Half of all adult small intestinal myenteric neurons are derived from a non-neural crest lineage.

(a) Enteric glia, labeled with GFAP (green) in the myenteric plexus from an adult Wnt1- cre:Rosa26-tdTomato mouse. Inset is magnified in color segregated panels, showing GFAP+ glia co-expressing tdTomato (red, red arrows). Nuclei are labeled with DAPI (blue). Scale bar = 10 µm.

(b) 2 dimensional representation of a 3 dimensional stack of images shows enteric neurons, labeled with Hu (green), in the myenteric plexus from an adult Wnt1-cre:Rosa26-tdTomato mouse. Enteric neurons either express tdTomato (red arrow) or not (white arrows). Nuclei are labeled with DAPI (blue). Scale bar = 10 µm.

(c) Orthogonal views of the image (b) shows the lack of tdTomato expression in a particular Hu-immunolabeled enteric neuron.

(d) Quantification of tdTomato+ and tdTomato- neurons in the myenteric ganglia of 6 P60 Wnt1- cre:Rosa26-tdTomato mice. Data represents Mean ± SEM.

Mesoderm-lineage-tracing and marker expression provide evidence of the mesodermal-derivation of half of all adult small intestinal myenteric neurons.

(a) Small intestinal LM-MP from adult male Tek-cre:Hprt-tdTomato shows the presence of Tek- derived and hence mesoderm-derived tdTomato+ (red) Hu+ (green) neurons (white arrows) and non-mesoderm-derived tdTomato- neurons (green arrows). Vascular cells depict higher fluorescence intensity (yellow arrow) than cells in the ganglia. Nuclei are labeled with DAPI. Scale bar = 10 µm.

Cardiac and small intestinal tissues from (b) heart and (c) small intestinal LM-MP layer of adult male Mesp1-cre:Rosa26-tdTomato mice shows the presence of variable tdTomato expression in cells of these tissues. While cardiomyocytes in heart and cells within the myenteric ganglia (white arrows) exhibit lower tdTomato fluorescence, similar to the tissues from Tek-cre:Hprt-tdTomato mice in (a), the vascular cells exhibit higher fluorescence intensity (yellow arrows). Scale bar = 10 µm.

(d) Small intestinal LM-MP from adult male Mesp1-cre:Rosa26-tdTomato mouse shows the presence of mesoderm-derived tdTomato+ (red) Hu+ (green) neurons (white arrows) and non-mesoderm-derived tdTomato- neurons (green arrows). Nuclei are labeled with DAPI. Scale bar = 10 µm.

(e) Small intestinal LM-MP from adult male Mesp1-cre:Rosa26-EGFP mouse, when immunostained with antibodies against GFP and against Hu, shows the presence of mesoderm- derived EGFP+ (green) Hu+ (red) neurons (white arrows) and non-mesoderm-derived EGFP- neurons (red arrows). Nuclei are labeled with DAPI. Scale bar = 10 µm.

(f) Quantification of tdTomato-expressing and non-expressing neurons in the myenteric ganglia of 3 P60 Mesp1-cre:Rosa26-tdTomato mice. Data represents Mean ± SEM.

MHCst immunolabeling exclusively marks all the (g) mesoderm-derived adult Hu+ (green) neurons in the small intestinal LM-MP of Mesp1-cre:Rosa26-tdTomato mice; and (h) all the non- NC-derived neurons in the small intestinal LM-MP of Wnt1-cre:Rosa26-tdTomato mice (white arrows). MHCst does not label tdTomato+ NC-derived cells (yellow arrows). Nuclei are stained with DAPI (blue). Scale bar = 10 µm.

Cellular and molecular phenotyping of MENs and NENs.

MET immunostaining (green) labels (a) Wnt1-cre:Rosa26-tdTomato- MENs and (b) a population of Mesp1-cre:Rosa26-tdTomato+ MENs (white arrow) while not labeling all MENs (red arrow).

(c) RET immunostaining (green) only labels Wnt1-cre:Rosa26-tdTomato+ NENs (green arrow) and not tdTomato- MENs (white arrows).

(d) NOS1 is expressed by both tdTomato+ NENs (red, white arrows) and tdTomato- MENs (blue arrows) in an adult Wnt1-cre:Rosa26-tdTomato mouse, but most MENs do not express NOS1 (marked by *).

(e) NEN lineage contains significantly higher proportions of NOS1+ neurons compared to MEN lineage in Wnt1-cre:Rosa26-tdTomato mice. Data represent mean ± S.E.M. p < 0.0001; Fisher’s exact test.

(f) ChAT is expressed by both tdTomato+ NENs (red, white arrows) and tdTomato- MENs (blue arrows) in an adult Wnt1-cre:Rosa26-tdTomato mouse.

(g) NEN lineage contains significantly higher proportions of ChAT+ neurons compared to MEN lineage in Wnt1-cre:Rosa26-tdTomato mice. Data represent mean ± S.E.M. p < 0.0001; Fisher’s exact test.

(h) Both tdTomato+ (red) and tdTomato- neurons in the myenteric plexus of an adult Wnt1- cre:Rosa26-tdTomato mouse (Hu, green) express CGRP (blue) Inset showing a tdTomato- CGRP+ neuron (white arrow) is magnified on the right.

(i) MEN lineage contains significantly higher proportions of CGRP+ neurons compared to NEN lineage in Wnt1-cre:Rosa26-tdTomato male mice. Data represent mean ± S.E.M. p < 0.0001; Fisher’s exact test.

Nuclei in (b) and (c) are labeled with DAPI (blue). Scale bar for all images denotes 10 µm.

scRNAseq-analyses identifies the distinct transcriptomic profile of the MENs.

(a) UMAP representation of 11,123 sequenced high-quality cells that were identified as meeting a 200 UMI minimum threshold with a mitochondrial read ratio of less than 20%. Clusters were annotated by markers that were found to be cell-type specific by searching UniProt, Allen Cell Atlas and Pubmed databases. Cells of the neural crest lineage were then identified as NENs by their expression of neural crest marker gene Ret and Neuroglia by their expression of Ncam1 and Sox10, or as MENs by co-expression of Calcb (CGRP), Cdh3, and Met genes.

(b) Visualization of expression of select MEN-specific and neuronal markers using quasi-violin plots. “Max value” represents the scale for the log-normalized expression of each gene.

(c) Validation of the MENs-specific marker genes discovered in the scRNAseq analyses by immunohistochemistry and confocal microscopy of small intestinal LM-MP from adult male Wnt1- cre:Rosa26-tdTomato mice. Expression of neuronal markers STX3, VSNL1, STMN2, HAND2, GPR88, PDE10A, ELAVL2, and TUBB2B (Gray, green arrows) was found in tdTomato- MENs. Immunostaining of the proteins AEBP1, CFTR, CLIC3, NT-3, SLPI, SMO, IL-18, SLC17A9, FMO2 and MYL7 (green; green arrows) was found to be localized to tdTomato- MENs. tdTomato+ (red) NC-cells did not immunostain for these markers (red arrows). Panel also shows Wnt1- cre:Rosa26-tdTomato tissue with no primary controls (stained with AlexaFluor 488 and AlexaFluor 647 antibodies). Figures with * annotations are 2D representation of 3D stacks of images. Nuclei in some panels were labeled with DAPI (blue). Scale bar denotes 10 µm.

MENs express the mesenchymal gene Decorin and the ENS-specific gene Phox2b.

(a) Sparkline representation plot of the top 90 percentile of expressed genes in the various scRNAseq cell clusters from the adult murine LM-MP tissue of a 6-month-old mouse shows that the MENs cluster, which expresses the genes Cdh3, Pde10a, and Hand2, also expresses significant amounts of the mesenchymal gene Decorin (Dcn). Darker colors of the sparkline plot represent higher expression.

(b) Sparkline representation plot of the neural crest-derived enteric neurons from the Zeisel et al. dataset shows that the Decorin gene is not found expressed by most neural crest-derived enteric neurons.

(c) 2 dimensional representation views and orthogonal view of a myenteric ganglion from the small intestinal LM-MP of adult male Wnt-cre:Rosa26-EGFP mouse, where tdTomato (red) is expressed by neural crest-derived cells, when immunostained with antibodies against DECORIN (grey) shows that the DECORIN-expressing myenteric cells (green arrows) do not express tdTomato and hence are non-neural crest-derived MENs. Nuclei are labeled with DAPI. Scale bar = 10 µm.

(d) Representative image of a myenteric ganglion from the small intestinal LM-MP of adult male Wnt-cre:Rosa26-EGFP mouse, where tdTomato (red) is expressed by neural crest-derived cells, when immunostained with antibodies against PHOX2b (green) shows that the PHOX2b- expression is found in myenteric cells that do not express tdTomato (green arrows, and hence are not neural crest-derived cells, or MENs) as well as in tdTomato-expressing neural crest- derived NENs (red arrow). Nuclei are labeled with DAPI. Scale bar = 10 µm.

(e) Volcano plot of gene expression profiles of Phox2b-expressing cells, which were sorted based on their CFP expression level, shows that the expression of MENs marker genes Smo, Aebp1, Cdh3, Fmo2, Il18, Slpi, Upk3b, Msln, and Hand2 is significantly enriched in the Phox2b-CFPlow fraction. Red dotted line shows the padjusted value of 0.05.

Top expressed genes by the Calcb-expressing clusters and the putative MENs cluster in data from May-Zhang et al, along with top-expressed genes in the MENs cluster in our data. Highlighted genes between MENs clusters from our data and data from May-Zhang et al and from Drokhlyansky et al show similar gene expression profiles between the clusters.

Computational analyses using projectR identifies MENs in publicly available murine and human transcriptomic datasets.

(a) Heatmap of cell weights for the patterns learned by NMF (k=50). Hierarchical clustering was calculated using Euclidean distance. Multiple clusters annotated as the same cell type are merged. The four most specific MENs patterns (16, 27, 32 and 41) are selected.

(b) In addition to using the four MEN-specific NMF patterns to label the MENs cluster in our scRNAseq dataset (plots in top row), two other transcriptomic datasets: May-Zhang et al.’s murine ileal ENS snRNAseq dataset50 (plots in second row), and Elmentaite et al.’s gut mesenchymal scRNAseq dataset81 (plots in third row), were projected into the defined four MEN-specific pattern space using ProjectR. Raw projected cell weights were visualized using previously learned UMAP embedding. The cell clusters that show high pattern usage are shown bounded by the red square. Plots in the final row shows projection of healthy post-natal mesenchymal cells from Elmentaite et al’s data81 into the four MEN-specific NMF patterns, which again show a population of cells showing high MEN-specific pattern utilization.

Detection of MENs in human single cell RNA sequencing data.

(a) UMAP representation of scRNAseq data from Elmentaite et al.’s healthy post-natal mesenchymal cells from the human gut81.

(b) Breakdown for previously published annotated features of the post-natal health subset of the gut cell atlas as presented in Elmentaite et al.81.

(c) Expression of the MENs marker gene Dcn in the UMAP representation of healthy post-natal data from Elmentaite et al.81., which was annotated as clusters of mesenchymal cells.

(d) Expression of neuronal marker genes Slc17a9, Stx3, Uchl1, Tubb2b, Pde10a, and Hand2 in the UMAP representation of healthy post-natal data from Elmentaite et al.81, which was annotated as clusters of mesenchymal cells.

Validation of SNAP-25 as a NENs marker

(a) SNAP-25 expression is restricted to the neural crest lineage in the adult myenteric ganglia.

SNAP-25 expression (green) co-localizes with tdTomato (red) but not with the MENs marker MHCst (cyan) as observed in 2D and orthogonal views of a myenteric ganglia from a Wnt1- cre:Rosa26-tdTomato mouse that was immunolabeled with antibodies against MHCst and SNAP- 25. Nuclei were labeled with DAPI (blue). Scale bar denotes 10 µm.

(b) In the adult male Snap25-Gcamp knock-in mice, the expression of Snap25-Gcamp/GFP (green) is restricted to a subset of Hu-expressing (red) myenteric neurons in the adult murine small intestinal myenteric ganglia (yellow arrows), while many neurons (red arrows) do not show any detectable expression of Snap25. Nuclei were labeled with DAPI (blue). Scale bar denotes 10 µm.

(c) Orthogonal views of z-stack of an image of the myenteric ganglion from the small intestinal LM-MP tissue from an adult male Snap25-Gcamp knock-in mice shows that the expression of Snap25-Gcamp/GFP (green) is exclusive of the expression of the MENs marker MHCst (red). Nuclei were labeled with DAPI (blue). Scale bar denotes 10 µm.

Observation of MENs marker expression in the adult human myenteric ganglia

(a) Hu-expressing small intestinal myenteric neurons (green) from the normal human duodenal tissue, when immunolabeled with antibodies against the MENs marker DECORIN (red) identifies putative human MENs (red arrows) and NENs (green arrows) by presence or absence of DECORIN immunolabeling. Scale bar denotes 10 µm.

(a) Hu-expressing small intestinal myenteric neurons (green) from the normal human duodenal tissue, when immunolabeled with antibodies against the MENs marker MHCst (red) and MET (blue) identifies putative human MENs (red arrows) and NENs (green arrows) by presence or absence of these MENs markers. Scale bar denotes 10 µm.

(c) Immunolabeling myenteric tissue with antibodies against the MENs markers SLC17A9 (green, green arrows) and MHCst (red) shows SLC17A9 expression in a subset of MHCst-expressing (red) neurons. Subset of MHCst-expressing cells do not express SLC17A9 (red arrow). Nuclei are labeled with DAPI (blue) in the normal human duodenal tissue. Scale bar denotes 10 µm.

GDNF and HGF signaling regulate age-dependent changes in NENs and MENs proportions.

(a, b) Immunostaining myenteric plexus tissue from juvenile and mature Wnt1-cre:Rosa26- tdTomato mice with antibodies against the pan-neuronal marker Hu (green).

(c) Age-associated loss of NENs and gain of MENs in the small intestinal LM-MP of maturing and aging Wnt1-cre:Rosa26-tdTomato mice. Data represent mean ± S.E.M.

(d) Mean ± SEM of the four MEN-specific pattern weights in the human mesenchymal cell populations from Elmentaite et al.81, wherein data from age ranges of 4 – 12 years was clubbed together as Juvenile (Juv), data from age ranges of 20 – 50 years was clubbed together as Adult (Ad), and data from age ranges of 60 – 75 years was clubbed together as Aged (Ag). Every datapoint refers to mean projected pattern weight for cells within a defined age or age range, where Juv data comprise of ages 4, 6, 9, 10, 12 years; Ad data comprise of age ranges 20 - 25, 25 – 30, and 45-50 years; and finally Ag data comprise of age ranges 60 – 65, 65 – 70, and 70 – 75 years. One-way ANOVA ** = p < 0.01).

(e) Western blot analyses of GDNF (green) and the house-keeping protein β-actin (red) expression in LM-MP tissues from mice of ages P10, P30, and P90. (n = 3 mice per group; each sample is a biological replicate). Fluorescent intensities of the two bands of GDNF (that correspond to ∼25 kD bands of protein marker) were quantified together. The lower band of GDNF is present only in the P10 tissues and disappears in P30 and P90 adult murine tissues.

(f) Western blot analyses of HGF (green) and the house-keeping protein β-Actin (red) expression in LM-MP tissues from mice of ages P10, P30, and P90. (n = 3 mice per group; each sample is a biological replicate). Fluorescent intensities of the two bands of HGF (that are between 50 kD and 37 kD bands of the protein marker) were quantified together.

(g) The normalized fluorescent intensity of GDNF protein to house-keeping protein β-Actin compared between the three age groups. GDNF presence was highest in P10 group and was significantly reduced in P30 and P90 groups. Data represent mean ± S.E.M. One-way ANOVA **** p < 0.001.

(h) Age-dependent decrease in Gdnf mRNA transcript expression (normalized to the house-keeping gene Hprt) in the myenteric plexuses of P30 and P90 old mice. Data represent mean ± S.E.M. Student’s t-test *p < 0.05.

(i) The normalized fluorescent intensity of HGF protein to house-keeping protein β-Actin was compared between the three age groups. HGF expression significantly increased from P10 through P90. Data represent mean ± S.E.M. One-way ANOVA *p < 0.05.

(j) Age-dependent increase in Hgf mRNA transcript expression (normalized to the house-keeping gene Hprt) in the myenteric plexuses of P10, P30, and P90 old mice. Data represent mean ± S.E.M. One-way ANOVA * p < 0.05, *** p < 0.01, **** p < 0.001.

(k) Percent proportions of tdTomato- MENs and mean Hu+ neurons/ganglia in LM-MP of cohorts of Wnt1-cre:Rosa26-tdTomato mice that were dosed with GDNF or Saline from P10 to P20 age. Data represent mean ± S.E.M. Student’s t-test ** p < 0.01.

(l) Percent proportions of tdTomato- MENs and mean Hu+ neurons/ganglia in LM-MP of cohorts of Wnt1-cre:Rosa26-tdTomato mice that were dosed with HGF or Saline from P10 to P20 age. Data represent mean ± S.E.M. Student’s t-test * p < 0.05.

Ect2-expression labels a population of cycling MENs

(a) UMAP representation of the cells of the MENs, NENs, and Neuroglia clusters from scRNAseq of cells from the small intestinal LM-MP tissue of two post-natal day 21 (P21) mice.

(b) Tricycle analyses of ENS cell-types with their cell cycle positions. The continuous cell cycle position (theta) is measured as the angle from the origin. Anti-clockwise representation of cells that are represented between 1.5π and 0.5π space on the embedding are increasingly present within the quiescent G1/G0 cell cycle phase. In contrast, cells that are represented between the 0.5π and 1π space on the embedding are present within the S phase, and those within the 1π and 1.5π space are present within the G2M phase.

(c) Cells within the P21 MENs cluster were projected into Tricycle software. 213 of the 504 P21 MENs were present between the 0.5π and 1.5π space on the embedding and hence were inferred to be undergoing cell cycling.

(d, e) Expression of cell-cycle correlated genes Top2a and Ect2 in MENs shows significant expression of these two genes in cells of the P21 MENs cluster that are present between the values of 0.5π and 1.5π in the Tricycle embedding. Loess curve fittings (black) represent the dynamics of gene expression over the cell cycle (theta).

Orthogonal views of small intestinal myenteric ganglia from a P21 wildtype mouse when immunostained with antibodies against ECT (green) and against MENs markers (f) DECORIN (red) and (g) MHCst (red), shows that ECT2 is expressed in a subset of MENs at this age. Nuclei are labeled with DAPI (blue). Scale bar denotes 10 µm.

Reduced RET signaling accelerates ENS aging to cause pathology

(a) Hu immunostaining (green) LM-MP tissues from 16-week-old Ret+/CFP (Ret+/-) mouse shows mutually exclusive expression of Ret-CFP (cyan, white arrow) and MHCst (red, red arrow) MENs. Nuclei are stained with DAPI (blue). Scale bar = 10 µm.

(b) Quantification of Ret-CFP+ neurons from 9- and 16-week-old Ret+/- mice show age-associated loss of Ret-CFP+ neurons. Data represent mean ± S.E.M. Student’s t-test * p < 0.05.

(c) Quantification of MHCst+ MENs shows significant increase in their proportions in Ret+/- mice but not in Ret+/+ mice with age. Data represent mean ± S.E.M. One way ANOVA * p < 0.05, ** p < 0.01, *** p < 0.001).

(d) Measures of whole gut transit time (WGTT) in cohorts of Ret+/- and Ret+/+ mice MENs show significant slowing of whole gut transit of Ret+/- but not Ret+/+ mice with age. Data represent mean ± S.E.M. One way ANOVA * = p < 0.05.

GDNF normalizes altered intestinal motility by increasing NENs proportions in the aging gut.

(a) Measures of whole gut transit time (WGTT) in GDNF (treated with GDNF) and Control (treated with Saline) cohorts of 17-month-old mice taken before the start of treatments and after the end of 10 consecutive days of treatment shows that the two groups are matched in their transit times before treatment, but GDNF treatment significant decreases average transit times when compared to the control cohort. Data represent mean ± S.E.M. Student’s t-test *** = p < 0.001.

(b) Quantification of percent MHCst+ MENs per Hu-labeled neurons in myenteric ganglia in the GDNF and Control cohorts shows significant decrease in their proportions in GDNF-treated cohort but not in saline-treated control cohort. Data represent mean ± S.E.M. Student’s t-test ** p < 0.01.

(c) Quantification of numbers of RET+ NENs per 40X field views of myenteric ganglia shows significant increase in their numbers in GDNF cohort mice when compared with Control cohort mice. Data represent mean ± S.E.M. Student’s t-test * p < 0.05.

Patients with chronic gut dysmotility show significant shifts in their normal proportions of the two neuronal lineages.

(a) ProjectR-based projection of bulkRNAseq data from intestinal specimens of patients with normal motility and patients with obstructive defecation (OD), a chronic condition of intestinal dysmotility, into the 50 different NMF patterns learnt earlier, shows that the MEN-specific NMF patterns 32 and 41 were significantly upregulated in bulk-RNAseq of OD patients compared to controls. Data represent mean ± S.E.M. One way ANOVA. Data mined from raw data generated by Kim et al.103

(b) OD patients show a significant decrease in the expression of important NENs-associated genes such as Ret, Gdnf, Snap-25, Nos1, Klhl1, and Chat, while showing a significant increase in the expression of important MENs-specific genes such as Clic3, Upk3a, Cdh3, Slpi, and Slc17a9. Data taken from Kim et al.103