Nerve fibres and the respective neurotransmitters are known to contribute to intestinal inflammation.1 2 The entire wall of the human colon is richly innervated by extrinsic sympathetic nerve fibres and by fibres expressing the neuropeptide substance P (SP).3 The neurotransmitters of different nerve fibre types exert different effects on inflammatory processes: sympathetic neurotransmitters such as norepinephrine (via β-adrenoceptors), adenosine (via A₂ adenosine receptors), and endogenous opioids (via mu and kappa opioid receptors) exert anti-inflammatory effects when present in micromolar concentrations as in the vicinity of sympathetic nerve terminals (eg, they inhibit secretion of tumour necrosis factor (TNF)).4^–6 Loss of the sympathetic nerve fibres leads to lower concentrations of these anti-inflammatory neurotransmitters, and this is unfavourable in the chronic phase of an inflammatory disease.7 In contrast, the neuropeptide SP exerts pro-inflammatory effects by inducing secretion of pro-inflammatory cytokines such as TNF, interleukin 6 (IL6) and IL8.8^–10 Due to differential effects of neurotransmitters from sympathetic and SP⁺ nerve fibres, a parallel investigation of the two nerve fibre types using identical techniques should shed light on the overall balance in chronically inflamed tissue.11

A marked loss of sympathetic nerve fibres was recently demonstrated in the dextran sodium sulfate (DSS) colitis in mice using immunohistochemical and functional tests.12 In addition, studies in arthritis and type 1 diabetes research demonstrated loss of sympathetic nerve fibres in inflamed area of synovium and pancreatic islets, respectively,11 13 and SP⁺ nerve fibres sprout into inflamed tissue.11 The gut is one of the most abundant sources of SP in the body, where it is expressed in the myenteric and submucosal plexus and in intrinsic and extrinsic sensory neurons (reviewed by Holzer and Holzer-Petsche14). Increased SP expression has been observed in both tissue and nerve fibres in the colon of patients with colitis.15 16 In addition, the neurokinin-1 receptor is upregulated in the colon of patients with Crohn’s disease.15 Increased SP levels in patients with ulcerative colitis are correlated with disease activity.15 Blockade of neurokinin-1 receptors in experimental colitis significantly ameliorated the disease.17 Thus, a loss of sympathetic nerve fibres and an increase of SP⁺ nerve fibres similar to that in rheumatoid arthritis would lead to a preponderance of SP over sympathetic neurotransmitters in colonic tissue.

In arthritis research, there is no explanation for the differential loss of sympathetic nerve fibres versus SP⁺ nerve fibres during inflammation. Nerve repellent factors, which are specific for the different types of nerve fibres, might play an important role;18^–20 eg, semaphorins are a group of evolutionarily highly conserved, surface or locally secreted nerve repellent factors with a specific repulsive action on either sympathetic or primary afferent sensory fibres via distinct surface receptors, neuropilin-1 and neuropilin-2.21 Semaphorins are defined by the presence of an approximately 500-amino acid Sema domain at their NH₂ termini. Since sympathetic and primary afferent sensory nerve fibres reside in very different areas of the body, their specific guidance by these repellent factors is of major importance through embryonic development.20 Specific hard wiring of these two types of nerve fibre is achieved by different types of semaphorins: semaphorin 3A is a repulsive cue for primary afferent sensory fibres, and semaphorins 3B, 3C and 3F are repellents of the sympathetic nerve fibres (reviewed by Chen et al20 and Mark et al22). We recently demonstrated that macrophages and fibroblasts in inflammatory lesions produce semaphorin 3C (SEMA3C).23 Thus, SEMA3C may be a good candidate for the differential behaviour of sympathetic as compared to SP⁺ nerve fibres in inflammatory lesions. However, it is not known whether or not this nerve repellent factor is expressed in inflamed colon of patients with Crohn’s disease or diverticulitis.

This present study aimed to investigate the sympathetic and SP⁺ nerve fibre innervation in different layers of the colon wall in patients with Crohn’s disease and diverticulitis compared to control subjects. This investigation should shed light on the differential behaviour of these different types of nerve fibres. In addition, we investigated the expression of the sympathetic nerve repellent factor SEMA3C. In order to demonstrate a functional role of the sympathetic nervous system in colonic inflammation, we investigated the effect of abrogation of the sympathetic nervous system in acute and chronic DSS colitis and in Il10^−/− colitis in mice.

MATERIAL AND METHODS

Control subjects, patients with Crohn’s disease and diverticulitis

A total of 22 control subjects with no intestinal inflammation and colon cancer, six patients with diverticulitis, and 13 patients with Crohn’s disease were included. Table 1 gives the location of colon samples obtained. All patients were informed about the purpose of the study and gave written consent. Basic characteristics of these patients are given in table 1. Variables such as C-reactive protein, haematocrit, and blood leukocyte count were measured by standard techniques. Since nerve fibre density decreases with age (at least in rats and mice24), a bias for an age-related lower density of nerve fibres in control subjects might have influenced the results of the study (expectation: lower levels in control subjects but the opposite was found, see Results). In addition, the different locations of colon samples might have influenced the results. However, the investigation of control subjects with left-sided versus right-sided colon samples did not reveal any difference for all measured variables (data not shown).

Tissue preparation of human specimens

Colon tissue samples were taken immediately after surgery. In control subjects, a piece of colon tissue (entire intestinal wall) was obtained from the histologically confirmed tumour-free non-inflamed area. In Crohn’s disease patients, a piece of colon tissue (entire intestinal wall) was obtained from the macroscopically determined inflamed and macroscopically determined non-inflamed ileocaecal region. In diverticulitis patients, specimens of the resected piece of the colon were investigated. On average, specimens had a size of approximately 20 cm2. Samples intended for haematoxylin–eosin (HE) staining and detection of nerve fibres were fixed for 12–24 h in phosphate-buffered saline (PBS) containing 3.7% formaldehyde and then incubated in PBS with 20% sucrose for 12–24 h.11 Thereafter, tissue was embedded in Tissue Tek and quick-frozen. All tissue samples were stored at −80°C.

Histological evaluation and determination of colonic innervation in Crohn’s disease

Frozen tissue samples were cut into 5 μm thick sections and densities of neutrophils, eosinophils and lymphocytes were determined using standard HE staining of about six non-consecutive sections from at least two different tissue samples per patient. The cell density was determined by estimating the density of neutrophils, eosinophils and lymphocytes in high power fields of view (×400) using a semiquantitative score according to a recent publication:25 0 = no cells present, 1 = moderate infiltration, and 2 = strong infiltration. The presence of neutrophils, eosinophils and lymphocytes was separately investigated in the mucosal, submucosal and muscular layers of the colon.

For determination of colon innervation, six to eight cryosections (5 μm thick) of the formaldehyde/sucrose-fixed tissue samples were used. We used non-consecutive sections of two tissue blocks in order to better represent the relative large tissue to be analysed. Immunohistochemistry was done with a primary antibody against tyrosine hydroxylase (TH, the key enzyme for norepinephrine production in sympathetic nerve endings (cat. no. AB152; Chemicon, Temecula, CA, USA) and against SP (cat. no. AB1566; Chemicon). An Alexa 546 conjugated secondary antibody (cat. no. A-121010 against rabbit IgG; Molecular Probes, Leiden, The Netherlands) was used to achieve immunofluorescent staining of sympathetic and SP⁺ nerve fibres (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Representative immunohistochemistry of sympathetic tyrosine hydroxylase nerve-positive fibres (A, B) and substance P-positive nerve fibres (C,D) (arrow heads) in control patients. Micrographs were taken at ×400 magnification.

The numbers of TH⁺ sympathetic and SP⁺ single nerve fibres per square millimetre were determined by averaging the number of stained nerve fibres (minimum length 50 μm, determined through a micrometer eyepiece) in 17 randomly selected high power fields of view (×400). We controlled the positive nerve fibre staining by incubating the tissue with polyclonal control antibodies which always yielded a negative result. We investigated single nerve fibres because they reflect the neuronal innervation of non-neuronal target cells; we did not investigate neuronal plexus elements in the different layers of the colon. For determination of colonic innervation, the density of TH⁺ sympathetic and SP⁺ nerve fibres was separately investigated in the mucosal, submucosal and muscular layers of the colon (in each layer, 17 high power fields). In our hands, for single nerve fibre determination, automatic or semi-automatic image analysis is not feasible.

Development of anti-human SEMA3C polyclonal antibodies

Two potentially antigenic amino acid sequences within the SEMA3C protein were selected using Gene Runner (Hastings Software, Hastings, NY, USA) and DNASIS (Hitachi Software Engineering, San Francisco, CA, USA), two programs for prediction of protein structure and antigenicity. The selected sequences of the two peptides were as follows: QGDESQKMRGDYGK (amino acids 718–731), and ESGKMRGDYGKLKA (amino acids 721–734). The two peptides were synthesised, hapten-conjugated, and a rabbit was immunised with a mixture of the two peptides. Following a typical 87-day immunisation period, we obtained rabbit serum containing SEMA3C antibodies and control serum collected prior to the initial immunisation. The antibodies were precipitated through addition of 45% saturated ammonium sulfate solution (Sigma, Deisenhofen, Germany). The purified SEMA3C antibodies were stored in 0.05 mol/l Tris-buffered saline solution (TBS) with 0.02% sodium azide (Sigma, Deisenhofen, Germany). The antibody was further purified through affinity chromatography. This particular antiserum has been described in detail recently.23

Detection of SEMA3C-positive cells in Crohn’s disease

Cryosections (7 μm) of at least two different formaldehyde-fixed tissue samples from each patient were air-dried for 1 h and then rehydrated in 0.05 mol/l TBS. Unspecific binding sites were blocked with 0.05 mol/l TBS containing 10% fetal calf serum, 10% bovine serum albumin and 10% normal goat serum for 1 h at room temperature. After 3×5 min washes with TBS, the sections were incubated with the polyclonal SEMA3C antibodies for 12–18 h at room temperature. The sections were washed 3×5 min and then incubated with an Alexa 546 conjugated secondary antibody (cat. no. A-121010 against rabbit IgG; Molecular Probes). Control stainings with pre-immunisation serum and with peptide-neutralised SEMA3C antibody solution were carried out in parallel as recently described.23 In this analysis, we found that epithelial cells were the only visible source of SEMA3C (see Results). Thus, we used the following method to quantify epithelial cells of crypts positive for SEMA3C: if epithelial cells of one-third of the circumference of a crypt stained positive for SEMA3C, the crypt was defined as SEMA3C-positive. We used 17 randomly selected high power fields of view (×400) of horizontally cut crypts with an obvious circular shape to calculate the positive crypts in relation to all crypts (unit: percent of all crypts).

In situ hybridisation of SEMA3C in Crohn’s disease

A mixture of three antisense semaphorin 3C mRNA-specific oligodeoxynucleotides (5′-AAAGTGGTCCACTTGTTGACAAGGCTACGCAGTCCACCAGTGTCA-3′, 5′-GAAGGGGTTTCCCAGGTATCTCTGCACCGCTGCCACATCTATGGT-3′, and 5′-CTGTTCAAGTAAGTACAGACAGGACTGAAAGCGCCACTC CCACAG-3′) was labelled with XXXXX (DIG) using the DIG Oligonucleotide Tailing Kit, 2nd generation (Roche, Mannheim, Germany) according to the manufacturer’s instructions. The DIG-labelled oligonucleotides were purified and tested before using for in situ hybridisation.

Tissue samples from patients with Crohn’s disease were sectioned in 10 μm thick slices. These were then hybridised with approximately 20 ng each per section of the oligonucleotide probes for 12–18 h. A detailed description of the hybridisation protocol can be found in the publication Non-radioactive In situ Hybridization Application Manual (Roche, Mannheim, Germany). The hybridisations were visualised through an immunohistochemical (anti-digoxygenin-AP, Fab fragments, Roche) and immunofluorescent (anti-digoxygenin–fluorescein isothiocyanate (FITC), Fab fragments, Boehringer, Mannheim, Germany) staining directed against the DIG label.

Assessment of colon length and histological score in dextran sodium sulfate colitis in mice

Female BALB/c mice (8–10 weeks; 20–22 g; Charles River, Sulzfeld, Germany) were used. The exact numbers for animals used are given in the legend to fig 5. Animals were fed a standard laboratory chow diet.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5 Bimodal influence of the sympathetic nervous system on dextran sodium sulfate (DSS) colitis. (A and B) Tyrosine hydroxylase-positive (TH⁺) sympathetic nerve fibres in the colon wall of a normal mouse (A) and a sympathectomised mouse (B). Sympathetic nerve fibres are lost in sympathectomised animals (B). (C and D) Colon length and histological score in acute DSS colitis without/with sympathectomy (Syx). (E and F) Colon length and histological score in chronic DSS colitis. In each subgroup, at least eight mice were studied in two independent experiments. The small graphs below panels D and F give the time points of sympathectomy induction and sacrifice in relation to DSS treatment. For an explanation of the box boundaries see the legend to fig 2. Co, control animals without sympathectomy; H₂O, water treatment; Syx, sympathectomy.

The model of acute and chronic colitis has been described by our group.26 27 DSS (MW 36 000–44 000) was purchased from ICN (Eschwege, Germany). Acute colitis was induced by feeding DSS for 7 days, whereas chronic colitis was provoked by four cycles of DSS feeding. One cycle consisted of feeding 1.5% DSS in drinking water for 7 days followed by a period of 10 days drinking water without DSS. In the acute model, mice were sacrificed on day 8, whereas in chronic colitis, animals were killed 30 days after the fourth cycle. The colon was removed for further investigation.

Inflammatory reduction of colon length is a well-known feature of chronic DSS-induced colitis and was previously used as a macroscopic parameter to describe the severity of intestinal inflammation.28 For determination of colon length, the colon was removed and measured to 0.5 cm precision. Histological examination was performed by two independent investigators blinded to the source of treatment. The distal third of the colon was removed and used for histological analysis scoring both inflammatory infiltration (0–4) and epithelial cell damage (0–4) (total histological score 0–8) as described previously.29

The model of colitis in interleukin 10 deficient mice

Interleukin 10 (IL10) deficient mice were shown to develop a spontaneous form of colitis.30 31 This model was shown to have many common pathophysiological aspects as compared to human inflammatory bowel disease.30 31 C3H/HeJBir.129P2-Il10^tm1Cgn (Il10^−/−) mice were 16 weeks old when sympathectomy was carried out with three intraperitoneal injections of 6-hydroxydopamine (80 mg/kg body weight) on three consecutive days. Since the onset of intestinal inflammation in these mice occurs at weeks 8–10, colitis was already at the chronic inflammation stage. One week after ablation of the sympathetic nervous system mice were sacrificed in order to investigate histological changes by blinded investigators as mentioned above. Scoring for both inflammatory infiltration (0–4) and epithelial cell damage (0–4) (total histological score 0–8) was carried out as described previously.29

Isolation and stimulation of mesenterial lymph node cells in dextran sodium sulfate colitis

Mesenterial lymph nodes (pooled from each group) from mice were collected under sterile conditions and incubated over 24 h in anti-CD3-coated wells as described before.32 After incubation supernatants were harvested and cytokine levels were measured by enzyme-linked immunosorbant assay (ELISA) (Endogene, Woburne, MA, USA).

Quantitative polymerase chain reaction of tumour necrosis factor and intereron-γ from colon samples in experimental colitis

A colonic tissue specimen (1 cm) from the distal third of the colon was harvested for quantitative real-time polymerase chain reaction (RT-PCR), and RNA was isolated and transcribed as described previously.27 Amplicons were defined by gene-specific oligonucleotide primers (see below). Within these amplicons, an oligonucleotide probe labelled with a reporter dye 6-carboxy fluorescein (6-FAM) for cytokines or VIC for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) covalently linked at the 5′ end, and a quencher dye, 6-carboxy-tetramethyl-rhodamine (TAMRA), was linked to the 3′ end of the probe. The sequences of the primers and fluorigenic probes were designed using Primer Express 1.5 software (PE Applied Biosystems, Forster City, CA, USA) and synthesised by MWG Biotech (Ebersberg, Germany); primers for GAPDH were by PE Applied Biosystems. The optimised reaction mix consisted of qPCR MasterMix (Eurogentec, Köln, Germany), 750 nmol/l of each forward and reverse primer, 250 nmol/l fluorigenic probe, and the cDNA template (50 ng/well) isolated from tissue samples. All reactions were performed in triplicate in a 384-well plate (Abgene, Epsom, UK). All PCR reactions were carried out using the ABI PRISM 7700 Sequence Detection System (PE Applied Biosystems). Cycle conditions were: 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of 95°C for 15 s (melting step), and 60°C for 1 min (anneal/extend step). Fluorigenic probes and primer sets were as follows: for IFN-γ, forward: 5′ TGC TGA TGG GAG GAG ATG TCT 3′; reverse: 5′ TGC TGT CTG GCC TGC TGT TA 3′; probe: 5′CAC TCC GGG CCA GC 3′; for TNF, forward: 5′ CAC AAG ATG CTG GGA CAG TGA 3′; reverse: 5′ TCC TTG ATG GTG GTG CAT GA 3′; probe: 5′ CTG GAC TGT GGG CCT 3′. The amount of reporter dye fluorescence was monitored using Sequence Detector software (SDS version 2.1, PE Applied Biosystems). Reporter dye fluorescence versus PCR cycles was plotted. Fluorescence values of GAPDH were subtracted from those of cytokines.

Presentation of data and statistical analysis

All data are given as means with the SEM. Data in figures are given as box plots (Sigma Plot, V.12.0). Two independent groups were compared by the non-parametric Mann–Whitney U-test (SPSS). Dependent data (inflamed vs non-inflamed in Crohn’s disease) were compared by the non-parametric Wilcoxon signed-rank test for paired data (SPSS), and correlations were calculated by Spearman rank correlation analysis (SPSS). p<0.05 was the level of significance.

RESULTS

Degree of inflammation

Since nerve fibre density may be related to histologically confirmed inflammation,33 we evaluated the density of neutrophils, eosinophils and lymphocytes in the different layers of the colon in patients with Crohn’s disease. A detailed analysis of density of neutrophils, eosinophils and lymphocytes in the mucosa and submucosa clearly demonstrated increased inflammation in Crohn’s disease patients as compared to control subjects (table 2). Such a difference between the two groups was not observed in the muscular layer (table 2). Parameters were not significantly different between the macroscopically determined non-inflamed and inflamed area in Crohn’s disease patients, which indicates that macroscopic judgment by the surgeon or pathologist does not reflect the true character of cellular infiltration (table 2).

Density of sympathetic and SP⁺ nerve fibres

In all layers of the colon of Crohn’s disease patients, the density of sympathetic nerve fibres was markedly lower as compared to control subjects, which was independent of the macroscopically determined inflammatory status in patients with Crohn’s disease (fig 2A–C). In contrast, the density of SP⁺ nerve fibres was markedly increased in the mucosa of Crohn’s disease patients compared to control subjects (fig 2D). No such increase was observed in the submucosal layer (fig 2E). A similar increase of SP⁺ nerve fibres was observed only in the non-inflamed muscle layer of patients with Crohn’s disease (fig 2F).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Density of sympathetic and substance P-positive (SP⁺) nerve fibres in control subjects (Co), patients with Crohn’s disease (CD), and diverticulitis (Div). Patients with Crohn’s disease were subdivided according to macroscopically determined signs of inflammation (CDn: macroscopically non-inflamed, CDi: macroscopically inflamed). The different layers are indicated below the graph. The boundary of the box closest to zero indicates the 25^th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75^th percentile. Whiskers (error bars) above and below the box indicate the 90^th and 10^th percentiles. p-Values given by the Mann–Whitney test are mentioned. SP, substance P; TH, tyrosine hydroxylase (sympathetic).

In patients with diverticulitis, sympathetic nerve fibres were lower than in controls in the submucosa and in the muscular layer but not in the mucosa (fig 2A–C). In diverticulitis, the density of SP⁺ nerve fibres was higher than in controls in the submucosa but not in the mucosa and in the muscular layer (fig 2D–F).

We next studied the direct relation between the two types of nerve fibres by calculating the ratio: density of sympathetic nerve fibres/density of SP⁺ nerve fibres. This analysis demonstrated that in all colon layers this ratio was decreased in Crohn’s disease as compared to control subjects (fig 3A–C). In patients with diverticulitis, this ratio was decreased only for submucosal tissue (fig 3B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Tyrosine hydroxylase-positive (TH⁺) and substance P-positive (SP⁺) nerve fibres and vessels in tissue in controls and patients with colon inflammation. (A to C) Ratio of density of TH⁺ divided by SP⁺ nerve fibres in control subjects (Co), patients with Crohn’s disease (CD), and diverticulitis (Div) in different layers of the colon. Since the differences between macroscopically determined non-inflamed and inflamed areas in Crohn’s disease patients was not statistically significant (see fig 2), the ratio is only given for macroscopically determined non-inflamed tissue (CDn: non-inflamed). For further explanations of this figure see legend to fig 2. (D to F) Submucosal vessels positive for TH (vessels with sympathetic nerve fibres) and density of neutrophils in control subjects (Co) and patients with Crohn’s disease. (D) Density of submucosal vessels positive for sympathetic nerve fibres are given in % of all vessels found. For further explanations of this figure see legend to fig 2. (E and F) Inter-relation between density of mucosal or submucosal neutrophils and percentage of sympathetically innervated vessels in patients with Crohn’s disease. The Spearman rank correlation coefficient and the p-value are given within the panel. Density of neutrophils was not investigated in all Crohn’s disease patients due to technical problems.

Sympathetically innervated vessels and neutrophils

During wound healing, sympathetic innervation of vessels appears in the final phase of wound repair indicating termination of wound healing. Since sympathetic nerve fibres are not present during the neo-angiogenic process, the number of vessels positive for sympathetic nerve fibres might be a good indicator of ongoing inflammation. This investigation was only carried out in Crohn’s disease patients and controls: a lower number of vessels positive for sympathetic nerve fibres was detected in the submucosal tissue as compared to control subjects (fig 3D). This was more pronounced in the macroscopically inflamed area. The number of vessels positive for sympathetic nerve fibres only tended to be decreased in the macroscopically non-inflamed area, which might link the macroscopic judgment of inflammation to the number of vessels positive for sympathetic nerve fibres (vasodilatation vs vasoconstriction). This was not investigated in other colonic layers because vascularity was too sparse. In addition, the density of mucosal or submucosal neutrophils was inversely related to the number of vessels positive for sympathetic nerve fibres (fig 3E,F), which demonstrates that presence of inflammation was related to sympathetic nerve fibre loss.

Expression of SEMA3C and sympathetic innervation

SEMA3C was identified as a sympathetic nerve repellent factor specific for this nerve fibre type.23 Analysis of the immunofluorescence of all colonic layers identified epithelial cells as major source of SEMA3C in patients with Crohn’s disease (fig 4). In some patients, SEMA3C staining was positive throughout the entire axis of the crypt (fig 4D), whereas in most patients SEMA3C staining appeared in the basolateral part of the crypt (fig 4E,F). No other cell type stained positive for SEMA3C in the intestinal wall. Patients with Crohn’s disease, irrespective of macroscopically determined inflammation, demonstrated an increased percentage of SEMA3C-positive crypts (fig 4G). It was remarkable that SEMA3C was almost not present in control tissue and in diverticulitis samples (except one control subject, fig 4G). In order to detect SEMA3C mRNA in the tissue, we applied in situ hybridisation similar to that in a previous study.23 Unexpectedly, we did not detect significant amounts of SEMA3C mRNA in the tissue (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4 Expression of semaphorin 3C (SEMA3C) in the colon of control subjects, patients with Crohn’s disease (CD), and diverticulitis. (A to F) Colonic tissue from a control subject (A) and patients with Crohn’s disease (B to F) are demonstrated. The red staining in panels B and C (arrow heads) demonstrates large vesicles within epithelial cells positive for SEMA3C (blue is the nuclear staining with 2,6-diamido-2-phenylindole (DAPI). (D to F) Crypts cut in a longitudinal direction. Micrographs were taken at ×400 magnification. (G) Comparison of SEMA3C-positive crypts of control subjects (Co), patients with Crohn’s disease, and diverticulitis (Div). Further explanations of symbols and abbreviations are given in legend to fig 2. (H) Inter-relation of density of tyrosine hydroxylase-positive (TH⁺) sympathetic nerve fibres and percentage of SEMA3C-positive crypts. Black triangles (black diamonds) demonstrate Crohn’s disease patients with macroscopically determined non-inflamed (inflamed) tissue and white circles reflect control subjects without inflammatory conditions (patients with diverticulitis are not included). The vertical and horizontal lines demonstrate the mean of all patients of TH-positive nerve fibres/mm² and the mean of all patients of SEMA3C-positive crypts, respectively. These lines demonstrate the four quadrants of parameter distribution.

In a further analysis comprising all investigated subjects, the relation of density of tyrosine hydroxylase-positive sympathetic nerve fibres and density of S3C-positive crypts was tested (fig 4H). Patients with Crohn’s disease (black symbols) irrespective of macroscopically determined inflammation are at the lower end of the x-axis (sympathetic nerve fibres) and had an increased percentage of SEMA3C-positive crypts, whereas control subjects had higher densities of sympathetic nerve fibres and lower densities of SEMA3C-positive crypts (white symbols) (patients with diverticulitis not included) (fig 4H).

The bimodal role of the sympathetic nervous system in acute and chronic DSS colitis

In order to understand the role of the sympathetic nervous system in colonic inflammation, we subjected mice with acute and chronic colitis to chemical sympathectomy with 6-hydroxydopamine (6-OHDA), 80 mg/kg body weight, intraperitoneally in sodium chloride on three consecutive days; sympathectomy was carried out 10 days before induction of acute DSS colitis and 20 days after the last cycle in chronic DSS colitis; the technique was recently described by Härle et al7). Mice subjected to chemical sympathectomy demonstrated a severe loss of sympathetic nerve fibres in the entire wall of the colon (compare fig 5A with 5B). Small numbers of sympathetic nerve fibres were found only in the outer layer of the muscularis and the adventitia along vessels (fig 5B).

In acute DSS colitis, sympathectomised animals demonstrated a less pronounced inflammatory reduction of colon length as compared to non-sympathectomised animals, which indicates a better outcome in sympathectomised mice (fig 5C). Similarly, the blinded histology score was markedly lower in sympathectomised than in control mice (fig 5D). In marked contrast, in chronic DSS colitis, sympathectomy led to an exacerbation of colitis as demonstrated by a further reduction in colon length and a higher histology score as compared to control mice (fig 5E,F). This clearly demonstrates a bimodal influence of the sympathetic nervous system on colon inflammation.

As the chronic DSS model is more relevant for studies related to inflammatory bowel disease (IBD) than the acute DSS model, we additionally analysed pro-inflammatory cytokine secretion from draining mesenteric lymph node cells in chronic colitis. In fact, along with the exacerbation of colitis pro-inflammatory cytokine secretion from anti-CD3 activated mesenteric lymph node cells was significantly increased in chronic DSS colitis after sympathectomy as compared to non-sympathectomised conditions: We found an approximately 4-fold increase in IL6 [with sympathectomy: 87.4 (SD 8.3) pg/ml vs without sympathectomy: 20.1 (SD 2.4) pg/ml, p = 0.03] and a 2-fold increase in IFN-γ secretion [32 957 (SD 5561) pg/ml vs 15 146 (SD 3646) pg/ml, p = 0.03]. In addition, quantitative PCR of TNF and IFN-γ from colon samples demonstrated a marked increase of TNF and IFN-γ mRNA in sympathectomised compared to control animals (fig 6A). This argues for a strong immune activation in this chronic model of intestinal inflammation induced by the lack of sympathetic nervous influence.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6 Colonic cytokine mRNA content in dextran sodium sulfate (DSS) colitic and Il10^−/− colitic mice, and histological score of interleukin 10 (IL10) deficient animals subjected to sympathectomy. (A and B) Content of mRNA for tumour necrosis factor (TNF) and interferon-γ (IFN-γ) in control mice (Co) and sympathectomised mice (Syx) with either chronic DSS colitis (A) or chronic colitis on the basis of IL10 deficiency (B). Colon material of at least six mice was used for these analyses. NS, not significant. (C) Blinded histological score of IL10 deficient animals subjected to control treatment (Co: sodium chloride i.p.) or sympathectomy (Syx). In each subgroup, six IL10 deficient mice were studied 1 week after intervention.

The role of the sympathetic nervous system in colitis of IL10 deficient mice

In order to confirm our results in another model of chronic experimental colitis, we used 16-week-old IL10 deficient mice. In these mice, ablation of the sympathetic nervous system increased inflammation as objectified by TNF mRNA content in the colon and by blinded histology (fig 6B,C). Colon levels of IFN-γ mRNA were not different in this model; however, the levels were very low in relation to control GAPDH mRNA (fig 6B).

DISCUSSION

This study revealed a marked preponderance of SP⁺ nerve fibres in relation to TH⁺ sympathetic nerve fibres in the different layers of the colon in patients with Crohn’s disease. This was particularly obvious in the mucosa and it was accompanied by a marked increase of SEMA3C expression in epithelial cells. In addition, this study demonstrates that the sympathetic nervous system confers a pro-inflammatory effect in acute colitis but a marked anti-inflammatory effect in chronic DSS colitis. In addition, sympathectomy increased inflammation in Il10^−/− mice with chronic intestinal inflammation.

To our knowledge, there has only been one study that demonstrated low numbers of sympathetic nerve fibres in the circular muscle layer of the ileum, which was not markedly different in control subjects as compared to Crohn’s disease patients.34 These authors found a high frequency of enlarged varicosities in the myenteric ganglia and/or nerve fibres of the circular muscle layer of Crohn’s ileum,34 which is not in contrast to our findings because the cytoplasm of repelled mucosal and submucosal nerve fibres might be taken up by myenteric neurons. Thus, small nerve fibres can be lost whereas myenteric neurons might become enlarged. Unfortunately, in this detailed study the sympathetic innervation of the submucosa and the mucosa was not investigated. Dvorak and Silen35 demonstrated axonal loss of autonomic nerves in the surgically resected ileum of patients with Crohn’s disease. They suggested that affected axons belong to the sympathetic nervous system but they did not use specific staining. Others have demonstrated that gut infection with Toxoplasma gondii resulted in colonic pseudo-obstruction due to selective sympathetic denervation.36 In a recent study, we demonstrated a marked loss of sympathetic nerve fibres in all layers of the colon in the DSS colitis in mice using immunohistochemical and functional tests.12 This present study confirmed the loss of TH⁺ nerve fibres in all layers of the human colon in patients with Crohn’s disease and in the submucosa and muscular layer in patients with diverticulitis. It also demonstrated that the number of submucosal vessels positive for TH⁺ nerve fibres was low in inflamed tissue of Crohn’s disease patients as compared to control subjects.

A disadvantage of our study is the fact that we can not be absolutely certain that the TH⁺ fibres investigated are extrinsic sympathetic fibres (with the main neurotransmitter norepinephrine) because we did not perform double staining of TH with dopamine β-hydroxylase (DBH). Thus, a proportion of the nerve fibres might be intrinsic dopaminergic nerve fibres, which lack DBH and produce mainly dopamine.37^–39 Nevertheless, some of the TH⁺ nerve fibres detected are certainly sympathetic and extrinsic in nature. Since the detection problem is probably similar in controls compared to Crohn’s disease patients, the functional consequences would be decreased release of norepinephrine in Crohn’s disease patients compared to controls, and this is most probably pro-inflammatory in the chronic phase of the disease.

Since neurotransmitters of the sympathetic nervous system exert marked anti-inflammatory effects at high concentrations and SP has strong pro-inflammatory effects,4^–6 8^–10 differential analysis of both types of nerve fibre, by using the same techniques, might shed light on the balance of these two pathways. To our knowledge, no similar comparative study has been carried out. In contrast to sympathetic nerve fibres, SP⁺ nerve fibres sprout into inflamed tissue as described in rheumatoid arthritis and inflammatory bowel diseases.3 11 Since sympathetic neurotransmitters normally inhibit release of SP, the preponderance of the SP⁺ nerve fibre system over the sympathetic system is likely an additional inflammatory factor. This study showed the preponderance of SP⁺ nerve fibres in relation to sympathetic nerve fibres in all layers of the colon in Crohn’s disease and in the submucosa of patients with diverticulitis. The question that arises concerns why this differential behaviour of sympathetic and SP⁺ nerve fibres can appear.

In searching for a possible reason for the differential loss of sympathetic nerve fibres in relation to SP⁺ nerve fibres, we focused on nerve repellent factors, which are specific for the different types of nerve fibre.18^–20 It has been demonstrated that a specific group of nerve repellent factors exerts repulsive actions on sympathetic nerve fibres only; these are the semaphorins 3B, 3C and 3F (reviewed by Chen et al20 and Mark et al22). In contrast, sensory SP⁺ nerve fibres are repelled by semaphorin 3A (also reviewed by Chen et al20 and Mark et al22). We recently demonstrated that macrophages and fibroblasts in inflammatory lesions produce nerve repellent factors specific for sympathetic but not for SP⁺ nerve fibres.23 This present investigation revealed, for the first time, that epithelial cells within crypts are positive for the nerve repellent factor SEMA3C, which was not found in other cell types in the mucosa, submucosa and muscular layers. This was only present in patients with Crohn’s disease but not in patients with diverticulitis or tumour controls. We applied in situ hybridisation to detect SEMA3C mRNA in the colon. However, in situ hybridisation did not reveal SEMA3C mRNA staining, which was unexpected on the basis of substantial protein staining in epithelial cells. However, it might well be that SEMA3C mRNA has a short half-life and is rapidly translated to SEMA3C protein so that the intracellular levels of mRNA always remain low and were not detectable by the method we used.

At present, our study gives no direct proof that sympathetic nerve fibres are lost because the measured decrease could be due to degeneration of nerve fibres (as neurodegenerative processes have been described in Crohn’s disease and animal models of inflammatory bowel disease), downregulation of tyrosine hydroxylase expression by inflammation, or nerve fibre repulsion by SEMA3C. Nevertheless, in Crohn’s disease, we found a very similar inter-relation between density of sympathetic nerve fibres and expression of SEMA3C as compared to the inflamed tissue in patients with arthritis (fig 4H).23 This suggests that repellent factors and decreased density of nerve fibres might be causally linked. Specific repulsion of sympathetic nerve fibres but not of SP⁺ nerve fibres would lead to an overall pro-inflammatory microenvironment. The question remains, whether in a situation with low numbers of sympathetic nerve fibres in chronic colitis, a further loss of sympathetic innervation would lead to an even more pro-inflammatory situation.

This question can only be studied in an animal model of experimental colitis. This study demonstrates for the first time in two different chronic colitis models (DSS colitic mice and IL10 deficient colitic mice) that elimination of the sympathetic nervous system led to a higher degree of colonic inflammation. In view of these data, it seems that the remaining sympathetic nerve fibres might still have an anti-inflammatory influence on the local immune process. A very similar finding was recently reported in chronic experimental arthritis.7 In contrast, abrogation of the sympathetic nervous system before inducing acute colitis conferred an anti-inflammatory effect, which indicates that the sympathetic nervous system exerts pro-inflammatory effects at the beginning of the disease. Again, these findings in experimental colitis corroborate a recent study in experimental arthritis.7 This clearly demonstrates that the sympathetic nervous system exerts pro-inflammatory effects at the beginning of tissue inflammation while it confers anti-inflammatory effects in the chronic phase of inflammation.

In conclusion, the preponderance of SP⁺ nerve fibres over sympathetic nerve fibres yielding high amounts of the pro-inflammatory neuropeptide SP over sympathetic neurotransmitters is most probably a pro-inflammatory signal in chronic Crohn’s colitis. The presence of nerve repellent factors such as SEMA3C probably contributes to this differential loss of sympathetic nerve fibres in the inflamed colon. Whether modulation of SEMA3C expression might be a favourable therapeutic option in Crohn’s colitis remains to be determined. The animal model of DSS and Il10^−/− colitis corroborates an anti-inflammatory effect of the sympathetic nervous system in chronic inflammation.