Parts of the results were presented in abstract form (poster) at DDW 2007 in Washington DC, May 23.
Alain M. Schoepfer MD, Department of Gastroenterology, Inselspital, Bern University Hospital, 3010 Bern, Switzerland. Tel: 01141 31 632 21 11; fax: 01141 31 632 97 65; e-mail: email@example.com
Abstract One of the several possible causes of irritable bowel syndrome (IBS) is thought to be low-grade mucosal inflammation. Flagellin, the primary structural component of bacterial flagellae, was shown in inflammatory bowel disease patients to activate the innate and adaptive immunity. It has not yet been conclusively established if IBS patients show reactivity to luminal antigens. In 266 patients [112 IBS, 61 Crohn’s disease (CD), 50 ulcerative colitis (UC) and 43 healthy controls (HC)], we measured antibodies to flagellin (FAB, types A4-Fla2 and Fla-X), anti-Saccharomyces cerevisiae antibodies (ASCA) (both ELISA), antipancreas antibodies (PAB) and perinuclear antineutrophil cytoplasmatic antibodies (p-ANCA) (both IF). All IBS patients had normal fecal calprotectin (mean 21 μg mL−1, SD 6.6) and fulfilled the ROME II criteria. Frequencies of antibodies in patients with IBS, CD, UC and HC, respectively, are as follows (in per cent): antibodies against A4-Fla2: 29/48/8/7; antibodies against Fla-X: 26/52/10/7; ASCA: 6/59/0/2; p-ANCA: 0/10/52/0; and PAB: 0/28/0/0. Antibodies against A4-Fla2 and Fla-X were significantly more frequent in IBS patients than in HC (P = 0.004 and P = 0.009). Antibodies to A4-Fla2 and Fla-X were significantly more frequent in IBS patients with antecedent gastroenteritis compared to non-postinfectious IBS patients (P = 0.002 and P = 0.012). In contrast to ASCA, PAB and p-ANCA, antibodies against A4-Fla2 and Fla-X were found significantly more often in IBS patients, particularly in those with postinfectious IBS, compared to HC. This observation supports the concept that immune reactivity to luminal antigens has a putative role in the development of IBS, at least in a subset of patients.
Irritable bowel syndrome (IBS) has a multifactorial aetiology and is heterogeneous in its clinical presentation and pathogenesis. In recent years, attention has been directed to the putative role of low-grade mucosal inflammation. A variety of mechanisms may contribute to the low-grade mucosal inflammation in IBS patients, including episodes of infectious enteritis,1 genetic factors,2 undiagnosed food allergies3 and changes in intestinal microflora.4 There is an increasing evidence to support the concept of postinfectious IBS (PI-IBS),5,6 and prospective studies have further clarified this concept.7,8
The link between prior infection and chronic inflammation has been supported by the detection of a persistent increase in rectal mucosal enteroendocrine cells, T lymphocytes and gut permeability in patients suffering from postdysenteric IBS.1 Different groups have found an increase in mast cells in IBS patient, probably attributable to a prior infection.9,10 These studies have examined the consequences of an acute enteric infection on the intestine; others suggest that more chronic alterations of the enteric flora may influence the development of IBS. These exciting results lead to the appraisal that IBS and inflammatory bowel disease (IBD) not only have a certain symptom overlap, but that certain mechanisms in pathogenesis can overlap as well.11
Reactivity to different microbial antigens has been widely published in IBD. So far, anti-Saccharomyces cerevisiae antibodies (ASCA) and antineutrophil cytoplasmatic antibodies (ANCA) are the most frequently studied markers.12 Anti-Saccharomyces cerevisiae antibodies, occurring mainly in CD patients, recognize distinct carbohydrate epitopes of yeast cell wall mannan.13,14 The ANCA are predominantly found in UC and present as atypical perinuclear ANCA (p-ANCA).15 Antibodies to exocrine pancreas (PAB) are highly specific for Crohn’s disease (CD), however, the sensitivity is only moderate.16 The exact antigen for PAB has not yet been elucidated. Flagellin, the primary structural component of bacterial flagella, is recognized by Toll-like receptor 5 and activates the innate as well as adaptive immunity.17 Flagellins are dominant antigens in CD.18 Phylogenic analyses have predicted the origin of the flagellins to be the Clostridium phylogenetic cluster XIVa.19 Duck et al.20 have isolated and characterized a number of flagellated bacteria from the Clostridium cluster XIVa. One of these bacteria, A4, has a flagellin gene highly comparable to the Fla-X flagellin to which CD patients have often seroreactivity.18 Analysis of the 16S rDNA of A4 has placed them in the family Lachnospiraceae (domain Bacteria, phylum Firmicutes, class Clostridia, order Clostridiales, family Lachnospiraceae). The flagellins A4-Fla2 and Fla-X are 97% identical in the amino acid sequence. Fla2 is one of the six flagellins of the A4 organism.
We hypothesized that IBS patients have signs of activated mucosal immunity against common microbial antigens compared to HC.
In this study, we primarily wanted to answer the following question:
• Are antibodies against microbial antigens more frequent in IBS patients in comparison to HC?
Our secondary goal was to answer the following questions:
• Is the antibody concentration to flagellins and oligomannan different in IBS patients compared to IBD patients and HC?
• Is there a correlation between antibodies against flagellins and antecedent gastroenteritis in IBS patients?
Materials and methods
We analysed serum samples from 266 patients [112 IBS, 61 CD, 50 ulcerative colitis (UC) and 43 healthy controls (HC)]. Outpatients and inpatients from the Department of Gastroenterology of the University Hospital Bern were enrolled between January 2005 and October 2007. The diagnosis of IBD or IBS was already known at the time of inclusion or was established during the inclusion period. All research-related activities were conducted with the patients’ consent and with the approval of the local ethical committees.
The diagnosis for each patient was based on standard clinical, endoscopic, histological and radiographic features.21
Inclusion criteria Complete colonoscopy with intubation of the terminal ileum including biopsies (at least two biopsies in the terminal ileum and three colonic biopsies), age 18–85 years. For understandable reasons, the persons in the HC group had no endoscopic workup.
Exclusion criteria Incomplete ileo-colonoscopy, infectious entero-colitis, microscopic colitis, colorectal cancer, colorectal polyps, history of colorectal or small bowel surgery, and regular intake of aspirin and/or NSAID (≥2 tablets per week). The two participating gastroenterologists (board certified) who established the diagnosis were not informed about the results of serum markers.
Diagnosis of colonic disease was established based on clinical history and examination, laboratory findings (haematogram, electrolytes, CRP, ASAT, ALAT, GGT, AP, bilirubin, lipase, creatinin, glucose), abdominal ultrasonography and ileo-colonoscopy also including biopsies of endoscopically normal regions. Additionally, all IBS patients had an endoscopy of the upper gastrointestinal (GI) tract with normal histological findings of the duodenum and stomach.
Patients were allocated to three groups based on the above mentioned findings: IBS, IBD (CD and UC) and HC.
For the diagnosis of IBS, the following criteria were applied: exclusion of infectious diseases (fecal microbiology), celiac disease (small bowel biopsies), chronic IBD (endoscopy, histology), diverticulosis or diverticulitis (endoscopy and/or CT scan), microscopic enterocolitis (histology), or ischaemic and medication induced colitis. All IBS patients fulfilled the ROME II criteria, there were no alarm symptoms such as anaemia or weight loss, the endoscopic and histological workup was normal. The decision to perform a systematic examination of the jejunum and proximal ileum was left to the judgment of the treating gastroenterologist. In total, 39 IBS patients (35%) had a small bowel examination, 26 (23%) by hydro-CT scan, and 13 (12%) by an MR enteroclysis. All IBS patients had normal fecal calprotectin.
The IBD group included patients with CD or UC. Diagnosis was established by symptoms and clinical examination, endoscopic findings, histological analysis, radiological workup and laboratory tests (exclusion of infectious enterocolitis).
The controls were healthy individuals from the clinical and laboratory staff willing to provide blood and fecal samples. All HC were free of symptoms and had a normal clinical examination and abdominal ultrasonography. Except for some women taking birth control pills, these controls were not taking any medication on a regular basis.
Antibodies against microbes and auto-antibodies
The following antibodies were measured: Antibodies to the flagellins A4-Fla2 and Fla-X, ASCA, PAB and p-ANCA. The scientists in the laboratory (TS, BS and SM) were blinded to the patient diagnosis and the study hypothesis. All ELISA were read at an OD of 450 nm (BioTek Instruments, Winooski, VT, USA). Antibody concentrations in patient sera are reported as OD (optical density).
Flagellin ELISA We evaluated two flagellins. Both flagellin A4-Fla2 and Fla-X were provided by CO Elson, University of Alabama in Birmingham. The cutoff was determined by measuring 43 sera of healthy people. The mean of the OD of these HC sera with two times the standard deviation (SD) was determined as cutoff for counting as positive for antibodies against flagellin. The concentration of antibodies in the patient sera was quantified using the following methods: values ranging between two to four times SD above the mean of the HC group were defined as low positive, values ranging from above four up to six times SD above the mean of the HC were defined as intermediate positive, and values ranging above six times SD above the HC mean were defined as high positive antibodies. The ELISA plates were coated either with the A4-Fla2 or with the Fla-X. Coating concentration with flagellins was 0.5 μg mL−1, sera were diluted 1 : 1000. Antibodies against flagellin in patient sera were detected using a peroxidase-marked antihuman immunoglobulin A (IgA) and immunoglobulin G (IgG) goat antibody respectively (Sigma, Buchs, Switzerland). A patient was counted positive when either antiflagellin IgG or IgA or both were present.
ASCA ELISA Phosphopeptidomannans from baker’s yeast (Hefe Vital Gold, Deutsche Hefewerke, Nürnberg, Germany) were extracted as previously described.22 Ninety-six well ELISA plates (MaxiSorbTM; Nunc, Wiesbaden, Germany) were coated with 100 μL of 0.25 μg mL−1 phosphopeptidomannans in carbonate–bicarbonate buffer, pH 9.6 and incubated overnight at 4 °C. Serum diluted 1/1000 in PBS + 0.5% BSA was applied to the coated plates in triplicates. The plates were incubated for 1 h at 37 °C. Secondary antibody was added and plates were incubated for 1 h at room temperature. The secondary antibodies (Sigma) were as follows: goat antihuman IgA (alpha-chain specific) peroxidase conjugate, and goat antihuman IgG (Fc specific) peroxidase conjugate, both at 1/5000 diluted in PBS + 0.5% BSA + 0.05% Tween 20. The plates were developed using TMB substrate (Sigma) and the reaction was stopped with sulphuric acid. The absorbance was read at 450 nm. The mean absorbance value for the HC sera plus two SD was used to discriminate between positive and negative subjects. Absorbance units equal to or above this value were considered positive for ASCA antibodies. A patient was considered ASCA+, when positive for ASCA IgG, IgA, or both. One serum highly positive for ASCA was used to obtain a standard curve at dilutions ranging form 1/100 to 1/102 400. The curve was fitted using a four-parameter logarithmic function.
PAB-immunofluorescence Sections of human pancreas (blood group 0) were used. Patient sera were incubated for 30 min in a moist chamber at room temperature with 50 μL serum samples diluted 1 : 100 in PBS. After washing in PBS, slides were incubated with polyvalent fluorescein-conjugated rabbit antihuman immunoglobulin (Dako, Germany) (diluted 1 : 100 in PBS) for 30 min. After washing, the sections were embedded in Immuno-Mount (Shandon, Pittsburgh, PA, USA) and read using a Nikon fluorescence microscope. All slides were coded for anonymity and read independently by two investigators.
p-ANCA testing p-ANCA testing was performed on cytospins of neutrophils as described elsewhere.23
The scientist (BS) performing the analyses was blinded to the patient diagnosis and the study hypothesis. All fecal samples were processed within 48 h after collection. PhiCalTM Test was purchased from Medical Instrument Corporation, Solothurn, Switzerland, Art-No. 006, the test is produced by CALPRO AS, Oslo, Norway. This sandwich ELISA measures quantitative calprotectin. Fecal specimens were diluted at 1 : 2500. The assay was performed according to the test instructions.
Data were listed on an Excel sheet (Microsoft Excel 2003; Microsoft Switzerland Ltd Liab. Co., Wallisellen, Switzerland). For statistical analysis, data were imported into a statistical package program (stata Vs 9.0, College Station, TX, USA). Results of numerical data are presented as mean ± SD and range. Categorical data were summarized as the percentage of the group total. Normal distribution of data was tested using a Normal-QQ-Plot. Fisher’s exact test (two-sided) or the chi-squared test was used to explore associations of categorical data in two independent groups. The Wilcoxon rank sum test was used to explore associations of numerical data in two independent groups. A P < 0.05 was considered statistically significant. A Bonferoni adjustment was performed where appropriate. For graphical evaluation, box plots were calculated.
A total of 266 patients were included, the demographic details of these patients are shown in Table 1. Mean fecal calprotectin was significantly higher in CD and UC patients compared to IBS patients or HC (P < 0.0001). All IBS patients and all HC had a normal laboratory workup (especially normal C-reactive protein and normal leucocytes).
Table 1. Patient characteristics
HC, healthy controls; IBS, irritable bowel syndrome; SD, standard deviation; SES-CD, simplified endoscopic score for Crohn’s disease.
Antibodies to flagellins are more frequently found in IBS than in HC
The frequencies of antibodies against flagellin A4-Fla2 and Fla-X, ASCA, PAB and p-ANCA are listed in Table 2. Antibodies against A4-Fla2 and Fla-X were found significantly more frequently in IBS patients compared to HC (P = 0.004 and P = 0.009 respectively).
Table 2. Frequency (numbers and per cent) of antibodies in different patient groups
Cutoff for positive samples (ELISA): 2 SD above mean of HC. ASCA, Anti-Saccharomyces cerevisiae antibodies; p-ANCA, perinuclear anti-neutrophil antibodies; CD, Crohn’s disease; HC, healthy controls; IBS, irritable bowel syndrome; PAB, antibodies against exocrine pancreas; UC, ulcerative colitis.
IBS (n = 112)
CD (n = 61)
UC (n = 50)
HC (n = 43)
Crohn’s disease patients had antibodies against A4-Fla2 and Fla-X significantly more frequently than IBS patients (P = 0.016 and P = 0.001) and compared with UC patients (P < 0.0001 for both).
Anti-Saccharomyces cerevisiae antibodies were found in a comparable frequency in HC and in IBS patients (P = 0.311). The p-ANCA were significantly more frequent in UC compared to CD patients (P < 0.0001).
Agreement between the anti-A4-Fla2 and Fla-X antibodies
The agreement between antibodies (IgG and/or IgA) to flagellins A4-Fla2 and Fla-X was overall 78%. In CD patients, the agreement was 79%, in UC patients 68%, compared to 67% in IBS patients and 100% in HC. In all patient groups, the frequency of IgA antibodies to flagellins was below 20%. In the IBS group, 32 patients were positive for A4-Fla2, of these, 30 (94%) had IgG, 6 (19%) had IgA and 2 (6%) had IgA only without IgG. The results for antibodies to Fla-X in the IBS group are as follows: 29 positive for Fla-X, of these, 28 (97%) had IgG, 4 (14%) had IgA and only 1 (3%) had IgA without IgG.
Antibody concentrations against flagellins A4-Fla2 and Fla-X in IBS and CD patients
We determined three groups of antibody concentrations: low positive, intermediate and high positive (see Flagellin ELISA in Materials and methods). We hypothesized that a long-lasting chronic active inflammation as can be seen in CD patients would produce higher antibody concentrations than in marker-positive IBS patients with postulated low-grade mucosal inflammation. The results of the antibody concentrations against A4-Fla2 and Fla-X are shown in Table 3.
Table 3. Frequency (in raw numbers and per cent) of antibody titres (low = 2–4 SD above mean, intermediate ≥ 4–6 SD above mean, and high ≥ 6 SD above mean) against the flagellins A4-Fla2 and Fla-X in the different patient groups
IBS (n = 112)
CD (n = 61)
UC (n = 50)
HC (n = 43)
Values in parentheses are percentages. Cutoff for positive samples: 2 SD above mean of HC. ASCA, anti-Saccharomyces cerevisiae antibodies; CD, Crohn’s disease; HC, healthy controls; IBD, inflammatory bowel disease; IBS, irritable bowel syndrome; UC, ulcerative colitis.
In summary, IBS patients were found to have significantly more frequently low than intermediate/high antibody concentrations against A4-Fla2 as well as against Fla-X (P < 0.0001 for both). In contrast to IBS patients, CD patients were found to have significantly more frequently intermediate/high than low antibody concentrations against A4-Fla2 (P = 0.003) and Fla-X (P = 0.007).
In the A4-Fla2 positive cohort, CD patients were found to have significantly more frequently high antibody concentrations compared to the IBS patients (14/29 vs 1/32, P < 0.0001). The same was observed in the Fla-X positive cohort where CD patients had significantly more frequently high antibody concentrations compared to the IBS patients (15/32 vs 2/29, P = 0.001).
The antibody concentrations in different patient groups are illustrated in Fig. 1. In summary, high antibody concentrations against the flagellins were predominately found in CD patients.
Frequency of antecedent gastroenteritis in IBS patients
It is known that IBS symptoms can start in a subset of patients after an episode with gastroenteritis.
We identified 27 IBS patients (24%) reporting an antecedent gastroenteritis. The mean time interval between gastroenteritis and development of IBS symptoms was 2.5 ± 1.9 months (range 1–9 months). In eight patients fecal cultures were performed, in two patients Campylobacter jejuni and in one patient Salmonella was identified as the causative agent; in other words, the causative agent in 24/27 (89%) was not identified.
Of this PI-IBS cohort, 14 (52%) patients were A4-Fla2 positive and 12 (44%) were Fla-X positive. Antibodies against flagellin A4-Fla2 and Fla-X were found significantly more often in IBS patients reporting an antecedent gastroenteritis (A4-Fla2: 14/27 vs 18/85, P = 0.002; Fla-X: 12/27 vs 17/85, P = 0.012) compared to patients without.
In this study, we evaluated the frequency of antimicrobial and antiself antibodies in IBS patients compared to IBD patients and HC.
While the cause of IBS remains obscure, most hypotheses on the pathogenesis deal with one of the following subjects: immune activation, altered intraluminal milieu, enteric neuromuscular dysfunction and brain–gut axis dysregulation. There is, especially in patients with postinfectious complaints (PI-IBS), accumulating evidence to suggest an association between IBS and immune activation. The incidence of PI-IBS is highly variable and ranges from 8% to 29%.24,25 We have identified 24% of patients reporting an antecedent gastroenteritis, this frequency is in accordance to already published data. Colonic biopsies in PI-IBS patients have demonstrated in some cases frank inflammation and in others more subtle signs of immune activation.26 The suggestion that inflammation may play a putative role in IBS is further supported by the presence of low-grade lymphocytic infiltration in the myenteric plexus.27 Our findings with elevated antibodies against flagellins in IBS patients, especially those suffering from PI-IBS, are indicative of an immune activation. Flagellins nowadays are recognized to be potent GI antigens capable of activating the innate as well as adaptive immunity, the inflammatory pathway is driven by the activation of TLR5.17 Worth noting is that we found antiflagellin antibody concentrations in CD patients to be significantly higher than in IBS patients. This may indicate that the antibodies in IBS patients represent an immunological answer to a low-grade inflammatory status or a temporary past mucosal break with immunological memory, in contrast to IBD patients where the chronic, more intense inflammation leads to more immunological contacts with flagellins, and is represented by the higher levels of antiflagellin antibodies. Furthermore, antibodies to flagellins were identified in 26% and 29% of IBS patients, predominately in PI-IBS; in other words, the majority of IBS patients did not have these antibodies. This indicates that flagellin-specific immune reaction plays a pathogenic role only in a subset of IBS patients and that in other IBS subgroups specific immune responses against different microbial antigens may be less important.
To the best of our knowledge, we have evaluated the frequency of antibodies to flagellin, ASCA, p-ANCA and PAB in IBS patients for the first time, therefore, a comparison with the results of other groups is not possible in this specific question. Since the first description of the CBir1 flagellin,18 many other flagellins have been identified in the meantime.20 The frequencies of A4-Fla2 and Fla-X antibodies in different patient groups have not yet been published. However, our antiflagellin antibody frequencies are comparable to the results with CBir1 in CD, UC patients and HC.17,28,29 Our ASCA, p-ANCA and PAB frequencies in the different patient groups are in accordance to other studies.30–34
Evidence accumulates that inflammation plays a role, at least in a subset of IBS patients.35 Although inflammation of a higher intensity is the hallmark of IBD, it is clear that IBS and IBD are not the same disorder nor do they represent opposite components of the same spectrum. Inflammatory bowel disease provides evidence for a role of inflammation in motor dysfunction, some of these motor abnormalities may persist following apparent resolution of inflammation. This may explain why some IBD patients can develop an IBS-type syndrome during periods of IBD remission. This suggestion is supported by studies associating IBS-type symptoms with impaired quality of life among IBD patients in remission.36,37 Simren et al.37 evaluated patients suffering from UC or CD who had been in remission of IBD for at least 1 year: 57% of the CD patients and 23% of the UC patients reported IBS-type symptoms.
If inflammatory response against microbes plays a role, at least in a subset of IBS patients, the question arises if anti-inflammatory therapy, for example, antibiotics or probiotics, might proof helpful in symptom therapy. Unfortunately, specific anti-inflammatory treatments have not been systematically evaluated and as yet there is no evidence of benefit.38
One limitation of our study lies in the problematic definition of the disease entity ‘IBS’. Of note, it is neither recommended to perform a systematic small bowel examination in these patients, nor to look for small bowel inflammation, e.g. by the use of fecal leucocyte markers such as calprotectin and lactoferrin that have shown to possess a high negative predictive value for excluding intestinal inflammation.39–41 Also, in our cohort, the decision to perform a small bowel examination was left to the clinician’s judgment, which represents common clinical practice. Therefore, the clinician is challenged with a remaining diagnostic uncertainty of whether a small bowel CD in this particular patient group has been missed. We are aware of the fact that the diagnostic workup in our IBS patients is above the usual, however, we have tried to further diminish the diagnostic uncertainty by routinely documenting a normal fecal calprotectin. Furthermore, a relatively high proportion of our IBS patients also had a small bowel examination.
Because of the high accuracy in discriminating IBD from IBS,42 we recommend to measure routinely fecal calprotectin or lactoferrin in IBS patients. However, looking at the current clinical recommendations and also the prevalence of IBS with its socio-economic burden, it is certainly not justified to perform a systematic small bowel examination on every IBS patient.
Another limitation may be a recall bias regarding the patient subgroup with PI-IBS. This problem could be avoided by the identification of pure PI-IBS patients in a large prospective study where patients are followed through an infectious outbreak. Besides, a higher proportion of culture proven patients with acute gastroenteritis developing later PI-IBS would also be desirable.
In summary, in contrast to ASCA, PAB and p-ANCA, antibodies against A4-Fla2 and Fla-X were found significantly more often in IBS patients compared to HC. Postinfectious IBS patients had more frequently antibodies to flagellins compared to the ‘non-PI-IBS’ group. Our findings support the accumulating evidence that reactivity to luminal antigens has a putative role, at least in a subgroup of IBS patients. So far, it remains unclear whether these antibodies represent an immunological memory of a former breakdown in mucosal integrity or whether flagellin acts as an active stimulus contributing to low-grade inflammation. Long-term studies are needed to determine if IBS patients with antibodies to flagellin develop differently from the ones being antiflagellin negative, furthermore, controlled studies should be performed to evaluate the role of anti-inflammatory therapy (e.g. probiotics) for symptom improvement in IBS patients.
The authors thank Professor CO Elson, MD, Division of Gastroenterology and Hepatology, Department of Medicine, University of Alabama at Birmingham, for critical discussion and the donation of the flagellins. The authors also acknowledge Miss L Bolzern for her thorough review.