Characterization of Neurochemical Signature Alterations in the Enteric Nervous System in Autoimmune Encephalomyelitis

Abstract

To date, it has remained unclear whether gastrointestinal symptoms, which are frequently observed in patients with multiple sclerosis (MS), are accompanied by pathology of the enteric nervous system (ENS). Here, the neurotransmitter signature of ENS neurons and morphological alterations of interstitial cells of Cajal (ICCs) were studied in patients with MS and mice with experimental autoimmune encephalomyelitis (EAE), which is an animal model of MS. Immunohistochemical analysis was performed on colonic whole mounts from mice with EAE and on paraffin-embedded sections of intestinal tissue from patients with MS. Antibodies against neurotransmitters or their enzymes (including vasoactive intestinal peptide (VIP), neuronal nitric oxide synthase (nNOS), and choline acetyltransferase (ChAT)) were used in conjunction with pan-neuronal markers. In addition, the presence of anoctamin 1 (ANO1)-expressing ICCs was studied. ENS changes were observed in the myenteric plexus, but they were absent in the submucosal plexus of both EAE mice and patients with MS. There was a significant decrease in the percentage of ChAT-positive neurons in EAE mice as opposed to a trend toward an increase in patients with MS. Moreover, while ANO1 expression was decreased in EAE mice, patients with MS displayed a significant increase. Although additional studies are necessary to accomplish an in-depth characterization of ENS alterations in MS, our results imply that such alterations exist and may reveal novel insights into the pathophysiology of MS.

1. Introduction

Multiple sclerosis (MS) is an inflammatory disease that causes demyelination of the central nervous system (CNS) [1]. Worldwide, 2.8 million people are affected by MS, which is the leading cause of neurological disability in young adults [2]. Depending on the location and intensity of inflammation in the CNS, patients with MS experience a variety of symptoms, with impairment of their visual, sensory, motor, cognitive, or autonomic functions [3]. About two-thirds of patients with MS experience gastrointestinal symptoms such as constipation, dyspepsia, dysphagia, and fecal incontinence [4,5].

While some studies suggest that these gastrointestinal symptoms derive from pathologies within the CNS—e.g., gut dysfunction might be caused by spinal cord injury [6]—other studies have shown that many patients with MS display bowel symptoms even years before a demyelinating event in the CNS [7].

The enteric nervous system (ENS) is often referred to as the little brain [8] not only because the human ENS contains between 200 and 600 million neurons [9] but also because of its unique ability to function autonomously after all connections to the CNS have been severed or even after removal of the gut from the body [10,11]. The neurons of the ENS are organized into two major plexuses: the myenteric plexus (MP), which is located between the longitudinal and circular muscle, and the submucosal plexus (SP), which is located within the submucosal connective tissue layer [12]. Intrinsic primary afferent neurons (IPANs) detect mechanical or chemical stimuli (e.g., distention), and communicate with motor neurons (MNs), either directly or via interneurons. The MNs of the MP affect intestinal motility [9]. Inhibitory MNs express enzymes to produce the neurotransmitters nitric oxide (NO) and vasoactive intestinal peptide (VIP), whereas in excitatory MNs, choline acetyltransferase (ChAT) synthesizes acetylcholine (ACh) [13,14,15,16]. In the SP, MNs control secretion and vasodilatation [17,18] and are either cholinergic or non-cholinergic [19]. In addition, interstitial cells of Cajal (ICCs) that express the chloride channel anoctamin 1 (ANO1) are the pacemakers of the ENS [20,21].

The number of ENS cells and their attributes, such as neurotransmitter signatures, have been shown to change as a result of many diseases of the gut, including achalasia [22] and inflammatory bowel diseases, e.g., ulcerative colitis [23,24], and also due to aging [25,26] and in the context of neurodegenerative disorders of the CNS such as Parkinson’s disease [27,28].

Wunsch et al. [29] recently conducted a series of experiments in C57BL/6J mice with experimental autoimmune encephalomyelitis (EAE), which is a common mouse model of MS. Mice were immunized with the fusion protein MP4, which consists of myelin basic protein (MBP) and the three hydrophilic domains of proteolipid protein (PLP) [30,31]. The experiments revealed a loss of myenteric axons before neurological signs of EAE and CNS pathology as well as increased expression of glial fibrillary acidic protein (GFAP) in the ENS—a sign of gliosis [29]. In addition, a decrease in gastrointestinal transit time was reported, and ENS pathology in MP4-induced EAE was antibody-mediated [29]. Spear et al. [32] used EAE induced by myelin oligodendrocyte glycoprotein (MOG_35–55), proteolipid protein (PLP_139–151), or spinal cord homogenate to study ENS pathology. They observed gastrointestinal dysmotility that was accompanied by decreased GFAP levels and also suggested a B cell-dependent mechanism [32]. Interestingly, Spear et al. did not observe neuronal loss in the ENS, which contrasted with the data of Wunsch et al. Overall, these findings suggest a link between the ENS, the pathogenesis of EAE, and eventually MS. Yet, it remains unclear whether the cell types of the ENS show alterations in their numbers and phenotypes, including their neurotransmitter signature. In addition, the mechanisms by which any such alterations occur need to be elucidated in greater detail. The aim of this study was to further characterize ENS pathology in EAE mice and patients with MS with a particular focus on neurotransmitter signatures and ICCs.

2. Materials and Methods

2.1. EAE Induction

For the analysis of neurotransmitter signatures, female C57BL/6J mice (n = 15) were obtained from Envigo (Horst, Netherlands) and kept under specific pathogen-free (SPF) conditions at the animal facility of the Center for Dental, Oral and Jaw Health (Zentrum für Zahn-, Mund- und Kiefergesundheit), University of Würzburg, Würzburg, Germany. For the analysis of ICCs, female C57BL/6J mice (n = 6) were obtained from Charles River (Sulzfeld, Germany) and kept under SPF conditions at the House for Experimental Therapy 4 (Haus für Experimentelle Therapie 4, “Maushaus”) at the University Hospital Bonn, Bonn, Germany. Female mice were used for all experiments to account for the gender bias in MS [2] and because female mice could be housed more easily in groups compared to male mice. All mice were between seven and nine weeks of age at the time point of EAE induction.

In preparation for EAE induction, complete Freund’s adjuvant (CFA) was obtained by combining nine parts paraffin oil (Sigma-Aldrich, St. Louis, MI, USA) and one part mannide monooleate (Sigma-Aldrich) with 5 mg/mL Mycobacterium tuberculosis H37 Ra (BD Difco Laboratories, Franklin Lakes, NJ, USA).

EAE was induced by MP4, which has previously been shown to induce ENS pathology in the MP [29]. All EAE mice (n = 13; n = 10 for the analysis of neurotransmitters and n = 3 for the analysis of ICCs) were immunized subcutaneously in both flanks with a total dose of 200 µg MP4 (Alexion Pharmaceuticals, New Haven, CT, USA) emulsified in 200 µL CFA. On the day of immunization and 48 h later, mice received an intraperitoneal injection of 200 ng pertussis toxin from List Biological Laboratories (Campbell, CA, USA; experiments in Würzburg) or 120 ng pertussis toxin from Hooke Laboratories (Lawrence, MA, USA; experiments in Bonn), respectively, in 250 µL sterile phosphate-buffered saline (PBS; Thermo Fisher, Waltham, MA, USA).

For the analysis of neurotransmitter signatures, an additional cohort of mice (n = 5) was immunized with 200 μg hen egg lysozyme (HEL) (Sigma-Aldrich) in CFA as a control group, following the same protocol as for MP4 immunization, while for ICC analysis, the control group (n = 3) remained non-immunized. In order to assess the neurotransmitter signature of the ENS over the course of the disease, two groups of EAE mice were analyzed: (i) five mice were analyzed 12 days after immunization, before the clinical onset of disease; (ii) five mice were euthanized during late-stage EAE, i.e., 25 days after the onset of clinical symptoms, corresponding to 40 days after immunization. To assess alterations of ICCs, only late-stage EAE mice were analyzed.

All animal experiments were performed according to protocols that were approved by the Government of Lower Franconia (Regierung von Unterfranken; file nos. 55.2.2531.01-114/13 and 55.2.2532.01-91/14) or by the State Office for Nature, Environment and Consumer Protection North Rhine-Westphalia (LANUV; file no. 81-02.04.2021.A146) and complied with the German Law on the Protection of Animals, the “Principles of Laboratory Animal Care” (NIH publication no. 86-23, revised 1985), and the ARRIVE guidelines for reporting animal research [33].

2.2. Whole-Mount Staining

For the analysis of neurotransmitter signatures, one approximately 1 cm long segment was immediately removed from the distal colon of each mouse, cleared of fecal matter, and a peripheral venous catheter was placed into the colonic lumen. After washing with phosphate-buffered saline (PBS) (Roth, Karlsruhe, Germany and Merck, Darmstadt, Germany), the colon was filled with 4% paraformaldehyde (PFA) solution (Roth), the ends tied, and then placed into 4% PFA for 30 min. Subsequently, the colon was cut open longitudinally, and the segments were then subjected to whole mount preparation for immunohistochemical (IHC) staining of the ENS. The whole mounts were preincubated at room temperature for 2 h with 0.05 M tris-buffered saline (TBS; pH 7.4) (Roth) containing 1% bovine serum albumin (BSA) (Roth), 0.5% Triton X-100 (Merck Millipore, Burlington, MA, USA), 0.05% thimerosal (Roth), and 5% normal donkey serum (Dianova, Hamburg, Germany). After rinsing the whole mounts in TBS, they were incubated at 4 °C for 96 h with the primary antibodies diluted in TBS containing 1% BSA, 0.5% Triton X, and 0.05% thimerosal. The four primary antibodies were an Alexa Fluor 555-conjugated anti-HuC/D pan-neuronal antibody (HU) (Thermo Fisher; dilution 1:200; labeled with an antibody-labeling kit from Thermo Fisher), and antibodies against ChAT (Sigma-Aldrich; 1:40), neuronal nitric oxide synthase (nNOS) (University of Graz, Austria; 1:1000), and VIP (Biomedicals, Augst, Switzerland; 1:200).

Specificity of the nNOS antibody was confirmed by preabsorption control experiments in previous studies at the Institute of Anatomy and Cell Biology in Erlangen [34,35].

Afterwards, the whole mounts were washed in TBS and incubated at room temperature for 4 h with the following three secondary antibodies: donkey anti-guinea pig DyLight 488 (Dianova; dilution 1:500), donkey anti-goat Alexa Fluor 647 (Thermo Fisher; 1:1000), and donkey anti-rabbit DyLight 405 (Dianova; 1:200). After washing with TBS, all whole mounts were mounted with TBS-glycerol (Roth; 1:1; pH 8.6).

For ANO1 staining, colonic tissue was rinsed for clearance of fecal matter and threaded onto a glass rod. The remaining mesentery was removed, and tissue was incised carefully lengthwise using tweezers. Using a cotton swab soaked in buffer, longitudinal muscle layer and attached myenteric plexus (LMMP) were separated from the remaining tissue by carefully rubbing the cotton swab around the tissue. LMMP was cut into approximately 1 × 0.5 cm² pieces and transferred into a 96-well plate. Tissue was blocked with 10% donkey serum (Biozol, Eching, Germany) in a buffer consisting of TBS + 0.5% Triton-X 100 (MP Biomedicals, Irvine, CA, USA), 1% BSA (Merck) and 0.02% sodium azide (Merck) (whole-mount buffer) for 5 h at room temperature on a plate shaker.

Primary antibodies against GFAP (chicken polyclonal to GFAP, Abcam, Cambridge, UK; 1:400) and ANO1 (rabbit polyclonal to TMEN16a, Abcam; 1:200) were diluted in whole-mount buffer and applied for 72 h at 4 °C on a plate shaker. The tissue was washed in TBS + 0.05% Tween 20 (VWR, Radnor, PA, USA) (TBST) before the secondary antibodies donkey anti-chicken Cy2 (Jackson ImmunoResearch, West Grove, PA, USA, 1:200) and donkey anti-rabbit Cy3 (Jackson ImmunoResearch, 1:200) were applied and incubated for 2 h at room temperature on a plate shaker. After another washing step using TBST, DAPI (Roche, Basel, Switzerland) was diluted 1:1000 in TBST and applied for 15 min. After rinsing in TBS, whole mounts were mounted using a 1:1 mixture of TBS and glycerol (Merck).

2.3. Human Gut Tissue Samples

Formalin-fixed paraffin-embedded blocks of human colon were obtained from the Department of Pathology, University Hospital Erlangen (Prof. Dr. med. Arndt Hartmann), the Community Pathology Practice (Gemeinschaftspraxis Pathologie) in Amberg (Dr. med. Lothar Mandl), and the Institute of Pathology, Caritas-Krankenhaus Bad Mergentheim (PD Dr. med. Matthias Woenckhaus), all in Germany. In total, samples from 19 patients with MS and 14 control patients without MS (CTRLs), matched for age, sex, and gastrointestinal comorbidities, were collected and subsequently screened for the presence of ganglia and absence of tumor or inflammatory cells. Only sections from patients with MS and CTRL patients that met these criteria were used for subsequent analysis, leading to the final inclusion of seven patients with MS and all CTRL patients.

Table 1. Composition of the Cohort of Patients with Multiple Sclerosis.

Patient ID	Sex	Age (Years)	Reason for Surgery	Tissue Source
1	F	25	I	Appendix
2	F	46	I	Neoileocolic anastomosis
3	M	67	C	Colon
4	M	68	C	Ileum and rectum
5	F	37	O	Ileum and colon
6	F	63	O	Colon
7	F	65	O	Colon

C, cancer; F, female; I, inflammatory disease; ID, identification number; M, male; O, other.

and

Table 2. Composition of the Non-Multiple Sclerosis Control Patient Cohort.

Patient ID	Sex	Age (Years)	Reason for Surgery	Tissue Source
1	F	61	I	Ascending colon
2	M	52	I	Ascending colon
3	M	61	I	Colon
4	F	56	I	Ileum
5	M	42	I	Ileum
6	F	25	I	Sigmoid colon
7	F	35	C	Sigmoid colon
8	F	67	C	Sigmoid colon
9	F	32	C	Sigmoid colon
10	M	54	C	Sigmoid colon
11	M	74	C	Descending colon
12	M	58	C	Sigmoid colon
13	F	38	C	Sigmoid colon
14	M	69	C	Sigmoid colon

C, cancer; F, female; I, inflammatory disease; ID, identification number; M, male; O, other.

provide detailed information on both study groups. As the quality and quantity of the tissue samples differed among individuals, some sections were excluded during image acquisition and analysis: for example, ganglia were detected that did not contain any HU-positive (HU⁺) neurons, background signals were too high, or the paraffin section was folded. This resulted in varying cohort numbers for the different IHC analyses, which are shown in detail in

Table 3. Overview of the MS and Non-MS CTRL Patient Cohorts.

Analysis	MS	CTRL	p Value
ChAT and HU (in MP), nNOS and HU
(in MP), nNOS and HU (in SP)
Number of patients (patient IDs)	4 (1, 2, 3, 4)	14 (all)
Age, years—median (min-max)	56.5 (25–68)	55 (25–74)	0.9399 *
Sex—% female	50	50	1 †
Reason for surgery—% C/% I	50/50	57/43	1 †
ChAT and HU (in SP)
Number of patients (patient IDs)	5 (1, 2, 3, 4, 5)	14 (all)
Age, years—median (min-max)	46 (25–68)	55 (25–74)	0.8056 *
Sex—% female	60	50	1 †
Reason for surgery—% C/% I	40/40 (20% other)	57/43	1 †
ANO1-positive ICCs
Number of patients (patient IDs)	5 (1, 2, 4, 6, 7)	10 (3–7, 9, 10, 12–14)
Age, years—median (min-max)	63 (25–68)	48 (25–69)	0.3866 *
Sex—% female	80	50	0.5804 †
Reason for surgery—% C/% I	20/40 (40% other)	60/40	0.5594 †

ANO1, anoctamin 1; C, cancer; ChAT, choline acetyltransferase; CTRL, control; HU, anti-HuC/D antibody; I, inflammatory disease; ICC, interstitial cell of Cajal; ID, identification number; MP, myenteric plexus; MS, multiple sclerosis; nNOS, neuronal nitric oxide synthase; SP, submucosal plexus. * Mann–Whitney U Test. ^† Fisher’s exact test.

.

Ethical permission was obtained from the Ethics Committees of the Universities of Erlangen-Nürnberg (file nos. 2550 and 299_17B) and Würzburg (file no. 81/14).

2.4. Immunohistochemistry

Sections of human colon were cut at 5 μm using either a SM2000 R microtome or a Histocore MULTICUT microtome (both Leica, Wetzlar, Germany) and transferred onto HistoBond slides (Roth). Subsequently, two sections per individual were dehydrated in an ascending series of ethanol (Brenntag, Essen, Germany) and xylene (Roth). For the investigation of enteric neurons, Trilogy solution (Sigma-Aldrich) was used for heat-induced epitope retrieval. Sections were blocked using TBS containing 1% BSA (Roth), 0.5% Triton X (Merck Millipore), and 5% normal donkey serum (Dianova). All sections were double-stained with an antibody against HU (Thermo Fisher; dilution 1:40) and an antibody directed either against ChAT (Sigma-Aldrich; 1:20) or nNOS (University of Graz; 1:1500); all primary antibodies were diluted in a solution of TBS, 1% BSA, and 0.5% Triton X. Sections were incubated at room temperature for 24 h. After washing with TBS, secondary antibodies were applied and incubated for 1 h. Secondary antibodies comprised donkey anti-rabbit Alexa Fluor 488 (Thermo Fisher; 1:1000), donkey anti-goat Alexa Fluor 647 (Thermo Fisher; 1:1000), biotinylated donkey anti-mouse immunoglobulin (Ig)G (Dianova; 1:200), and streptavidin-Cy3 (Dianova, 1:1000). For each staining and combination of secondary antibodies, two slides were incubated in the absence of primary antibodies, thus serving as negative controls. Hoechst 33258 (Sigma-Aldrich, dilution of 1:1000 in TBS) was used as a nuclear counterstain. All sections were coverslipped with TBS and glycerol (Roth) (dilution 1:2, pH 8.6).

For analysis of ICCs, a solution of Tris and ethylenediaminetetraacetic acid (EDTA) (both Roth) was used for heat-induced epitope retrieval. Sections were blocked in TBS containing 0.05% Tween 20 (TBST) (Sigma-Aldrich) and 5% milk powder (Heirler, Radolfzell, Germany). Antibodies were diluted in TBST + 0.5% milk powder. The primary antibody rabbit anti-ANO1 (Abcam; dilution 1:200) was incubated at 4 °C overnight. Additional slides were incubated in the absence of primary antibodies, thus serving as negative controls. After washing with TBST, biotinylated goat anti-rabbit IgG (Abcam; 1:500) was applied, and the slides were incubated at room temperature for 1 h. Again, TBST was used for washing before incubation with streptavidin–horseradish peroxidase (HRP) (Abcam; 1:2500 in TBST) at room temperature for 30 min. Subsequently, sections were washed with TBST and developed with a 3,3′-diaminobenzidine (DAB) substrate kit (Vector Laboratories, Burlingame, CA, USA) according to the manufacturer’s protocol. Hematoxylin (Waldeck/Chroma, Münster, Germany) counterstaining was used to identify the MP. Sections were dehydrated and coverslipped with Entellan (Sigma-Aldrich).

2.5. Image Analysis

Murine whole mounts stained for neurotransmitters were analyzed using a confocal laser scanning microscope system (Nikon Eclipse E1000-M; Nikon Digital Eclipse C1; Nikon, Tokyo, Japan) fitted with 488 nm and 543 nm solid-state lasers (Sapphire 488LP, Sapphire 561-50; both from Coherent, Santa Clara, CA, USA), a 405 nm diode laser (CUBE 405-100C; Coherent), and a 642 nm diode laser (Melles Griot, Carlsbad, CA, USA). A BIO1 filter set (DAPI/Cy5 for C1-Detector; AHF Analysentechnik, Tübingen, Germany) reduced nonspecific background fluorescence. Images were taken as z-series (z-steps: 1.0 μm). Only ganglia and neurons that could be fully captured were considered. During analysis, detailed attention was paid to neurons located in more than one z-position to avoid multiple false-positive counts. For each mouse, ten images were analyzed from both the SP and MP. Most images displayed more than one ganglion. The mean number of neurons per ganglion was calculated as was the percentage of neurons containing either ChAT, nNOS, or VIP in relation to the total number of HU⁺ neurons. Exemplary single-optical sections were selected from the z-series for representation in the figures.

Murine tissue stained for ANO1 was analyzed using a Leica DMI8 Thunder imager equipped with a K5-14401790 camera (Leica) and the Leica filter cube DFT51010 with DAPI (excitation filter (EX) 391/32, dichroic mirror (DC) 415, suppressor filter (SF) 435/30), Cy2 (EX 479/33, DC 500, SF 519/25), and Cy3 (EX 554/24, DC 572, SF 594/32). For each whole mount, six z-stacks (z-steps: 2.4 μm) were acquired from different locations. For each z-stack, one single-optical section that displayed the MP best was selected and analyzed. GFAP was used as a marker for the MP because ICCs are located at the level of the MP and within the longitudinal and circular muscle layer [21].

The ANO1-positive (ANO1⁺) area was determined by setting a grayscale threshold in the corresponding channel after images underwent instant computational clearing using the thunder function of the LAS X software (Leica).

After IHC staining, sections of human colon were analyzed using a Leica DM6 B fluorescence microscope, equipped with a Leica DFC3000G camera and the following filters: DAP (EX 350/50, DC 400, SF 460/50), L5 (EX 480/40, DC 505, SF 527/30), RHO (EX 546/10, DC 560, SF 585/40), and Y5 (EX 620/60, DC 660, SF 700/76). For the determination of neurotransmitter signatures, tissue from five patients with MS and 14 CTRL patients (Table 3) was analyzed. All HU⁺ ganglia were imaged for IHC analysis.

Light microscopic ANO1 staining was analyzed using a Leica DM6 B fluorescence microscope, equipped with a Leica DMC2900 camera and the filter block ANT (Leica). Staining for ANO1⁺ ICCs was performed on sections from five patients with MS and 10 CTRL patients (Table 3); only tissue containing MP was included. Owing to their small cell bodies and numerous processes [9], ICCs cannot be counted as easily as neurons. Instead, entire sections of colon were scanned and, using ImageJ version 1.53c (National Institutes of Health, Bethesda, MD, USA), the areas of longitudinal and circular muscle tissue, MP, and ANO1⁺ ICCs were measured and compared.

Image analysis was performed using Fiji software version 2.0.0-rc-68/1.52 h (National Institutes of Health) [36]. Exemplary images were selected for representation in the figures, which were prepared using Inkscape 1.0 (Inkscape Project, Boston, MA, USA).

2.6. Statistical Analysis

GraphPad Prism version 9.1.2 (GraphPad Software, Inc., San Diego, CA, USA) was used for statistical analysis. Normality and lognormality were determined by Shapiro–Wilk test. Differences between groups were evaluated using a one-way analysis of variance or Kruskal–Wallis test in mice and a Mann–Whitney U test or unpaired Student’s t test in humans. The Mann–Whitney U test (age) and Fisher’s exact test (sex, reason for surgery) were used for comparison of demographic data sets. Median values calculated from the mean values of individual mice and patients were displayed graphically. p-values < 0.05 were considered to be statistically significant: * p < 0.05, ** p < 0.01 and *** p < 0.001.

3. Results

3.1. Decrease in the Percentage of ChAT⁺ Neurons in the MP of Early- and Late-Stage EAE Mice

The neurochemical signature of both the SP and MP was studied in EAE, which is the common mouse model of MS. Mice were studied during early- and late-stage disease, and results were compared with a control-immunized group that had received HEL as the immunizing antigen. In HEL-immunized control mice, 181 ganglia were analyzed in the SP and 54 in the MP; in early-stage EAE mice, 194 ganglia were analyzed in the SP and 72 in the MP; in late-stage EAE mice, 186 ganglia were analyzed in the SP and 81 in the MP. The number of HU⁺ neuronal perikarya was investigated in both the SP (Figure 1) and the MP (Figure 2).

The percentages of neurons containing nNOS, VIP, or ChAT were calculated in each group. In the SP of HEL-immunized mice, ganglia contained a median number of 7.58 (min–max: 5.93–8.95) neurons, whereas 6.48 (4.41–10.83) neurons per ganglion were observed in the early-stage EAE group and 8.00 (6.58–10.34) in the late-stage EAE group (Figure 1b). The median percentage of VIP-positive (VIP⁺) neurons barely varied among the groups (HEL: 51.81% (min–max: 25.78–82.73%); early-stage EAE: 47.97% (32.84–58.45%); late-stage EAE: 55.28% (40.19–73.73%)) (Figure 1c). The same observation applied to nNOS-positive (nNOS⁺) neurons (HEL: 46.46% (min–max: 12.55–68.72%); early-stage EAE: 31.63% (8.7–46.22%); late-stage EAE: 32.97% (3.25–58.65%)) (Figure 1d). The number of submucosal neurons staining positively for ChAT (ChAT⁺) was consistently low (HEL: 11.91% (min–max: 6.34–14.77%); early-stage EAE: 8.05% (3.88–17.11%); late-stage EAE: 6.20% (3.13–10.93%)) (Figure 1e). Overall, no notable differences were observed in the SP between HEL-immunized mice, early-stage, and late-stage EAE.

In the MP, the median number of neurons per ganglion was slightly higher in HEL-immunized (39.17 (min–max: 25.67–44.1)) and early-stage EAE mice (38.85 (25.71–48.44)) compared with late-stage EAE (27.55 (18.93–30.56)) (Figure 2b). VIP⁺ neurons were scarce (HEL 5.30% (min–max: 2.47–5.78%); early-stage EAE 6.61% (4.65–10.48%); late-stage EAE 4.31% (3.25–6.76%)) (Figure 2c). The percentage of nNOS⁺ neurons was similar in all three groups (HEL 28.64% (min–max: 16.25–38.37%); early-stage EAE 32.00% (27.86–46.47%); late-stage EAE 25.65% (16.79–28.85%)) (Figure 2d). Thus, there were no statistically significant alterations in the total number of neurons per ganglion and in the percentages of VIP⁺ and nNOS⁺ neurons in the MP. However, a statistically significant lower percentage of ChAT⁺ neurons was found in early-stage EAE (9.96% (min–max: 6.73–10.28%)) compared with the HEL control group (14.04% (11.46–32.36%)) (p < 0.05; Figure 2e). A strong trend toward a decrease was observed in late-stage EAE (10.15% (min-max: 4.96–11.85%)).

3.2. Decrease in the Percentage of the ANO1⁺ Area in the MP of Late-Stage EAE Mice

Having observed changes in the neurochemical signature of ENS neurons in EAE mice, we next investigated potential alterations on the level of the ICCs using ANO1 as a marker. EAE mice were studied during late-stage disease, and results were compared with non-immunized control mice. For each mouse, six images from different locations within the MP were analyzed by calculating the percentage of the ANO1⁺ area in relation to the MP. As shown in Figure 3, the median ANO1⁺ area was higher in non-immunized controls compared to late-stage EAE mice (1.18% (0.72–1.23%) vs. 0.50% (0.39–0.78%), while statistical significance was not reached (p = 0.07).

3.3. Increase in the Percentage of ChAT⁺ Neurons in the MP of Patients with MS

To translate the results obtained in the EAE model to MS itself, intestinal resectates and biopsies were studied. As shown in Table 1 and Table 2, patients with MS and matched CTRL patients underwent abdominal surgery for various medical reasons, which were divided into two groups: cancer versus inflammatory diseases, such as appendicitis, Crohn’s disease or ulcerative colitis. In contrast to the whole-mount preparations of murine colonic tissue, the 5 μm-thick sections of human tissue did not allow for the acquisition of three-dimensional z-stacks. In addition, and particularly in CTRL patients, human colon was often cut in a longitudinal direction—i.e., parallel to longitudinal muscle layers and orthogonal to circular muscle layers—showing only a fraction of the broad web that makes up the MP. This explains why the number of neurons per ganglion was lower in the human ENS compared with the murine ENS.

For the IHC analyses, all HU⁺ ganglia were imaged, which resulted in 156 ganglia in the SP (Figure 4) and 86 in the MP of patients with MS (Figure 5). In the CTRL group, 289 SP ganglia (Figure 4) and 331 MP ganglia (Figure 5) were analyzed. In the SP, a median of 3.02 (min–max: 2.31–4.67) neurons per ganglion were counted in CTRL patients compared with 3.89 (1–8.5) in patients with MS (Figure 4b). VIP was dismissed from analysis because in the human MP, inhibitory VIP⁺ neurons are generally also nNOS⁺ [16]. While nNOS was barely detectable in the neurons of the SP in both groups (CTRL: 0.00% (min–max: 0.00–28.89%); MS: 2.28% (0.00–5.28%)) (Figure 4c), the percentage of ChAT⁺ cells was higher (CTRL: 23.96% (7.37–60.24%); MS: 18.90% (0.00–75.00%)) (Figure 4d). Overall, the neurochemical signature of the SP showed no significant difference between patients with MS and CTRL patients.

In the MP, a median of 4.25 (min–max: 2–8.09) neurons per ganglion were detected in CTRL patients compared with 6.25 (3.29–11.41) in patients with MS (Figure 5b). In CTRL patients, 20.40% (min–max: 0.00–63.89%) of neurons were nNOS⁺ compared with 30.81% (8.33–55.00%) in patients with MS (Figure 5c); and 22.51% (0.00–66.67%) of neuronal cells were ChAT⁺ in CTRL patients compared with 45.16% in patients with MS (Figure 5d). While no statistically significant observations were made comparing CTRL and MS patients, a trend toward an increase in ChAT⁺ neurons in patients with MS was noted. Interestingly, the percentage of ChAT⁺ neurons in patients with MS showed the smallest variation of all human data sets, ranging only from 44.44% to 48.57%.

3.4. Significant Increase in the Area of ANO1⁺ ICCs in Relation to the Area of Muscle Tissue in Patients with MS

The relative areas of longitudinal and circular muscle tissue, MP, and ANO1⁺ ICCs were investigated (Figure 6). Overall, when the area of the MP was compared with the area of muscle tissue, the MP covered a median area of 0.65% (min–max: 0.28–2.3%) in CTRL patients compared with 0.98% (0.51–1.196%) in patients with MS (Figure 6d). In relation to the MP area, the ICC area was larger in patients with MS than in CTRL patients: a median of 16.19% (min–max: 5.61–57.89%) of the size of MP area was ANO1⁺ in CTRL patients compared with 61.7% (6.67–108.82%) in patients with MS (Figure 6e), yet, no statistical significance was reached. However, relative to the muscle tissue area, the ICC area was significantly higher in patients with MS than in CTRL patients: median 0.56% (min–max: 0.05–0.92%) versus 0.09% (0.03–0.41%) (p < 0.05; Figure 6f).

4. Discussion

The key findings of this study were a significant loss of ChAT⁺ neurons in the MP of mice with EAE as opposed to an increase in ChAT⁺ neurons in the MP of patients with MS. Interestingly, while these changes pertained to the MP, no pathological alterations were evident in the SP. In addition, there was a significant increase in the expression of ANO1⁺ ICCs in patients with MS and a decrease in the MP of mice with late-stage EAE.

A decrease in the number of cholinergic neurons within the MP has previously been reported in the ENS of aged rats [26]. The same observation was made in colonic tissue samples from elderly people, where a decline in ChAT⁺ neurons was accompanied by an increased number of nNOS⁺ neurons, which might serve as an explanation for the higher prevalence of constipation with age [25]. Along these lines, ChAT⁻/⁻ mice exhibited colonic dysmotility [37], and myenteric ChAT⁺ neurons were diminished in patients with slow-transit constipation [38]. Of note, constipation is also a frequent symptom observed in patients with MS [4,5], which may be linked to ENS alterations.

Conflicting with these findings and the observations made in EAE, we found a trend toward an increase in ChAT⁺ neurons in patients with MS vs. CTRL patients. There may be several reasons for the discrepancy between EAE and MS. On the one hand, the MS patient sample size was limited in our study. Our cohort consisted of only few individuals from different age groups and with different comorbidities. Important information regarding the clinical subtype of MS and MS treatment prior to abdominal surgery was not available. While relapsing–remitting MS is characterized by predominant inflammation, the progressive stages of MS mainly show degenerative pathology [39]. This switch in pathomechanisms may also affect the ENS differentially so that it will be worthwhile to distinguish between different disease stages of MS in future studies on ENS involvement in MS. The same applies to the impact of MS treatment on the ENS, since some therapies are primarily immune modulatory and suppressive, while others may also have beneficial effects on neurodegeneration [40]. On the other hand, our methods of analysis differed between EAE and MS. While only paraffin-embedded tissue sections were available from MS patients, we were able to study whole mounts from EAE mice. The whole mount technique allows for three-dimensional analysis and hence provides a more complete picture of the ENS. Thus, it is likely to be better suited for studying ENS alterations compared to paraffin-embedded tissue sections. Finally, it needs to be considered that EAE can be induced by immunization with different CNS antigens. While MP4 was used in our study, immunization with, e.g., whole spinal cord homogenate may lead to results that are more similar to what we have observed in MS patients. It will be subject to future studies to dissect the involvement of different target antigens in ENS pathology.

Alternatively, our results obtained in MS patients could be observed from a different, more theoretical perspective, and one could speculate about the occurrence of neurogenesis in the ENS, which is currently heavily debated. Postnatal neurogenesis originating from enteric glial cells has been reported in mice after chemical injury [41] as well as in zebrafish under physiological circumstances [42]. In addition, in the context of colitis, Belkind-Gerson et al. showed the expression of markers indicative of neurogenesis in the human ENS, with subsequent studies of a corresponding mouse model suggesting that the newborn neurons were predominantly excitatory [43]. Compared to mice with early-stage EAE, which were analyzed 12 days after immunization and before the appearance of any symptoms, patients with MS in our study had been affected by the disease for a prolonged time. Hence, it is conceivable that an initial loss of neurons in patients with MS might be compensated for by the preferential genesis of new excitatory neurons.

In addition to ChAT, we detected nNOS⁺ neurons in the ENS. In CTRL patients, a median of 20.4% of myenteric neurons were nNOS⁺, and a median of 22.5% neurons were ChAT⁺. The MP of patients with MS contained slightly more nNOS⁺ neurons (median 30.8%) and markedly more ChAT⁺ neurons (median 45.2%). Earlier studies of the human MP have described equal proportions of nitrergic and cholinergic neurons, with each of the two populations accounting for approximately 50% of the total ENS neuronal population [13,14,15,38]. However, in prior studies, sample sizes were larger, and whole mounts were used instead of sections of paraffin-embedded tissue [13,14,15,38], which may explain why our results differ from the literature. Similar to our findings, Neunlist et al. reported divergent numbers, detecting only about 35% ChAT⁺ neurons in their subjects, for which they suspected different strategies of tissue processing before staining as an explanation [23]. In fact, in two of the studies cited above, tissue was cultured before fixation and staining in order to enhance ChAT immunoreactivity [14,38].

In our study, the MP seemed to be affected more than the SP in both EAE and MS. Indeed, there are other diseases in which alterations are predominantly shown in the MP. Among such diseases is enteric ganglionitis, which is characterized by the inflammation and infiltration of lymphocytes into the ENS and the development of circulating anti-neuronal antibodies targeting the ENS [44]. Anti-MP autoantibodies were recently also proposed as the origin of idiopathic achalasia, which is characterized by disturbed esophageal peristalsis [22].

A further main result of our study was that the area of ICCs, which are located at the level of the MP and in the muscle layers surrounding it [21], was increased in patients with MS, while it was decreased in EAE mice. Increased networks of ICCs were found in early stages of type 1 diabetes, where they were associated with accelerated emptying of the stomach [45]. On the contrary, in conditions such as constipation, in which a slower passage of the gastrointestinal system prevails, ICCs are typically less abundant [21]. Consistent with our findings in the murine colon, a decrease in the number of ICCs was previously observed in the bladder of EAE mice [46]. Interestingly, ICCs are not independent in their role as intestinal pacemakers, but they interact, inter alia, with inhibitory and excitatory neurons [21]. A study in nNOS^-/^- mice demonstrated a significant reduction in the volume of ICCs in the MP, while incubation in vitro with an NO donor increased the number of ICCs [47]. Another study showed that the terminal endings of excitatory neurons were close to networks of ICCs, which have receptors for ACh [48]. Thus, interrelations between ICCs and this type of neuron are conceivable. Considering the increase in ChAT⁺ cells in the MP of patients with MS, as described in our study, it is possible that alterations in different ENS cell types occur at different times during the course of the disease, which subsequently or simultaneously affects other cell types.

In summary, to our knowledge, this is the first study to evaluate neurotransmitter signatures and ICC morphology in the context of EAE/MS. While our data show a decrease in the percentage of ChAT⁺ neurons and the ANO1⁺ area in the MP of EAE mice and an increase in the percentage of ChAT⁺ neurons and the ANO1⁺ area in the MP of patients with MS, further research is needed not only to unravel the underlying pathological mechanisms but also to rule out the methodological limitations of this pilot study. Nevertheless, our data imply that a better understanding of the phenotype of ENS neurons and ICCs may result in intriguing and novel insights into the pathophysiology of MS, its etiology, and future treatments.