Enteric glia promote intestinal mucosal healing via activation of focal adhesion kinase and release of proEGF

Abstract

Wound healing of the gastrointestinal mucosa is essential for the maintenance of gut homeostasis and integrity. Enteric glial cells play a major role in regulating intestinal barrier function, but their role in mucosal barrier repair remains unknown. The impact of conditional ablation of enteric glia on dextran sodium sulfate (DSS)-induced mucosal damage and on healing of diclofenac-induced mucosal ulcerations was evaluated in vivo in GFAP-HSVtk transgenic mice. A mechanically induced model of intestinal wound healing was developed to study glial-induced epithelial restitution. Glial-epithelial signaling mechanisms were analyzed by using pharmacological inhibitors, neutralizing antibodies, and genetically engineered intestinal epithelial cells. Enteric glial cells were shown to be abundant in the gut mucosa, where they associate closely with intestinal epithelial cells as a distinct cell population from myofibroblasts. Conditional ablation of enteric glia worsened mucosal damage after DSS treatment and significantly delayed mucosal wound healing following diclofenac-induced small intestinal enteropathy in transgenic mice. Enteric glial cells enhanced epithelial restitution and cell spreading in vitro. These enhanced repair processes were reproduced by use of glial-conditioned media, and soluble proEGF was identified as a secreted glial mediator leading to consecutive activation of epidermal growth factor receptor and focal adhesion kinase signaling pathways in intestinal epithelial cells. Our study shows that enteric glia represent a functionally important cellular component of the intestinal epithelial barrier microenvironment and that the disruption of this cellular network attenuates the mucosal healing process.

Fig. 1.

Enteric glial cells (EGC) are a major cellular component of the intestinal epithelial barrier (IEB) microenvironment. A: S-100β-immunoreactive EGC form a dense cellular network surrounding the intestinal crypts (human colonic mucosa; ×100; scale bar 100 μm). B: electron microscopy highlights the close proximity (≅ 1 μm) between EGC ensheathing axons (a) and intestinal epithelial cells (IEC), the basal membrane (bm), as well as myofibroblasts (f) (human colonic mucosa; ×20,000; scale bar 1 μm). C: S-100β-immunoreactive EGC are also present in the periglandular region of the human colonic crypts (×200; scale bar 50 μm). D: immunofluorescence labeling of human colonic mucosa using antibodies directed against S-100β (green) and α-smooth muscle actin (red) shows an absence of colocalization of these two markers. Nuclei were counterstained with 4,6-diamidino-2-phenylindole (DAPI; blue) (×200; scale bar 50 μm). E, left: colonic mucosal damage induced by 7 days of dextran sulfate sodium (DSS) treatment was increased in transgenic (Tg) mice compared with their nontransgenic (NTg) littermates (hematoxylin and eosin-stained colonic specimens; scale bar 200 μm; Veh, vehicle). Right: histological scoring confirms the increase in the severity of the DSS-induced colitis in Tg mice (open bars) compared with NTg mice (solid bars) (n = 5; P = 0.032; Mann-Whitney rank sum test). F, left: histopathological characterization of the effects of EGC ablation on diclofenac sodium salt (DCF)-induced ulceration. Mucosal damage in the 3rd quartile of the small intestine in Tg was similar compared with their NTg littermates 18 h after DCF treatment. In contrast, 48 h after DCF treatment, mucosal damage was significantly lower in NTg mice compared with their Tg littermates (scale bars: 200 μm). Right: whole-mount scoring of ulcer numbers in the 3rd quartile was significantly lower in NTg (solid bars) compared with Tg (open bars) 48 h after DCF treatment (n = 4–8; 2-way ANOVA; *P < 0.05 compared with Tg 48 h; #P < 0.05 compared with their respective control at 18 h).

Fig. 2.

EGC promote intestinal epithelial wound healing. A: wound surface areas of IEC monolayers were significantly reduced after coculture with EGC compared with controls (scale bar 200 μm). B: EGC or EGC-conditioned medium (EGC-CM) significantly reduced wounded surface areas compared with controls (n = 10; P < 0.001; t-test and n = 11; P = 0.002; paired t-test, respectively). A nontransformed EGC line (JUG2) also increased epithelial restitution (n = 3; P < 0.05; t-test) as well as the human colonic fibroblast cell line CCD18Co (n = 6; P < 0.001; t-test). C: EGC and EGC-CM as well as JUG2 induced a significant increase in transepithelial electrical resistance (TER) recovery. Data are presented as a percentage after normalization to controls (n = 7; P = 0.003; t-test, n = 6; P = 0.014; paired t-test and n = 3; P < 0.05; t-test, respectively). *Significantly different.

Fig. 3.

EGC increase IEC spreading. A, left: the cell surface areas of the first 5 circumferential cell layers surrounding the wounded areas were evaluated following staining for zonula occludens-1 (ZO-1) (scale bar 200 μm). Right: EGC and EGC-CM induced a significant increase in IEC spreading in the first 3 and 2 cell layers respectively compared with controls (a and b, respectively; n = 6; P < 0.05; t-test). By contrast, CCD18Co induced a significant increase in IEC spreading only in the first cell layer compared with controls (c; n = 3; P < 0.05; t-test). B: photomicrograph of Caco-2 cells stained with ZO-1 antibody revealed a significant increase in cell surface area induced by EGC (+EGC) compared with controls (−EGC) (scale bar 50 μm). C: quantitative analysis of IEC surface areas showed a significant increase in cell surface area induced by EGC and EGC-CM compared with controls (n = 4; P < 0.001; t-test and n = 3; P = 0.009; paired t-test, respectively). The nontransformed EGC line JUG2 induced a significant increase in IEC cell surface area (n = 4; P < 0.05; t-test). The human colonic fibroblast cell line CCD18Co had no effect on IEC spreading. *Significantly different.

Fig. 4.

EGC increase epithelial restitution via induction of FAK-dependent pathways. A: EGC induced a significant increase in FAK autophosphorylation (Y397P-FAK) in Triton X-soluble (left) and Triton X-insoluble (right) fractions of IEC compared with IEC cultured alone (n = 4; P < 0.05, Mann-Whitney test). B: in the presence of PP2, the effects of EGC-CM on cell spreading were significantly reduced (n = 4; P = 0.008; paired t-test). C: transfection of IEC with FRNK-GFP (FRNK/GFP Caco-2) inhibited EGC effects on IEC spreading, whereas IEC transfection with mock-GFP plasmid (mock/GFP Caco-2) had no effect. Induction of IEC spreading by EGC (arrow) was significantly reduced in FRNK-GFP-transfected cells (arrowheads) compared with mock-GFP-positive cells (n = 5; P = 0.007; t-test; scale bar 50 μm). *Significantly different; NS, not significant.

Fig. 5.

EGC promote epithelial restitution via the regulation of FAK expression. A: EGC induced a significant increase in FAK expression in Triton X-soluble (left) and Triton X-insoluble (right) fractions of IEC compared with controls (n = 4; P < 0.05, Mann-Whitney test). B: real-time quantitative PCR analysis reported a significant increase in FAK mRNA expression in IEC cultured with EGC or EGC-CM (n = 5; P = 0.006, t-test and n = 5; P = 0.008, paired t-test, respectively). C, top: transfection of IEC with FAK short hairpin RNA (shRNA) significantly decreased FAK protein expression compared with IEC transfected with nonspecific target control shRNA or noninfected IEC (control). Middle and bottom: transfection of IEC with FAK shRNA significantly inhibited EGC-CM effects on IEC spreading compared with control shRNA (n = 4; P = 0.013; paired t-test; scale bar 50 μm). D: real-time quantitative PCR analysis showed that in vivo ablation of EGC reduced the expression of FAK mRNA expression in intestinal fragments from Tg mice (Tg) compared with controls (NTg) (n = 7; P = 0.021; t-test). E: real-time quantitative PCR analysis of human colonic mucosal biopsies revealed a significant positive correlation between FAK and S-100β mRNA expression (n = 16; Pearson's correlation coefficient: 0.605; P = 0.013). *Significantly different.

Fig. 6.

EGC promote intestinal epithelial wound healing via EGFR-dependent pathways. A: addition of an EGFR neutralizing antibody (anti-hEGFR, 2 μg/ml) significantly blocked the increase in epithelial restitution induced by EGC-CM (n = 3; P = 0.007; paired t-test). Similarly, PD153035 (1 μM), a specific inhibitor of EGFR kinase, significantly reduced the EGC-CM-induced epithelial restitution (n = 6; P = 0.035; paired t-test). B: EGC-mediated increase in IEC spreading was significantly decreased by the addition of anti-hEGFR (2 μg/ml) (n = 4; P = 0.009; paired t-test; scale bar 50 μm). *Significantly different.

Fig. 7.

EGC secrete proEGF, and EGF neutralizing antibody and metallo-matrix proteinase (MMP) inhibitor decrease the EGC-CM-induced IEC spreading. A: EGF neutralizing antibody (16 μg/ml) and the MMP inhibitor GM6001 (10 μM) reduced the EGC-CM-induced IEC spreading (n = 3; P = 0.011; paired t-test and n = 3; P = 0.038; paired t-test, respectively; scale bar 50 μm). B: human proEGF (100 ng/ml), human EGF (1 ng/ml), and rat EGF (20 ng/ml) increased epithelial restitution (n = 3; P = 0.001, P = 0.01 and P = 0.008; t-test; respectively). C: human proEGF (100 ng/ml), human EGF (1 ng/ml), and rat EGF (20 ng/ml) significantly increased IEC spreading (n = 4; P < 0.001, P = 0.029 and P < 0.001; t-test, respectively). D: PCR studies demonstrating EGF mRNA accumulation in EGC. E: Western blot experiments showing immunoreactive proEGF in EGC-CM (lane 1: recombinant human proEGF: 138–145 kDa; lane 2: EGC-CM). *Significantly different.