Mast cells and cellularity of the colonic mucosa correlated with fatigue and depression in irritable bowel syndrome
- T Piche1,2,
- M C Saint-Paul3,
- R Dainese1,
- E Marine-Barjoan1,
- A Iannelli4,
- M L Montoya1,
- J F Peyron2,
- D Czerucka2,
- F Cherikh1,5,
- J Filippi1,
- A Tran1,
- X Hébuterne1
- 1Fédération d’Hépato-Gastroentérologie et de Nutrition Clinique, Hôpital de l’Archet, Pôle digestif, CHU de Nice, Nice France
- 2Unité INSERM 526, Nice, France
- 3Laboratoire d’Anatomie Pathologique, Hôpital Pasteur, CHU de Nice, Nice, France
- 4Service de Chirurgie digestive et de Transplantation hépatique, Hôpital de l’Archet, Pôle digestif, CHU de Nice, France
- 5Psychiatry of liaison, CHU de Nice, Nice, France
- Dr T Piche, Hôpital de l’Archet 2, Pôle digestif, Service de Gastroentérologie, BP 3079, 06202 Nice Cedex 3, France;
- Revised 26 November 2007
- Accepted 28 November 2007
- Published Online First 14 January 2008
Background: A subset of patients with irritable bowel syndrome (IBS) have an increased number of mast cells (MCs) in the colonic mucosa. Psychological factors are believed to contribute to the course of IBS.
Aims: To examine associations between fatigue, depression and MCs of the colonic mucosa in IBS.
Methods: Colonic biopsies were taken from 50 Rome II IBS patients, 21 healthy controls and 11 depressed/fatigued patients without IBS. The cellularity of the lamina propria was determined as the number of inflammatory cells per high power field (hpf) through a 400× microscope. The Fatigue Impact Scale (FIS) and the short form Beck Depression Inventory (BDI) evaluated the severity of fatigue and depression.
Results: IBS patients had a significant increase in the cellularity of the lamina propria compared with controls or with depressed patients (mean (SD) 94.5 (48–110) vs 68 (58–82) and 78 (87–90) cells per hpf, p = 0.005 and p = 0.05, respectively), in particular of MCs (9.3 (5.6–11.7) vs 4.0 (2.7–6.8) and 4.3 (2.8–7.8) cells per hpf, p = 0.001 and p = 0.005, respectively). Both the FIS and BDI scores were significantly higher in IBS or in depressed patients than in controls (p<0.001). In IBS, the FIS score correlated significantly with the cellularity of the lamina propria (r = 0.51, p<0.0001) and MCs (r = 0.64, p<0.0001). In IBS, the BDI score correlated significantly with MCs (r = 0.29, p = 0.03).
Conclusions: Elevated MCs counts are a key feature of the low-grade inflammatory infiltrate in the caecal mucosa of IBS. Fatigue and depression are associated with mucosal cell counts, in particular MCs, suggesting that psychological factors are associated with the low-grade inflammatory infiltrate in IBS.
Irritable bowel syndrome (IBS) is defined as a group of functional bowel disorders in which abdominal discomfort/pain is associated with defecation or a change in bowel habit without an identifiable disease process. Currently, the pathophysiological model of IBS involves a theoretical relationship among psychological factors, inflammation of gut mucosa and disrupted physiology modulated by central nervous system (CNS)–enteric nervous system (ENS) interactions (eg, the brain–gut axis) (for a review, see Collins et al1).
A growing literature indicates that immune activation and inflammation contribute to the pathophysiology of IBS, a subset of patients having significantly increased numbers of activated immunocompetent cells in the colonic mucosa.2 3 Among these, mast cells (MCs) are important candidates in IBS since they release several potent inflammatory and immune-modulating mediators that can alter nerve4 and muscle function.5 Quantitative assessments have identified infiltrates of MCs in the ileum6 7 and colon2 8 9 of IBS patients. It has also been shown that MC infiltration and release of mediators in the proximity of mucosal innervation may contribute to abdominal pain perception in IBS patients.8 Although the causes of the development of chronic low-grade inflammation in IBS are unclear, psychological factors are believed to play an important role.
The influence of psychological factors on IBS symptoms and outcome is becoming increasingly recognised. Indeed, IBS patients commonly have increased levels of anxiety, depression, fatigue or hypochondriasis, and stressful life events frequently precede the onset or exacerbation of IBS symptoms.10–14 Experimental evidence suggests that stress, either acute or chronic, can modulate inflammation at both systemic and tissue levels.15 16 Moreover, it was demonstrated that antidepressants can attenuate increased susceptibility to colitis in a murine model of depression.17 Also, emerging evidence suggests that the effect of depression may promote inflammatory processes.18 In addition, it was shown that MCs play an important role in the occurrence of mucosal damage induced by both acute and chronic stressors.19 20 To date, the relationship between psychosocial factors, which are relevant in IBS, and colonic immune activation, in particular MCs, has been poorly investigated. The presence of sustained anxiety or depression at the time of an acute gastroenteritis has been shown to increase the risk of developing IBS together with persistent mucosal inflammation.21–23 In one study involving postinfectious IBS (PI-IBS) patients, however, anxiety, levels of depression and fatigue did not correlate with mucosal cell counts.23 Finally, the relationship between immune activation and fatigue, an important determinant of quality of life of IBS patients14 24 that relates to well-being, psychological and gastrointestinal symptoms,13 has received little attention.
We aimed to determine if there was a relationship between the effects of fatigue/depression on inflammatory cell counts, in particular MCs, of caecal biopsy specimens in a consecutive population of IBS patients, healthy controls and depressed/fatigued patients without IBS.
PATIENTS AND METHODS
Fifty consecutive patients who met the Rome II criteria for IBS25 were seen in our gastrointestinal department between November 2004 and February 2007. This required the presence of abdominal pain or discomfort relieved by defecation or associated with change in stool form or frequency, together with at least two of the following symptoms: altered stool frequency (>3 per day or <3 per week), altered stool passage, passage of mucus, bloating and/or abdominal distension. These symptoms had to be present more than 25% of the time after exclusion of organic disease. Patients were not included if their symptoms occurred after a history of gastrointestinal infection or clinical patterns of adverse reactions to food. IBS subtypes were divided as follow: predominant constipation (IBS-C) (ie, infrequent hard stools or increased straining) or predominant diarrhoea (IBS-D) (ie, frequent loose or watery stools and urgency). Patients with alternating symptoms (IBS-A) of constipation and diarrhoea were also included with at least one of the criteria required for the diagnosis of IBS-C or IBS-D. The mean number of stool emissions was determined using a 15 day diary. Patients were thoroughly examined for other organic diseases including: blood tests (eg, erythrocyte count, C-reactive protein, leucocyte count, electrolytes, liver function tests, amylase, coeliac antibodies, HIV and hepatitis C virus, and thyroid function tests); a faecal culture; an abdominal ultrasound; a gastroscopy with duodenal biopsies; and finally a colonoscopy with staged biopsies to exclude microscopic colitis. The severity of abdominal pain/discomfort was graded from 0 to 4 according to the impact on the patient’s activities:8 0, absent; 1, mild (not influencing activities); 2, relevant (diverting from usual activities); 3, severe (influencing usual activities markedly enough to urge modifications); and 4, extremely severe (precluding daily activities). The severity of IBS was also evaluated on a visual analogue scale measuring the impact of abdominal well-being on usual activities that ranged from 0, absent to 10, extremely severe.
Twenty-one subjects without IBS enrolled in a colonoscopic surveillance programme for polyps formed the control group. These subjects were considered normal on the basis of medical history, physical examination, blood screening and the absence of any digestive complaint.
Eleven patients without any digestive complaint but with clinical depression diagnosed by our liaison psychiatrist (FC) were also included. All these patients reported fatigue. They were enrolled if a colonoscopy was indicated, in particular for the supervision of polyps.
Participants had not taken any substance (including alcohol and tobacco) or drugs (eg, β-blockers, antidepressants, sedatives, thyroid extracts, steroids and non-steroidal anti-inflammatory compounds) that could interfere with psychological distress or mucosal inflammation for 1 month prior to the investigations. The study was performed according to the Declaration of Helsinki, and all participants gave their written informed consent.
Colonic biopsies and cell counting methods
A colonoscopy was performed in all patients and controls, also examining the terminal ileum, and 4 of 10 staged biopsies were taken at the caecum site. Biopsies were limited to the caecum in controls and in depressed patients. All participants were prepared using polyethylene glycol.
Biopsy specimens were fixed in formol solution, and histopathological, morphometric and immunohistological studies were performed on paraffin-embedded 4 μm thick sections. Slides were coloured by May–Grunwald–Giemsa (MGG). For immunohistochemistry, slides were incubated for 1 h with mouse monoclonal antibodies specific for CD-3 (eg, T lymphocytes) and CD-117 (eg, MCs). Sections were reincubated with secondary biotinylated anti-mouse rabbit antibodies (Dako StreptABC complex) followed by streptavidin/horseradish peroxidase conjugate. Positive and negative controls were performed. Biopsies were orientated and cut so that the full mucosal height was visible and the extent of inflammatory cells deep in the lamina propria was appreciated, and fields containing lymphoid aggregates were avoided for counting. Cell counts were performed twice by a single pathologist (MC-SP) who was blinded to the diagnosis.
Lamina propria cellularity was calculated as the total number of inflammatory cells within five non-overlapping high power fields (hpfs) at 400× final magnification (0.0125 mm2) (Olympus Bx41 light microscope), with MGG staining, using a graticule eyepiece. When the dot landed on inflammatory cells, they were classified as plasma cells, lymphocytes, neutrophils, eosinophils or MCs, according to their typical morphology and staining properties. Fields containing lymphoid aggregates were excluded.
Conventional CD-117 immunostaining was applied on the sections to calculate the absolute number of MCs within five non-overlapping hpfs. The repeatability coefficient for two readings of the same hpf and its 95% CI were 40% higher using CD-117 staining (1.042 (6.29–8.37)) than using MGG (0.612 (6.73–7.96)), explaining the choice of CD-117 to express the number of MCs. The limits of agreement between the two staining methods were calculated and plotted on a Bland–Altman plot.26
Intraepithelial lymphocytes (IELs) were enumerated as CD-3+ cells per 100 epithelial cells, after counting at least 200 cells, and more if possible, in separate sections. Transitional lymphocytes traversing the basement membrane were not included.
The use of five fields for cell counting gave the following average coefficient of variation between hpfs: MCs, 4.4%; lamina propria T lymphocytes, 10.8%; plasmocytes, 15%; neutrophils, 1.3%; eosinophils 7.2%. The coefficient of variation of IELs was 5.8%. The reproducibility of two measures within 5 hpfs (eg, SD expressed as a percentage of mean value) gave the following average coefficients: MCs, 3.77%; lymphocytes, 2.67%; plasmocytes, 4.15%; neutrophils, 5.23%; eosinophils, 2.72%. The coefficient for IELs was 3.44%.
Psychological factors: fatigue, depression and adverse life events
The severity of fatigue and depression and the recall of adverse life events (eg, presence or absence of physical or sexual abuse, and marital status) served to evaluate psychological factors.
The severity of fatigue was measured using the French version of the Fatigue Impact Scale (FIS).27 28 This questionnaire has been subsequently used in patients with chronic hepatitis C infection29 30 and also in a cohort of IBS patients.14 FIS is a self-report questionnaire consisting of 40 statements that describes manifestations of fatigue: statements are divided in three categories (cognitive (n = 10), physical (n = 10) and psychosocial (n = 20)). Each item is rated on a 5-point scale, ranging from 0 (“no problem”) to 4 (“extreme problem”), with a maximum of 160 points.
Depression was assessed using the short version of the Beck Depression Inventory (BDI), a well-standardised 13-item measure of cognitive, affective and somatic symptoms of depression31 available in French.32 The short 13-item BDI contains only pertinent items that correlate with both the global score of the original 21-item form and the clinical evaluation of depression by doctors.33 Four sentences are included in each item, rating from 0 to 3 according to symptom severity. The summing of each item gives a global score allowing an evaluation of depression: 0–4 absence, 5–7 minor, 8–15 moderate and ⩾16 severe depression.
Univariate analysis of the characteristics of the population (mean (SD)) was conducted using analysis of variance (ANOVA) and Fisher’s exact test. The study parameters were compared using χ2 for qualitative variables, and Kruskall–Wallis and Mann–Whitney non-parametric tests for quantitative variables (medians, and 25th and 75th quartiles). The association between psychological characteristics and histological findings was a decision made a priori using linear regression. A p value <0.05 was considered statistically significant.
Demographic and clinical data
Among 50 IBS patients included, 58% (n = 29) and 42% (n = 21) were considered IBS-C and IBS-D, respectively. Four patients with alternating symptoms were finally classified as IBS-D because diarrhoea was predominant. The clinical features of our study population are described in table 1. The characteristics of the IBS population were similar to those usually reported, showing a female predominance and a higher incidence of constipation among women. The duration of symptoms was similar between IBS subtypes. The colonic mucosa was macroscopically normal in all patients. Polyps were resected in four IBS patients, two controls and three patients with depression.
Histological findings of IBS, controls and depressed patients are detailed in table 2. The cellularity of the lamina propria was significantly higher in IBS patients than in controls or in depressed patients (p = 0.005 and p = 0.05, respectively).
The number of MCs per hpf was significantly increased in IBS patients compared with controls or with depressed patients (p = 0.001 and p = 0.005, respectively). Figure 1 illustrates a CD-117 immunostaining of MCs in the lamina propria of a patient with IBS. In IBS patients, but not in controls or in depressed patients, MCs were frequently localised deep in the lamina propria (fig 2). The number of lymphocytes per hpf was significantly higher in IBS patients than in controls (p = 0.008) and there was a statistical trend when compared with depressed patients (p = 0.07).
Histological findings between IBS subtypes were not significantly different (table 2). Comparisons of histological findings in relation to gender, age and duration of IBS revealed no significant difference. None of the patients presented with microscopic colitis.
Fatigue, depression and adverse life events
Data pertaining to psychological distress (eg, FIS and BDI scores) in IBS patients, controls and depressed patients are given in table 3. Overall, 30 of 50 IBS patients (60%) reported fatigue. Fatigue was reported by one control (4,7%) and by all depressed patients. The FIS score and all three domains of fatigue were significantly higher in IBS patients than in controls (p<0.001) and similar to that of depressed patients. The FIS score was similar between IBS subtypes.
Seventeen of 33 (51.5%) IBS patients had a BDI score >4 and could be classified as “depressed”. The BDI score was significantly higher in IBS patients than in controls (p<0.001) and similar to that of depressed patients. BDI scores were not significantly different with respect to IBS subtypes. In IBS patients, the FIS score was associated with the BDI score (r = 0.57, p = 0.04). In IBS patients, the FIS score was associated with the BDI score (r = 0.57, p = 0.04).
The percentage of adverse life events was not significantly different between IBS patients and controls or depressed patients, nor was it different between IBS subtypes (data not shown).
Correlations between the severity of psychological factors and histological findings are detailed in table 4. In IBS, but not in controls or depressed patients, the cellularity of the lamina propria correlated significantly with FIS score (r = 0.51, p = 0.0001), physical (r = 0.56, p<0.0001), cognitive (r = 0.56, p<0.0001) and psychosocial (r = 0.46, p = 0.0009) domains. However, the cellularity of the lamina propria did not correlate with the BDI score (r = 0.002, p = 0.88).
In IBS, but not in controls or depressed patients, MCs per hpf correlated significantly with the FIS score (r = 0.64, p<0.0001), physical (r = 0.69, p<0.0001), cognitive (r = 0.63, p<0.0001), psychosocial (r = 0.64, p<0.0001) domains and the BDI score (r = 0.29, p = 0.03).
Data according to IBS subtypes are given in table 4. Adverse life events were not significantly associated with histological changes in IBS patients, controls or depressed patients (data not shown).
The severity of abdominal pain/discomfort and the general abdominal well-being of IBS patients were not associated with FIS and BDI scores or MCs per hpf (data not shown). However, a slight significant correlation between the severity of abdominal pain/discomfort and FIS score was observed in IBS-C (r = 0.38, p = 0.04).
This study confirms that elevated numbers of MCs are key features of the low-grade inflammatory infiltrate existing in the caecal mucosa of IBS patients. In IBS patients, we also found significant associations between the severity of fatigue and depression and the cellularity of the lamina propria, in particular MCs.
Although the majority of IBS studies have methodological differences concerning sample size, biopsy site and histological staining that flaw the interpretation, an inflammatory infiltrate has been regularly demonstrated in a subset of IBS patients. Indeed, the cellularity of the lamina propria,2 34 35 the number of MCs in the colon2 8 9 or the terminal ileum6 7 36 and the numbers of IELs2 have been found to be significantly increased in IBS patients compared with healthy subjects. Interestingly, in several of these studies, these changes were more marked in the ileum or right colon, whereas the rectum often tended to be normal.6 9 34 Accordingly, our findings obtained from biopsies taken at the caecum site confirm the potential for MC involvement in the pathophysiology of IBS. The observation of a more frequent localisation of MCs deep in the lamina propria in IBS patients also suggests the role of neuroimmune interactions in the pathogenesis of IBS.8 In the present study, we did not found the gender-dependent differences in immune response, including tissue MC infiltration, that may be one possible explanation of the higher prevalence of IBS among women.8 37 However, as there were only a few men in the present study, sampling errors may have accounted for the absence of difference between males and females. Similarly, immunopathological parameters of IBS patients were not influenced by age or the length of history. To date, there is no agreement on the type of bowel habit associated with immunopathological patterns in IBS; some authors have found increased MCs in IBS-C2 whereas other have reported higher MCs in both IBS subtypes.7 8 In the current study, the similarity of histological abnormalities between IBS subtypes provides a role for MCs in the pathophysiology of both constipation and diarrhoea. This is further supported by the fact that patients with mastocytosis, a symptom complex not dissimilar from IBS including abnormal bowel habit, abdominal pain and fatigue, may have both diarrhoea and constipation.38
There are limited data on the role of chronic psychological factors in gastrointestinal inflammation, in particular MC responses and/or gut functional correlates in IBS patients. From data obtained in animals, it was shown that psychological stress can contribute to the initiation and reactivation of experimental gastrointestinal inflammation39 40 and the increase of intestinal permeability.41 Subsequent experimental studies have also linked stress-related MC activation to changes in gut function.42 43 In patients with inflammatory bowel diseases, it was shown that long-term sustained stress and stressful life events increased the number of relapses in ulcerative colitis44 45 whereas environmental stress altered the course of Crohn’s disease.46 Also, it was shown that stress-induced activation of the brain–gut axis and of mucosal MCs was important in the initiation and/or flare up of inflammatory bowel disease.47 The interactions between psychological factors, behavioural factors and immune activation have also been highlighted in patients with PI-IBS, in whom the presence of sustained psychological factors at the time of the gastroenteritis was shown to influence the development of IBS symptoms.21 48 In patients with non-PI-IBS, we found a significant association between the FIS score and both the cellularity of the lamina propria and MC counts in the colonic mucosa. We also observed the same associations between BDI score and MC counts. These histological abnormalities were found to be similar in both IBS-C and IBS-D patients. The apparent discrepancies between these data and the findings reported by Dunlop et al,23 in which the authors failed to detect any correlation between anxiety (that we did not evaluate), depression and mucosal cell counts, may be due to a low rate of psychological distress usually reported in selected PI-IBS populations.49 Moreover, data obtained in PI-IBS cannot be generalised to non-PI-IBS populations because of qualitative differences in the low-grade inflammatory response, non PI-IBS patients exhibiting elevated lamina propria T lymphocytes and MCs but normal enterochromaffin cell counts.6 9 34
Whether depression/fatigue and low-grade inflammation are related to each other in the context of IBS has not been investigated in the present study. From the literature obtained from coronary heart disease, one possible scenario involves a bidirectional relationship in which depression may enhance inflammation, that could subsequently increase depression (for a review, see Santos et al18). However, in our study, depressed/fatigued patients without IBS showed no evidence of increased MCs or other immune cells compared with controls, suggesting that psychological factors per se are not potent triggering factors of gut inflammation in the case of normal epithelial barrier function (eg, integrity). This is further supported by a recent study in which adult maternally separated mice showed evidence of depressive-like behaviour and no evidence of spontaneous colitis compared with unseparated mice. Moreover, maternally separated animals were more vulnerable to dextran sulphate sodium colitis, and antidepressants reversed this susceptibility but had no effect in unseparated animals.17
Although these significant associations only suggest a link and do not demonstrate the underlying mechanism nor the sequence of these interactions, our findings strongly suggest that chronic psychological factors may be linked to gut inflammation, in particular MC infiltration, in the colonic mucosa of IBS patients. Also, studies aimed at evaluating the effect of selective serotonin reuptake inhibitors, which have been shown to be effective on both depression and IBS symptoms,50 should also target gut immune activation.
Our study has several limitations. Whether depression and fatigue can be regarded separately when using both the FIS and BDI has not been investigated in the present study. As previously reported,28 29 we have observed a significant correlation between the severity of fatigue and depression in this sample of IBS patients. In fact, we believe that these constructs are complementary to evaluate fatigue/depression and we acknowledge a partial overlap in the evaluation of these psychological symptoms. We did not evaluate MC activation nor the spontaneous release of histamine and tryptase from mucosal biopsy specimens, both of these being candidates involved in the process leading to stress-induced mucosal abnormalities.19 20 Numerous factors may also contribute to enhance the inflammatory response of the gut in IBS. Although we did not include IBS patients reporting IBS symptoms as a consequence of food ingestion, we cannot exclude that occult food allergies may have played a minor role.51 Emerging evidence suggests the contribution of a genetic predisposition in IBS patients having, for example, significantly reduced frequencies of the interleukin 10 and transforming growth factor β high producer allele that could increase the release of proinflammatory cytokines from MCs.52 Finally, our patients did not complete specific digestive symptoms or anxiety questionnaires which may be relevant to evaluate the whole clinical spectrum of IBS.8
In conclusion, we have shown that IBS patients have a significant increase of the cellularity of the lamina propria, in particular of MCs. We found significant associations between the severity of fatigue and depression and both the cellularity of the lamina propria and MC counts, suggesting that psychological factors may contribute to the development/perpetuation of the low-grade inflammatory infiltrate of MCs in IBS patients.
We are very grateful to Professor Pierre Michel Huet for his help in the use of the French version of the Fatigue Impact Scale. We also thank Dr Frédéric Berthier (Department of Statistical Analysis, CHU Nice) for his helpful contribution to the statistical analysis of the findings.