old WT mice (Figure 5N and Supplemental Figure 2, B and C). Cells within the neurospheres were predominantly (83.2%) Plp1+ glia/progenitors, while 13.6% were neurons, as shown by Baf53btdT expression. Interestingly, within the neurospheres before transplantation, a small population of cells (3.4%) was double-positive for Plp1GFP and Baf53btdT (Figure 5N and Supplemental Figure 2, A–A′′′, arrows), and only a small percentage of neurons expressed subtype markers (0.56% nNOS+ and 0.66% calretinin+, Figure 5N). After transplantation into aganglionic colon in vivo, the transplant-derived neo-ganglia were examined. The proportion of cells expressing Plp1GFP was essentially unchanged (88.4%, Figure 5N), and differentiated enteric neurons, expressing nNOS or calretinin, were present. However, the proportions of nNOS- and calretinin-expressing neurons within the neo-ganglia were significantly lower compared with those within normal enteric ganglia (7.1% nNOS and 3.0% calretinin in neo-ganglia vs. 13.1% NOS and 24.0% calretinin in “Enteric ganglia of large bowel,” Figure 5N). Since ENSCs were isolated from the small intestine, we also examined the ENS composition of enteric ganglia in the small bowel of 1- to 2-month-old mice and found that it contains 64.7% Plp1+ enteric glia, 10.1% nNOS+ neurons, and 19.5% calretinin+ neurons (Figure 5N and Supplemental Figure 2, D and E). This analysis suggests that neurosphere culture leads to an expansion of the Plp1+ glial population, which is known to include enteric neuronal progenitors, and a significant reduction in terminally differentiated enteric neuronal subtypes. These changes in cell proportions largely persist at 2 weeks following cell transplantation (Figure 5N), the latest time point we examined.

Colonic dysmotility in the aganglionic colon is restored by ENSC transplantation. We next evaluated whether ENSC transplantation restores colonic motility, as assessed by spatiotemporal mapping. This approach complements and extends the EFS and optogenetic experiments as, unlike those methods, it provides a quantitative whole-organ assessment of coordinated colonic contractility, which is a critically important endpoint for a regenerative cell therapy to treat HSCR. Two weeks following cell transplantation in vivo, the colon was removed, placed in an organ bath, and colonic migrating motor complexes (CMMCs) were evaluated.

Kymographs were generated from 10-minute video recordings of colons from Ednrb WT (Supplemental Video 2), Ednrb-KO (Supplemental Video 3), and Ednrb-KO + cells (Ednrb-KO mice with ENSCs transplanted into the mid-colon) (Supplemental Video 4) mice. Representative recordings are shown in Figure 6A. Analysis of the kymographs demonstrated that the number of CMMCs was significantly decreased in Ednrb-KO colons compared with Ednrb WT colons (5.0 ± 0.6 CMMCs in Ednrb WT vs. 0.5 ± 0.2 CMMCs in Ednrb-KO, P < 0.001, Figure 6C), and this was significantly improved following cell transplantation (8.2 ± 0.6 CMMCs in Ednrb-KO + cells, P < 0.001, Figure 6C). The significant reductions in CMMC velocity (0.4 ± 0.2 mm/s in Ednrb-KO vs. 2.3 ± 0.3 mm/s in Ednrb WT, P < 0.01, Figure 6D) and distance (9 ± 3.4 mm in Ednrb-KO colon, Figure 6E) were also improved by cell transplantation (Figure 6, D and E).

ENSC transplantation restores colonic motility in mice with HSCR and prolongs their survival. Representative spatiotemporal map kymographs generated from video recordings of colonic motility from Ednrb WT (n = 6), Ednrb-KO (n = 5), and Ednrb-KO + cells (n = 3) mice 2 weeks after cell transplant, depicting colonic contraction (red) and relaxation (yellow) along the length of the colon over time. The propagating CMMCs observed in WT mice are absent in KO mice but are partially restored following cell transplantation (Ednrb-KO + Cells) (A). Simultaneous intraluminal pressure recordings show effective colorectal contractility in WT mice, minimal pressure generation in Ednrb-KO mice, and significant restoration after cell transplant (B and F). CMMC frequency (C), velocity (D), and distance propagated (E) are all markedly increased in the Ednrb-KO + Cells group compared with the Ednrb-KO group. (G) Survival curve of Ednrb-KO mice that underwent ENSC transplantation (n = 3) or no treatment (n = 5). Statistical significance was determined by log-rank (Mantel-Cox) test (G). All the values represent the mean of 2–4 animals for each group, repeated 2–3 times. Data are shown as the mean ± SEM. Statistical significance was determined by the 1-way ANOVA with a post hoc Tukey’s test (C–F). *P < 0.05, **P < 0.01, and ***P < 0.001 are statistically significant.

We also measured luminal pressure in the colon to determine whether the CMMCs were associated with a pressure change in the colonic lumen, a process necessary for the propagation of fecal contents. In Ednrb WT, a sharp increase in luminal pressure was observed during each CMMC, and this was absent in Ednrb-KO colon (Figure 6B). However, following ENSC transplant, luminal pressures were restored in the aganglionic colon, and these corresponded to the CMMCs that were observed in WT (50.2 ± 3.0 mmHg in Ednrb WT vs. 11.0 ± 2.0 mmHg in Ednrb-KO vs. 28.7 ± 2.7 mmHg in Ednrb-KO + Cells, Figure 6F). These findings indicate that ENSC transplantation significantly improves colonic motility in the aganglionic colon of mice with HSCR.

ENSC transplant increases survival of Ednrb-KO mice via amelioration of enterocolitis. Finally, to determine whether the restoration of dysmotility in Ednrb-KO mice by cell transplant has an effect on overall animal survival, we assessed survival time (days) following ENSC transplantation and compared these data with findings in naive Ednrb-KO mice. We found that ENSC transplantation significantly prolonged animal survival (median age; 10 days in naive, n = 5 vs. 26 days in ENSC transplant, n = 3, P = 0.02, Figure 6G). Because severe gut inflammation is a common and lethal complication of HSCR and is associated with early death of Ednrb-KO mice, we asked whether ENSC transplant could play a role not only in restoring gut motility, but in reducing colonic inflammation. Histological examinations of distal colon were performed to assess severity and depth of inflammation (Supplemental Figure 4A), and the degree of colonic inflammation was evaluated using enterocolitis scoring (29) (Supplemental Figure 4B). Consistent with previous reports (29), colonic inflammation was observed in Ednrb-KO mice (Supplemental Figure 4B), whereas ENSC transplantation markedly reduced the colonic inflammation score (Supplemental Figure 4B).