Vasoactive Intestinal Polypeptide and Mast Cells Regulate Increased Passage of Colonic Bacteria in Patients With Irritable Bowel Syndrome

Abstract

Background & Aims

Irritable bowel syndrome (IBS) is associated with intestinal dysbiosis and symptoms of IBS develop following gastroenteritis. We aimed to study passage of live bacteria through the colonic epithelium, and determine the role of mast cells and vasoactive intestinal polypeptide (VIP) in barrier regulation in IBS and healthy individuals.

Methods

Colon biopsies from 32 women with IBS and 15 age-matched healthy women (controls) were mounted in Ussing chambers; we measured numbers of fluorescently labeled Escherichia coli HS and Salmonella typhimurium that passed through from the mucosal side to the serosal side of the tissue. Some biopsies were exposed to agents that block the VIP receptors (VPAC1 and VPAC2) or mast cells. Levels of VIP and tryptase were measured in plasma and biopsy lysates. Number of mast cells and mast cells that express VIP or VIP receptors were quantified by immunofluorescence. Biopsies from an additional 5 patients with IBS and 4 controls were mounted in chambers and Salmonella were added; we studied passage routes through the epithelium by transmission electron microscopy and expression of tight junctions by confocal microscopy

Results

In colon biopsies from patients with IBS, larger numbers of E coli HS and Salmonella passed through the epithelium than in biopsies from controls (P<.0005). In transmission electron microscopy analyses, bacteria were found to cross the epithelium via only the transcellular route. Bacterial passage was reduced in biopsies from patients with IBS and controls after addition of antibodies against VPACs or ketotifen, which inhibits mast cells. Plasma samples from patients with IBS had higher levels of VIP than plasma samples from controls. Biopsies from patients with IBS had higher levels of tryptase, larger numbers of mast cells, and a higher percentage of mast cells that express VPAC1 than biopsies from controls. In biopsies from patients with IBS, addition of Salmonella significantly reduced levels of occludin; subsequent addition of ketotifen significantly reversed this effect.

Conclusions

We found that colonic epithelium tissues from patients with IBS have increased translocation of commensal and pathogenic live bacteria, compared with controls. Mechanisms of increased translocation include mast cells and VIP.

Introduction

Irritable bowel syndrome (IBS) is characterized by chronically recurring abdominal pain and disturbed bowel habits and affects 10–15% of the general population in industrialized countries, with a 2:1 female predominance¹. Even though the pathophysiology is incompletely understood, increasing evidence supports the concept of altered brain-gut microbiome interactions in IBS etiology^{2, 3}.

For many years it has been known that IBS-like symptoms develop following an acute gastroenteritis in about 10% of affected individuals^{4, 5}. Specifically, it has been confirmed in several studies that enteric infection with Salmonella, such as the foodborne pathogenic gram-negative Salmonella typhimurium, is a significant risk factor for the development of IBS^{6, 7}. Evidence also suggests that IBS can be associated with an altered gut microbiota composition or dysbiosis⁸, including the presence of Escherichia (E.) coli^{9, 10}. Some of the observed changes in species richness or community composition may reflect alterations in gastrointestinal transit¹¹. There is increasing evidence demonstrating enhanced intestinal permeability and altered tight junction patterns^12–14 both in colon and small bowel of IBS^15–19. In intestinal barrier function studies, IBS patients typically have been divided into subgroups based on predominant bowel habit (IBS-diarrhea (IBS-D), IBS-constipation (IBS-C), IBS-mixed (IBS-M))²⁰, but the association between bowel habit subtype and mucosal barrier function remains unclear^{15, 21}. There is some evidence that visceral hypersensitivity, a common feature in IBS, may be linked to disturbances in barrier function²², and such changes have been observed in an animal model of gastrointestinal infection²³.

The gut barrier function is controlled by several physiological mechanisms. Vasoactive intestinal polypeptide (VIP) is a neuropeptide found in both immune cells, including lymphocytes²⁴ and mast cells (MCs)²⁵, and in enteric neurons of the gastrointestinal tract. These neurons innervate the intestinal epithelium and regulate ion and fluid secretion, as well as epithelial homeostasis^{26, 27}. A few studies have also implicated VIP in the regulation of intestinal permeability^28–30. Our group previously showed that MCs and VIP regulate the ileal barrier of healthy humans and during stress in rats via VIP-receptors (VPAC1/VPAC2) on MCs²⁵. Increased plasma levels of VIP have been reported in IBS patients³¹ and in an IBS-animal model³², and there appears to be a positive correlation between an increased intestinal transcellular permeability and amount of mucosal MCs in IBS-D³³. In addition, increased MC numbers and tryptase levels have been found in biopsies of caecum³⁴ and jejunum³⁵ of IBS patients, and increased paracellular permeability and higher MC numbers were demonstrated in colon of IBS-D patients¹⁸. To our knowledge there are no reports exploring the intestinal epithelial response to live bacteria and its regulation by VIP and MCs in the colonic mucosa of IBS subjects.

The aims of the study were to investigate 1) the passage of live commensal and pathogenic microbes, through the colonic mucosa of women with IBS 2) the role of MCs and VIP in the regulation of bacterial passage, and 3) the distribution of MCs, VIP/VPACs in the colon of IBS subjects. We present evidence of increased bacterial passage through the colonic mucosa in IBS, with MCs and VIP involved as modulating factors. Our findings indicate a potential link between altered intestinal barrier function, and the increased transepithelial passage of bacteria in IBS, and suggest an involvement of VIP and MCs in this mechanism.

Material and Methods (See also Supplementary Methods)

Subjects and endoscopy

Subjects

Thirty-seven women with IBS, mean age 32.2 years (range 19–55), meeting Rome III criteria, were recruited from the Gastroenterology Department, University Hospital, Linköping. IBS patients had a moderate-severe IBS with mean symptom severity score of 347 (range 167–480)³⁶. Twenty healthy age-matched women mean age 29.9 years (range 20–48) without medical history of gastrointestinal symptoms or complaints, were recruited as healthy controls (HCs). IBS subjects were classified according to predominant bowel habit into IBS-M (n=21), IBS-C (n=8) and IBS-D (n=8). Exclusion criteria for both groups included self-reported allergy, organic gastrointestinal disease, metabolic or neurological disorders and nicotine or NSAID intake. The committee of human ethics, Linköping, approved the study and all subjects gave their written informed consent.

IBS subgroups

IBS patients were subgrouped based on the predominant stool consistency according to the Rome III questionnaire²⁰.

Sigmoidoscopy

A flexible sigmoidoscopy was performed after 8 hours of fasting without sedation and with scope insertion approximately 30–40 cm orally from linea dentata. Colonic biopsies were taken with a biopsy forceps without a central lance, and directly put in ice-cold oxygenated Krebs buffer³⁷. Prior to sigmoidoscopy, venous blood samples were collected for measurements of VIP concentrations.

Ussing chamber experiments

Barrier function studies

Colonic biopsies from 32 IBS patients and 15 HCs were mounted in Ussing chambers³⁸ as previously³⁷. After 20 min biopsies were treated with 1 μM anti-VPAC, 1 μM of the MC-blocker ketotifen or vehicle (Krebs). After another 20 min, 34 μCi/ml of the paracellular probe⁵¹chromium (Cr)-EDTA was added to the mucosal side of each chamber. Live green fluorescent protein (GFP)-labelled E. coli HS or Salmonella typhimurium, prepared as previously³⁷, was added to the mucosal sides of separate chambers at a final concentration of 10⁸ CFU/ml. Serosal samples were collected and bacterial and ⁵¹Cr-EDTA passage was measured. The transepithelial potential difference (PD), short-circuit current (I_sc) and the transepithelial resistance (TER) across the tissues were monitored throughout the experiments to ensure tissue viability.

Electron microscopy

To identify bacterial translocation and MC ultrastructure, colonic biopsies from 5 IBS patients and 4 HCs were mounted in Ussing chambers and exposed to bacteria. Anti-VPACs or ketotifen was added as described above followed by addition of GFP-labelled Salmonella. One biopsy was added Krebs, as control to evaluate the eventual effects on tissue structure during incubation. After 30 min, biopsies were fixed and processed following standard protocols for scanning and transmission electron microscopy (SEM, TEM). MC-degranulation was also quantified in basal biopsies, not mounted in Ussing chambers, from 20 IBS patients and 8 HCs.

Quantification of tight junctions

To study the effects of Salmonella, anti-VPACs and ketotifen on tight junctions, biopsies from 4 IBS patients and 4 HCs were mounted in Ussing chambers and added anti-VPACs, ketotifen and Salmonella. After 30 min, biopsies were fixed in the chambers and prepared for immunofluorescence of tight junctions (see below).

Giemsa staining of biopsies

To identify the presence of intestinal bacteria at baseline, one paraffin-embedded biopsy/individual, was stained with routine Giemsa according to a previous protocol³⁹. Bacteria in 10 areas of 0.033 mm²/biopsy were counted and results are presented as bacteria/mm².

Measurements in biopsies and plasma

Preparation of biopsy lysates

Lysates from frozen biopsies were prepared to a final concentration of 1 mg protein/ml, according to the procedure described in Supplementary Methods.

Quantification of VIP in plasma and biopsies by EIA

To examine VIP levels in plasma and biopsies from IBS patients and HCs, a VIP-EIA kit was used following manufacturer’s instructions (Phoenix Pharmaceuticals, Germany).

Quantification of MC-tryptase in biopsies by Western Blot

Biopsy lysates with protein concentrations of 20μg/μl were loaded on gels and further processed as described in Supplementary Methods. Values were normalized according to the β-actin loading control and values are given as fluorescence units.

Microscopy

Quantification of MC-tryptase, VIP and VPACs

Biopsies from 32 IBS patients and 15 HCs were stained following the same protocol previously described for tight junctions. Primary antibodies used were MC tryptase+ VPAC1/VPAC2 or VIP. The total number of MCs, and co-localization with either VPACs or VIP were quantified manually and results are presented as the percentages of VIP/VPAC1/VPAC2 positive MCs from patients with IBS and HCs respectively.

Evaluation of mucosal MC ultrastructure

MCs were identified based on morphological characteristics in a minimum of 20 ultrathin sections from each sample. Ultrastructural changes indicative of activation were described and a quantitative method was used to assess MC-activation, based on the degranulation area. For each MC, the cytoplasm and degranulated area (loss of granular content) were measured using the ImageJ Fiji software (NIH, Bethseda, MD). Degranulation was expressed as the percentage of degranulated area respect to the total area of the cytoplasm.

Evaluation of bacterial passage routes

After incubation, the presence of bacteria on the epithelial surface was observed by SEM and TEM. Interaction of Salmonella with epithelial cells and bacteria crossing through the colonocytes were identified by TEM.

Immunofluorescence of tight junctions

Biopsies were stained for occludin and ZO-1, evaluated by confocal microscopy and images were further processed in ImageJ Fiji software. Results were normalized and presented as % of the expression in vehicle biopsies set to 100%.

Statistic analysis

Parametric data are given as mean±SEM and comparisons between groups were done with Student t test and ANOVA test. Non-parametric data are given as median (25–75th interquartile range) and comparisons between groups were done with Kruskal-Wallis and Mann-Whitney U tests. Wilcoxon matched-pairs signed rank test was used for paired comparisons. Differences with p<0.05 were considered significant. Correlation testing was done with Spearman Correlation test.

Results

Mucosal barrier function in IBS and HCs

Electrophysiology

Ussing chamber experiments showed stable PD after equilibration in all biopsies. The active net ion transport, assessed as I_sc, was similar in IBS and HCs throughout the experiments. There were no effects on I_sc of E. coli HS, Salmonella typhimurium, anti-VPACs or ketotifen (data not shown).

In vehicle chambers there was a lower TER in IBS compared to HCs when experiments were started (22 ohm/cm² (19–27) vs. 25 (21–32), p<0.05), at 30 min (21 (18–25) vs. 24 (22–29), p<0.05), and at 60 min (20 (17–24) vs. 24 (20–26), p<0.05). At 90 min from start, TER remained lower in IBS (20 (17–25)) compared to HCs (24 (19–26)), however, this difference no longer reached statistical significance (p=0.07). In IBS biopsies added Salmonella there was a significantly decreased TER at 30, 60 and 90 min compared to vehicle, p<0.05 (Table 1). In chambers pre-incubated with ketotifen, the Salmonella-induced decrease in TER was inhibited, p<0.05, while anti-VPACs had no significant effect. In HC, Salmonella had no significant effect on TER, neither at 30, 60 or 90 min (data not shown). Biopsies mounted in chambers with added E. coli HS showed the same pattern for TER as vehicle chambers in both IBS and HC (data not shown).

Table 1.

Effects of live Salmonella typhimurium on transepithelial resistance over time.

Time after start	Vehicle	Salmonella (% decrease from start)	Salmonella+Anti-VPACs	Salmonella+ketotifen
30 min	3.4 [1.0–6.6]	12.4 [1.3–25.7]*	8.1 [1.5–14.2]	4.4 [0.0–11.8]#
60 min	5.1 [3.1–11.1]	19.6 [4.8–33.6]*	10.8 [0.0–17.8]	7.8 [0.0–16.4]#
90 min	6.7 [3.5–10.4]	15.1 [3.1–30.0]*	10.0 [0.3–21.6]	8.8 [2.8–19.9]

Colonic biopsies from 32 patients with irritable bowel syndrome (IBS) were mounted in Ussing chambers and added buffer (vehicle), Salmonella, Salmonella+vasoactive intestinal polypeptide receptor blocker (anti-VPACs) or Salmonella+ketotifen. Transepithelial resistance (TER) was monitored at 30, 60 and 90 min from start and values are given as % decrease in TER from start.

^*= p<0.05 vs. vehicle at 30, 60, 90 min, respectively,

^#=p<0.05 vs. Salmonella at 30 and 60 min, respectively. Comparisons were done with Mann–Whitney U test and values given as (median (25^th–75^th percentile)).

Increased bacterial and paracellular passage in IBS

The transepithelial passage in biopsies of IBS subjects was nearly 2 times higher for E. coli, and 2.8 times higher for Salmonella compared with HCs, p<0.0005, and fluorescence microscopy confirmed uptake of both E. coli and Salmonella in Ussing chambers (Fig. 1 A–B). The passage of the two bacteria strains was not significantly correlated.

Fig. 1.

Bacterial passage through colonic biopsies from 32 patients irritable bowel syndrome (IBS) and 15 healthy controls (HC) mounted in Ussing chambers for 120 min. (A) Passage of live E. coli HS and Salmonella typhimurium. Comparisons were done with Mann-Whitney U test, ^***p<0.0005 vs. HC. (B). Representative scanning electron micrographs of E. coli and Salmonella on the surface of the colonic epithelium and representative confocal micrographs illustrating bacterial uptake (white arrows) into the mucosa of IBS colon, after incubation. (C) Transmission electron micrographs illustrating interaction of Salmonella (black arrows) and microvilli (white arrows); a transversal section where interaction is shown; and bacterial uptake through endocytosis (note double membrane) at the apical side of the epithelium. (D) Intestinal bacteria quantified by Giemsa staining. Arrows indicate two bacteria just entering the epithelium in an IBS sample.

Permeability to ⁵¹Cr-EDTA was higher in IBS compared to HCs (1.10 cm/s x10⁻⁶ (0.74–1.50) vs. 0.80 (0.53–1.10), p<0.05). Addition of E. coli or Salmonella had no significant effect on ⁵¹Cr-EDTA permeability. No significant correlations were identified between ⁵¹Cr-EDTA, Salmonella, and E coli passage.

Bacterial passage route

E coli and Salmonella were detected on the epithelial surface and in the lamina propia after incubation (Fig. 1B). Salmonella was observed to interact with the microvilli of colonocytes and translocation was identified by endocytosis through the transcellular pathway (Fig. 1C). No bacteria were observed to interact at the tight junctions in any sample. No differences were identified between groups and treatments, in terms of bacteria interaction with the epithelial surface and internalization to the colonocytes.

Intestinal bacteria identified by Giemsa

Giemsa staining of biopsies that had not been experimentally exposed to bacteria revealed that bacteria within the epithelium were only scarcely seen in both IBS and HCs (Fig. 1D).

Regulation of barrier function by VIP and MCs

Bacterial passage was significantly diminished in both IBS subjects and HCs after blocking with anti-VPACs and ketotifen (Fig. 2A–B).

Fig. 2.

Influence of vasoactive intestinal polypeptide (VIP) and mast cells (MCs) on bacterial passage through colonic biopsies from 32 patients with irritable bowel syndrome (IBS) and 15 healthy controls (HC) mounted in Ussing chambers. Figure indicates passage at 120 min of E. coli HS (A) and Salmonella typhimurium (B) before (vehicle) and after addition of VIP-receptor blocker (anti-VPACs) or MC-stabilizer (ketotifen) to the chambers. Comparisons were done with Mann-Whitney U test, ^**p<0.005, ^***p<0.0005 vs. vehicle.

Tight junction proteins

Confocal microscopy showed close to significantly decreased baseline levels of occludin in IBS (346.7 fluorescence units (308.2–371.9)) compared to HCs (420.8 (368.8–609.7)), p=0.057, but no difference in ZO-1 expressions (IBS: 530.4 (345.5–599.9), HCs: 398.7 (280.9–463.4)). The addition of Salmonella to the biopsies decreased occludin expression in the apical side of the colonocytes compared to vehicle (set to 100%) in both IBS (42.9 [21.8–80.1]) and HCs (47.3 [10.7–105.8]), but it did not reach statistical significance. Notably, in IBS biopsies pre-incubated with ketotifen, the Salmonella-induced decrease in occludin expression was significantly inhibited (130.5 [100.2–154.6], p<0.005), while anti-VPACs had no significant effect (61.6 [29.0–88.4]). There was no effect on expression or distribution of ZO-1 by Salmonella, ketotifen, or anti-VPACs. For more detailed results see Supplementary Table 1. Representative photographs of occludin staining during the different conditions are shown in Fig. 3 and for ZO-1 in Supplementary Fig. 1.

Fig 3.

The effect of live Salmonella typhimurium on occludin. Photographs illustrate staining for occludin in colonic biopsies from a patient with irritable bowel syndrome mounted in Ussing chambers. Biopsies were added buffer only (vehicle), Salmonella, Salmonella+vasoactive intestinal polypeptide (VIP)-receptor blocker (anti-VPACs) or Salmonella+ketotifen. After 30 min, biopsies were removed and stained for occludin by immunofluorescence.

Increased VIP in plasma of IBS patients

VIP in plasma of IBS patients was higher than in HCs (2.1±0.09 ng/ml vs. 1.5±0.09, p<0.0005). However, there were no significant differences in biopsy lysates (0.57±0.10 ng/mg protein vs. 0.32±0.05). There was no significant correlation between VIP levels in plasma and in biopsies. No significant correlation was observed between E. coli/Salmonella and VIP levels and between ⁵¹Cr-EDTA and VIP levels.

Increased number of MCs, MCs expressing VPAC1, and MC-activation in IBS

Colonic IBS biopsies showed increased number of MCs compared to HCs (104.0±8.0 MCs/mm² vs. 73.7±11.6, p<0.05). Furthermore, the percentage of MCs expressing VPAC1 was higher in IBS compared to HCs (46±4.2 % vs. 31±6.2, p<0.05). However, there was no difference in the percentage of MCs expressing VPAC2 (30±4.4 % vs. 26±6.1). When comparing microscopy results patient by patient it was revealed that only 75% of the IBS patients had elevated MC numbers compared to the mean of HCs, and 78% showed higher percentage of MCs expressing VPAC1. Representative photographs of MCs expressing VPAC1 in an IBS patient and HC are shown in Fig. 4A–B. MCs expressing VPAC2 in IBS colon is illustrated in Fig 4C. There was a trend to higher amounts of MCs containing VIP in IBS compared to HCs, however it did not reach statistical significance (24±1.0 % vs. 20±1.3, p=0.06). A VIP-containing MC of an IBS patient is shown in Fig. 5. In all subjects, MC density correlated with the percentage of MCs expressing VPAC1 (R=0.35, p=0.01). In addition, the percentage of MCs expressing VPAC1 strongly correlated with the percentage of MCs expressing VPAC2 (R=0.67 p<0.0001).

Fig 4.

Staining of mast cell (MC)-tryptase (red) and vasoactive intestinal polypeptide receptors (VPACs) (green) in colon of a patient with irritable bowel syndrome (IBS) and healthy control (HC). Photographs show MCs±VPAC1 expressions in an IBS patient (A) and HC (B). Arrows indicate individual MC or VPAC1 staining and arrowheads indicate co-localization i.e. MCs expressing VPAC1 (C) Photograph showing MCs expressing VPAC2 (arrowheads) in IBS colon. Arrows indicate individual MC or VPAC2 staining. Overview photographs; 400Xmagnification, zoomed photographs; 1000X.

Fig 5.

Staining of mast cell (MC)-tryptase (red) and vasoactive intestinal polypeptide (VIP) (green) in colonic biopsy of a patient with irritable bowel syndrome (IBS). Photograph shows a MC expressing VIP (arrowhead). Window in upper right corner shows a MC in close proximity to secreted VIP, which was a common feature seen in IBS patients, but could also be observed in healthy controls. Overview photographs; 400Xmagnification, zoomed photographs; 1000X.

Analysis of biopsy lysates by Western blot revealed higher levels of MC-tryptase in IBS compared to HCs (2.26±0.46 fluorescence units vs. 1.42±0.46, p<0.05) (Fig. 6A).

Fig 6.

Mucosal mast cell (MC) activation: tryptase, ultrastructure and degranulation in irritable bowel syndrome (IBS) and healthy control (HC) subjects. (A) Expression of MC-tryptase (mw 31–36 kDa) and β-actin as loading control. Lysates from colonic biopsies of 32 women with IBS and 15 HCs were analyzed by Western Blot. Panels show a representative blot from 5 IBS subjects and 5 HCs. (B) Degranulation of MCs at baseline of IBS (n=19) and HC (n=8) (C) Degranulation of MC in biopsies from IBS patients (n=5) and HCs (n=4) after incubation with buffer only (vehicle), Salmonella, Salmonella+vasoactive intestinal polypeptide (VIP) receptor blocker (anti-VPACs), or Salmonella+ketotifen. (D) Micrographs of MCs from IBS and HC from the different experimental groups. Note signs of MC-activation such as irregular plasma membrane with emission of pseudopodes (black arrow) and piecemeal degranulation (black arrow head). Fusions of granules are predominantly detected in MCs from the IBS-group (white arrow).

In TEM, mucosal MCs showed signs of activation in samples from all subjects, with variable granular electron density and piecemeal degranulation. However, in MCs from IBS patients, distribution of the cytoplasmic granules to the periphery, close to the plasma membrane, was identified, together with higher frequency of membrane projections to the extracellular matrix, indicative of active release of granular content. Anaphylactic degranulation was not observed in any sample, although some MCs in the IBS-group also displayed fusion of several granules. MCs from IBS showed higher degree of degranulation, however this difference did not reach statistical significance, p=0.113 (Fig. 6B). Exposure to vehicle did not affect MC ultrastructure, as they were similar to MCs in non-incubated tissues. Salmonella increased MC degranulation, and pre-incubation with anti-VPACs or ketotifen maintained a more stable granular structure in IBS and HCs. However, degranulation was not statistically significant between experimental conditions (Fig. 6C–D).

There was no correlation between MC-tryptase in biopsy lysates and bacterial passage (E. coli and Salmonella), however in all subjects, MC density correlated with MC-tryptase expression in biopsy lysates (R=0.34, p=0.03).

Relation between gut findings and clinical data

There was no correlation between age of subjects with IBS/HC and any of the parameters investigated (results not shown).

There was no significant correlation between symptoms (severity scoring system, IBS-SSS) and bacterial passage, MC numbers, or MC-tryptase. There was a negative correlation between ⁵¹Cr-EDTA (R=−0.41, p<0.05) and MC degranulation (R=−0.59, p<0.01) with IBS-SSS. Significant differences were found between the subgroups IBS-M and IBS-C in terms of MCs. MC density was higher in IBS-M compared with IBS-C (127±48 MCs/mm² vs. 75±23, p<0.005) and the proportion of MCs expressing VPAC1 and VPAC2 was significantly higher in IBS-M compared with IBS-C (VPAC1: 55.7±4.3 % vs. 21.5±6.3, p<0.005; VPAC2: 38.4±4.9 % vs. 8.4±3.2, p<0.0005). For IBS-D the numbers were 46.3±10.3% for VPAC1 and 32.1±12.9 % for VPAC2.

DISCUSSION

In the present study we demonstrated that female IBS subjects show a greater colonic epithelial permeability to live bacteria and that this barrier dysfunction may be associated with an upregulation of MCs- and VIP-related signalling mechanisms. IBS patients had an increased passage of both live E. coli HS, a commensal bacteria that normally colonize the colon,⁴⁰ and of the pathogenic bacteria Salmonella typhimurium. In IBS and HCs, bacterial passage of both strains was diminished by blocking the VIP-receptors as well as after stabilizing MCs with ketotifen. Ketotifen is a histamin H1-receptor antagonist and it has been reported that thresholds for discomfort were increased in IBS when taking this drug compared with placebo⁴¹. We also found that the colonic mucosa of IBS patients had an enhanced number of MCs and that MCs of IBS patients expressed more VPAC1 than HCs. Further, biopsy lysates from IBS patients demonstrated elevated tryptase levels and the MCs displayed a more activated morphological/ultrastructural phenotype when compared with HCs. Even though disturbances of the epithelial barrier function have previously been described in IBS^{19, 42}, to our knowledge, this is the first study demonstrating an increased transepithelial passage of live bacteria in human subjects with IBS, and identifies VIP as a key regulatory molecule of this mechanism.

The role of immune activation and its involvement in IBS clinical symptoms remains unclear. Evidence for mucosal immune activation in IBS has been reported in some, but not all studies^{3, 43}. Increased MC numbers, T-cell activation and cytokine polymorphism have also been reported in IBS colon by numerous groups^44–46. For example, Cremon et al⁴⁷ found increased MC and T cell numbers in the colon of 50% of IBS subjects. In the present study we found increased MC numbers in 75% of IBS subjects, however, we could not find any correlation between MC-infiltration and symptoms.

There is growing evidence for altered microbiota composition in some IBS patients⁴⁸. The gut microbiota is involved in several aspects of gastrointestinal function, including motility, sensitivity and immune function². Recent data also suggest that the human gastrointestinal microbiota is able to communicate with the central nervous system⁴⁹. E. coli K12 and HB101 have been used as commensal-like bacteria to study luminal uptake/passage into the intestinal mucosa^{37, 39}. In the present study we chose the human commensal E. coli HS as a non-pathogenic E. coli, which has similar properties as K12 and HB101⁵⁰. Live E. coli HS passed through the colonic mucosa of IBS patients to a significantly higher extent compared with HCs. These results suggest that commensal bacteria and its metabolites can potentially affect and interact with intra- and submucosal targets in IBS to a greater degree than in HCs.

There are multiple pathways how bacteria and bacterial products can interact with immune-neuroendocrine or neural constituents^{5, 8, 49, 51}. However, it is important to remember that gut permeability is determined not only by the transepithelial component, but also by the inner compartment of the mucus layer⁵². Ex vivo, this mucus layer is easily removed from the epithelial cells due to flow in the Ussing chambers⁵³, which is important to take under consideration when interpreting the results and translating it to the clinical situation. As there was very little overlap between IBS and HCs in terms of transepithelial bacterial passage, and evidence for immune activation has not consistently been found in colon of IBS subjects, despite the fact that all IBS patients had moderate to severe symptoms, it is conceivable that in the intact colon, live bacteria such as E. coli gain access to the underlying submucosa only in those patients who also have a compromised mucus layer. This could also explain why we found IBS symptom severity to be inversely correlated with paracellular permeability, and why there were no significant correlation between symptom severity and bacterial passage and MCs, limited though by a small number of subjects when divided into IBS-subgroups. These findings strongly suggest that alterations in transepithelial permeability alone cannot explain the formation of symptoms in IBS.

Monitoring of electrophysiology showed no difference in I_sc between IBS and HCs. Previous studies on I_sc and IBS are mostly lacking, however, recently Peters et al⁵⁴ showed a similar I_sc in IBS-C and HCs, which is in line with our findings and indicates an increased bacterial uptake in IBS, even in circumstances where ion transport is unchanged. To further determine changes in permeability due to tight junction components, we chose to study occludin and ZO-1 expression, as main constituents of the apical junctions, based on previous reports demonstrating its deregulation in the colon of IBS¹³ Our study confirms previous data and extends its contribution to bacteria-mediated epithelial alterations. Although the role of occludin still remains unknown, our results suggest its regulation by mechanisms that involve MC-activation. Since bacteria was not observed entering the epithelium through the paracellular space, we can speculate that occludin deregulation is secondary to epithelial invasion and, probably, immune activation, by e.g. TNF-α which has been related with the internalization of occludin, inducing secretion of water, barrier dysfunction and diarrhea⁵⁵

The well-described onset of IBS symptoms after gastroenteritis in a number of affected individuals is another strong indicator of the importance of microbial factors in IBS pathophysiology^{48, 56, 57}. It has also been demonstrated that having Salmonella infection during childhood was a risk factor for the development of long-standing IBS symptoms in adulthood⁵⁸. Salmonella typhimurium is a pathogenic gram-negative bacteria that is well known to cause gastroenteritis in humans and other mammals⁵⁹. The toxicity is based on an outer membrane with lipopolysaccharides protecting the bacteria from the environment. In the present study we found an almost 3 times higher passage of live Salmonella typhimurium through IBS mucosa compared to HCs. This indicates that there is a disturbance in IBS patients leading to increased invasion of pathogenic bacteria into the epithelium and the underlying submucosa. One could speculate that the enhanced translocation of Salmonella could lead to increased antigen load and for example activation of MCs and other immune cells, which might facilitate local immune activation. Interestingly, the passage of E. coli and Salmonella did not correlate, indicating different uptake mechanisms and diverse translocation patterns.

Only a small decrease in baseline TER was found in IBS compared to HCs, which may be due to the fact that conventional TER measurements are not sensitive enough to fully detect differences between biopsies. Another noteworthy point is that neither Salmonella nor E. coli significantly affected ⁵¹Cr-EDTA permeability, even though Salmonella had a significant effect on TER. Several pathogenic E. coli strains have shown to affect the paracellular pathway in intestinal epithelia⁶⁰; however, others and us observed no effects on paracellular permeability by non-pathogenic E. coli strains^39,61. For Salmonella one could have expected an effect on ⁵¹Cr-EDTA permeability, since previous studies indicate that Salmonella decreases TER and cause tight junction redistribution in vitro⁶². Though, in the present study there was only a trend to increased ⁵¹Cr-EDTA permeability by Salmonella, despite a significant effect on TER, which probably refers to different regulations of permeability pathways i.e. the leak and the pore pathways⁶³.

Functional experiments in the present study demonstrated that bacteria passage was significantly diminished after blocking the VPACs and after ketotifen administration, an effect that was not only seen in IBS patients, but also in HCs, suggesting the involvement of VIP and MCs in the physiological regulation of intestinal barrier function. However, the fact that both VIP levels, and MCs expressing VPACs, are increased in IBS relative to health, further points at a novel role of this mediator in bacterial translocation in IBS. There are only a few reports examining the influence of VIP in IBS. Zhang et al³¹ showed increased VIP levels in plasma of IBS patients compared to controls, and in an IBS-stress model, Wu et al³² showed colonic MC-activation and increased VIP levels in stressed animals. Interestingly, electro-acupuncture inhibited these events, indicating that the interactions of MCs and VIP could be key mechanisms in IBS pathophysiology.

In addition we found increased VIP levels in plasma of IBS patients compared to controls. In biopsy lysates there was only a trend to increased VIP levels, however, IBS patients showed almost 2 times higher VIP-concentrations compared to controls. Since IBS patients also had elevated VIP-receptor density on their MCs in combination with an increased MC density we presume that the physiological VIP-MC-dependent regulation of bacterial passage might be upregulated in IBS.

MCs are innate immune cells involved in protection against pathogens and in allergic responses. When activated, they undergo a degranulation process, which leads to the release of mediators, among which histamine and tryptase have proved to modulate visceral hypersensitivity and epithelial permeability. Our study identified increased MC-tryptase together with more pronounced ultrastructural signs of MC-activation from IBS, even though the quantification method used did not corroborate our observations, which might refer to a large variability within the IBS-group at baseline. Salmonella did induce MC degranulation from both IBS and HCs, and treatments used reduced MC-activation; however a larger cohort should be studied in order to better prove these effects. Most studies showing differences in MC-density. Moreover, in contrast to the majority of previous reports, subjects with known allergies or atopic diseases were excluded in the current study, reducing the risk of allergy and atopic disease as an important confounding factor when studying MCs in IBS⁶⁴. In fact, a combination of allergic factors, such as positive skin tests, elevated IgE or eosinophil levels, personal or family history of allergy, significantly correlated with MC count and enhanced intestinal permeability⁶⁵. Recently, a positive correlation between increased intestinal permeability and number of mucosal MCs was found in IBS-D patients³³. However, in the present study we found no significant correlation between MC count and mucosal permeability.

In conclusion, the present study demonstrates an increased transepithelial passage of live commensal and pathogenic bacteria in the colon of IBS subjects. Both MCs and VIP appear to be important factors in the regulation of this bacterial translocation through the mucosa, and theses mechanisms seem to be upregulated in IBS. Future studies are required to elucidate the relationship between these epithelial abnormalities, mucosal immune activation and symptom generation in IBS subjects.

Abbreviations

⁵¹Cr-EDTA

⁵¹chromium (Cr)-EDTA

E. coli

Escherichia coli

HCs

healthy controls

GFP

green fluorescent protein

IBS

irritable bowel syndrome

IBS-C

IBS with constipation

IBS-D

IBS with diarrhea

IBS-M

IBS with mixed stool consistency

IBS-SSS

IBS severity scoring system

I_sc

short-circuit current

MCs

mast cells

P_app

apparent permeability coefficient

PD

transepithelial potential difference

RT

room temperature

SEM

scanning electron microscopy

TER

transepithelial resistance

TEM

transmission electron microscopy

VIP

vasoactive intestinal polypeptide

VPAC1

VPAC2, VIP receptors types 1 and 2