Impairment of stromal-epithelial regenerative cross-talk in Hirschsprung disease primes for the progression to enterocolitis

Abstract

Hirschsprung disease (HSCR) is a congenital condition characterized by the improper migration of enteric neural crest cells, leading to aganglionosis most commonly in the rectosigmoid colon. This severe and life-threatening disorder often results in the development of Hirschsprung-associated enterocolitis (HAEC), which can occur either before or after surgical resection of the affected bowel segment. Using colonic tissue from patients with HSCR alongside the well-established endothelin receptor B knockout mouse model, we investigated epithelial regeneration dynamics and stromal-epithelial cross-talk in the distal ganglionic colon, a critical site for HAEC development. In individuals with HSCR but without epithelial damage, the distal ganglionic colon displayed impaired epithelial regeneration and alteration of intestinal stem cell dynamics, characterized by the reduction of leucine-rich repeat-containing G protein-coupled receptor 5 (LGR5⁺) epithelial stem cells. This phenomenon was consistent in the mouse model, where impaired regenerative ability preceded HAEC when epithelial damage occurred on site. Patients with HSCR also exhibited remodeling in stromal cells in this distal ganglionic colon region, with fewer primary sources of Wingless-related integration site (Wnt) signal-releasing stromal cells and the exclusive presence of proinflammatory (matrix metalloproteinase 1⁺) stromal cells. Stromal cells from the HSCR distal ganglionic colon failed to sustain the growth of colonic organoids. However, ibuprofen suppressed the proinflammatory stromal cells, leading to effective restoration of epithelial organoid growth. These observations underscore the crucial role of impaired stromal-epithelial cross-talk in HSCR and the pathogenesis of HAEC and suggest potential therapeutic targets for the prevention or treatment of the condition.

Graphic summary

Fig. 1. Study Design and analysis overview in Single-Cell RNA Sequencing

(A) Overall study design with flowchart of single-cell analysis. (B) Markers of cell type specific genes used for cell annotation shown as fraction of expressing cells (circle size) and mean expression (color) of gene markers (rows) across major cell types (columns). (C) Illustrations showing the patients corresponding to the major cell types used for single cell RNA sequencing. (D) Uniform Manifold Approximation and Projection (UMAP) embedding of single cell transcriptomes of cells from eight major cell types.

Fig. 2. Impaired intestinal regeneration occurs in the colons of HSCR individual even in the absence of epithelial injury.

(A) UMAP plot visualizing clusters of epithelial cell subtypes. (B) Dot plot of gene expression of markers in inflammation and cell cycle, with color indicating the fold change (log2 scale) comparing HSCR to control, and size indicating the adjusted p-value from Wilcoxon rank sum test. (C) Pie chart visualizing the proportion of cycling cells of different epithelial cell subtypes from different groups. (D) Representative H&E-stained histomicrographs of the distal colon and rectum from control/HSCR samples. (E) Immunofluorescence micrograph of Ki67 expression of the distal colon and rectum from control/HSCR samples, with DAPI nuclei counterstaining and quantification of Ki67 positive cells per crypt in all colonic regions. (F) Trajectory of epithelial cell subtypes in control and HSCR inferred by Monocle2. (G) Epithelial differential pseudotime in control and HSCR inferred by Monocle2. (H) Dynamic time warping of pseudotime trajectories of control and HSCR allows for comparison of the dynamics of epithelial differentiation process along a common axis. Hierarchical clustering of kinetic curves for dynamically regulated genes that vary significantly between control and HSCR epithelial trajectories. (I) MUC2 gene expression across warped pseudotime in control and HSCR samples. Blue line as control and red line as HSCR. (J) Immunofluorescence micrograph of MUC2 expression of the distal colon and rectum from control/HSCR samples, with DAPI nuclei counterstaining and quantification. (K) Histological injury grading of the proximal colon, distal colon and rectum from 2-week-old wildtype and Ednrb^-/- mice. (L) Representative immunofluorescence staining for the proliferation marker Ki67 in the proximal colon, distal colon, and rectum from wildtype and Ednrb^-/- mice at 2 weeks, with DAPI nuclei counterstaining and quantification of Ki67+ cells per crypt. Scale bars are shown in all the images. HSCR patients n=6 and control patients n=3. Wildtype mice n=6 and Ednrb^-/- mice n=6. Experiments were repeated independently 3 times, with similar results. Each dots represented average value of each individual. Data are presented as mean ± SD and compared using one-way ANOVA with post-hoc tests. The primary p-value is indicated directly on the graph.

Fig. 3. Alterations of intestinal stem cell dynamic occur in the distal colons of HSCR patients.

(A) UMAP plot visualizing intestine stem cell populations. (B) Dot plot visualizing the average expression of ISC function-related markers. (C) PHATE plot visualizing intestine stem cell populations and overlaid with RNA velocity streams (arrows). (D) PHATE embedding of stem cell clusters in the distal and rectum of control and HSCR patients. (E) Representative micrographs of fluorescence staining for Clu (Green), Ki67 (Red) in the distal and rectum of control and HSCR patients, with DAPI nuclei counterstaining. (F) Quantification of Olfm4 mRNA expression in the distal colon and rectum from control and HSCR samples. (G) Quantification of ASCL2 mRNA expression in the distal colon and rectum from control and HSCR samples. (H) Representative fluorescence in situ hybridization (ISH) staining for Lgr5 in the proximal colon, distal colon, and rectum from 2-week-old wildtype and Ednrb^-/- mice, with DAPI nuclei counterstaining and representative photomicrographs of organoids derived from proximal colon, distal colon and rectum of 2-week-old wildtype and Ednrb^-/- mice after 5 days in culture. (I) Quantification of Lgr5 positive ISH cells per crypt and Lgr5 mRNA expression in the proximal colon, distal colon and rectum from 2-week-old wildtype and Ednrb^-/- mice. (J) Quantification of organoids number in each image and percentage of increase in surface area from day 2 to day 5 cultures. Quantification of surface area increase comparing day 5 to day 2 of all groups from 2-week-old wildtype and Ednrb^-/- mice. Scale bars are shown in all the images. HSCR patients n=6 and control patients n=3. Wildtype mice n=6 and Ednrb^-/- mice n=6. Experiments were repeated independently 3 times, with similar results. Each dots represented average value of each individual. Data are presented as mean ± SD and compared using one-way ANOVA with post-hoc tests. The primary p-value is indicated directly on the graph.

Fig. 4. Impairment of intestinal regeneration in the colon of HAEC patients and mice with HAEC.

(A) Schematic representation of HAEC pathology/diagnosis, as determined by histological assessment. (B) Representative neuron marker TUJ1 stained, H&E-stained, Ki67-stained and Lgr5-stained histomicrographs in the proximal, distal and aganglionic rectum from HAEC patients. (C) Histological injury grading scores, quantification of Ki67 positive cells per crypt and quantification of Lgr5 positive cells per crypt of the proximal, distal and rectum from HAEC patients. (D) Representative images of mouse morphology and dissected whole gastrointestinal tract of Ednrb^-/- mouse at 3-weeks of age. Black box: proximal colon, red box: dilated distal ganglionic colon, and green box: rectum; and their corresponding segments in wildtype colon. (E) Histological injury grading scores and pro-inflammation maker TNFa mRNA expression of the proximal, distal and rectum of Ednrb^-/- mouse at 3-weeks of age. (F) Immunofluorescence staining for the proliferation marker Ki67, in situ hybridization (ISH) staining for Lgr5, intestinal stem cell marker Lgr5 in the proximal colon, distal colon and rectum from wildtype and Ednrb^-/- mice at 3-week-old, with DAPI nuclei counterstaining, and representative photomicrographs of organoids derived from proximal colon, distal colon and rectum of wildtype and Ednrb^-/- mice after 5 days in culture. (G) Quantification of Ki67+ cells per crypt in the proximal colon, distal colon and rectum from wildtype and Ednrb^-/- mice. (H) Quantification of Lgr5 ISC positive cells per crypt and quantification of Lgr5-GFP+ cells per crypt in the proximal colon, distal colon and rectum. (I) Quantification of organoids number in each image and percentage of increase in surface area from day 2 to day 5 cultures of all groups from wildtype and Ednrb^-/- mice. Scale bars are shown in all the images. HSCR patients n=6 and control patients n=3. Wildtype mice n=6 and Ednrb^-/- mice n=6. Experiments were repeated independently 3 times, with similar results. Each dots represented average value of each individual. Data are presented as mean ± SD and compared using one-way ANOVA with post-hoc tests. The primary p-value is indicated directly on the graph.

Fig. 5. Wnt-releasing Stromal 2 cell population was reduced in the colon of HSCR patients.

(A) Dot plot visualizing the interaction strength between ligands expressed by other cell types with receptors expressed by epithelial cells. (B) UMAP plot visualizing clusters of stromal cell subtypes. (C) Heatmap visualizing gene expression of stromal cell subtype specific markers. (D) Proportion of stromal cell subtypes in control and HSCR samples. (E) Representative histomicrographs of Stromal 2 cell marker F3 of the distal colon and rectum from control and HSCR samples. (F) Protein expression of Stromal 2 cell marker BMP4 of control and HSCR samples (n=3). (G) Dot plot of WNT signaling gene expression in different stromal cell subtypes. (H) Dot plot of ligand-receptor pairs in WNT signaling between stromal cells and epithelial cells. (I) GSEA analysis shows WNT signalling is impaired in stem cells/TAs in HSCR group. (J) Quantification of tissue expression of Wnt pathway genes in the distal and rectum from control and HSCR. (K) Representative micrographs of fluorescence ISH staining for Wnt5a in the proximal, distal, and rectum from 2 weeks old wildtype and Ednrb^-/- mice, with DAPI nuclei counterstaining and quantification of ISH of Wnt5a positive dots per cells. (L) Representative micrographs of fluorescence ISH staining for Wnt5a and quantification of ISH of Wnt5a positive dots per cell in the proximal colon, distal colon and rectum from 3 weeks old wildtype and Ednrb^-/- mice. Scale bars are shown in all the images. HSCR patients n=6 and control patients n=3. Wildtype mice n=6 and Ednrb^-/- mice n=6. Experiments were repeated independently 3 times, with similar results. Each dots represented average value of each individual. Data are presented as mean ± SD and compared using nonparametric unpaired Student’s t-tests or one-way ANOVA with post-hoc tests as appropriate. The primary p-value is indicated directly on the graph.

Fig. 6. Pro-inflammatory Stromal 4 cells were highly expressed in the distal colon of HSCR patients

(A) Heatmap visualizing the enrichment of specific stromal cell subtype among the different groups studied, estimated by R_o/e. (B) Violin plots showing activity of signaling pathway inferred by UCell in different stromal cell subtypes. (C) Bar chart showing the percentage of Stromal 4 cells to total number of stromal cells in different groups. (D) Violin plots showing MMP1 expression in different stromal cell subtypes. (E) Representative immuno-staining micrographs of Pdgfrα (red) and MMP1 (green) in the colon from non-HSCR control and HSCR patients, with DAPI nuclei counterstaining and (F) quantification of MMP1 positive cells per crypt of control and HSCR patients. (G) Representative immuno-staining micrographs of Pdgfrα (red) and MMP1 (green) in the biopsy of obtained from the distal ganglionic colon and aganglionic rectum during pull-through surgery from HSCR patients following colostomy (DAPI nuclei counterstaining). (H) Representative photomicrographs of stromal cells derived from distal colon from non-HSCR control and HSCR patients at day 1 and day 5 of culture. (I) Growth curve of stromal cells derived from distal colon of non-HSCR control and HSCR patients, measured by cell number quantified daily for 5 days. (J) Immunofluorescence micrograph of F4+ Stromal 2 cells and MMP1+ Stromal 4 cells in the stromal cell culture derived from distal colon of non-HSCR and HSCR patients, with DAPI nuclei counterstaining. (K) Mean fluorescence intensity (MFI) of MMP1 in the stromal cell culture derived from distal colon of non-HSCR and HSCR patients. (L) Representative immunofluorescence micrographs of stromal cells 3 marker RSPO3 (Green) and KCNN3 (Red) in the distal and rectum from HSCR patient and non-HSCR control patient, with DAPI nuclei counterstaining. Scale bars are shown in all the images. HSCR patients n=6 and control patients n=3. Experiments were repeated independently 3 times, with similar results. Each dots represented average value of each individual. Data are presented as mean ± SD and compared using nonparametric unpaired Student’s t-tests or one-way ANOVA with post-hoc tests as appropriate. The primary p-value is indicated directly on the graph.

Fig. 7. The failure of Stromal cells from the distal colon of HSCR patients to support the growth of epithelial organoids can be rescued by inhibiting Stromal 4 cells

(A) Representative photomicrographs of distal colon epithelial organoids derived from non-HSCR control and HSCR patients, cultured in media enriched with recombinant “WENR” growth factors Wnt3a, epidermal growth factor (EGF), Noggin, and Rspo1. (B) Representative photomicrographs of EDU and TUNEL staining for distal colon epithelial organoids derived from non-HSCR control and HSCR patients cultured with WENR growth factors. (C) Representative photomicrographs of distal colon epithelial organoids derived from non-HSCR control and HSCR patients, co-cultured with the distal colon stromal cell derived from non-HSCR control and HSCR patients in the absence of WENR growth factors. (D) Quantification of organoids number/ image in each culturing conditions. (E) Drug2cell analysis prediction of specific drugs inhibiting Stromal 4 cells (Wilcoxon rank sum test, q-value: FDR-adjusted p-value). (F) Immunofluorescence micrograph of MMP1+ Stromal 4 cells derived from HSCR patient’s distal colon treated with or without ibuprofen (Ibu, 100µg/mL), with DAPI nuclei counterstaining. (G) Quantification percentage of MMP1 positive cells over total number of HSCR distal colon stromal cells treated with or without ibuprofen. (H) Mean fluorescence intensity (MFI) of MMP1 in the HSCR distal colon stromal cell treated with or without ibuprofen. (I) Cytokines released from stromal cells from HSCR proximal colon, distal colon treated with or without ibuprofen as well as the non-HSCR control distal colon. (J) Representative photomicrographs of non-HSCR control epithelial organoids co-cultured with HSCR distal colon stromal cells pretreated with or without ibuprofen in the absence of WENR growth factors and quantification of organoids number in each culturing condition. (K) Representative photomicrographs of HSCR organoids co-cultured with HSCR distal colon stromal cells pretreated with or without ibuprofen in the absence of WENR growth factors and quantification of organoids number in each culturing condition. Scale bars are shown in all the images. HSCR patients n=6 and control patients n=3. Experiments were repeated independently 3 times, with similar results. Each dots represented average value of each individual. Data are presented as mean ± SD and compared using nonparametric unpaired Student’s t-tests or one-way ANOVA with post-hoc tests as appropriate. The primary p-value is indicated directly on the graph.