Microbiota controls the homeostasis of glial cells in the gut lamina propria

Abstract

The intrinsic neural networks of the gastrointestinal tract are derived from dedicated neural crest progenitors that colonize the gut during embryogenesis and give rise to enteric neurons and glia. Here, we study how an essential subpopulation of enteric glial cells (EGCs) residing within the intestinal mucosa is integrated into the dynamic microenvironment of the alimentary tract. We find that under normal conditions colonization of the lamina propria by glial cells commences during early postnatal stages but reaches steady-state levels after weaning. By employing genetic lineage tracing, we provide evidence that in adult mice the network of mucosal EGCs is continuously renewed by incoming glial cells originating in the plexi of the gut wall. Finally, we demonstrate that both the initial colonization and homeostasis of glial cells in the intestinal mucosa are regulated by the indigenous gut microbiota.

Figure 1

The Network of mEGCs Develops after Birth (A) S100β immunostaining of a vibratome cross-section from the ileum of an adult wild-type mouse. In addition to the myenteric (mp) and submucosal plexus (smp), EGCs (arrows) are also found within the lamina propria around the crypts and within villi. (B) Highly branched GFP⁺ glial cells (arrows) within the mucosa in the ileum of Sox10::Cre;MADM^GR/RG mice. (C–E) S100β immunostaining of cryosections of the mucosa of P0 (C), P10 (D), and adult (E) mice. Arrows in (D) and (E) point to mEGCs. (F) Quantification of glia⁺ VC units (GFP and S100β antibodies on sections from Sox10::Cre;R26REYFP mice) demonstrating that the network of mEGCs develops postnatally. Data are represented as mean of all glia+ VC units ± SEM. One-way ANOVA, p value < 0.0001: Tukey post hoc test showed that comparison of E16.5 to P0, P10 to P18–P27, and P27–P38 to Adult was not significant (NS). However, comparison of E16.5 to P10, P18–P27, P27–P38, and Adult; P0 to P10, P18–P27, P27–P38, and Adult; P10 to P27–P38 and Adult; and P18–P27 to Adult was significant (^∗∗∗∗p < 0.0001). Comparison of P18–P27 to P27–P38 was also significant (^∗∗∗p = 0.0003). The F (DFn, DFd) value is 227.6 (5, 12). Scale bars: 100 μm (A and B); 50 μm (C–E).

Figure 2

Inducible Lineage Tracing of mEGCs Distribution of confetti⁺ cells in the ileum of tamoxifen-treated Sox10::CreER^T2;R26RConfetti mice at T0 (A), T15 (B), and T90 (C). Boxed areas in (A)–(C) correspond to panels (A′)–(C′), respectively. (D and D′) Confetti⁺ cells in a flat-mount preparation of myenteric ganglia from Sox10::CreER^T2;R26RConfetti mice at T0. Arrows in (D) and (D′) indicate Sox10⁺Confetti⁺ glial cells. Asterisks in (D) indicate the position of confetti-negative enteric neurons. (E) Quantification of VC units with confetti⁺ glial cells at T0, T15, and T90. Data represented as mean ± SEM. Significant differences between the ages have been obtained with the two-way ANOVA, p < 0.0001, Tukey post hoc test, ^∗∗∗∗p < 0.0001 (T0–T15;T0–T90), p = 0.7173 (T15–T90). Although VC units with monochromatic glia showed no significant change in number (p > 0.1), VC units with polychromatic glia showed significant differences: ^∗∗∗∗p < 0.0001 (T0–T15; T0–T90), p = 0.6222 (T15–T90). The F(DFn, DFd) and p values for Factor 1: T0 Vs T15 Vs T90 is 29.94 (2, 18) with p = 0.08; Factor 2: polychromatic Vs monochromatic is 3.36 (1, 18) with ^∗∗∗∗p < 0.0001 and interaction of Factor 1 with Factor 2 is 8.104 (2, 18) with ^∗∗p = 0.0031. cm, circular muscle layer; lm, longitudinal muscle layer; mp, myenteric plexus. Scale bars: 100 μm (A–C); 50 μm (A′–C′); 20 μm (D and D′).

Figure 3

Formation of the mEGC Network Depends on Microbiota Cross-sections from the ileum of CONV (A), GF (B), and CONV-D (C) mice immunostained for S100β. Arrows indicate glial cells. (D) Quantification of S100β⁺ cells in the three conditions. One-way ANOVA, p value = 0.0093, Tukey post hoc test, ^∗p = 0.0128 (CONV versus GF), ^∗p = 0.0260 (GF versus CONV-D). Glial populations were not significantly different between CONV and CONV-D mice. The F(DFn, DFd) value is 7.086 (2, 12). (E) Percentage of VC area positive for S100β is represented in all three conditions. One-way ANOVA, p value < 0.05, Tukey post hoc test, ^∗∗p = 0.002 (CONV versus GF), ^∗p = 0.018 (GF versus CONV-D). The F(DFn, DFd) value is 10.70 (2, 12). Note that the glial cell number and distribution and the area of VC units positive for S100β was similar between CONV and CONV-D animals, but significantly different from GF mice. Scale bars: 100 μm (A–C). CONV = conventional; GF = germ-free; CONV-D = conventionalized.

Figure 4

Microbiota Controls the Continuous Supply of Glial Cells to the Intestinal Mucosa (A and B) Sections of the ileum of control (A) and antibiotic-treated (B) wild-type adult mice. Arrows in (A) indicate glial cells. (C–F) Distribution of confetti⁺ glial cells in the ileum of tamoxifen-treated 8- to 12-week-old Sox10::CreER^T2;R26RConfetti mice at T0 (C and D) and T15 (E and F) in the absence (C and E) or presence (D and F) of antibiotics. Arrows indicate confetti⁺ glial cells in the lamina propria. (G) The average number of confetti⁺ cells in VC units at T15 is reduced in antibiotic-treated animals. Data is represented as mean ± SEM. The distribution of the number of labeled glial cells per VC unit has been plotted using the non-linear paradigm, representing a different curve for each dataset as the best fit with a ^∗∗∗p value = 0.0004. Using the two-way ANOVA and Sidak’s multiple comparisons test, significant differences were observed in the 0 and 2 cells per VC unit categories with ^∗∗p values = 0.0014 and 0.0081, respectively. The F(DFn, DFd) and p values for Factor 1: With Vs Without antibiotic is 0 (1, 36) p > 0.99; Factor 2: number of glial cells in VC units is 21.47 (5, 36) ^∗∗∗∗p < 0.0001 and interaction of Factor 1 with Factor 2 is 6.353 (5, 36) with ^∗∗∗p = 0.0003. Scale bars: 100 μm (A–F).