Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021;12(2):507-545.
doi: 10.1016/j.jcmgh.2021.03.004. Epub 2021 Mar 16.

Lack of Mucosal Cholinergic Innervation Is Associated With Increased Risk of Enterocolitis in Hirschsprung's Disease

Affiliations

Lack of Mucosal Cholinergic Innervation Is Associated With Increased Risk of Enterocolitis in Hirschsprung's Disease

Simone Keck et al. Cell Mol Gastroenterol Hepatol. 2021.

Abstract

Background & aims: Hirschsprung's disease (HSCR) is a congenital intestinal motility disorder defined by the absence of enteric neuronal cells (ganglia) in the distal gut. The development of HSCR-associated enterocolitis remains a life-threatening complication. Absence of enteric ganglia implicates innervation of acetylcholine-secreting (cholinergic) nerve fibers. Cholinergic signals have been reported to control excessive inflammation, but the impact on HSCR-associated enterocolitis is unknown.

Methods: We enrolled 44 HSCR patients in a prospective multicenter study and grouped them according to their degree of colonic mucosal acetylcholinesterase-positive innervation into low-fiber and high-fiber patient groups. The fiber phenotype was correlated with the tissue cytokine profile as well as immune cell frequencies using Luminex analysis and fluorescence-activated cell sorting analysis of colonic tissue and immune cells. Using confocal immunofluorescence microscopy, macrophages were identified in close proximity to nerve fibers and characterized by RNA-seq analysis. Microbial dysbiosis was analyzed in colonic tissue using 16S-rDNA gene sequencing. Finally, the fiber phenotype was correlated with postoperative enterocolitis manifestation.

Results: The presence of mucosal nerve fiber innervation correlated with reduced T-helper 17 cytokines and cell frequencies. In high-fiber tissue, macrophages co-localized with nerve fibers and expressed significantly less interleukin 23 than macrophages from low-fiber tissue. HSCR patients lacking mucosal nerve fibers showed microbial dysbiosis and had a higher incidence of postoperative enterocolitis.

Conclusions: The mucosal fiber phenotype might serve as a prognostic marker for enterocolitis development in HSCR patients and may offer an approach to personalized patient care and new therapeutic options.

Keywords: Cholinergic Nerve Fibers; Enterocolitis; Macrophages; Microbiome; Neuroimmunology; Th17 Cells.

PubMed Disclaimer

Figures

None
Graphical abstract
Figure 1
Figure 1
Distribution of neuronal innervation and fiber scoring of HSCR patients. (A) Schematic representation of neuronal innervation in non-HSCR and HSCR patients. Immunohistochemical anti-β3tubulin staining of ganglionic DC region (non-HSCR) or anti-AChE staining of aganglionic rectum region (HSCR). Scale bar: 50 μm. (B and C) Anti-β3tubulin immunohistochemical staining of Swiss roll cryosections from HSCR fiber-low phenotype (B) and fiber-high phenotype (C). Close-up views show mucosal region from 1, rectum; 2, sigmoid colon; and 3, descending colon. Scale bar: 2 mm; Scale bar close-up views: 20 μm.
Figure 2
Figure 2
Fiber innervations grades for fiber scoring. (A) AChE immunohistochemistry of 5-μm cryosections of distal aganglionic colon showing 4 different mucosal innervation grades. Grades 1 and 2 were grouped into fiber-low and 3 and 4 were grouped into fiber-high HSCR patients. (B) Using brightfield microscopy and CellSens Dimension Software, the innervation grade was quantitatively confirmed in 3 representative patients per group. Scatter dot plots show means ± standard error of the mean. Scale bar: 50 μm
Figure 3
Figure 3
Fiber scoring of distal colon of HSCR patients using tubulin and AChE immunohistochemistry. Tubulin and AChE immunohistochemistry of 5-μm cryosections of distal aganglionic colon of 2 fiber-high (A) and 2 fiber-low (B) HSCR patients. Scale bar: 50 μm.
Figure 4
Figure 4
Cholinergic innervation in distal colon of HSCR patients. (A) Immunofluorescence of epithelial/mucosal region from aganglionic and ganglionic colonic tissue from HSCR and control patients using tubulin (Alexa647, green) and AChE (Cy3, red). DAPI (blue) shows cell nuclei. Encircled areas mark the epithelial crypt/lumen. Scale bar: 50 μm. DC control images: scale bar: 100 μm. DC AG: aganglionic descending colon; DC G: ganglionic descending colon; DC control: ganglionic descending colon from control patients. (B) Positive control: immunofluorescence of myenteric plexus ganglia from ganglionic control tissue. Negative control: immunofluorescence of epithelial/mucosal region from aganglionic fiber-high colonic tissue using secondary antibody controls mouse IgG2b-biotin/SA-Cy3 together with mouse IgG2a-A647. Scale bar: 50 μm. (C) Immunofluorescence of epithelial/mucosal region from aganglionic fiber-low and fiber-high colonic tissue from HSCR patients using tubulin (Alexa647, green) and VAChT (A555, red). DAPI (blue) shows cell nuclei. Encircled areas mark the epithelial crypt/lumen. Positive control: immunofluorescence of myenteric plexus ganglia from ganglionic control tissue. Scale bar: 50 μm. Negative control: immunofluorescence of epithelial/mucosal region from aganglionic fiber-high colonic tissue using secondary antibody controls mouse IgG1-A555 together with mouse IgG2a-A647. Scale bar: 20 μm.
Figure 5
Figure 5
AChE expression in colonic epithelial cells. After tissue digestion colonic epithelial cells were purified from the interphase of 20% and 40% Percoll Gradient. Epithelial cells were immediately lysed, total RNA was isolated, and cDNA was synthesized. Relative gene expression was calculated using the 2-ΔCT method, with β2-microglobulin as the housekeeping gene. Scatter plots with bar show means ± standard error of the mean. Significance was determined by multiple comparison using one-way analysis of variance. Fiber-low, n = 27; fiber-high, n = 13; DC (AG), n = 10; DC (G), n = 11; DC control, n = 6. AG, aganglionic; G, ganglionic.
Figure 6
Figure 6
Immunofluorescence analysis of fiber-high HSCR tissue for neuronal marker. (A) Immunofluorescence (5-μm Swiss roll cryosections) of aganglionic rectum region from a fiber-high patient using tubulin (Alexa647, green), a general neuronal marker, combined with VIP (A555, red), labeling peptidergic neurons; TH (A555, red), labeling dopaminergic and adrenergic neurons; NOS (A555, red), labeling nitrinergic neurons, and S100B (S100 calcium-binding protein B; A555, red), labeling glia cells. DAPI (blue) shows cell nuclei. Images were taken by fluorescence microscopy (original magnification, ×20; scale bar, 100 μm) and processed by Fiji software. Close-up views (white boxes) show mucosal/epithelial region. Negative control: immunofluorescence (5-μm Swiss roll cryosections) of aganglionic rectum region from a fiber-high patient using secondary antibody controls. (B) Immunofluorescence of myenteric plexus ganglia from ganglionic control tissue using the antibody combinations from (A). Scale bar, 50 μm; scale bar for NOS images, 30 μm.
Figure 7
Figure 7
Cytokine profile and immune cell status in colon of HSCR patients. (A) Luminex analysis of full-thickness colonic tissue isolated from different aganglionic (AG) and ganglionic (G) colonic segments of HSCR and control patients. IL23 was determined by enzyme-linked immunosorbent assay. Fiber-low (n = 30); fiber-high (n = 13); AG DC (n = 12); G DC (n = 12); control DC (n = 5). IL23 protein expression was not detectable in 5 fiber-low, 3 fiber-high, 3 aganglionic DC, 5 ganglionic DC, and 3 control DC samples. (B–E) FACS analysis of mononuclear cells isolated from different aganglionic and ganglionic colonic segments of HSCR and control patients. Frequencies of IL17+ Th17 T cells (B), CD127-Foxp3+ Treg cells (C). Fiber-low (n = 30); fiber-high (n = 13); AG DC (n = 10); G DC (n = 10); control DC (n = 5). (D) Ratio Th17/Treg cells. (E) Frequencies of CD11c+ and CD14+ MΦ. Fiber-low (n = 26); fiber-high (n = 13); AG DC (n = 10); muscle (n = 10); G DC (n = 10); control DC (n = 5). (F) Immunofluorescence of CD64+ MΦ from a colonic muscle region (left), a mucosal region from a fiber-low patient (center), and a mucosal region from a fiber-high patient (right). Scale bar: 10 μm and 30 μm (muscle). Images from 3 fiber-high and 3 fiber-low patients were analyzed for MΦ shape. A total of 100–200 CD64+ cells per patient were counted, and the percentage of bipolar cells was calculated (bar graph). Scatter plots (with and without bars) show means ± standard error of the mean. Significance was determined using one-way analysis of variance multiple comparison analysis (A–F) and an unpaired nonparametric Mann-Whitney test (∗P ≤ .05; ∗∗P ≤ .01; ∗∗∗P ≤ .001). Table 10 shows P values. Figure 15 shows the FACS gating strategy.
Figure 8
Figure 8
Immune cell populations in colonic tissue from HSCR and control patients and in vitro T-cell conversion assay. (A) Flow cytometric analysis of mononuclear cells isolated from different aganglionic and ganglionic colonic segments of HSCR and control patients. Frequencies of total cells are shown. IL17+ Th17 T cells were gated in viable CD3+ CD4+ lymphocytes restimulated with PMA/ionomycin. Treg cells were defined as CD127- Foxp3+ cells in unstimulated viable CD3+ CD4+ lymphocytes. Fiber-low (n = 30); fiber-high (n = 13); AG DC (n = 10); G DC (n = 10); control DC (n = 5). CD14+ MΦ were gated in viable HLA+ CD64+ cells. Fiber-low (n = 26); fiber-high (n = 13); AG DC (n = 10); muscle (n = 10); G DC (n = 10); control DC (n = 5). Exact P values indicated in Table 10. (B) Frequencies of NCR+ (Nkp44+) and NCR- (Nkp44-) ILC3 (CD117+) subsets. Fiber-low (n = 17); fiber-high (n = 9); AG DC (n = 6); G DC (n = 6); control DC (n = 3). Figure 15 shows flow cytometric gating strategy. Scatter plots with bar show means ± standard error of the mean. Significance was determined by multiple comparison using one-way analysis of variance (∗P ≤ .05; ∗∗P ≤ .01, ∗∗∗P ≤ .001, ∗∗∗∗P ≤ .0001); exact P values indicated in Table 10. (C) Quantitative RT-PCR analysis of full-thickness colonic tissue or isolated muscle tissue from different aganglionic and ganglionic colonic segments of HSCR and control patients (CX3CR1, circles and CCR2, squares). Significance was determined by using Wilcoxon matched-pairs signed-rank test. Fiber-low (n = 31); fiber-high (n = 13); AG DC (n = 12); G DC (n = 11); control DC (n = 5). (D and E) Representative brightfield microscopic images (original magnification, ×10) of M2 (D) and M1 (E) after 10 days of differentiation. Scale bar: 20 μm. (F and G) Flow cytometric analysis of blood-derived M1 and M2 macrophages from adult healthy donors. Histograms show CD14 (F) and CX3CR1 (G) expression in M1 (green) and M2 (blue) macrophages as well as in unstained controls (filled histograms). Representative analysis out of 3 independent experiments is shown. (H) Th17 and Treg conversion assay using blood-derived M1 and M2 macrophages. Frequencies of Th17 and Treg cells were determined, and Th17/Treg ratio was calculated. One out of 3 independent experiments is shown. Significance was determined using unpaired t test. Figure 15 indicates FACS gating strategy.
Figure 9
Figure 9
CD64+LPMΦ co-localize with AChE+fibers. (A) Immunofluorescence of epithelial region in cryosections of representative fiber-high and fiber-low patients. Anti-AChE (red); anti-CD64 or anti-CD3 (green); DAPI (blue). Scale bar: 10 μm (original magnification, ×63 images) and 30 μm (original magnification, ×20 images). (B) Confocal immunofluorescence image of an intercrypt region from a fiber-high patient. Overlapping pixels (yellow, arrows) were visualized on a Z-projection image. Scale bar: 7 μm. See Supplementary Video 1. (C) Quantification of pixel intensity of CD64+ MΦ and AChE or CD3+ T cells and AChE using a Fiji software macro. Confocal images from rectosigmoid colon regions of fiber-low (n = 4 for CD64; n = 3 for CD3) and fiber-high (n = 5 for CD64; n = 3 for CD3) patients as well as from colonic muscle regions of HSCR patients (n = 2) were taken. The following negative staining controls were used: CD64-A488 or CD3-A488 with biotin + streptavidin-Cy3 and IgG1-A488 + biotin/streptavidin-Cy3 (n = 2). Per patient (dot), 200–300 cells were analyzed. Scatter plots with bar show mean raw integrated density (RawIntDen) ± standard error of the mean. Significance between CD64/AChE and CD3/AChE within a group was determined using an unpaired nonparametric Mann-Whitney test. (D) CD64+ MΦ from 3 fiber-high patients were investigated with respect to bipolar (n = 170) or round/stellate (n = 152) morphologic phenotype. Raw integrated density of co-localized pixels between CD64 and AChE are shown. Significance was determined using an unpaired nonparametric Mann-Whitney test (∗∗∗∗P ≤ .0001).
Figure 10
Figure 10
CD64+LPMΦ associate with neuronal AChE+fibers in HSCR colon. (A) Immunofluorescence of epithelial/mucosal region from aganglionic and ganglionic colonic tissue from HSCR and control patients using CD64 (Alexa488, gray), AChE (Cy3, red), and tubulin (Alexa647, green). DAPI (blue) shows cell nuclei. Scale bar: 50 μm. (B) Close-up view of images from (A). Scale bar: 10 μm. Encircled areas mark the epithelial crypt/lumen. DC AG, aganglionic descending colon; DC G, ganglionic descending colon; DC control, ganglionic descending colon from control patients.
Figure 11
Figure 11
MΦ isolated from fiber-high colonic tissue express diminished IL23 levels. RNA sequencing was performed on sorted viable colonic MΦ and blood-derived M1 and M2 MΦ. (A) PCA plot shows similarity between the different populations based on their first 2 principal components. The top 500 genes, selected by highest row variance, were used. Fiber-low (green, n = 4); fiber-high (blue, n = 5); muscle (red, n = 3); M1 (black, n = 3); M2 (gray, n = 3). (B) Normalized sequence reads of selected macrophage-related genes. Scatter plots with bar show means ± standard error of the mean. (C) Global transcriptional changes between macrophages isolated from fiber-low colonic tissue and fiber-high colonic tissue visualized by a volcano plot. Comparison of gene expression between the groups of samples was performed with the package DESeq2. The Wald test was used to generate P values and log2 fold changes. Genes with an adjusted P value <.05 and log2 fold change greater than 1 (red dots) and less than -1 (blue dots) were regarded as differentially expressed genes. (D) Heat map of differentially expressed genes. Each column represents 1 patient. (E) Normalized sequence reads of selected genes. Scatter plots with bar show means ± standard error of the mean. (F and G) IL23+ CD64+ MΦ (arrows) were assessed by immunofluorescence in 5-μm colonic cryosections. Representative images of mucosal/epithelial region from fiber-low tissue (F) and fiber-high tissue (G) are shown. Scale bar: 30 μm. CD64 (A647, green); IL23p19-PE (red). (H) Quantitative analysis of IL23 expression in macrophages evaluated in images from 4 fiber-high and 4 fiber-low patients. Total of 900–4700 CD64+ cells per patient were analyzed for IL23 expression, and percentage of IL23+ CD64+ cells (% of total macrophages) was calculated. Scatter plots with bar show means ± standard error of the means. Significance was determined using unpaired nonparametric Mann-Whitney test.
Figure 12
Figure 12
Receptors possibly involved in macrophage-neuron crosstalk. (A–E) Heat maps of selected genes possibly involved in macrophage-neuron crosstalk. Normalized reads are shown. RNA sequencing was performed on sorted viable colonic MΦ (CD45+ HLA+ CD64+) from mucosa of fiber-low (n = 4) and fiber-high (n = 5) colonic tissue as well as colonic muscle tissue (n = 3). Heat maps show mean values of each group. (A) Expression of nicotinic (CHRNA) and muscarinic (CHRM) ACh receptors. (B) Expression of serotonergic (HTR) and catecholaminergic (ADRB) receptors. (C) Expression of receptors for neuromedin U (NMUR1), VIP (VIPR1), somatostatin (SSTR2), and calcitonin gene-related peptide (CALCRL/RAMP1). (D) Expression of neurotrophins (BMP and GDF) and neurotrophin receptors (NTRK2 binds brain-derived neurotrophic factor [BDNF] and neurotrophin [NT]-4; GFRA2 and RET bind GDNF and neurturin). (E) Expression of axon-guidance proteins and respective receptors. SEMA bind to plexins or NRP1. UNC-5 represents the receptor for netrin. Receptor proteins ROBO1 and 3 bind slit proteins. Neogenin (Neo) receptor binds to repulsive guidance protein (RGM). Ephrins bind to Eph receptors EPHA and EPHB. (F) Normalized sequence reads of selected genes. Scatter plots with bar show means ± standard error of the mean.
Figure 13
Figure 13
Microbial dysbiosis between fiber-low and fiber-high colonic tissue. (A and B) 16s rRNA FISH analysis was performed on 5-μm cryosections of distal colon (rectum and sigmoid colon) of fiber-low (n = 27) and fiber-high (n = 8) HSCR patients using a bacteria-specific probe (EUB338). (A) Representative images of fiber-low and fiber-high colonic tissue probed with Cy3-labelled EUB338 and NON338 (negative control). Scale bar: 50 μm. (B) Frequencies of translocated bacteria (16sRNA+/DAPI+) were determined using CellProfiler software and are shown as percentage of total DAPI+ cells. (C) qRT-PCR analysis of isolated colonic epithelial cells from different aganglionic (AG) and ganglionic (G) colonic segments of HSCR and control patients. Fiber-low (n = 28); fiber-high (n = 12); AG DC (n = 10); G DC (n = 10); control DC (n = 5). Occludin was not detectable in 2 fiber-low colonic tissue samples. (D) 16S rRNA analysis of colonic tissue from aganglionic (fiber-low, n = 14; fiber-high, n = 11) and ganglionic (DC, n = 4) segments of age-matched HSCR patients. Beta diversity 2-dimensional plot shows patterns of intersample relations using detrended correspondence analysis (DCA).
Figure 14
Figure 14
Colonic microbial dysbiosis in fiber-low colonic tissue from HSCR patients. (A–D) 16S rRNA analysis of colonic tissue from aganglionic (fiber-low, n = 14; fiber-high, n = 11) and ganglionic (DC ganglionic, n = 4) segments of age-matched HSCR patients. (A) Alpha diversity analysis shows effects on observed OTUs and estimated OTU richness by Shannon and Simpson indices. (B) Pie charts show microbial composition in each group on phylum level. (C) Normalized reads of bacterial classes dedicated to respective phyla. Scatter blots with bar show mean values ± standard error of the mean. Significance was determined using one-way analysis of variance multiple comparison analysis. (D) Differential bacterial family analysis of fiber-low versus fiber-high groups. Asterisks mark significantly changed families with adjusted P value <.05 (∗∗∗P = 4.01 × 10-15; ∗∗P = 2.08 × 10-4; ∗P = 0.022). Families are assigned to respective phyla by color code.
Figure 15
Figure 15
Gating strategy for flow cytometric analysis and cell sorting. (A–D) Representative gating strategies for FACS analysis of LP mononuclear cells from HSCR and control patients. (A) Th17 cells were evaluated in viable CD3+ CD4+ lymphocytes restimulated with phorbol 12-myristate 13-acetate (PMA)/ionomycin in the presence of GolgiStop and Brefeldin A. Unstimulated lymphocytes (Golgi-stop/Brefeldin A only) served as negative control. (B) CD127-Foxp3+ Tregs were gated in viable CD3+ CD4+ lymphocytes. (C) Subsets of ILC3 were detected in viable lin-CD127+ CD161+ CRTH2- lymphocytes. ILC3 NCR+ were defined as c-kit+NKp46+ and ILC3 NCR- as c-kit+NKp46-. (D) Macrophage gating strategy for FACS analysis as well as for cell sorting. HLA+ CD64+ macrophages were detected in viable CD45+ leukocytes.

Comment in

References

    1. Demehri F.R., Halaweish I.F., Coran A.G., Teitelbaum D.H. Hirschsprung-associated enterocolitis: pathogenesis, treatment and prevention. Pediatr Surg Int. 2013;29:873–881. - PubMed
    1. Mattar A.F., Coran A.G., Teitelbaum D.H. MUC-2 mucin production in Hirschsprung's disease: possible association with enterocolitis development. J Pediatr Surg. 2003;38:417–421. discussion 421. - PubMed
    1. Imamura A., Puri P., O’Briain D.S., Reen D.J. Mucosal immune defence mechanisms in enterocolitis complicating Hirschsprung's disease. Gut. 1992;33:801–806. - PMC - PubMed
    1. Thomas D.F., Fernie D.S., Malone M., Bayston R., Spitz L. Association between Clostridium difficile and enterocolitis in Hirschsprung’s disease. Lancet. 1982;1:78–79. - PubMed
    1. Bruder E., Terracciano L.M., Passarge E., Meier-Ruge W.A. [Enzyme histochemistry of classical and ultrashort Hirschsprung's disease] Pathologe. 2007;28:105–112. - PubMed

Publication types

MeSH terms