Article Text
Abstract
Objective To determine the role of colonic barrier defects and low-grade inflammation in irritable bowel syndrome (IBS)-like symptoms in quiescent inflammatory bowel disease (IBD).
Design Caecal biopsies were collected from 51 IBS, 49 quiescent IBD (31 Crohn's disease (CD) and 18 ulcerative colitis (UC)) patients and 27 controls. IBS was assessed using the Rome III criteria and the IBS severity score. Epithelial barrier integrity was evaluated by determining the paracellular permeability of biopsies mounted in Ussing chambers and the mRNA expression of tight junction proteins (ZO-1, α-catenin and occludin). Low-grade inflammation was evaluated by counting cells, including intraepithelial lymphocytes (IELs), eosinophils and mast cells, and by determining the mRNA and protein expression of tumour necrosis factor (TNF)-α in biopsies and culture supernatants.
Results IBS-like symptoms were present in 35.4 and 38% of CD and UC patients, respectively. Paracellular permeability was significantly increased in both quiescent IBD with IBS-like symptoms and IBS compared with quiescent IBD without IBS-like symptoms (p<0.01, respectively) or controls (p<0.01, respectively). Significantly lower expression of ZO-1 and α-catenin was detected in IBS and quiescent IBD with IBS-like symptoms. IELs and TNF-α were significantly increased in quiescent IBD with IBS-like symptoms, but not in IBS.
Conclusions In quiescent IBD, IBS-like symptoms related to persistent subclinical inflammation associated with increased colonic paracellular permeability. A persistent increase in TNF-α in colonic mucosa may contribute to the epithelial barrier defects associated with abdominal pain in quiescent IBD, but not in IBS. Optimisation of anti-inflammatory therapy may be considered in quiescent IBD with IBS-like symptoms.
- ABDOMINAL PAIN
- INFLAMMATORY BOWEL DISEASE
- IRRITABLE BOWEL SYNDROME
- TNF-ALPHA
- TIGHT JUNCTION
Statistics from Altmetric.com
Significance of this study
What is already known about this subject?
-
IBS-like symptoms are common in patients with quiescent IBD.
-
IBS-like symptoms have a negative impact on the patient's QOL.
-
Quiescent IBD patients with IBS-like symptoms have increased faecal calprotectin, a marker of inflammation.
What are the new findings?
-
IBS-like symptoms in quiescent IBD are associated with ongoing inflammation and epithelial barrier disruption, which is the case for ‘true’ IBS.
-
Persistent increased expression of TNF-α in colonic mucosa may contribute to epithelial barrier defects found in IBD with IBS-like symptoms, but not in IBS.
How might it impact on clinical practice in the foreseeable future?
-
These findings have important clinical consequences in terms of optimisation of anti-inflammatory therapy in IBD patients considered to be in remission.
Introduction
Inflammatory bowel diseases (IBD), including ulcerative colitis (UC) and Crohn’s disease (CD), are thought to result from the activation of the mucosal immune system and the disruption of the epithelial barrier by the intestinal microbiota, which are likely to be influenced by genetic factors.1 ,2 Current drug development for IBD targets predominantly the control of symptoms by modulating the intestinal immune system. However, a substantial proportion of IBD patients continue to suffer from abdominal pain and distension, which mimics irritable bowel syndrome (IBS) while in remission (eg, IBS-like symptoms).3 Several studies have reported IBS-like symptoms in approximately a third of UC and a half of CD patients in remission, but symptoms cannot be attributed to an identifiable inflammatory disease activity.4
Whether IBS-like symptoms associated with quiescent IBD reflect the coincident occurrence of IBS or persistent, but clinically undetected, low-grade inflammation is still unclear. Today, many studies support a role for mucosal inflammation resulting in sensorimotor dysfunction, which may persist after the mucosal inflammation flare-up has been resolved. In patients with quiescent IBD, Keohane et al5 identified high levels of faecal calprotectin, a surrogate marker of colonic low-grade inflammation, which supports the notion that subclinical immune activation is more likely to explain residual abdominal pain than IBD coexisting with ‘true’ IBS. An inflammatory-induced visceral hypersensitivity has also been reported in quiescent IBD with IBS-like symptoms, showing a 3.9-fold increase in transient receptor potential vanilloid type 1 (TRPV1)-immunoreactive nerve fibres, which correlated with the severity of abdominal pain.6
There is also mounting evidence demonstrating the contribution of epithelial barrier defects (eg, ‘leaky gut’) to the development of visceral hypersensitivity and abdominal pain in both IBS and IBD. Increased paracellular permeability has been documented in the acutely inflamed and chronically damaged intestinal mucosa of IBD and IBS patients.7–9 Disruptions in the IBS-induced epithelial barrier have been associated with an altered expression of tight junction (TJ) proteins, including ZO-1 and occludin.2–10 However, the role of α-catenin, an important protein that localises to adherent junctions and is linked to E-cadherin, has never been explored in IBS. Several studies have demonstrated that patients with clinically inactive CD have an increased permeability of their small intestine and these patients are at elevated risk of flare-up.11 ,12 Numerous soluble factors released by mucosal immune cells have been shown to contribute to the pathology in these patients. For example, in IBS patients, we and others have shown previously that mast cells (MCs) were located in close vicinity to mucosal nerves and that their secreted mediators (tryptase and histamine) caused altered epithelial permeability and neural function, thus increasing abdominal pain.2 ,13 ,14 In contrast to IBS, several pro-inflammatory cytokines, including tumour necrosis factor (TNF)-α, are likely to be responsible for breakdown of TJs and increased intestinal permeability in IBD patients.15–18 Together, these data suggest that TNF-α can alter the TJ permeability via cytoskeletal contraction.19 ,20
To identify whether inflammatory mechanisms contribute to IBS symptoms and IBS-like symptoms observed in quiescent IBD is important for future treatment of these patients. We therefore aimed to determine the contribution of the activation of the mucosal immune system and of the epithelial barrier disruption to the occurrence of IBS-like symptoms in patients with IBD while on prolonged remission.
Methods
Design of the study
In this prospective study conducted between January 2008 and March 2012, we enrolled consecutive patients who required colonoscopy for the purpose of IBD in remission (31 CD and 18 UC), IBS (n=51) and healthy controls (n=27). Colonoscopy also examined the terminal ileum and 12 biopsy specimens were taken at the caecum. Large bowel preparation was similar for all patients and controls (ie, polyethylene glycol). All subjects completed questionnaires evaluating the severity of IBS symptoms. None had any clinical traits of depression or anxiety. The integrity of the colonic epithelial barrier was assessed by measuring the paracellular permeability of biopsies and the mRNA expression of the mucosal TJ proteins. Mucosal low-grade inflammation was evaluated by immunochemistry and by TNF-α mRNA expression assessed by quantitative real-time PCR (qPCR).
Participants
Information collected from IBD patients included disease extent, duration, current medication, symptoms and smoking status. IBD patients were in remission when meeting all of the following criteria: (i) Physicians Global Assessment (based on history and physical examination); (ii) C reactive protein<10 mg/L, erythrocyte sedimentation rate <20 mm/h, platelets <450×1012/L and white cell count <11×109/L; (iii) no use of corticosteroids over the last 12 months; (iv) the CD activity index ≤150 or an UC activity index ≤3; and (v) macroscopically normal mucosa on colonoscopy, including the terminal ileum in CD (scores Mayo=0 and Crohn's Disease Endoscopic Inflammation Score (CDEIS)≤4).21 ,22 Patients with stenotic disease or complicated CD were excluded. Patients with another coexisting disease (eg, malignancy, unstable cardiovascular, hepatic, renal or psychiatric disease) were excluded. The Rome III criteria for IBS23 were used to define the IBS-like symptoms in quiescent IBD and were subdivided into four groups: CD with and without IBS-like symptoms and UC patients with and without IBS-like symptoms.3
IBS patients were selected according to the Rome III criteria based on the presence of abdominal pain and/or discomfort, relieved by defecation or associated with a change in stool frequency and consistency. Symptoms had to be present more than 25% of the time and were evaluated daily using the Bristol Stool Scale, which was noted in a diary for 10 days. IBS subtypes were defined according to the predominance of constipation (IBS-C), diarrhoea (IBS-D) or mixed symptoms (IBS-M). These criteria had to be present for at least 6 months before the interview and to last for at least 3 months. Patients were thoroughly examined for other organic diseases before entering the study. None of the IBS subjects were of postinfective origin (ie, IBS symptoms did not begin after a bout of gastroenteritis). Moreover, none of the IBS patients presented with a history of gastrointestinal (GI) infection over the past 6 months. None were taking anti-inflammatory drugs or steroids. The control group included asymptomatic patients who had colonoscopy for polyp surveillance. The clinical characteristics of subjects are detailed in table 1.
The study was approved by the ethics committee (CCPPRB N° 07.047, Programme Hospitalier de Recherche Clinique 2007–2008) and all patients gave written informed consent.
Severity of IBS symptoms and quality of life
Each individual completed the validated French IBS severity scoring system developed by Francis et al24 and used in a large cohort of patients in France.2 ,3 The questionnaire asked about abdominal pain and bloating, as well as the intensity and impact of symptoms. The total score of the questionnaire ranged from 0 to 500.
Histological assessment
Cell counts were performed twice by a single pathologist blinded to the diagnosis, as previously described.13 Two biopsies were fixed in formol and histopathology, morphometry and immunohistology were performed on paraffin-embedded 4 μm-thick sections. Briefly, slides were incubated for 1 h with prediluted (1/500) mouse monoclonal antibodies specific for CD3 (1/2 h, IR503) and CD-117 (1 h, A4502) (Dako Corporation). Sections were then incubated with secondary biotinylated anti-mouse rabbit antibodies (Dako StreptABC complex) and then with a streptavidin/horseradish peroxidase conjugate. Conventional CD-117 immunostaining was done on the sections to calculate the absolute number of MCs within three non-overlapping high power fields (hpfs) at 400×(0.0125 mm2) (Olympus B×41 light microscope). Cells were classified by the pathologist as eosinophils or mastocytes. Intraepithelial lymphocytes (IELs) were enumerated as CD3 cells per 100 epithelial cells, after counting a minimum of 200 cells in at least three separate sections. Transitional lymphocytes traversing the basement membrane were not included.
Colonic paracellular permeability
The colonic paracellullar permeability was measured in the caecum as previously described.2 ,25 Four biopsies were mounted in adapted Ussing Chambers (Transcellab; TBC, Paris, France) with 0.0314 cm2 of exposed surface. Biopsies were bathed on each side with 3 mL of Ham's Nutrient Mixture (HAM's F12 Glutamax; Invitrogen) and continuously oxygenated (95% O2/5% CO2) and maintained at 37°. After a 15 min baseline or equilibration period, 150 μl of apical medium was replaced by the same volume of fluorescein-5.6 sulfonic acid (FITC-sulfonic acid; 10 mg/mL, 400 D, Invitrogen). The fluorescence level was measured at 30 min intervals for 180 min using a fluorimeter (Tecan Infinite F500; Tecan SA, Lyon, France). Fluorescence values were converted to fluorescein concentrations (ng/mL) calculated from a standard curve. Values for permeability were the mean value of four biopsies measured at 180 min with a calculated individual coefficient of variability less than 5%.
TNF-α assay in the supernatant of cultured biopsies
We prepared conditioned medium from colonic biopsies as previously described.2 ,25 Briefly, two colonic biopsies per patient were immersed in a tube containing 1 mL of Hank's buffered saline solution (HBSS; Invitrogen, Cergy-Pontoise, France) and continuously oxygenated (95% O2/5% CO2) at 37°C. After 20 min, the solution was removed and centrifuged at 200 g for 10 min before filtration through a centrifuge filter (200 μm, SPIN-X; Corning, New York, USA) to remove bacterial components. Conditioned medium stored at –20°C was obtained from 16 IBS, 34 CD, 14 UC and 7 controls from the same cohort and was used for Ussing permeability assays and qPCR. The conditioned media were tested for TNF-α using a human TNF-α high sensitivity ELISA (eBioscience, Vienna, Austria). Experiments were performed in duplicate and testing was carried out according to the manufacturer's recommendations. TNF-α concentrations were normalised to the biopsy weight and expressed in pg/mg.
To analyse the TNF-α response to lipopolysaccharide (LPS) stimulation, we collected additional colonic biopsies from 10 quiescent IBD with (n=6) or without (n=4) IBS-like symptoms. Colonic biopsies were cultured for 2 h as described above, with or without the addition of LPS 1 ng/mL.26 Conditioned media were then collected, centrifuged and tested in duplicate for TNF-α by ELISA.
Quantitative real-time PCR
Immediately after endoscopy, four biopsy specimens were transferred to lysing matrix tubes (Mpbiomedicals, Illkirch, France) containing 350 µl of RLT lysis buffer (Qiagen, Maryland, USA) and the RNA was extracted using Nucleospin RNAII kit (Macherey-Nagel, Hoerdt, France). cDNA was obtained using the high cDNA reverse transcription kit (Applied Biosystems, California, USA) according to the manufacturer's instructions. DNA templates were amplified with specific primers designed to span exon–intron junctions to prevent amplification of genomic DNA (Invitrogen, Saint Aubin, France).
Primer sequences:
Human ZO-1, ZO-1_F: CAGTGCCTAAAGCTATTCCTGTGA, ZO-1_R: CTATGGAACTCAGCACGCCC;
Human occludin, OCLN_F:GCAGGAAGGTCAAAGAGAACAGA OCLN_R:ATATTCCCTGATCCAGTCCTCCTC;
Human α-catenin, CTNNA1_F:AGAGCAAGCTGGATGCTGAAG CTNNA1_R:CCGATGTATTTTTGAGTGGTCCTT;
Human TNF-α, TNF-α_F: ATCTTCTCGAACCCCGAGTGA TNF-a_R: GGAGCTGCCCCTCAGCTT;
Human GAPDH F: CCATCAATGACCCCTTCATTG R: CTTGACGGTGCCATGGAATT Human ubiquitin, F: CCGACCACAGTGGCTATGC R:CCTCTTTTAATATCTCCAGGCTTGA;
Human S6, S6_F: CCAAGCTTATTCAGCGTCTTGTTACTCC S6_R: CCCTCGAGTCCTTCATTCTCTTGGC.
The qPCR analyses were performed using a 7900HT Sequence Detector System (Applied Biosystems) and the SYBR Green kit (Invitrogen) as previously described.27 ,28 Amplification and quantification of products were performed as follows: 2 min at 50°C to remove uracil residues, 2 min at 95°C followed by 40 cycles of denaturation (15 s at 95°C) and annealing/extension (1 min at 60°C). Experiments were performed in duplicate. Expression of target genes was measured after normalisation of RNA concentration to four different housekeeping genes (Glyceraldehyde-3-phosphate dehydrogenase (GADPH), actin, S6 and ubiquitin) and the variation in expression of target genes between ΔCT (sample) and ΔCT (CTRL) was expressed as a fold increase in expression above the control.29 ,30
Statistical analysis
The parameters of the study population were compared using the χ2 test for qualitative variables and the Kruskal–Wallis and Mann–Whitney non-parametric tests for quantitative variables (medians and 25th and 75th quartiles). Association between permeability, MCs and the severity of IBS was a decision made a priori using linear regression analysis. Data are given as the median and total ranges. A p value <0.05 was considered statistically significant.
Results
Among the 127 individuals recruited, there were 51 IBS, 49 quiescent IBD (31 CD and 18 UC) patients and 27 controls (table 1). Among the IBD patients who were in remission, as based on predefined criteria, 35.4% (11/31) with CD and 38% (7/18) with UC fulfilled the Rome III criteria for IBS diagnosis and were considered as having IBS-like symptoms. The presence of IBS-like symptoms in quiescent IBD was not influenced by the nature of treatment of IBD, disease extent, the time since the last relapse over the 6 months of remission, gender or age (data not shown).
Clinical characteristics of patients and controls
The severity of IBS was significantly more marked in patients with IBS compared with CD and UC patients and healthy controls (table 1). The severity of IBS-like symptoms was not statistically different between quiescent CD and UC patients (p = 0.1). The severity of IBS symptoms, the IBS subtypes and the presence of IBS-like symptoms in CD and UC patients are given in figure 1.
CD patients with IBS-like symptoms had a significantly higher CD activity index than those without (95 (54–150) vs 84 (27–118), p=0.03). UC patients with IBS-like symptoms had a higher UC activity index than those without (1.2 (0.9–2) vs 0.9 (0.2–1.6), p=0.01).
The severity of IBS-like symptoms was not influenced by the disease extent or the disease duration neither in CD nor in UC (data not shown).
Epithelial barrier integrity
The paracellular permeability was significantly higher in IBS (60(29–229) ng/mL) than in controls (38(3–95) ng/mL, p<0.0001) and similar among IBS subtypes (72(31–198) in IBS-M, 57(34–152) in IBS-D, 62(29–229) in IBS-C ng/mL, p=NS) (figure 2). IBD patients with IBS-like symptoms had a significantly higher paracellular permeability than those without (74(38–98) vs 46(27–72) ng/mL in CD, p=0.04; 86(36–124) vs 53(28–81) in UC, p=0.03). IBD patients with IBS-like symptoms and IBS patients had similar increased values for paracellular permeability (figure 2). In contrast, IBD patients without IBS-like symptoms had a similar paracellular permeability compared with healthy controls. Paracellular permeability significantly and positively correlated with the IBS severity score in patients with IBS (r=0.42, p=0.002), CD (r=0.44, p=0.01) and UC (r=0.50, p=0.03) (figure 3).
The mRNA expression of ZO-1 and α-catenin was significantly decreased in both IBS and IBD patients compared with controls. The decrease was more pronounced in IBD patients with IBS-like symptoms (figure 4). The mRNA expression of occludin was significantly decreased in both IBD and IBS compared with controls without any significant difference between patients with or without IBS-like symptoms (figure 4).
Mucosal low-grade inflammation
Cell count
The histological findings are given in table 2. The number of MCs was significantly higher in patients with IBS (12.0(9–17 n/hpf), CD (12.0(4–16 n/hpf) or UC (11.5(5–15 n/hpf) than controls (6.0(2–9) n/hpf, p<0.001, respectively). In patients with CD or UC, the number of MCs was not significantly different between patients with or without IBS-like symptoms. The number of IELs was significantly higher in patients with CD (13.0(5–23) n/hpf) and UC (13.0(5–25) n/hpf) compared with IBS (7.5(4–12) n/hpf, p<0.001, respectively) and controls (5.0(3–12) n/hpf, p<0.001, respectively). In patients with CD and UC, the number of IELs was significantly increased in patients with IBS-like symptoms. The number of eosinophils was significantly higher in UC patients compared with controls (15(5–19) vs 12(5–14) n/hpf, p=0.01).
TNF-α mRNA and protein
Figure 5 shows the mRNA and protein expression of TNF-α in the IBS subtypes and according to the presence or not of IBS-like symptoms in patients with CD and UC. Quiescent IBD patients with IBS-like symptoms had a significantly higher mRNA TNF-α expression compared with those without (7.0(1.2–25) vs 1.8(0.1–20)-fold increase in CD, p=0.0001; 8.6(2.9–23) vs 2.3(0.1–5)-fold increase in UC, p=0.004) (figure 5A). The expression of TNF-α mRNA was similar between IBS patients and controls. The expression of TNF-α mRNA was not influenced by the disease extent or the disease duration in CD or in UC patients (data not shown).
TNF-α protein expression was determined from the biopsy cultures from 16 IBS, 34 CD, 27 UC patients and 7 controls. Data showed a trend favouring higher TNF-α expression in quiescent IBD patients with IBS-like symptoms (figure 5B), with a positive correlation to the mRNA and protein expression (r=0.35, p<0.01).
As the correlation was weak, we also tested the TNF-α response to LPS stimulation in an additional cohort of 10 IBD patients in remission (five CD and five UC), six out of the 10 had IBS-like symptoms. TNF-α levels measured in the absence of LPS stimulation showed the same trend observed with the first cohort, with higher TNFα levels secreted by the biopsies of patients with IBS-like symptoms (figure 6). After LPS stimulation, the TNF-α response was significantly higher in patients with IBS-like symptoms (p<0.01) (figure 6). There were no significant differences between CD and UC, although the sample size was low (data not shown).
In quiescent IBD patients, the TNF-α mRNA expression positively and significantly correlated with the paracellular permeability (r=0.34, p=0.01) (figure 7).
Discussion
In the present study, we report the novel finding that quiescent IBD patients with IBS-like symptoms had persistent epithelial barrier defects, involving increased paracellular permeability and alterations to TJs, which are similar to those seen in patients with IBS. Moreover, we found a significant association between an increase in paracellular permeability and the severity of IBS symptoms. We also found that mild mucosal immune activation (eg, increased IELs and expression of TNF-α) was more likely to be associated with IBS-like symptoms in quiescent IBD, but not in IBS. Therefore, our findings strongly suggest that the persistence of IBS-like symptoms in quiescent IBD may be more related to ongoing but undetectable inflammation due to IBD rather than the coexistence of ‘true’ IBS.
In the present study, we confirmed that a substantial proportion of quiescent CD or UC patients (35% to 38%) had persistent IBS-like symptoms.3 The severity of IBS-like symptoms was similar between quiescent IBD and ‘true’ IBS patients. Moreover, the presence of IBS-like symptoms in quiescent IBD was associated with impaired quality of life (QOL).
While confirming that IBS-like symptoms were common among IBD patients in remission,3 this study showed that IBS-like symptoms were related to persistent subclinical and undetected IBD activity. In fact, our findings are in line with those of Keohane et al5 who recently found significant high levels of faecal calprotectin, a marker of gut inflammation, in 54/106 quiescent IBD patients with IBS-like symptoms. In our study, mucosal inflammation was studied by counting immune cells in the caecum (eg, IELs and MCs) and by measuring the mucosal mRNA expression of pro-inflammatory TNF-α as well as secretion of the protein into the culture supernatant. We observed that increased TNF-α mRNA was associated with increased TNF-α protein expression at baseline and in response to LPS stimulation. Interestingly, the number of IELs was significantly increased in quiescent IBD compared with IBS patients or controls, particularly in quiescent IBD with IBS-like symptoms. The number of IELs was similar between ‘true’ IBS and controls, as previously reported.13 In addition, the number of MCs was significantly increased in both IBS and quiescent IBD patients with IBS-like symptoms, suggesting that the pattern of mediators released by MCs may be more important than absolute cell numbers in contributing to abdominal pain. Furthermore, TNF-α mRNA expression in biopsies was significantly higher in quiescent IBD patients with IBS-like symptoms compared with those without symptoms or with ‘true’ IBS and the increase in TNF-α correlated with an increase in paracellular permeability. Moreover, TNF-α secretion was higher in colonic biopsies from patients having IBD with IBS-like symptoms. Therefore, these findings strongly suggest that an incomplete remission of IBD is more likely to be explained by the presence of IBS-like symptoms than the existence of ‘true’ IBS. These findings are also in accordance with previous studies on IBS patients that showed a marked increase in infiltration of colonic mucosa by IELs and MCs13 ,14 ,31; producers of large amounts of soluble mediators, including proteases, histamine and prostaglandins, but not pro-inflammatory cytokines, such as TNF-α. Furthermore, quiescent IBD patients with IBS-like symptoms had a significantly higher disease activity, although it was considered to be in the normal range. Our findings should therefore question whether IBD remission is sufficiently defined using currently available endoscopic and clinical tools in clinical practice. Taken together, our study suggests that low-grade inflammation constitutes a plausible underlying explanation for IBS-like symptoms in patients with IBD while on prolonged remission, but that strong mechanistic differences in immune activation exist between traditional IBS and IBS-like symptoms overlapping with quiescent IBD.
Disruption of the intestinal epithelial barrier is a feature of gut inflammation in humans and is a pathogenic factor in IBD32 and more recently described in IBS.2 ,33 ,34 We showed previously that the paracellular permeability of colonic biopsy specimens of the caecum was significantly increased in all IBS subtypes and associated with IBS severity.2 These epithelial barrier defects have also been reported in IBS-D patients for jejunal mucosa that displayed disrupted apical junction complex integrity associated with MC activation and clinical symptoms.35 In the present study, using the same methodology, patients having IBS or quiescent IBD with IBS-like symptoms showed a similar degree of epithelial barrier disruption. Interestingly, both IBS and quiescent IBD patients with IBS-like symptoms showed a similar level of expression of TJ proteins (ZO-1, α-catenin and occludin). Therefore, our study confirmed that the underlying basis of abdominal pain/discomfort in IBS and quiescent IBD may be due to the presence of a ‘leaky’ epithelial barrier that may result in aberrant exposure to luminal antigens.
Importantly, our findings help to explain the differences in epithelial disruption between IBS and IBD patients. However, soluble mediators from MCs and an increase in the proteolytic activity have been shown to play a critical role in epithelial barrier disruption in IBS patients.36 ,37 The pro-inflammatory TNF-α cytokine is likely to be an important contributor, as suggested by our results showing a significant association between TNF-α expression and colonic permeability in IBD patients. Our data are in accordance with several in vitro and in vivo experiments that support the role of TNF-α in barrier function.38 ,39 Our data, together with the demonstration that TNF-α neutralising antibodies can restore barrier function in CD patients,40 suggest that there is putative clinical benefit in optimising treatments with anti-inflammatory agents for quiescent IBD patients with IBS-like symptoms.
Although the magnitude and the nature of the alterations in TJs were similar in both IBS and quiescent IBD with IBS-like symptoms, these alterations may affect the outcome of the disease in different ways. Indeed, CD patients with increased intestinal permeability have been found to be at a higher risk of subsequent flare-up.33 In contrast, the increased intestinal permeability documented in IBS patients was associated with only mild activation of the immune system both locally (eg, intestinal mucosa) and systemically (eg, plasma, serum and peripheral blood mononuclear cells) and not with IBS flare-up. Obviously, the relatively weak correlation between the severity of IBS and paracellular permeability indicate that other routes for allergen signalling (eg, transcellular permeability) may contribute to IBS symptoms. Finally, whether the increased intestinal permeability observed in IBD patients is the cause of the GI disorder or a consequence of an existing inflammatory response remains to be clarified.
An important limitation in our study needs to be discussed so as to clarify the interpretation of our findings. In particular, it is important to acknowledge that the substantial overlap of peripheral abnormalities observed in patients compared with controls strongly suggests that mucosal inflammation and epithelial barrier disruption are not the only explanation for the occurrence of IBS-like symptoms in patients with quiescent IBD. Indeed, a number of central mechanisms, which we did not measure in the current study, including psychosocial stress, somatisation and coping, are well known factors that may contribute to symptom perception in both IBS and IBD.41 ,42 Finally, the possibility of studying IBD patients with active inflammation above the terminal ileum cannot be completely excluded despite a thorough evaluation of their small intestine either by video capsule or by enteroMRI when optimisation of treatment was required before remission.
In conclusion, our findings have shown that the presence of IBS-like symptoms in patients with quiescent IBD is related to the persistence of subclinical and ongoing inflammation associated with increased colonic paracellular permeability, which is similar to that seen in IBS patients. A persistent increase in expression of TNF-α in colonic mucosa may contribute to the epithelial barrier defects associated with abdominal pain in quiescent IBD, but not in IBS. These findings have important clinical implications for optimisation of anti-inflammatory therapy in IBD patients considered to be in remission.
References
Footnotes
-
Contributors MV-N contributed to the study design, data analysis and interpretation and revision of the manuscript. GF-M contributed to technical support and acquisition of data. HB performed experiments. RD was involved in acquisition of data and recruitment of patients. GB was involved in manuscript revision. RA participated in acquisition of data and technical support. MC S-P performed histological assessments. MKT was involved in Ussing experiments, writing and revision of the manuscript. VV was involved in manuscript preparation. XH and JF contributed to acquisition of material support and patient recruitment. TP was involved in the study design, support, manuscript preparation and study supervision.
-
Funding Programme Hospitalier de Recherche Clinique (PHRC 2007–2008) of the CHU of Nice.
-
Competing interests None.
-
Patient consent Obtained.
-
Ethics approval The study was approved by the ‘Comité de Protection des Personnes’ dated 6 November 2007 (N° 07.047).
-
Provenance and peer review Not commissioned; externally peer reviewed.