SEARCH

SEARCH BY CITATION

Keywords:

  • diverticula;
  • enteric nerves;
  • galanin;
  • substance P;
  • symptoms

Abstract

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

Abstract:  Some patients with colonic diverticula suffer recurrent abdominal pain and exhibit visceral hypersensitivity, though the mechanism is unclear. Prior diverticulitis increases the risk of being symptomatic while experimental colitis in animals increases expression of neuropeptides within the enteric nervous system (ENS) which may mediate visceral hypersensitivity. Our aim was to determine the expression of neuropeptides within the ENS in diverticulitis (study 1) and in patients with symptomatic disease (study 2). Study 1 – Nerves in colonic resection specimens with either acute diverticulitis (AD, n = 16) or chronic diverticulitis (CD, n = 16) were assessed for neuropeptide expression recording % area staining with protein gene product (PGP9.5), substance P (SP), neuropeptide K (NPK), pituitary adenylate cyclase activating polypeptide (PACAP), vasoactive intestinal polypeptide (VIP) and galanin. Study 2 – Seventeen symptomatic and 15 asymptomatic patients with colonic diverticula underwent flexible sigmoidoscopy and multiple peridiverticular mucosal biopsies. Study 1– Neural tissue, as assessed by PGP staining was increased to a similar degree in circular muscle in both AD and CD. The CD specimens showed significant increases in the immunoreactivity of SP, NPK and galanin in both mucosal and circular muscle layer compared with controls. Study 2 – Mucosal histology was normal and PGP9.5 staining was similar between groups however patients with symptomatic diverticular disease demonstrated significantly higher levels of SP, NPK, VIP, PACAP and galanin within the mucosal plexus. Patients with symptomatic diverticular disease exhibit increased neuropeptides in mucosal biopsies which may reflect resolved prior inflammation, as it parallels the changes seen in acute and chronic diverticulitis.


Introduction

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

As our population ages, diverticulosis, which is found in two-thirds of those aged over 65 years in the United Kingdom, becomes an increasingly important cause of morbidity (64 289 hospital admissions) and mortality (3662 deaths) in 2004 (Department of Health Hospital Episode Statistics; http://www.dh.gov.uk). However this is only part of the picture, as many more who are never hospitalized will suffer chronic abdominal pain and disordered bowel habit without the abscesses, perforation or bleeding which account for many of the hospital admissions. Interestingly only 15% of individuals with colonic diverticula are symptomatic.1 What determines who develops symptoms is not currently understood, as there is a poor correlation between symptoms and the apparent severity as judged by radiological appearances.

A previous survey of symptoms in patients with diverticulitis attending University Hospital Nottingham, showed that previous episodes of presumed acute diverticulitis doubled the risk of recurrent, often daily, transient (1–3 h) bouts of pain.2 A follow up study using multivariable analysis identified a history of acute diverticulitis (odds ratio 3.98) and a raised score on the Hospital Anxiety and Depression Scale (odds ratio 2.40) as the best predictors of recurrent pain.3 Several recent studies have documented that pain in diverticulitis is related temporally to contractions in the sigmoid colon, particularly high pressure propagated contractions (HAPCs).4,5 This is most obvious after a meal,6 when many patients experience pain. However, there is also an element of visceral hypersensitivity, as HAPCs of similar amplitude do not evoke pain in controls.5 Recent studies using rectal balloon distension confirm visceral hypersensitivity in symptomatic but not asymptomatic diverticular disease.7

These clinical observations have been supported by numerous animal studies which show long lasting visceral hypersensitivity associated with pronounced circular muscle hypertrophy following induction of colitis with the hapten, trinitrobenzene sulphonic acid (TNBS), an effect which can last more than 100 days.8,9 Given the similarities with the pathology in diverticular disease which also exhibits muscular hypertrophy and hypersensitivity, we hypothesized that in painful diverticular disease postinflammatory changes in enteric nerves and muscles contribute to the development of symptoms.

Previous studies in ulcerative colitis and Crohn’s disease suggest postinflammatory alterations in tackykinins and VIP.10–14 Our own studies using the animal model of TNBS colitis have shown that this induces a remodelling of enteric nerves and a change in the neurochemical coding with increased expression of a number of markers of neural injury including tachykinins, pituitary adenylate cyclase activating polypeptide (PACAP), vasoactive intestinal peptide (VIP) and galanin in both mucosa and myenteric plexus.15 These mucosal changes gradually returned to normal over 14 weeks apart from a persistent elevation of mucosal galanin. However the increase in myenteric and mucosal tachykinins substance P (SP) and neuropeptide K (NPK) persisted for at least 14 weeks as did the increase in circular muscle thickness. These studies suggested that mucosal tachykinins and galanin might be good mucosal markers of both previous postinflammatory neural damage and persistent increase in myenteric tachykinins, which might in turn relate to visceral hypersensitivity as recent animal studies have suggested.16–18

As in most cases patients do not come to colonic resection, one of the aims of our study was to validate the use of the more readily available mucosal markers as an indication of neuropeptide changes within the myenteric plexus. With this aim in mind, we undertook two studies. The first used colonic specimens from patients undergoing colonic resection for both acute and chronic diverticular disease and examined how the changes observed in myenteric plexus related to those in the mucosa. The second part of our study compared mucosal biopsies from patients with symptomatic and asymptomatic diverticular disease for the expression of the neuropeptides, SP, NPK, PACAP, VIP and galanin which we had previously shown to change their expression after acute inflammation.

Methods

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

Study 1

Resection specimens from patients with colonic diverticula were studied. Our control group, Group A had undergone a colonic resection for a non-obstructing colonic malignancy or polyp. Group B (acute diverticulitis; AD) had undergone emergency colonic resection for acute perforated diverticular disease with associated peritonitis. Sections demonstrating diverticulitis were selected. Group C (chronic complication of diverticular disease; CD) had undergone an elective colonic resection for a chronic inflammatory complication of diverticular disease, namely stricture, fistulation or abscess formation.

Study 2

Patient selection  Patients were invited to participate in Study 2 based on their initial symptoms recorded as part of our previous symptom survey.2

Flexible sigmoidoscopy  Four peri-ostial biopsies were obtained with standard endoscopic forceps (FB-13K-1, Olympus, Japan) from the affected diverticular segment and three mid rectal biopsies.

Samples obtained were fixed for 24 h in 10% formalin and then embedded in paraffin.

Assessment of inflammation  The sections were dewaxed and underwent immunoperoxidase staining using an automated stainer with 3′3 diaminobenzidine tetrahydrochloride as the chromagen. The sections were stained for mast cells, T lymphocytes and enterochromaffin cells using primary antibodies to mast cell tryptase (Dako, Glostrup, Denmark), serotonin (Dako) and CD3 (Dako). These cells were then counted.

Assessment of enteric nerves  Morphological markers of enteric nerves [PGP9.5 and neurofilament protein (NFP)] were assessed using the avidin–biotin method. Neuropeptides (SP, NPK, PACAP, VIP and galanin) were stained using fluorescent antibodies following the same antigen retrieval process. The neuronal location of the functional neuropeptides in all experiments was confirmed by performing double labelling experiments on selected slides using combination of Texas red and fluorescein conjugated secondary antibodies to identify the neuropeptide stain under study and a selected morphological marker, either PGP9.5 or NFP (Fig. 1).

image

Figure 1.  (A, B, C) double label staining of myenteric plexus for Neuropeptide K (NPK) and Neurofilament protein (NFP). Texas red conjugated fluorescent antibodies to Neuropeptide K were used to stain nerves in A and fluorescein conjugated antibodies to NFP in B. The images have been combined in C. As can be seen, NPK staining is confined to linear neural structures.

Download figure to PowerPoint

Quantification of immunostaining  All sections were examined using the ×40 objective (final magnification ×400) using a Leica DMLB microscope (Leica, IL, USA). As the image was viewed on the computer screen the lamina propria or circular muscle was highlighted as the area of interest. The results were expressed as the percentage area of axonal staining. Sections stained with immunofluorescent antibodies were assessed using an Olympus BX60 fluorescence microscope and digital camera (Olympus, Hamburg, Germany). Immunofluorescence was highlighted with Microsoft Paint software (Microsoft Corporation, Redmond, WA, USA) and analyzed using an identical method to the ABC method with SigmaScan Pro Version 5.0.0 software (SPSS Inc., Chicago, IL, USA). Results were expressed as the percentage area axonal staining of the circular muscle or lamina propria. Nerve diameter was assessed using PGP9.5 stained transverse sections of longitudinal muscle. The number of nerves in the selected area were counted and results expressed as the median number of nerves per unit area (0.2 mm2 each). Nerve angulation was assessed in PGP stained transverse sections of circular muscle using analySIS Pro version 3.1 software (Soft Imaging Systems, Münster, Germany) in which a straight line was drawn along the long axis of the muscle and used as a reference point. Colonic wall thickness was measured using a digital camera JVC KY-F55 (Victor Company of Japan, Japan) and haematoxylin and eosin (H and E) stained slides. For the resection specimens the circular muscle thickness was measured within 5 mm of either side of the neck of the diverticula where the muscle thickness was uniform. Conventional histology was performed on H and E stained, paraffin-embedded sections derived from mucosal biopsies obtained at flexible sigmoidoscopy which were assessed using standardized published criteria.19 The archived diverticular disease specimens were graded macroscopically and/or microscopically using a previously described scoring system ranging from 0 (normal) to 5 (extensive transmural inflammation and inflammation involving the entire section).20

Statistical analysis21,22

For study 2, in order to detect a 40% difference at a 5% level of significance with a power of 90% a sample size of at least 12 per group was required.

Results

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

Study 1

From the histopathology database, 416 specimens met the coding criteria. Following study of the histopathology reports, and assessing the availability and quality of these archived specimens, 102 were found to be suitable for the study, 44 group A, 30 group B and 28 group C. Sixteen specimens for each group were selected at random to be studied and the clinical diagnosis was confirmed by cross referencing with patient case notes.

Histological changes

As expected, there was a spectrum of changes seen on the H & E sections. The microscopic damage score in Group A specimens was 0 (normal) compared with 4.7 ± 0.26 (P < 0.0001) in Group B specimens which demonstrated acute transmural inflammation with marked ulceration. Group C specimens demonstrated a spectrum of histological changes from minor focal inflammation to chronic inflammatory changes with fibrosis, microscopic damage score being 2.1 ± 0.43, P < 0.0001 compared with both Groups A and B. There was also significant muscle thickening in all the diseased specimens, the circular muscle thickness (μm) being 2202 ± 182 for Group A, 4519 ± 308 and 2977 ± 208, respectively for Groups B and C. The difference in thickness between groups B and C vs A were significant at P < 0.0001 and P < 0.01 respectively and between group B and C at P < 0.001 (anova followed by Tukey’s post hoc analysis).

Immunohistochemistry

Morphological markers: PGP9.5 and NFP staining were identified in the mucosal, submucosal and myenteric plexi of the ENS. As a result of the severe mucosal ulceration and loss of architectural integrity within the acutely inflamed specimens, accurate immunohistochemical measurement in this area was not possible in the AD group. However, when comparing CD with control mucosa, there was a significantly increased expression of both the morphological markers PGP9.5 and NFP (Table 1, Fig. 2A,B). Within the circular muscle, a similar rise was seen in the CD specimens compared with controls. A rise in immunoreactivity was also seen in the muscular layer of the AD specimens although with NFP this rise was higher than that seen in the corresponding CD group. In the control specimens, the majority of the nerves within the circular muscle layer were orientated along the long axis of the muscle. This pattern was lost in the AD and CD specimens and the nerve processes appeared more disorganized as shown by the median (inter-quartile range) values for nerve angulation which were 3.3(2.3–4.8), 12.9(10.0–19.1) and 17.4(9.2–20.0) degrees for the controls, AD and CD respectively, both differences from control being significant at P < 0.0001 (Fig. 3A,B). The median nerve diameter (μm) in the longitudinal muscle of the AD group at 4.1(2.9–5.6) was significantly less than both the CD specimens a 5.1(3.5–6.6) and control values of 4.8(3.7–6.1), both differences P < 0.0001. Although there was a significant increase in density/0.2 mm2 in the AD specimens, compared with controls 25(22–28) vs 16(14.3–18), P < 0.0001, the higher density in the CD specimens 29(26–31) was significantly raised over both these groups, P < 0.003 and P < 0.0001 respectively.

Table 1.   Staining for morphological and neuropeptide markers in full thickness sections in patients with resections for acute diverticulitis and chronic diverticular complications compared to controls in Study 1
 GroupSignificance* (P)
A (control)B (acute diverticulitis; AD)C (chronic complication of diverticular disease; CD)
  1. % area staining. Median (inter-quartile range are shown) *anova Kruskal–Wallis, Mann–Whitney U-test for paired comparisons. a< 0.0001 compared with Group A. b= 0.004 compared with Group A. c= 0.001 compared with Group A. d= 0.01 compared with Group B. e= 0.018 compared with Group A. fP = 0.047 compared with Group B. g= 0.001 compared with Group B. hP = 0.086 compared with group A. i= 0.05 compared with Group B. j< 0.0001 compared with Group B.

Mucosa
 Morphological Markers
  PGP0.08 (0.075–0.115)n/a0.2 (0.183–0.210) <0.000
  NFP0.035 (0.012–0.057)n/a0.143 (0.119–0.223)0.002
 Neuropeptide Markers
  SP0.031 (0.019–0.072)n/a0.079 (0.06–0.108)0.032
  NPK0.054 (0.042–0.07)n/a0.096 (0.073–0.125)0.002
  PACAP0.008 (0.004–0.159)n/a0.009 (0.033–0.03)0.589
  VIP0.17 (0.132–0.211)n/a0.167 (0.135–0.259)0.593
  Galanin0.008 (0–0.033)n/a0.12 (0.098–0.18)<0.000
Muscle
 Morphological Markers
  PGP1.31 (0.93–1.72)2.5 (1.83–3.09)a3.3 (2.16–3.83)a<0.000
  NFP0.18 (0.15–0.25)0.99 (0.73–1.13)b0.56 (0.47–0.66)c,d<0.001
 Neuropeptide Markers
  SP0.01 (0–0.14)0.11 (0–0.37)0.41 (0.13–0.37)e,f0.038
  NPK0.007 (0.003–0.009)0.01 (0.002–0.017)0.02 (0.017–0.035)a,g<0.000
  PACAP0.005 (0.003–0.01)0.002 (0–0.005)h0.005 (0.003–0.021)i0.089
  VIP0.102 (0.07–0.12)0.094 (0.066–0.136)0.205 (0.177–0.243)a,j<0.000
  Galanin0.02 (0.003–0.03)0.07 (0.56–0.09)a0.07 (0.05–0.11)a<0.000
image

Figure 2.  (A and B) showing mucosal nerves stained for PGP9.5 from Study 1 in controls (A) and from patients with a chronic complication of diverticular disease (B).

Download figure to PowerPoint

image

Figure 3.  (A and B) showing the orientation of myenteric nerves stained with PGP9.5 in Study 1 in controls (A) and from patients with acute diverticulitis (B). Measurement of nerve angulation is measured in the lower panel. The dotted line represents the long axis of the muscle and the solid lines show the orientation of myenteric nerves. Angulation to the long axis of the muscle was recorded.

Download figure to PowerPoint

Neuropeptides (Table 1)

The CD specimens showed significant increases in the immunoreactivity of SP, NPK and galanin in both mucosal and circular muscle layer compared with controls. Vasoactive intestinal polypeptide immunoreactivity was also significantly increased compared with controls, but only in the circular muscle layer. Although there were rises in PGP9.5 and NFP in the circular muscle of AD specimens, galanin was the only neuropeptide found to demonstrate such a rise in this part of the results (0.02% in controls vs 0.07% in AD specimens, P < 0.0001).

Expressing the neuropeptide stains as a % of the morphological marker PGP9.5 allows detection of specific upregulation of the different markers. When this is done in the mucosa, only galanin showed a specific increase while other markers such as VIP and PACAP actually declined (Table 2, Fig. 4). By contrast in the muscle layer, there was a specific increase only in tachykinin staining (SP and NPK) in group C when compared with control values. There was also a specific rise in galanin within the circular muscle in the AD group and a significant fall in VIP and PACAP but these all were normal in Group C.

Table 2.   Neuropeptide markers staining in full thickness sections in patients with resections for acute diverticulitis and chronic diverticular complications compared with control in Study 1. Results expressed as % of PGP9.5 staining
 GroupSignificance* (P)
A (control)B (acute diverticulitis)C (chronic complication of diverticular disease)
  1. ND, not done as assessment unreliable owing to severe mucosal damage. *P-values for muscle stains are expressed as overall P-value using Kruskal–Wallis (K–W) non-parametric analysis of variance. a= 0.048 compared with Group A. b= 0.04 compared with Group B. c= 0.002 compared with Group A. d,eP < 0.0001 compared with Groups A & C. f= 0.006 compared with Group A.

Mucosa
 SP37.2% (22.8–86.4)ND39.3% (29.9–53.8)0.945
 NPK64.5% (49.9–83.7)ND47.9% (36.3–62.4)0.027
 PACAP9.7% (4.2–19.1)ND4.5% (1.6–14.9)0.297
 VIP203% (158.7–253.6)ND83.4% (67.4–129.7)0.000
 Galanin9.0% (0–38.9)ND60.1% (49.2–90.2)0.01
Muscle
 SP0.39% (0–3.3)0.42% (0–1.8)1.2%a (1.0–5.0)0.083
 NPK0.53% (0.23–0.7)0.38% (0.06–0.68)0.68%b (0.52–1.1)0.085
 PACAP0.37% (0.24–0.75)0.09%c (0–0.21)0.14% (0.1–0.64)0.01
 VIP7.8% (5.4–8.8)3.7%d,e (2.6–5.4)6.1% (5.3–7.3)0.000
 Galanin1.5% (0.2–2.2)2.9%f (2.2–3.6)2.1% (1.6–3.1)0.015
image

Figure 4.  Representative slides of galanin staining from study 1, (A: Mucosa, B: Muscle).

Download figure to PowerPoint

Study 2

Symptoms and clinical features of symptomatic and asymptomatic patients Demographics: There were 15 patients (nine females) in the asymptomatic group and 17 (12 female) in the symptomatic group. Their respective median ages were not significantly different at 73 (range 46–85) and 66 (45–78) years. No changes in their symptom status were noted at the pre-endoscopy interview compared with their initial postal questionnaire.

Chronic recurrent pain: Symptomatic patients replied yes to the question ‘do you experience recurrent short lived abdominal pain as part of your normal bowel habit’ while asymptomatic patients in contrast reported no episodes of abdominal pain, recurrent or prolonged. Symptomatic patients reported frequent pain on a median of 15 (2–31) [IQR7-25] days per month, lasting a median of 6 (1–12) [IQR2.5–10] hours.

Prolonged painful episodes: Over the preceding year, all the symptomatic patients also described episodes of pain lasting for one day or longer. The median duration of the longest attack each patient experienced was 4.5 (1–14) [IQR2.75–8.75] days. During these attacks the most common bowel disturbance was loose stools with a median frequency of defecation per day of 3 (1–6) [IQR1.75–4.25]. Bloating (77%), rectal bleeding (53%), fever (47%) and passage of mucus per rectum (47%) were also common. Emergency medical attention had been sought by 12/17 (70.6%) patients during these bouts a combined total of 27 times, and in 47% of patients, fever was noted. Six had called out their GP (10 episodes) and eight had visited the surgery (17 episodes). On nine of the 27 consultations, antibiotics had been prescribed for the pain. An additional three episodes resulted in referral to hospital where inpatient medical treatment for acute diverticulitis was received. The length of hospital stay was 1, 4 and 4 days. Hospitalization or antibiotic prescription was considered to indicate a presumptive diagnosis of acute diverticulitis. Using these criteria, 12 individuals were considered to have experienced acute diverticular inflammation in the last year. The incidence of these bouts of prolonged pain in the previous year was not related to the number of diverticula. Furthermore, the patients stated that these prolonged episodes were entirely different in character to the recurrent sharp pains experienced as part of their normal bowel habit.

Symptoms of bowel dysfunction  As expected, symptomatic patients experienced significantly more chronic bowel symptoms distinct from the acute episodes described above. These included significantly more loose stools [4/15 (26.6%) asymptomatic patients vs 10/17 (58.8%) symptomatic patients, P < 0.023 Mann–Whitney U-Test], bloating [4/15 (26.6%) vs 11/17 (64.7%), P < 0.031] sense of incomplete evacuation [3/15 (20%) vs 10/17 (58.8%), P < 0.026] and urgency [1/15 (6.7%) vs 6/17 (35.3%), P < 0.051]. They also had significantly more bowel movements per day, median (range) 2(0.6–4), compared with controls who had 1(1–4), P < 0.02 (Mann–Whitney U-Test).

Flexible sigmoidoscopy visit: The procedure was carried out in all patients without complication and confirmed the presence of sigmoid diverticula. Although some views were restricted due to the presence of faeces in the unprepared bowel, no gross abnormal findings were noted. In particular, there was no evidence of inflammation. At this time full blood count, and markers of inflammation including white cell count, erythrocyte sedimentation rate and C-reactive protein were within the normal range in all cases.

Histology: Using conventional criteria19 all H and E stained biopsies were reported by an experienced pathologist (DJ) as being within the normal range. All the antibodies used demonstrated satisfactory even staining intensity throughout the lamina propria.

Quantitative histology showed no difference in mast cells or T lymphocytes but a slight reduction in 5HT staining enterochromaffin cells in the symptomatic patients (asymptomatic patients 2.2 [1.2–3.5], (median, IQR) cells per high powered field (hpf) vs 0.4 [0.2–2.00] cells per hpf; P < 0.006 Mann–Whitney U-test).

Difference in neural staining between symptomatic and asymptomatic patients  The % area staining of PGP9.5 and NFP in the lamina propria was found to be equal in both groups studied (see Table 3). However, striking differences were found in the staining for the neuropeptides under study. The level of VIP, galanin, SP, PACAP and NPK staining was significantly increased within the mucosal nerves of the symptomatic compared with asymptomatic patients when expressed by the % area staining of lamina propria. In the combined study population, there was no correlation between galanin and SP expression and frequency of pain. However, there was a significant correlation between the expression of galanin in the mucosal nerves and the frequency of defecation. (Spearman ρ = 0.38, = 0.016).

Table 3.   Staining for morphological and neuropeptide markers in mucosal biopsies in Study 2
StainAsymptomaticSymptomaticMann– Whitney P
  1. Values expressed are % area staining of lamina propria; Median [inter-quartile range] are shown.

Morphological markers
 PGP0.112 [0.081–0.132]0.095 [0.075–0.142]1.0
 NFP0.033 [0.018–0.044]0.046 [0.029–0.052]0.122
Neuropeptides
 SP0.053 [0.03–0.073]0.13 [0.11–0.20]<0.0001
 NPK0.037 [0.027–0.048]0.118 [0.075–0.139]<0.0001
 PACAP0.012 [0–0.016]0.036 [0.021–0.056]<0.0001
 VIP0.085 [0.032–0.186]0.210 [0.1–0.35]<0.0001
 Galanin0.004 [0–0.02]0.162 [0.091–0.219]<0.0001

Discussion

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

Our studies clearly show important alterations in the enteric nervous system which can be related to the presence of symptoms in patients with diverticulitis. We report for the first time, a selective increase in a range of neuropeptides, the most notable increase being shown by galanin and the tachykinins. Importantly, in the chronic phase of the illness, the total amount of nerves as shown by staining for the pan-neuronal markers PGP 9.5 and neurofilament protein were unchanged in the mucosa, while levels of galanin were increased nearly 10-fold in symptomatic patients.

One potential weakness of using the ubiquitin carboxyl-terminal hydrolase, PGP9.5 as a pan-neuronal marker is that it may also be expressed by non-neural cells including enteroendocrine cells, cutaneous wound fibroblasts23 and, following injury, glial cells, at least in the transected sciatic nerve.24 We excluded the epithelial layer so enteroendocrine cells were not included but fibroblasts might have been included in the lamina propria. It was for this reason that we also used NFP, which we found to only stain linear structures which co-localized with neuropeptides and are highly likely to represent nerves. Neurofilament protein is quite different structurally and functionally from PGP 9.5, though as we and others have found, it is present in very small amounts in normal mucosa.25 However, our finding of a similar increase in the two different markers gives us confidence that the increase in neural tissue is genuine.

Interpretation of the significance of these findings is considerably aided by reference to the first part of the study in which full thickness sections of the diverticular containing colon were available. These show a close parallel between the neuropeptide changes in the mucosa and that within the circular muscle layer. As these were archival material, inevitably the clinical history was limited to what was recorded in the notes and thus linkage to clinical features could not reliably be performed. However, the main point of this part of the study was to establish the link between mucosal and deeper changes in the muscular layers. We were able to show that substance P and neuropeptide K were significantly increased in both mucosa and circular muscle layer of patients with chronic diverticular disease. Changes in mucosal galanin were likewise reflected by changes in the circular muscle layer, though interestingly VIP was not increased in the mucosa thought it was increased twofold in the circular muscle layer. Detailed morphology of the nerves suggested a very active process of remodelling. Both mucosal and circular muscle layer PGP9.5 was increased as was the nerve density within the longitudinal muscle. The muscle layer showed an increase in thickness (1.3-fold) so the total amount of nerves calculated from nerve density times muscle layer thickness would have been increased three- to fourfold. Thus as muscular hypertrophy occurs, so new nerves sprout to innervate them. This process, together with structural distortions found in chronic diverticular disease, may well account for the increased nerve angulation that we noted. In patients operated on for acute diverticulitis, there was an increase in small diameter nerves, though in patients with more chronic disease this was no longer a feature. This would be compatible with sprouting of new nerves during recovery from acute inflammation as is observed for example in the reparative process in dental abscesses 26 and also in the TNBS colitis model.27

Our observations of increased mucosal tachykinins agree with numerous studies suggesting that both in animals and man, mucosal inflammation initially destroys nerves but is then followed by a proliferation of tachykinin-containing axons. These are an important part of the inflammatory response, mediating both visceral hypersensitivity, vasodilatation and inflammatory changes via axonal reflexes. Tachykinin antagonists inhibit visceral hypersensitivity and in some models inhibit inflammation markedly.28,29 The changes in peripheral nerves are also paralleled by changes within the dorsal root ganglia30 and are likely to contribute to visceral hypersensitivity, as tachykinin antagonists can inhibit colitis induced visceral hypersensitivity.16,18 This is supported by the increased expression of SP in another condition of postinflammatory hypersensitivity, namely dental caries, where the % of PGP 9.5 positive nerves in dental pulp co-staining for SP has been shown to correlate with the occurrence of pain.31

The changes in galanin, VIP and PACAP expression are similar to those observed in the early time points in our studies of TNBS colitis in the rat.15 In these studies, we showed prolonged elevation of mucosal galanin, though the changes in VIP and PACAP were relatively short lived (3–6 weeks). The persistent abnormalities observed in the mucosa of our patients raises the possibility that there is ongoing low level chronic inflammation. However lymphocyte and mast cell numbers were not increased and all our biopsies were reported by an expert pathologist to be within the normal range.

The decrease in 5HT staining enterochromaffin cells has not been previously reported in diverticular disease but reduced serotonin concentration has been reported in inflammatory bowel disease.32 By contrast in patients with postinfective irritable bowel syndrome following a Campylobacter enteritis an increased turnover was noted with reduced 5HT content per cell and increased cell counts.33 Recent animal studies suggest that whether inflammation decreases or increases enterochromaffin cells depends on whether the immune response shows Th1 or Th2 predominance.34 Reduced enterochromaffin cell counts have been reported in Citrobacter rodentium infections35 associated with a Th1 predominant immune response while Trichinella spiralis infection associated with a predomiantly Th2 response causes an increase.36 Thus interpretation of this new finding in diverticular disease will require better characterization of the immune response profile in this particular condition.

The cause of increases in the neuropetide markers is not entirely clear but increased VIP has certainly been noticed in Crohn’s colitis.12 Both VIP and PACAP have general anti-inflammatory effects and may be part of the down regulation of inflammation.37 Vasoactive intestinal polypeptide also controls vasodilatation in the rectum and may be an important part of the mucosal defence,38 being markedly increased after radiotherapy affecting the rectal mucosa.39

The increase in galanin in both mucosa and circular muscle layer of patients with chronic diverticular disease in Study 1 and in the mucosa in symptomatic patients in Study 2 was the most striking abnormality. Galanin has a range of activities which are likely to be different in specific parts of the enteric nervous system. Galanin receptors coupled to inhibitory G proteins are predominately inhibitory in the CNS. Galanin is often co-secreted with substance P and calcium gene related peptide (CGRP) in sensory neurones and transported to the superficial lamina of the dorsal horn where it has a possible antinociceptive role. Upregulation in the circular muscle layer including the myenteric plexus may well have important effects on both spinal afferents and motor responses since galanin-1 receptor knockout mice fail to show increased spinal nociceptive reflex excitability in response to inflammation.40 Galanin has an important role in the increased intestinal chloride secretion which develops in response to inflammation,41 which upregulates the galanin-1 receptor acting via NFκB42. The persistent increase we observed in galanin in the mucosal biopsies did correlate with a significant increase in diarrhoeal symptoms and frequency of defecation, which may be causally relevant.

The main strength of this study is the strong correlation of objective measures with carefully documented symptoms. The fact that we were also able to show similar findings in both the TNBS model and the resection specimens also gives added confidence to our findings, as does the specific behaviour of the different neuropeptides which parallel what happens in the TNBS colitis model. Furthermore, the analysis of full thickness specimens justifies the use of endoscopic mucosal biopsies as markers of neuropeptide changes within the muscular layer in future studies.

Whether galanin is merely a useful marker of neural injury or contributes substantially to the symptoms of patients with diverticular disease remains uncertain but specific galanin antagonists are under development and could be used to answer such questions. While the relation between galanin and secretion may be direct, the relation with recurrent pain may be indirect, via a correlation with increased tachykinins in the circular muscle layer arising secondary to the same inflammatory neuronal damage which elevated the galanin.

Galanin antagonists have been developed and show little sedating effects. They might possibly therefore be a useful future treatment for painful diverticular disease. Alternatively if galanin is merely a marker for inflammatory injury, it could be used to target patients who might benefit specifically from anti-inflammatory treatments and hence improve the response to various combinations of antibiotics and mesalazine which uncontrolled trials have suggested may benefit chronic symptoms in patients with diverticular disease.43,44 Finally if increased mucosal galanin is a marker of increased tachykinins in the muscle layers then tachykinin antagonists might also benefit chronic symptoms. Cleary the next step is to use these readily available markers of previous injury or ongoing inflammation to subdivide patients with diverticular disease to see if they can identify those who would respond selectively to novel agents.

Grant

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

Supporting Grant – Medical Research Council, Clinical Training Fellowship (JS).

References

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