SEARCH

SEARCH BY CITATION

Keywords:

  • Crohn’s disease;
  • enteric nervous system;
  • immunohistochemistry;
  • ulcerative colitis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients
  5. Methods
  6. Results
  7. Discussion
  8. Conflict of interest statement
  9. References

Abstract  Various studies have described abnormalities of the enteric nervous system (ENS) in tissue samples from patients with chronic idiopathic inflammatory bowel diseases (IBD). The distribution of density of the different cell types of the ENS was however not studied in a systematic way. The aim of this study was to examine the density of neurons, enteroglial cells and interstitial cells of Cajal (ICC) in the different plexuses of the ENS in samples from patients with Crohn’s disease (CD), ulcerative colitis (UC) and controls. Tissue samples from 16 patients with CD (ileum) and 16 patients with UC obtained in involved and non-involved areas were studied using immunohistochemistry with antibodies directed against neuron-specific enolase, S100, C-Kit and CD3. Sections were analysed blindly by two pathologists and the number of positive cells was counted for each type. Overall, an increase was noted for neuronal cell bodies, enteroglia and ICC in the deep muscular plexus in CD. In uninvolved areas of CD patients, the number of enteroglial cells was decreased. In UC, an increase of ICC in the muscularis propria and enteroglial cells was observed in diseased tissue. The study confirms the presence of abnormalities of the different cells of the ENS in IBD. The presence of lesions in samples from uninvolved areas, such as a reduction of enteroglia, supports a pathogenetic role of the ENS.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients
  5. Methods
  6. Results
  7. Discussion
  8. Conflict of interest statement
  9. References

The enteric nervous system (ENS) is composed of a large collection of neuronal cells distributed along the length of the alimentary tract and organized in two major ganglionated plexuses, the myenteric (Auerbach’s) and submucosal (Meissner’s and Henle’s) plexus. Both plexuses contain various types of cells including functionally different neurons and enteric glial cells (EGC), which are both in close contact with the interstitial cells of Cajal (ICC).1

Patients with chronic inflammatory bowel diseases (IBD), i.e Crohn’s disease (CD) and ulcerative colitis (UC), display several structural and functional abnormalities of the ENS.2 Although it has been hypothesized that these abnormalities are secondary to the inflammatory process, i.e. to tissue injury or via the effects of soluble mediators of the inflammatory process itself (cytokines, arachidonic acid metabolites, oxygen-derived free radicals),2 the occurrence of lesions in the ENS in macroscopically uninvolved areas may actually suggest that they precede mucosal inflammation.3

The literature data on the morphology of the ENS data in IBD are heterogeneous and confusing. Routine pathology reports describe: (i) hypertrophy, hyperplasia and axonal damage of nerve fibres;4,5 (ii) a normal aspect, hypertrophy, hyperplasia or damage of neuronal cell bodies;6–8 (iii) EGC hyperplasia.9,10 Moreover, an increased number of mast cells and axonal necrosis of gut nerves have been reported with electron microscopic studies.11,12 Most studies show however abnormalities of the ENS. More recent studies with immunohistochemical methods and electron microscopy have shown conflicting results for the expression of vasoactive intestinal peptide and other mediators,2 an increased expression of nitric oxide synthase in neuronal cell bodies,8,13 a variable increase of EGC,3,14 decrease and ultrastructural injury of ICC15 and ganglioneuritis.16

However, most studies on the ENS in IBD have mainly focused on the presence of inflammation and the description of lesions but a systematic, comprehensive study of the numbers of the different cell types of the ENS involved has not usually been performed. Thus, the purpose of the present investigation was that of evaluating, by means of immunohistochemical methods, the abnormalities of the ENS (neurons, EGC and ICC) in IBD patients to confirm the involvement of these cells in the disease process and eventually to see whether one cell type is more involved than the others.

Patients

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients
  5. Methods
  6. Results
  7. Discussion
  8. Conflict of interest statement
  9. References

Archival full thickness specimens from 32 IBD patients (16 from the ileum of CD patients, eight men and eight women, age range 28–47 years; 16 from the colon of UC patients, 10 men and six women, age range 35–52 years) were obtained from patients undergoing surgery for severe disease refractory to medical treatment in the period June 2006–June 2007. The diagnosis of CD or UC was based on clinical, radiological and endoscopic examination and histological findings. All IBD patients had been treated with 5-amino-salicylic acid and immunosuppressive drugs.

The samples studied were taken from macroscopically involved and non-involved areas. For each patient we chose the samples better oriented and more representative of disease.

Control specimens were obtained from patients undergoing colonic (n = 10, five men and five women, age range 41–59 years) or ileal (n = 15, 10 men and five women, age range 35–50 years) resection for neoplastic disease. The samples studied were taken at least 3 cm from the resection margin in tumour-free areas. The same methods were used for the evaluation of these control samples.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients
  5. Methods
  6. Results
  7. Discussion
  8. Conflict of interest statement
  9. References

Transversal sections were obtained from samples of the resected ileum or colon. For conventional histology 5 μm paraffin sections were stained with haematoxylin-eosin, PAS and trichrome stain.

Immunohistochemistry

At least 40 slides (20 from involved and 20 from non-involved areas) for each patient were processed for immunohistochemistry. To evaluate markers of the ENS, we used monoclonal antibodies directed against neuron-specific enolase (NSE, NCL-NSE2, Novocastra laboratories, dilution 1 : 50; Newcastle upon Tyne, UK) acting as a marker of neuronal cell bodies in the ganglia, and the glial marker protein S100 (S-100, dilution 1 : 50; Dako, Carpinteria, CA, USA) for enteroglial cells.17,18 Since ICC express Kit, an antiKit antibody (CD117, rabbit polyclonal antibody, IgG, dilution 1 : 50; Dako) was used to detect these cells, as previously reported.19 The presence of T lymphocytes was assessed by means of monoclonal mouse antihuman CD3 antibody (Dako Cytomation, dilution 1 : 40).

Neuron-specific enolase, S-100 and CD3 immunostaining was carried out using a peroxidase-based visualization kit (Dako LSAB®), following the manufacturer’s recommendations. Diaminobenzidine tetrahydrochloride was used as chromogen. The slides were then counterstained with Mayer’s haematoxylin for 5 s, dehydrated and mounted in Clarion (Biomeda, Foster City, CA, USA). To account for non-specific staining, peptides that blocked polyclonal antibody bindings (passage with normal goat serum) were used, or sections were incubated in the absence of primary antibody. In these cases, no immunostaining was detected.

Expression of kit

Consecutive formalin-fixed, paraffin sections were dewaxed and rehydrated through decreasing alcohol series up to distilled water. Sections were then subjected to heat-induced epitope retrieval by immersion in a heat-resistant container filled with citrate buffer solution (pH 6.0) placed in a pressure cooker and microwaved for 20 min. Endogenous peroxidase activity was suppressed by incubation with 3% solution of H2O2 for 5 min. Kit immunostaining was carried out using a peroxidase-based visualization kit (Dako EnVision™), following the manufacturer’s recommendations.

Data analysis

All slides were coded and analysed blindly by two pathologists (VV and KG). For NSE, S100 and CD3 both the submucosal and the myenteric plexuses were taken into account by optical microscopy at ×200 magnification (Olympus BX 40; Olympus Corporation, Tokyo, Japan). For each patient, the number of immunopositive cells was calculated and expressed as the mean of cells on 10 well stained and well oriented randomly selected microscopic fields for each region of interest. We assessed macroscopically involved and not involved areas separately by the disease in the patients’ cohort. To be considered as positive, the intensity of cell immunostaining in relation to possible background had to be from moderate to strong, as described previously.20 The morphological evaluation only considered the cell bodies, whereas the nerve fibres were not analysed.

Due to the different anatomic distribution, the density of ICC was graded differently for the ileal (CD)21 and the colonic (UC)22 specimens, after the evaluation of 10 well-stained and well-oriented fields at ×400 magnification (see below for details).

Colonic ICC

The three previously identified populations of ICC were taken into consideration23: IC-SM, along the submucosal surface of the circular muscle bundle, IC-MY, within the intermuscular space between circular and longitudinal muscle layers (myenteric region, which displays the highest yield of ICC in normal tissue24) and IC-IM, within the muscle fibres of the circular and longitudinal muscle layers.

Ileal ICC

The three previously identified populations of ICC in the small bowel were taken into consideration25: IC-DMP (deep muscular plexus), located between the internal thin layer and the thick outer layer of the circular muscle, IC-MP (myenteric plexus), located between the circular and longitudinal muscle layers, and IC-IM (intramuscular), within the muscle fibres of the circular and longitudinal muscle layers.

The number of mast cells in the muscle layers, positive to CD117 but having a strikingly different shape from ICC, was also assessed. To avoid a possible overlap with ICC, mast cells were also counted at ×400 magnification.

Plexitis was evaluated on CD3 stained slides, and defined as the presence of one or more lymphocytes appositioned to or within an enteric ganglion and was graded based on the appearance of the most severely inflamed ganglion in the available tissue, according to previously described criteria.26 It was graded mild if this ganglion contained four or less inflammatory cells, moderate if four to nine cells were present, or severe if 10 or more cells were found.

Statistical analysis

Data from controls and patients (in the latter, from both involved and uninvolved areas) were compared by non-parametric methods, using the Wilcoxon test. Values of P < 0.05 were chosen for rejection of the null hypothesis. Data are expressed as mean (95% CI).

Ethical considerations

As this was a retrospective study, no individual patient identification was involved and no study-driven clinical intervention was performed; therefore no ethical approval was necessary.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients
  5. Methods
  6. Results
  7. Discussion
  8. Conflict of interest statement
  9. References

Tables 1 and 2 show the results obtained in CD and UC patients.

Table 1.   Enteric nervous system variables in ileal specimens of Crohn’s disease and controls. Data are expressed as mean (95% CI)
 CD-involved areasCD-uninvolved areasControlsP
  1. ICC, interstitial cells of Cajal; MP, myenteric plexus; IM, intramuscular; DMP, deep muscular plexus; NSE, neuron-specific enolase; SP, submucous plexus; MP, myenteric plexus; MC, mast cells.

ICC-MP219 (136–302)212 (126–299)274 (240–308)0.49 vs uninvolved 0.20 vs controls 0.17 vs controls
ICC-IM16 (10–22)14 (8–21)67 (60–74)0.29 vs uninvolved <0.0001 vs controls
ICC-DMP45 (35–54)39 (27–50)34 (28–39)0.02 vs uninvolved 0.05 vs controls 0.44 vs controls
NSE-SP50 (43–57)41 (35–47)42 (34–51)0.012 vs uninvolved 0.14 vs controls 0.81 vs controls
NSE-MP75 (66–84)64 (54–74)69 (58–80)0.048 vs uninvolved 0.38 vs controls 0.48 vs controls
S100-SP114 (98–130)78 (70–86)95 (80–110)0.0001 vs uninvolved 0.07 vs controls 0.03 vs controls
S100-MP238 (213–263)209 (185–234)207 (186–238)0.011 vs uninvolved 0.18 vs controls 0.89 vs controls
MC89 (31–147)19 (13–24)35 (28–41)0.016 vs uninvolved 0.09 vs controls 0.0001 vs controls
Table 2.   Enteric nervous system variables in colonic specimens of ulcerative colitis and controls. Data are expressed as mean (95% CI)
 UC-involved areasUC-uninvolved areasControlsP
  1. ICC, interstitial cells of Cajal; MY, myenteric plexus; IM, intramuscular; SM, submucous plexus; NSE, neuron-specific enolase; SP, submucous plexus; MP, myenteric plexus; MC, mast cells.

ICC-MY230 (197–263)221 (183–259)214 (154–274)0.70 vs uninvolved 0.58 vs controls 0.82 vs controls
ICC-IM45 (42–49)43 (40–46)38 (32–44)0.29 vs uninvolved 0.01 vs controls 0.1 vs controls
ICC-SM24 (20–29)34 (24–45)29 (23–35)0.04 vs uninvolved 0.17 vs controls 0.34 vs controls
NSE-SP47 (29–65)42 (22–63)58 (46–70)0.72 vs uninvolved 0.34 vs controls 0.18 vs controls
NSE-MP69 (60–78)58 (48–67)64 (55–74)0.07 vs uninvolved 0.44 vs controls 0.29 vs controls
S100-SP118 (103–134)114 (89–139)127 (93–161)0.74 vs uninvolved 0.56 vs controls 0.49 vs controls
S100-MP249 (228–270)216 (202–229)214 (190–238)0.014 vs uninvolved 0.03 vs controls 0.89 vs controls
MC36 (19–51)21 (15–27)36 (28–44)0.10 vs uninvolved 0.04 vs controls 0.01 vs controls

Crohn’s disease

The evaluation with conventional and histochemical stains (H&E, PAS Trichrome) was used only for the morphological characteristics of the disease. Microscopy confirmed the presence of classical lesions, especially transmural lymphoid hyperplasia, in sections obtained from diseased areas and normal morphology in sections from non-involved tissue in all cases.

Immunohistochemistry

The number of enteric neurons was not different between patients and controls in both the submucous and the myenteric plexus, whereas it was significantly decreased in uninvolved compared to involved areas of CD patients (Fig. 1A, B). The number of EGC was significantly increased in both the submucous and myenteric plexus in involved compared to non-involved areas of CD patients, but no differences were found between patients and controls, except for a decreased number in uninvolved areas of the submucosal plexus compared to controls (Fig. 1C, D). For ICC, no differences were found between patients and controls in ICC-MP (Fig. 2A), whereas ICC-IM were significantly decreased in both involved and uninvolved areas of CD patients compared to controls (Fig. 2B, C). In CD, the numbers of ICC-DMP were furthermore increased in these areas when compared with the controls. Yet overall, the numbers were significantly reduced in uninvolved areas compared to involved areas. The number of mast cells was significantly decreased in involved and uninvolved areas of CD patients compared to controls. Plexitis (Fig. 2D) was found in 12 (75%) cases, and it was severe in one, moderate in three and mild in eight of them. So overall, an increase is noted for ICC-DMP, neuronal cell bodies and EGC in involved areas in CD.

image

Figure 1.  (A) Myenteric plexus of a patient with Crohn’s disease (involved area) showing almost normal number of gangliar cells (arrows). NSE, original magnification ×400. (B) Myenteric plexus of a patient with Crohn’s disease (uninvolved area) showing only one residual gangliar cell (arrow). NSE, original magnification ×400. (C) Enteric glial cells (arrows) in the submucosal plexus of a patient with Crohn’s disease (uninvolved area). S100, original magnification ×400. (D) Enteric glial cells (arrows) in the submucosal plexus of a control. Note that cell bodies are more numerous than that seen in C. S100, original magnification ×200.

Download figure to PowerPoint

image

Figure 2.  (A) Interstitial cells of Cajal around the myenteric plexus and mast cells in the muscularis propria (arrows) in a patient with Crohn’s disease. Note the different shape of mast cells with respect to interstitial cells of Cajal. CD 117, original magnification ×200. (B) Intramuscular interstitial cells of Cajal (arrows) in a patient with Crohn’s disease. CD 117, original magnification ×200. (C) Intramuscular interstitial cells of Cajal (arrows) in a control. Note that cell bodies are more numerous than that seen in B. CD 117, original magnification ×200. (D) CD3 positive T lymphocytes (arrows) infiltrating the myenteric plexus in a patient with Crohn’s disease. Original magnification ×1000.

Download figure to PowerPoint

Ulcerative colitis

Again, the evaluation with conventional and histochemical stains (H&E, PAS Trichrome) was used only to assess the morphological aspects of the disease. Microscopy confirmed the presence of classical lesions in biopsies obtained from diseased areas and normal morphology in biopsies from non-involved tissue in all cases.

Immunohistochemistry

No differences were found concerning the number of enteric neurons between patients and controls, nor between involved and uninvolved areas in patients, in both the submucous and the myenteric plexus. Again, no differences between patients and controls (and involved and uninvolved areas in patients) were found for EGC in the submucous plexus. However, in the myenteric plexus EGC were significantly increased in involved areas of UC patients compared to both uninvolved areas and controls (Fig. 3A, B). For ICC, no differences were found between patients and controls in ICC-MY numbers, whereas ICC-IM were significantly increased in involved (but not in uninvolved) areas of UC patients compared to controls. ICC-SM were not different between patients and controls, the number was lower in samples from involved areas when compared with uninvolved areas. The number of mast cells was significantly decreased in involved and uninvolved areas of UC patients compared to controls, but it was increased in samples from involved areas when compared with uninvolved areas in UC patients (Fig. 3C, D). Plexitis was found in nine (56%) cases, and it was moderate in one, and mild in eight of them. So overall, an increase of ICC-IM and EGC cells is noted in diseased tissue while ICC-SM are decreased.

image

Figure 3.  (A) Enteric glial cells (arrows) in the myenteric plexus of a patient with ulcerative colitis, uninvolved area. S100, original magnification ×400. (B) Increased number of enteric glial cells (arrows) in an involved area of the same patient. S100, original magnification ×400. (C) Mast cells (arrows) in the muscularis propria of a patient with ulcerative colitis, involved area. CD 117, original magnification ×200. (D) Mast cells (arrows) in the muscularis propria of a control. Note that mast cell bodies are more numerous than that seen in C. CD 117, original magnification ×200.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients
  5. Methods
  6. Results
  7. Discussion
  8. Conflict of interest statement
  9. References

Overall, we found mixed abnormalities in all CD and UC patients for ICC and for enteric neurons and EGC. The ENS seems thus clearly involved in the disease process and all cell types are affected. The abnormalities we observed for ICC were usually confined to one of the subtypes or locations of the cells and consisted mainly in differences in numbers. These abnormalities were in agreement with those found in another study,16 even though we did not find atrophy and vacuolization of ICC in our patients. Ultrastructural abnormalities of ICC (swelling of mitochondria, decreased electron density, autophagosomes and partial depletion of the cytoplasm) were also described in previous studies.15,27 However, as we did not examine our biopsies with electron microscopy we may have missed ultrastructural abnormalities of otherwise seemingly normal cells.

Interestingly, in uninvolved areas of CD patients a decreased number of EGC was found in the submucosal plexus compared to controls. As experimental animal studies have shown that loss of glial cell function may increase gut permeability and alter the barrier function of the intestine,28,29 this abnormality might contribute to the pathogenesis of CD.30 Indeed, EGC alterations have been postulated to play a role in IBD, acting on both the barrier and the motility function of the gut.31,32 As EGC cells can also act as antigen presenting cells, the loss of these cells can be a link with the current hypothesis suggesting that CD may be due to an abnormal response to the intestinal flora.

Mast cells were significantly decreased in both CD and UC compared to controls. This was at variance with another study showing an increase of this cell population in ileal muscle specimens of CD patients.15 However, the higher number of subjects we recruited and differences in the patients’ population and treatment (all our patients received intensive and aggressive therapeutic regimens before surgery) could explain these differences. For instance, treatment with corticosteroids has been demonstrated to reduce mast cells numbers.33,34 As mast cells can produce membrane bound stem cell factor which is critical for the survival of both MC and ICC through binding to the c-kit receptor,35,36 a decrease of this cell population might also influence the fate of ICC.

An important finding in this study is the presence of myenteric plexitis in 75% of CD and 56% of UC patients. Plexitis is a common finding in CD, especially in samples from diseased areas, because CD is a transmural disease. In UC, it is supposed to be less common as UC is a mucosal disease. However, in severe, medically refractory UC, the inflammation may extend into the submucosa and the muscularis propria. The finding of plexitis in 56% of our UC cases may therefore not be unusual, as all patients were operated for refractory disease. The significance of plexitis remains unclear. It may be related to the development of the disease however. The presence of myenteric plexitis in proximal resection margins of ileocolonic resection specimens from patients with CD indeed been shown to be predictive of early endoscopic CD.26 The data were subsequently confirmed in CD and UC, even in macroscopically normal intestinal segments.16

We are presently unable to give an explanation for the modest but significant increase in neuronal cell bodies in patients with CD. Data from experimental animal models suggest that such an effect may be due to parenteral nutrition,37 used in most of our patients before surgery.

The results of the counting of cells can be influenced to a certain extent by the type of pathology examined in this study. Classically, specimens of CD patients are characterized by thickening of the wall of the affected bowel segments while, conversely, thinning of the colonic wall is observed in UC. These differences might influence the results, but this seems unlikely as our data also involve uninvolved areas and the abnormalities were mainly observed in these segments.

In fact, many of the above abnormalities we found in IBD patients were more evident in tissue samples macroscopically not involved by the inflammatory process, suggesting that ENS abnormalities may represent an early pathogenetic factor, and that the subsequent inflammation may mask some findings, due to an architectural distortion of the tissue with rearrangement and packing together of the enteric structures. On the other hand, experimental animal models have demonstrated that inflammation in the gut alters the cellular components of enteric reflex circuits in non-inflamed regions in ways similar to those at sites of active inflammation. These changes underlie altered function in non-involved regions during episodes of intestinal inflammation.38

Taken together, the literature data and our present results highlight the fact that the ENS is more or less deranged in patients with IBD, with its main components (enteric neurons, ICC, EGC) displaying several grades of abnormality in the various layers of the lower gut. It is likely that some of these abnormalities may act synergically (the decrease of mast cells and EGC, for instance, might cause a reduction of ICC by decreasing trophic factors for these cells15,31) by impairing the neurophysiological architecture of the involved intestinal segments while others (such as the myenteric plexitis) favour a pathogenetic role of the ENS in the spreading of CD and, probably, of UC.26

Further studies, also taking into consideration recent observations in experimental animal models on the ability of EGC to regulate tight-junction integrity, thereby influencing intestinal permeability,28,39 are however underway and for confirming the precise pathophysiological role of the ENS in patients with IBD.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients
  5. Methods
  6. Results
  7. Discussion
  8. Conflict of interest statement
  9. References