The G4-RFP transgene is active in both neurons and glia of the distal colon after birth.

a,b. Immunofluorescence staining of SOX10, HuC/D and RFP in the distal colon of untreated (a) and GDNF-treated (b) HolTg/Tg;G4-RFP at P10. Each black dot in the accompanying quantitative analysis corresponds to a microscopic field of view (7 to 10 fields of view per animal; white dots indicate the average per animal; N indicates the number of animals; n indicates the total number of counted cells). The dashed box in each diagram indicates the colon segment that was analyzed. The dashed outlines in the micrographs highlight extrinsic nerve fibers and/or induced ganglia. Scale bar, 25 μm.

The distal colon of HolTg/Tg;G4-RFP mice is enriched in SCPs and EGCs.

a. Overview of experimental procedure, from GDNF treatment regimen to scRNA-seq analysis of FACS-recovered RFP+ cells at P10. The dashed boxes in the diagram indicate the colon segment that was analyzed, here also including part of the transition zone. b,c. Uniform manifold approximation and projection (UMAP) visualization showing unsupervised clustering of SCPs, EGCs and enteric neurons from distal colon of P10 HolTg/Tg;G4-RFP mice that were treated (c) or not (b) with GDNF. d. Dot plot showing relative expression levels of a selection of marker genes in each cluster. e. UMAP visualization showing all clusters stratified by experimental group (untreated vs. GDNF-treated HolTg/Tg;G4-RFP mice). f. Comparative frequency plots showing relative proportion of each cluster in untreated vs. GDNF-treated HolTg/Tg;G4-RFP mice.

NCAM1 and phospho-FAK[Y397] are stimulated by GDNF in the distal colon of HolTg/Tg;G4-RFP mice.

a. Feature plots showing relative gene expression levels of signaling (Ret and Ncam1) and binding (Gfra1, Gfra2, Gfra3, Gfra4) GDNF receptors in scRNA-seq dataset of GDNF-treated HolTg/Tg;G4-RFP mice. b,c. Immunofluorescence staining of NCAM1 and its downstream signaling effector phospho-FAK[Y397] in most distal colon of HolTg/Tg mice at indicated time points before, during and after GDNF treatment (panel b shows representative images of N=3 animals per time point). Quantification of mean fluorescence intensity for each experimental group is shown in panel c. Each black dot corresponds to a single microscopic field of view (7–8 imaging fields per animal; white dots indicate the average per animal). ****P < 0.0001; Statistics were calculated from the datasets plotted as black dots using ordinary one-way ANOVA with post-hoc Tukey’s test. Dashed outlines in panel b highlight either extrinsic nerve fibers or individual ganglia. d. Immunofluorescence staining of NCAM1 and phospho-FAK[Y397] in mid-colon of untreated HolTg/Tg mice at indicated time points (representative images of N=3 animals per time point). Scale bar, 25 μm.

The neurogenic effect of GDNF in HolTg/Tg;G4-RFP mice is mediated by NCAM1, not RET.

a. Diagram of the experimental procedure showing that HolTg/Tg mice were administered GDNF enemas and were also injected with the indicated selective inhibitors (or vehicle only) for phospho-FAK[Y397] (PF-562271) or phospho-RET[Y1062] (BLU-667) from P4 to P8. b. Western blot analysis of both total and phosphorylated forms of FAK or RET in the proximal colon of P8 HolTg/Tg mice treated as indicated above. Each value in the accompanying quantitative analyses of phospho/total ratios correspond to an animal. ***P < 0.001; ****P < 0.0001; Statistics were calculated from the datasets plotted as black dots using ordinary one-way ANOVA with post-hoc Tukey’s test. c,d. Immunofluorescence staining of HuC/D and SOX10 (c) and cleaved Caspase 3 (d) in most distal colon of the same P8 HolTg/Tg mice that were used in panel b. Quantitative analysis of HuC/D+ neurons per mm2 is shown on the right, with each black dot corresponding to a microscopic field of view (7 to 10 fields of view per animal; white dots indicate the average per animal; N indicates number of animals per group; n indicates total number of counted cells). ****P < 0.0001; Statistics were calculated from the datasets plotted as black dots using ordinary one-way ANOVA with post-hoc Tukey’s test. Dashed outlines in panels c-d highlight either extrinsic nerve fibers or individual ganglia. Scale bar, 150 µm (c) and 25 μm (d).

Re-clustering based on Elavl4 expression reveals a stepwise differentiation mechanism connecting SCPs, EGCs and enteric neurons.

a. Feature plot showing Elavl4 expression levels in scRNA-seq dataset of GDNF-treated HolTg/Tg;G4-RFP mice. b. UMAP visualization of scRNA-seq dataset of GDNF-treated HolTg/Tg;G4-RFP mice after reclustering based on Elavl4 expression levels. c. Dot plot showing relative expression levels of a selection of marker genes in each cluster. d. Feature plots showing relative gene expression levels of pan-glial (Sox10), pan-neuronal (Elavl4 and Tubb3), SCP (Mpz and Dhh), cycling cell (Ubec2c) and EGC (Gfap and Slc18a2) markers. e. Pseudotime analysis of Elavl4-based reclustered scRNA-seq datasets of untreated and GDNF-treated HolTg/Tg;G4-RFP mice.

GDNF-induced neuronal differentiation of SCPs occurs in part via and EGC-like state.

a,b. Immunofluorescence staining of most distal colon samples from P10 GDNF-treated HolTg/Tg;Dhh-CreTg/+;R26YFP/+ mice showing several newly induced myenteric ganglion cells derived from SCPs (YFP+) that are undergoing neuronal differentiation (PHOX2B+ cells with large rounded nucleus) while also expressing either GFAP (a) or SLC18A2 (b) EGC markers (see yellow arrows; representative images of N=3 animals). c,d. Immunofluorescence staining of most distal colon samples from GDNF-treated HolTg/Tg;Dhh-CreTg/+;R26YFP/+ mice at indicated time points showing that proportion of SCP-derived (YFP+) myenteric ganglion cells also positive for the EGC marker GFAP decreases over time (c) and that GFAP is not expressed in mature neurons at P20 (d). Yellow and white arrows point to GFAP+ cells that are either YFP+ or YFP-, respectively. Quantitative analysis of percentage of double YFP+ GFAP+ myenteric ganglion cells is shown on the right of panel c, with each black dot corresponding to a microscopic field of view (7 to 10 fields of view per animal; white dots indicate the average per animal; N indicates number of animals per group; n indicates total number of counted cells). ****P < 0.0001; Statistics were calculated from the datasets plotted as black dots using ordinary one-way ANOVA with post-hoc Tukey’s test. Dashed outlines in panels a-d highlight individual ganglia. Scale bar, 25 μm.

GDNF-induced neurogenesis is more rapid from SCPs than from EGCs.

a. Diagram of time-course experimental procedure and involved mouse models for comparative genetic cell lineage tracing in the HolTg/Tg mutant background. b. Quantitative analysis of density of GDNF-induced myenteric neurons in most distal colon of HolTg/Tg mice as a function of time, regardless of cell lineage tracing tools. c-e. Immunofluorescence-based kinetic analysis of distal colon myenteric SOX10+ glia (yellow arrows) or HuC/D+ neurons (yellow arrowheads) that are also positive for YFP, which indicates whether they are derived from Dhh-expressing SCPs (c), Slc18a2-expressing EGCs (d) or Gfap-expressing EGCs (e). For all quantitative analyses, each black dot corresponds to a microscopic field of view (7 to 11 fields of view per animal; white dots indicate the average per animal; N indicates number of animals; n indicates total number of counted cells). *P < 0.5; ***P < 0.001; ****P < 0.0001; Statistics were calculated from the datasets plotted as black dots using ordinary one-way ANOVA with post-hoc Tukey’s test. Dashed outlines in panels c-e highlight individual ganglia. Scale bar, 25 μm.

GDNF predominantly promotes direct transdifferentiation of ENS progenitors into neurons.

a. Diagram of time-course experimental procedure and involved mouse models for combined EdU incorporation assays and genetic cell lineage tracing in HolTg/Tg mutant background. b. Quantitative analysis of percentage of GDNF-induced HuC/D+ myenteric neurons that did not incorporate EdU in most distal colon of HolTg/Tg mice as a function of time, regardless of cell lineage tracing tools. c-e. Immunofluorescence-based kinetic analysis of distal colon myenteric HuC/D+ neurons (yellow arrowheads) that have incorporated EdU and are also positive for YFP, which indicates whether they are derived from Dhh-expressing SCPs (c), Slc18a2-expressing EGCs (d) or Gfap-expressing EGCs (e). For all quantitative analyses, each black dot corresponds to a microscopic field of view (7 to 11 fields of view per animal; white dots indicate the average per animal; N indicates number of animals; n indicates total number of counted cells). *P < 0.5; ***P < 0.001; ****P < 0.0001; Statistics were calculated from the datasets plotted as black dots using ordinary one-way ANOVA with post-hoc Tukey’s test (c, d) or unpaired two-tailed Student’s t-test (e). Dashed outlines in panels c-e highlight individual ganglia. Scale bar, 20 μm.

GDNF-induced neurons derived from Dhh-expressing SCPs and Gfap-expressing EGCs are mostly cholinergic.

a-c. Immunofluorescence-based analysis of relative proportions of nitrergic (NOS1+) and cholinergic (ChAT+) neurons among GDNF-induced myenteric neurons (HuC/D+) in the most distal colon at P20, with YFP signal (yellow arrowheads) further indicative of their origin from Dhh-expressing SCPs (a), Gfap-expressing EGCs (b) or Slc18a2-expressing cells (c). For all quantitative analyses, each black dot corresponds to a microscopic field of view (7 to 11 fields of view per animal; white dots indicate the average per animal; N indicates number of animals; n indicates total number of counted cells). ****P < 0.0001; Statistics were calculated from the datasets plotted as black dots using unpaired two-tailed Student’s t-test. Dashed outlines highlight individual ganglia. Scale bar, 25 μm.

A neural crest-independent ENS progenitor subtype is required for balancing cholinergic vs. nitrergic differentiation of GDNF-induced myenteric neurons.

a. Immunofluorescence-based analysis of the contribution of Wnt1-expressing NCCs (YFP+) to overall pool of HuC/D+ neurons and SOX10+ EGCs in most distal colon of P20 WT-like control and GDNF-treated HolTg/Tg mutant mice. Empty arrowheads point to YFP-negative neurons in HolTg/Tg mutant background. Insets show results for the proximal colon of the same mouse as internal positive control b. Immunofluorescence-based analysis of relative proportions of nitrergic and cholinergic neurons among GDNF-induced myenteric neurons in the most distal colon at P20, with YFP signal further indicative of their origin from Wnt1-expressing NCCs. Empty arrowheads point to YFP-negative neurons. For all quantitative analyses, each black dot corresponds to a microscopic field of view (7 to 11 fields of view per animal; white dots indicate the average per animal; N indicates number of animals; n indicates total number of counted cells). ****P < 0.0001; Statistics were calculated from the datasets plotted as black dots using unpaired two-tailed Student’s t-test. Dashed outlines highlight individual ganglia. Scale bar, 25 μm.

NCAM1 is robustly expressed in human samples of aganglionic colon.

Immunofluorescence analysis reveals NCAM1 localization both within and outside βIII-Tubulin+ extrinsic nerves of the submucosa and muscularis tissue layers. Scale bar, 50 μm (submucosa) and 100 μm (muscularis).