Bacillary dysentery as a causative factor of irritable bowel syndrome and its pathogenesis
- Department of Gastroenterology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Beijing 100730, China
- Correspondence to:
Professor G-Z Pan
Department of Gastroenterology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, 1 Shuaifuyuan, Beijing 100730, The People’s Republic of China;
- Accepted 6 January 2004
Background and aims: The incidence of irritable bowel syndrome (IBS) or functional bowel disorders (FBD) after bacillary dysentery (BD) has not been extensively evaluated, and little is known of the pathogenesis of post-infective (PI) IBS. Therefore, we investigated the incidence of IBS and FBD in a Chinese patient population who had recovered from BD. To further elucidate its pathogenesis, neuroimmunological changes, including interleukins (IL), mast cells, neuropeptides, and the relationship between mast cells and intestinal nerves, were investigated.
Methods: A cohort study of 295 patients who had recovered from BD (shigella identified from stool in 71.4%) and 243 control subjects consisting of patient siblings or spouses who had not been infected with BD were included in the study. All subjects were followed up using questionnaires for 1–2 years to explore the incidence of FBD and IBS, as defined by the Rome II criteria. In 56 cases of IBS (PI and non-PI) from another source, the number of mast cells in biopsy specimens from the intestinal mucosa were stained with antitryptase antibody and counted under light microscopy. Also, the relationship of mast cells to neurone specific enolase (NSE), substance P (SP), 5-hydroxytryptamine (5-HT), or calcitonin gene related peptide positive nerve fibres was observed using double staining with alcian blue and neuropeptide antibodies. In 30 cases of IBS (PI-IBS, n = 15) taken at random from the 56 cases, expression of interleukin (IL)-1α, IL-1β, and IL-1 receptor antagonist (IL-1ra) mRNAs in intestinal mucosa were identified using reverse transcription-polymerase chain reaction. The above results were compared with 12 non-IBS controls.
Results: In the BD infected cohort, the incidences of FBD and IBS were 22.4% and 8.1% (in total)–10.2% (among those in who shigella were identified) respectively, which were significantly higher (p<0.01) than the incidences of FBD (7.4%) and IBS (0.8%) in the control cohort. A longer duration of diarrhoea (⩾7 days) was associated with a higher risk of developing FBD (odds ratio 3.49 (95% confidence interval 1.71–7.13)). Expression of IL-1β mRNA in terminal ileum and rectosigmoid mucosa was significantly higher in PI-IBS patients (p<0.01). The number of mast cells in the terminal ileum mucosa in PI-IBS (11.19 (2.83)) and non-PI-IBS patients (10.78 (1.23)) was significantly increased compared with that (6.05 (0.51)) in control subjects (p<0.01). Also, in the terminal ileum and rectosigmoid mucosa of IBS patients, the density of NSE, SP, and 5-HT positively stained nerve fibres increased (p<0.05) and appeared in clusters, surrounding an increased number of mast cells (p<0.01 compared with controls).
Conclusions: BD is a causative factor in PI-IBS. The immune and nervous system may both play important roles in the pathogenesis of PI-IBS.
- BD, bacillary dysentery
- FBD, functional bowel disorder
- IBS, irritable bowel syndrome
- IL, interleukin
- IL-1ra, interleukin 1 receptor antagonist
- PI, post-infective
- RT-PCR, reverse transcription-polymerase chain reaction
- NSE, neurone specific enolase
- SP, substance P
- CGRP, calcitonin gene related peptide
- 5-HT, 5-hydroxytryptamine
- SA-HRP, streptavidin-horseradish peroxidase
- PBS, phosphate buffered saline
- irritable bowel syndrome
- functional bowel disorder
- bacillary dysentery
- mast cells
- enteric nerve fibres
Previous studies revealed that the prevalence of functional bowel disorders (FBD) six months after infectious diarrhoea was 25%.1,2 A study in the UK indicated that during a one year follow up, the diagnostic rate of irritable bowel syndrome (IBS) was 4.4% in patients after an episode of bacterial gastroenteritis compared with a rate of 0.3% in the general population cohort.3 Our study provided further evidence that a previous history of dysentery was an important risk factor (odds ratio 3.00; p<0.001) of IBS in a randomised population study in Beijing.4 However, the prevalence of either IBS or FBD after bacillary dysentery (BD) has not been extensively evaluated.
Also, little is known of the pathogenesis of post-infective IBS (PI-IBS). The fact that only 25% of patients who have had infectious diarrhoea develop IBS-like symptoms suggests other prerequisites, such as interaction of the nervous and immune systems, are required for IBS symptoms to develop among those infected patients.5 It is believed that brain-gut interactions play an important role in the pathogenesis of IBS. Possible connections exist between enteric nerves and immune cellular components. In light of the present literature, it was suggested that mast cells could be a possible candidate connecting the local immune response to the neurohormonal system during acute intestinal infection.6,7
The aims of the present study were: (1) to explore the incidence of PI-IBS and FBD after BD; (2) to identify the intestinal immune response, particularly expression of interleukins (IL) and number of mast cells, in PI-IBS and non-PI-IBS patients; and (3) to study the relationship between mast cells and intestinal nerves in the pathogenesis of IBS.
PATIENTS AND METHODS
Source I: follow up study
This was a cohort study comprising patients who had suffered from BD and who attended the Dysentery Clinic of the Peking Union Medical College Hospital (PUMCH), from April to October 1998, and their family members who had not been infected with BD. All subjects were adult Chinese, followed up for a period of 1–2 years, interviewed mainly by telephone by answering a detailed symptom questionnaire. Acute BD was diagnosed by a positive stool culture or on the basis of the following criteria: (1) symptoms of lower abdominal pain, rectal burning, and acute diarrhoea; (2) microscopic examination of faecal effluent revealed polymorphonuclear leucocytes >15/per high power microscopy, and no protozoa or ova identified; and (3) cured by treatment with antibiotics and no relapse.
In the cohort study, patients’ siblings or spouses who did not have episodes of dysentery or acute infectious diarrhoea during the same period were enrolled as controls. Numbers of patients and sources are shown in fig 1.
Source II: subjects involved in mast cell/nerve fibre, and interleukin studies
Patients were from the Gastroenterology Clinic, PUMCH (fig 1). Fifty six cases of diarrhoea-type IBS who fulfilled the Rome II criteria and had acute symptoms were examined for number of mast cells and nerve fibre density in the intestinal mucosa. The group comprised 25 males and 31 females (mean age 43.3 years). Among them, 27 cases with a history of infectious diarrhoea successfully treated with antibiotics were considered to be PI-IBS (shigella were identified from their original stool culture in 11 cases). The other 29 cases without a history of infectious diarrhoea were grouped as non-PI-IBS; patients who had bowel symptoms preceding the onset of IBS were excluded. All subjects underwent the following to exclude an anatomical or biochemical explanation for their symptoms: detailed case history; physical examination; routine blood, urine, and faecal analyses; stool culture; blood chemistry tests; abdominal ultrasonography; and colonoscopy. During colonoscopy, one mucosal biopsy specimen was taken at each of the following locations: 5 cm proximal to the ileocaecal valve and at the rectosigmoid junction. Specimens were freshly suspended in formalin for Immunohistochemical study. In 30 cases taken at random from the 56 patients, an additional biopsy specimen was obtained at each location and stored in liquid nitrogen for investigation of IL mRNA expression. There was no difference (p>0.05) in onset to biopsy interval between PI-IBS (mean 1.32 (SD 0.6) years) and non-PI-IBS (1.57 (0.56) years) patients. Twelve subjects (five male and seven female; mean age 43.4 years) who had a negative screening colonoscopy as part of a health examination (n = 3), for rectal bleeding finally diagnosed as haemorrhoids (n = 6), or for post-polypectomy surveillance (n = 3), and who had no bowel symptoms, served as controls for both study groups. The control group underwent the same examination procedure as the study groups and two biopsies at each location were taken during colonoscopy. All biopsies were approved by the examinees.
Patients and control subjects in the follow up study answered the same questionnaire, as reported by Neal and colleagues2 with minor modifications. Questions concerned the presence of symptoms after the episode of dysentery, such as: abdominal discomfort or pain, form of stool, change in bowel habit, sense of urgency, straining or incomplete evacuation at the passage of stool, passage of slime or mucus, bloating, or feeling of abdominal distension. Question also concerned the patient’s general health, previous medical history, and bowel habits a year before dysentery. All questions were followed by “if yes, from what time, on how many days a week, and for how many months a year”?
The majority (90%) of patients and their family members were interviewed by telephone, and the remainder by mail. Diagnosis of FBD or IBS was made according to the Rome II criteria.8 FBD included other functional bowel disorders as well as IBS (see footnote to table 1).
Identification of expression of IL mRNAs
IL-1α, IL-1β, and IL-1 receptor antagonist (IL-1ra) mRNAs were identified as follows. Extraction of RNA from mucosal biopsies by Trizol (Gibco) reagent was based on the procedures provided by the company, and reverse-transcription polymerase chain reaction (RT-PCR) was according to the procedure of Jobin and Gauthier9 with minor modifications. The RT-PCR kit was from Boda Company (Beijing, China).
The RT reaction was as the follows: 1 μl dNTP, 1 μl M-MuLV, 1 μl 5×M-MuLV buffer, 1 μl random primer, 0.5 μl Rnase inhibitor, and 4 μl RNA sample, and then ddH2O was added to make a final volume of 20 μl. The reaction was set at 37°C for two hours and then at 95°C for five minutes.
For PCR, 4 μl of RT product were incubated with 1 μl dNTP, 0.5 μl Taq polymerase, 5 μl 10×PCR reaction buffer, 5 μl 15 mM MgCl2, and one of the following primers: 4 μl IL-1α (P1: 5′-GTC TCT GAA TCA GAA ATC CTT CTA TC-3′; P2: 5′-CAT GTC AAA TTT-CAC TGC TTC ATC C-3′, 420 bp product); 4 μl IL-1β (P1: 5′-ATA AGC CAC TCT ACA-GCT-3′; P2: 5′-ATT GGC CCT GAA AGG AGA GA-3′, 436 bp product); 4 μl IL-1ra (P1: 5′-GGC CTC CGC AGT CAC CTA ATC ACT CT-3′; P2: 5′TAC TAC TCG TCC TCC TGG-AAG TAG AA-3′, 521 bp product); or 1 μl β-actin (P1: 5′GAA TTC ATG TTT GAG ACC TTC AA3′; P2: 5′CCG GAT CCA TCT CTT GCT-CGA AGT CCA-3′, 318 bp product). dd H2O was added to make a final volume of 50 μl. The conditions for the PCR reactions for IL-1α, IL-1β, and β-actin were: 94°C for five minutes, then 35 cycles at 94°C for 50 seconds, 55°C for 50 seconds, and 72°C for one minute; re-extension was carried out at 72°C for six minutes. The conditions for the PCR reaction for IL-1ra were: 94°C for five minutes, 35 cycles at 94°C for one min, 55°C for one minute, and 72°C for three minutes; re-extension was carried out at 72°C for eight minutes.
After agarose gel electrophoresis of the PCR products, photos of the mRNAs of IL-1α, IL-1β, IL-1ra, and β-actin bands were expressed as their complementary cDNAs and were measured by Phoretix ID Advanced software. The intensity of the cDNA for IL-1α, IL-1β, or IL-1ra of each sample was normalised by the intensity of β-actin (relative IL-1 intensity = IL-1 intensity/β-actin intensity).
Immunohistochemical study of mast cells and nerve fibres
Rabbit antimast cell tryptase (Maixin Co., Fuzhou, China) was employed for the staining of mast cells in the mucosa. Tissue sections (4 μm) were incubated with rabbit antimast cell tryptase overnight at 4°C. Then, goat antirabbit biotinylated IgG and streptavidin-horseradish peroxidase (SA-HRP) were added for 30 minutes at 37°C. Staining was developed by diaminebenzidine. All slides were rinsed between each step with phosphate buffered saline (PBS).
The relationship between mast cells and nerve fibres was studied by double staining of mast cells and neuropeptides, according to the method of Stead and colleagues,6 with minor modifications. Rabbit antiserum antibodies for neurone specific enolase (NSE), substance P (SP), calcitonin gene related peptide (CGRP), and 5-hydroxytryptamine (5-HT) were from Maixin Co. (FuZhou, China). In addition to neutral formalin, biopsy samples were further fixed in acetic acid/ethanol (1:9) for eight hours. After tissue sections were stained with alcian blue 8GX (Sigma), rabbit antihuman antibody (NSE, SP, CGRP, or 5-HT) was added overnight at 4°C and goat antirabbit biotinylated IgG added for 10–30 minutes at 37°C. Slides were treated with SA-HRP. Staining was developed by 3-amino-9-ethylcarbazole. All slides were rinsed with PBS between each step. PBS was used instead of the primary antibody in the processing of control slides.
Mast cell counting and scoring of neuropeptide staining were performed single blind (that is, the examiner read the slides without knowing the patient’s status or diagnosis). Every patient’s sample was encoded with a number assigned at random; the code was broken after cell counting. The counting and scoring systems were as follows:
The frequencies of all items analysed were compared between the study group and the control group using the χ2 test. Risk factors were explored using regression analysis and analysis of variants.
Cell counts were expressed as the number of cells per field under high power microscopy (mean (SD)). Positive nerve fibre staining was expressed as the numeric score assessed per field under high power microscopy (mean (SD)). One way ANOVA was used in the comparison. Significant difference was taken as p<0.05; p<0.01 indicated very significant difference.
The follow up study
A total of 329 patients participated in the study (response rate 73.1%). Thirty four of the responders with a history of FBD were excluded. The remaining 295 cases without a history of FBD were enrolled in the study. Shigella were identified from stool culture in 71.4% of cases. As the incidence of IBS or FBD in the final analysis was not significantly different between the bacteria identifiable and non-identifiable groups (p>0.05), data were pooled in the analysis.
In the 295 cases of infectious diarrhoea, the male to female ratio was 1:1.34. Although the male to female ratio in the 243 controls was higher (1.53:1), there was no significant difference in the incidence of FBD or IBS between the sexes (p>0.05). Mean age of the infectious diarrhoea group was 40.74 (SD 14.94) years compared with 41.26 (14.76) years in the control group. Both groups were comparable in terms of age and economic status.
Incidence of FBD and IBS in infectious diarrhoea and control groups
In the BD group, after a follow up period of 1–2 years, 66 cases (22.4%) were found to have FBD and 24 cases (8.1% in total) developed IBS. The incidence of IBS was 10.2% (24/235) in patients with shigella infection. In the control group, the incidence of FBD was 7.4% and of IBS 0.8%, which was significantly lower (p<0.01) than that in the BD group.
Risk factors for FBD during infectious diarrhoea
Regression analysis of variants indicated that prolongation of the duration of infectious diarrhoea resulted in an increase in the incidence of FBD (table 1). If the duration of infection exceeded seven days, the patient had a 2.5-fold increased risk of future development of FBD compared with those whose duration was less than a week. Therefore, we conclude that duration of infectious diarrhoea is an important risk factor in FBD.
Sex or age had no significant effect on the incidence of FBD or IBS in our study (both p>0.05).
Expression of IL-mRNAs in intestinal mucosa
In PI-IBS patients, expression of IL-1β mRNA (expressed as cDNA identified at 436 bp by RT-PCR) in the terminal ileum mucosa or in the rectosigmoid mucosa was significantly higher (p<0.01) than that in non-PI-IBS patients or in control subjects (table 2, fig 2).
There were no differences in expression of IL-1α or IL-1ra mRNAs between the IBS groups (p>0.05), or between the IBS and control group (p>0.05).
Number of mast cells and the relationship between mast cells and nerve fibres in the intestinal mucosa of IBS patients
Number of mast cells in the intestinal mucosa of IBS patients and controls
The number of mast cells in the mucosa of the terminal ileum in PI-IBS (11.98 (2.83)) and non-PI-IBS (10.78 (1.23)) patients was significantly higher than that (6.05 (0.51)) in controls (p<0.01) (table 3). There was no significant difference in the number of mast cells in the rectosigmoid mucosa when either IBS group was compared with controls (p>0.05). In either location, there was no significant difference in the number of mast cells in PI-IBS and non-PI-IBS patients (p>0.05).
Scores for expression of NSE, SP, 5-HT, and CGRP positive fibres in intestinal mucosa and the relationship between nerve fibres and mast cells in IBS patients
As indicated in table 3, significantly higher scores for expression of NSE, SP, and 5-HT positive fibres in the mucosa of the terminal ileum or at the rectosigmoid junction were found in PI-IBS and non-PI-IBS patients compared with controls (p<0.05). But expression of CGRP positive fibres in the mucosa at either location was not significantly different between either IBS group and the control group (p>0.05). There was no difference in mucosal expression of NSE, SP, 5-HT, or CGRP positive fibres between PI-IBS and non-PI-IBS patients at either location (p>0.05).
Using the technique of double staining, it was possible to identify the relationship between mast cells and antibody stained nerve fibres. Nerve fibres stained bright red and appeared in clusters closely surrounding mast cells (fig 3). In IBS patients, positively stained nerve fibres around mast cells were significantly increased in density and also mast cells surrounded by those fibres were increased in number in the terminal ileum mucosa compared with controls (6.73 (1.02) v 4.25 (0.50); 6.84 (0.85) v 4.28 (0.40); 6.72 (0.81) v 4.00 (0.63); and 6.73 (0.82) v 4.33 (0.54), respectively, for NSE, SP, 5-HT, and CGRP stained fibres; all p<0.01). However, no difference was observed between the PI-IBS and non-PI-IBS groups (p>0.05).
Our study in 295 patients recovering from BD with no previous history of FBD indicated that after a follow up period of 1–2 years, 22.4% of patients developed symptoms of FBD, and 8.1% (in total)–10.2% (among whom shigella were identified) of cases developed IBS-like symptoms, as defined by the Rome II criteria. In the control cohort, consisting of 243 subjects with no episode of BD, and comparable in age, and economic and social status with the study cohort, the development of symptoms of FBD and IBS occurred in only 7.4% and 0.8% of cases, respectively, which was significantly lower than that of the study group (p<0.01). The results of our study are similar to those reported by Neal et al in the UK (FBD 25%, IBS 7%),2 despite the fact that their patients had food poisoning, chiefly resulting from campylobacter or salmonella infection, and our patients had suffered from BD, the majority of whom were infected with shigella. We also noticed that a higher incidence of FBD tended to occur in patients who had a longer duration of BD. This may be explained by delayed or inappropriate initial treatment, or to a more severe inflammation which resulted in deeper impairment of the underlining nerve fibres.10,11 It is suggested from the present study that treatment for acute enteric infection should be given initially and efficiently to prevent susceptible cases from subsequent development of FBD or IBS.
The notion that PI-IBS results from an enhanced inflammatory response is further supported by our observation that expression of IL-1β mRNA in the intestinal mucosa was significantly increased in PI-IBS patients. Our observation is similar to that reported by Gwee and colleagues.12 They found increased expression of IL-1β mRNA in the rectal mucosa of their patients who developed IBS symptoms at three months after acute gastroenteritis but our study provided evidence for a longer duration of inflammatory response after infection. We also detected an increase in the number of mast cells within the lamina propria in the terminal ileum of our IBS patients. This is consistent with the study of Sullivan and colleagues.13 In another study, we found that activation of the mast cell population in the intestinal mucosa was 85% in IBS patients during the active stage and only 15% in non-IBS controls (p<0.01) (to be published). The increase in number and activation of mast cells in the intestinal mucosa and release of its mediators (as represented by IL-1β) probably reflects enhancement of the immune response to previous inflammation in PI-IBS patients. Release of IL-1β may cause altered physiological function, such as diarrhoea, via its inhibitory effect on intestinal transport of water, electrolytes, and small particles.14,15 Also, IL-1β is a potent hyperalgesic agent which may be responsible for hypersensitivity to rectal stimulation in IBS.16 Recently, other authors17,18 have observed a markedly increased number of T lymphocytes in the colorectal mucosa of IBS patients, indicating persistence of the immune response in these patients. These observations together with the results of our study on increased numbers of mast cells and higher expression of IL-1β mRNA in PI-IBS patients strongly suggest that activation of the mucosal immune system as an inflammatory response may play an important role in the pathogenesis of PI-IBS.
Our study provides further evidence that a close attachment exists between mast cells and enteric nerves in the intestinal mucosa of IBS patients. Stead and colleagues6 indicated over a decade ago that there was a close relationship between mast cells and nerves in human mucosa from the ileum to the sigmoid colon in a number of colonic structural diseases. In the present study, we are able to further demonstrate in IBS patients that a higher density of positively stained nerve fibres (such as NSE, SP, and 5-HT fibres) were closely attached to mast cells in the intestinal mucosa. Also, mast cells surrounded by nerve fibres were significantly increased in number in IBS patients (in both PI-IBS and non-PI-IBS) compared with normal controls (p<0.01). Spiller and colleagues17 also identified increased numbers of enteroendocrine cells (including 5-HT containing cell) and T lymphocytes in the rectal mucosa of their PI-IBS patients following campylobacter enteritis. This linkage between immune cells/mast cells and enteric nerves is important as it provides the structural requirement for a dynamic interaction between immune cells and nerves to form integrated neural-immune regulation on gut function in IBS. This was speculated by other investigators recently.19,20 The hypothesis is further supported by identification and excitation of 5-HT3 receptors on sensory neurones in the rat colon.21,22 Neuropeptides, such as SP and CGRP, could modulate the activity of mast cells in the opposite direction.23 Therefore, factors influencing the nervous system may also participate in the pathogenesis of IBS through this pathway.
In conclusion, our study provides new evidence in support of BD as a causative factor of PI-IBS. Our observation that an enhanced inflammatory response occurred in PI-IBS, and the appearance of clustering of nerves and mast cells in the intestinal mucosa of IBS patients, indicates that the immune system and the nervous system both play important roles in the pathogenesis of PI-IBS. However, we cannot explain the fact that there was also an increased number of mast cells and the same increased expression in positively stained nerve fibres in non-PI-IBS and PI-IBS patients. We do not know whether PI-IBS represents a special form of IBS or if the development of all forms of IBS has to undergo the same pathogenetic mechanism. Further studies are needed to elucidate this question.