Wild-type FVB mice [FVB/NCrl; Strain code 207] were obtained from Charles River Laboratory, whereas the Dhh-Cre line [Jax stock #012929; FVB(Cg)-Tg(Dhh-cre)1Mejr/J] was obtained from the Jackson Laboratory. The Rosa26^{[FloxedSTOP]YFP} line [Gt(ROSA)26Sor^{TM1(EYFP)Cos]} was kindly directly provided by Dr. Frank Costantini (Columbia University). Both Dhh-Cre and Rosa26^{[FloxedSTOP]YFP} mice were maintained on the FVB background and genotyped by standard PCR using primers listed in Supplementary Table 1. Breeding couples were kept in individually ventilated cages and fed Charles River Rodent Diet #5075 (Cargill Animal Nutrition). All experiments were conducted in accordance with the guidelines of the Canadian Council on Animal Care (CCAC) and approved by the institutional committee (CIPA #992) of the Université du Québec à Montreal (UQAM). Euthanasia was performed either by decapitation for mice aged between postnatal day (P) 1 to P10, or by exposure to carbon dioxide (CO₂) following isoflurane anesthesia for P20 mice. Animals from both sexes were included for each time point (see supplemental Data Sheet 1 for more details).

After euthanasia, mice were pinned to a styrofoam slab to facilitate the removal of the skin and peritoneum, providing access to the gastrointestinal system. The small intestine and colon were then collected and kept in ice-cold 1X PBS (Phosphate-Buffered Saline) throughout the duration of the dissection. For both segments, we used only the most distal quarter, either just upstream of the cecum (distal ileum) or the anus (distal colon). Each sample was cut longitudinally along the mesentery, extensively washed in 1X PBS to remove stool, and pinned with mucosa side up onto Sylgard-coated (Dow Corning, Freeland, MI) Petri dishes. Pinned tissues were subsequently fixed overnight in 4% paraformaldehyde diluted in 1X PBS, followed by three washes of 10 min in 1X PBS. Tissues from P20 mice were further carefully microdissected to peel off the mucosal layer (including the submucosal plexus) while also preserving the muscle layers (including the myenteric plexus). All samples were stored in 1X PBS at 4°C prior to whole-mount immunofluorescence staining.

Ileum and colon samples were initially permeabilized for 2 h in blocking solution (5% Fetal Bovine Serum and 1% Triton-X100 in 1X PBS), followed by overnight incubation at 4°C with primary antibodies diluted in same blocking solution. After three washes of 10 min in 1X PBS, samples were then incubated with relevant secondary antibodies diluted in blocking solution for 1 h 30, followed by three additional washes of 10 min in 1X PBS. During the second wash, samples were counterstained with DAPI (4’,6-diamidino-2-phenylindole) diluted in 1X PBS. The specific antibodies used, and their dilution factors are listed in Supplementary Table 2. To enable the observation of plexuses via confocal microscopy, the stained tissues were mounted between two 24 x 50 mm glass coverslips in 100% Glycerol.

All immunofluorescence images were acquired using a Nikon A1 confocal microscope with Plan Fluor 20x/0.75 MImm and Plan Apo λ 60x/1.40 objectives. For each biological replicate, between 3 and 10 representative z-stack images were acquired, and then analyzed using the ImageJ software. EGC subtypes were manually quantified using the “Cell Counter” plugin, based on the location of SOX10+ cells and/or morphology of S100β+ cells (Supplementary Figures 1–3 and Supplementary Movies 1, 2). EGC proportions were calculated by including all relevant EGC subtypes in the total count (Type I to IV in the myenteric plexus / circular muscle layer; Type I to III in the submucosal plexus). The surface area of ganglia, extraganglionic space and interganglionic fibers were manually delineated and measured using the “freehand selections” tool, as shown in Supplementary Figure 1C.

Each experiment was performed using 3–4 biological replicates, and either 3–5 (for myenteric plexus / circular muscle) or 6–10 (for submucosal plexus) fields of view per replicate (see supplemental Data Sheet 1 for detailed information about each replicate, including sex and number of counted cells). For quantitative analysis, data are expressed as the mean ± standard deviation (SD), with the value of each imaging field also represented by a dot. Based on these single values, statistical significance of differences between groups was determined in GraphPad Prism 9.5.1, using either two-tailed Welch’s t-tests (when comparing two groups) or One-Way/Two-Way ANOVA with Tukey’s multiple comparison tests (when comparing more than two groups). Tests used are indicated in figure legends. Data were considered statistically significant when the P-value was less than 0.05.

To gain insight about the process of EGC diversification, we first analyzed the appearance of the four main topo-morphological subtypes of EGCs in the myenteric plexus and circular muscle of wild-type FVB mice. Using whole-mount immunostaining, we studied four time points (P1, P5, P10 and P20) spanning the postnatal period before weaning. The longitudinal muscle layer was excluded from this analysis because of interference from the juxtaposed serosa, which gave high background signal in immunofluorescence staining. Staining of EGCs and neuronal networks was performed using antibodies against SOX10 and βIII-Tubulin, respectively. SOX10 staining is great for cell counting, being also generally sufficient for identifying EGC subtypes based on topological criteria only (Supplementary Figure 1A and Supplementary Movies 1, 2)−as the unique morphology of each subtype is linked to a specific location, along both the x/y plane and the z-axis (Boesmans et al., 2015; Gulbransen and Sharkey, 2012; Charrier and Pilon, 2017; Hanani and Reichenbach, 1994). Notably, location along the z-axis is especially useful for distinguishing Type III from Type IV EGCs, as both subtypes are clearly not in the same plane (Supplementary Figure 1A and Supplementary Movies 1, 2). That said, we also co-stained all tissues for S100β to enable definitive identification of EGC subtypes based on gross morphological criteria as well. Although global S100β staining is not as good as sparce labeling approaches for revealing morphological details of EGCs (Boesmans et al., 2015; Hanani and Reichenbach, 1994; Rao et al., 2015), it nonetheless often proved to be useful for distinguishing between multipolar Type I and elongated Type II EGCs (from P5 onwards), when Type II EGCs are at the border of ganglia and oriented toward interganglionic fibers (Supplementary Figure 1B). Interestingly, our S100β staining data further revealed that each EGC subtype within the myenteric plexus / circular muscle gradually acquires its definitive gross morphological characteristics between P1 and P20 (Supplementary Figure 2). For the sake of simplicity, detailed S100β data are however not shown in the main figures. Given that as little as 2–6% of all SOX10+ cells from the myenteric plexus / circular muscle are negative for S100β during the investigated period (Table 1), we reasoned that the impact of this decision on our estimates of the relative proportions of EGC subtypes would be minimal.

At P1, the vast majority of SOX10+ cells (> 85%) can be classified as Type I EGCs, in both the distal ileum (Figures 1A–D) and the distal colon (Figures 2A–D). Most of the remaining EGCs fall into the Type II category (8–12%), while the other EGC subtypes are either marginally detected (2% for Type III) or absent (Type IV) in both bowel segments at this stage (Figures 1A–D, 2A–D). At P5, the proportions of Type III and Type IV EGCs are both markedly increased, with a steeper rise in Type III in the distal ileum (Figures 1A–D) and Type IV in the distal colon (Figures 2A–D). These increased proportions of Type III and Type IV are mostly made at the expense of Type I EGCs, while the proportion of Type II EGCs remains relatively stable in both bowel segments (Figures 1A–D, 2A–D). The pool of Type III and Type IV EGCs continue to grow at P10 and P20, with asymmetric patterns favoring Type IV in the distal ileum (Figures 1A–D) or Type III in the distal colon (Figures 2A–D). As the proportion of Type II remains stable at all analyzed time points, this expansion of the pool of Type III and Type IV EGCs continues to be made at the expense of Type I EGCs in both bowel segments. Nonetheless, Type I EGCs are still the most common subtype, representing about half (49 ± 10% in distal ileum; 45 ± 10% in distal colon) of all EGCs at P20 (Figures 1A–D, 2A–D). At this stage, Type III (20 ± 6% in distal ileum; 20 ± 5% in distal colon) and Type IV (17 ± 3% in distal ileum; 21 ± 8% in distal colon) EGCs are now equally represented, both outnumbering Type II EGCs (14 ± 5% in distal ileum; 14 ± 6% in distal colon) (Figures 1A–D, 2A–D).

We next verified the possibility that the sequential appearance of the four topo-morphological subtypes of EGCs described above could be related to structural changes that occur in the bowel wall between P1 and P20. To this end, we carefully measured the surface area of ganglia, extraganglionic space, and interganglionic fibers as well as the thickness of the circular muscle layer (Supplementary Figure 1C), in all samples used for EGC subtype quantification in both the distal ileum (Figure 1E) and distal colon (Figure 2E). In accordance with the observed loosening of the connections between myenteric ganglia in both bowel segments, these analyses revealed a significant stage-dependant decrease in ganglionic surface area accompanied by a commensurate increase in extraganglionic surface area (Figures 1E, 2E). When analyzed as a function of EGC subtype proportions in corresponding samples, these structural changes are strongly correlated with the proportions of Type I EGCs and Type III EGCs, respectively (Figures 1F, 2F). The surface area of interganglionic fibers was not found to vary much overall between P1 and P20 (Figures 1E, 2E), similar to the proportion of Type II EGCs (Figures 1D, 2D), but both parameters are nonetheless again positively correlated when analyzed as a function of each other in each sample (Figures 1F, 2F). Lastly, we noted that the thickness of the circular muscle layer generally increases with age in both bowel segments, peaking at P10 and then slightly decreasing at P20. However, this increase was found to be delayed in the distal ileum compared to the distal colon, being suddenly detectable at P10 and P5, respectively (Figures 1E, 2E). Remarkably, this particular kinetics mimics the abrupt increase in the proportion of Type IV EGCs detected in each bowel segment, hence resulting in a positive correlation once again (Figures 1F, 2F).

From this systematic characterization of the maturing myenteric plexus / circular muscle, we can thus conclude that the four topo-morphological subtypes of EGCs appear in numerical order, from Type I/II (co-occurring) to Type IV, regardless of the bowel segment examined. Moreover, our correlation data strongly support the notion that structural changes in the early postnatal bowel, including within the ENS, have a profound influence on the relative proportions of these four EGC subtypes.

To complement our analyses of the myenteric plexus / circular muscle, we then turned to the submucosal plexus (largely excluding the mucosa), using the exact same experimental design as described above. Here, DAPI staining (in conjunction with βIII-Tubulin staining) proved to be useful for delineating the tiny submucosal ganglia, based on the round nucleus of neurons (Supplementary Figure 3). We also found that the overlap between SOX10 and S100β stainings is much lower here than in the myenteric plexus at P1 (Table 1), but we validated that proportions of EGC subtypes are nonetheless very similar when using combined SOX10-S100β data instead of SOX10 data alone at both P1 (ileum) and P5 (colon) (Supplementary Figure 4). Of note, S100β data further revealed that acquisition of definitive gross morphological attributes is delayed in the submucosal plexus (Supplementary Figure 2B) compared to the myenteric plexus (Supplementary Figure 2A).

Our detailed SOX10 immunofluorescence analyses again revealed the sequential appearance of EGC subtypes between P1 and P20, but not in the same order as in the myenteric plexus and with more pronounced regional differences between distal ileum (Figures 3A–C) and distal colon (Figures 4A–C). At P1, Type II EGCs are clearly predominating in the submucosal plexus, being even virtually exclusive in the distal colon (Figures 4A–C). This pattern is inversely reflected in the detection of Type I EGCs, which are present in the distal ileum at P1 (Figures 3A–C) but are not detected until P5 in the distal colon (Figures 4A–C). This population of Type I EGCs then generally grows over time in both bowel segments (Figures 3A–C, 4A–C), until eventually outnumbering Type II EGCs in the distal ileum at P20 (49 ± 13% of Type I vs 38 ± 14% of Type II), but not in the distal colon where Type II remains the most common subtype at this stage as well (36 ± 15% of Type I vs 58 ± 14% of Type II). In both bowel segments (Figures 3A–C, 4A–C), Type III EGCs are not detected at P1, then forming a relatively marginal population until P20, at which stage they are more numerous in the distal ileum (13 ± 6%) than in the distal colon (6% ± 6%). However, it is important to bear in mind that these proportions only take into account Type III EGCs close to the submucosal plexus (a.k.a. Type III_[SMP]), and not those located deeper in villi (a.k.a. Type III_[mucosa]), which are difficult to visualize along the z-axis after whole-mount staining (as performed in the current study).

Comparison with morphometric parameters of the submucosal plexus again revealed a robust correlation between the surface area occupied by ENS ganglia and the proportion of Type I EGCs, in each sample from the distal ileum (Figures 3D–E) and distal colon (Figures 4D–E). Yet, at the difference of the myenteric plexus, where the ganglionic surface area decreases between P1 and P20 in both bowel segments (Figures 1E, 2E), the corresponding surface area in the submucosal plexus was found to generally increase over time (Figures 3D, 4D) – in line with the notion that the submucosal plexus initially develops via inward cell migration from the myenteric plexus (McKeown et al., 2001; Singh et al., 2015; Lasrado et al., 2017). Accordingly, the surface area of the extraganglionic space, which markedly increases over time in the myenteric plexus (Figures 1E, 2E), was here found to slightly decrease in the submucosal plexus from P1 onwards (Figures 3D, 4D). Despite this difference between plexuses, a positive correlation can again be observed when the extraganglionic surface area is analyzed as a function of the proportion of Type III_[SMP] EGCs in each sample from both bowel segments (Figures 3E, 4E). This correlation is however weaker in the distal colon, where fewer Type III_[SMP] EGCs were detected. A more striking regional difference was noted for the surface area occupied by interganglionic submucosal nerves between P1 and P20, resulting in a robust stage-dependent decrease in the distal ileum (Figure 3D) and, conversely, a modest stage-dependent increase in the distal colon (Figure 4D). Nonetheless, a positive correlation with the proportion of Type II EGCs can once again be noted in both bowel segments (Figures 3E, 4E), this correlation being again weaker in the distal colon.

In sum, our immunofluorescence data indicate that EGC diversification based on topo-morphological criteria takes a different path in the submucosal plexus, with Type II EGCs appearing first, followed by Type I and Type III, respectively. Also, our correlation data again support a model whereby EGC diversification is generally impacted by structural changes in the ENS.

To evaluate the specific contribution of SCPs to enteric gliogenesis during early postnatal development, we used the same genetic cell lineage tracing system as previously used to identify SCP derivatives in the bowel (Soret et al., 2020) and the inner ear (Bonnamour et al., 2021), comprising the transgenic Dhh-Cre driver (Jaegle et al., 2003) and the Rosa26^{[FloxedSTOP]YFP} reporter allele (Srinivas et al., 2001). When ENS progenitors directly derived from neural crest cells have just completed their colonization of the developing bowel (i.e., at embryonic day 14.5), this system specifically labels a subset of SOX10+ SCPs on mesentery-associated nerves, and none of the SOX10+ ENS cells already present in the ileum and colon (Supplementary Figure 5). A similar conclusion was previously drawn using the same Cre driver with a different Cre reporter allele, and this study further reported that SCPs start to colonize the developing bowel a few days later, just before birth (Uesaka et al., 2015). As also reported in this prior work focused on enteric neurogenesis (Uesaka et al., 2015), our detailed immunofluorescence analyses revealed that the contribution of YFP+ SCPs to the general pool of ENS cells after birth is much greater in the colon than in the ileum, at all stages examined between P1 and P20. As evidenced in the myenteric plexus at P5 (Figures 5A, B) we further found that this contribution is unevenly distributed, being apparently more extensive in the vicinity of extrinsic nerves that constitute the entry point for SCPs (see inset in Figure 5B). This patchy distribution led to widely distributed values in our global quantitative analysis of SCP-derived EGCs at P20 (Figure 5C). Interestingly, this analysis nonetheless highlighted the fact that the SCP contribution is also generally lower in the myenteric plexus/circular muscle than in the submucosal plexus, in both bowel segments (Figure 5C).

When analyzed as a function of topo-morphological subtypes, our immunofluorescence data then revealed that SCPs can contribute to every EGC subtype from both plexuses, as notably shown in the colon at P20 (Figures 5D–F). However, it is important to note that the final SCP contribution at P20 is clearly biased to particular EGC subtypes in a plexus-specific manner, favoring EGC Type IV in the myenteric plexus / circular muscle (40–53% of all SCP-derived EGCs; Figures 6A, B and Supplementary Figures 6A, B) and Type II in the submucosal plexus (58–67% of all SCP-derived EGCs; Figures 6C, D and Supplementary Figures 6C, D), in both bowel segments. Each bias seems to be established in accordance with the kinetics of appearance of their corresponding EGC subtype between P1 and P20, with the Type IV bias in the myenteric plexus / circular muscle (see Figures 6A, B vs. Figures 1C, 2C) appearing in a delayed manner relative to the Type II bias in the submucosal plexus (see Figures 6C, D vs. Figures 3C, 4C). Type I EGCs in the myenteric plexus / circular muscle from the distal colon initially constitute the highest contribution of SCPs, but this contribution then markedly decreases to ultimately stabilize around 26–32% in all locations at P20 (Figure 6 and Supplementary Figure 6). Type III EGCs constitute the lowest contribution overall, representing only 1–16% of SCP-derived EGCs at P20 (Figure 6 and Supplementary Figure 6). In conclusion, the SCP contribution to the EGC pool is not equal, being regionally different according to both the longitudinal (ileum vs. colon) and the radial (myenteric plexus / muscle vs. submucosal plexus) axes of the bowel.