Serotonergic reinforcement of intestinal barrier function is impaired in irritable bowel syndrome

Authors

  • D. Keszthelyi,

    Corresponding author
    1. Top Institute Food and Nutrition, Wageningen, The Netherlands
    2. Division of Gastroenterology-Hepatology, Department of Internal Medicine, Maastricht University Medical Centre+, Maastricht, The Netherlands
    • Correspondence to:

      Dr D. Keszthelyi, Division of Gastroenterology-Hepatology, Department of Internal Medicine, Maastricht University Medical Centre+, PO Box 5800, 6202 AZ Maastricht, The Netherlands.

      E-mail: daniel.keszthelyi@maastrichtuniversity.nl

    Search for more papers by this author
  • F. J. Troost,

    1. Top Institute Food and Nutrition, Wageningen, The Netherlands
    2. Division of Gastroenterology-Hepatology, Department of Internal Medicine, Maastricht University Medical Centre+, Maastricht, The Netherlands
    Search for more papers by this author
  • D. M. Jonkers,

    1. Top Institute Food and Nutrition, Wageningen, The Netherlands
    2. Division of Gastroenterology-Hepatology, Department of Internal Medicine, Maastricht University Medical Centre+, Maastricht, The Netherlands
    Search for more papers by this author
  • H. M. van Eijk,

    1. Department of Surgery, Maastricht University Medical Centre+, Maastricht, The Netherlands
    Search for more papers by this author
  • P. J. Lindsey,

    1. Department of Clinical Genetics and Bioinformatics, Maastricht University Medical Centre+, Maastricht, The Netherlands
    Search for more papers by this author
  • J. Dekker,

    1. Top Institute Food and Nutrition, Wageningen, The Netherlands
    2. Host-Microbe Interomics Group, Department of Animal Sciences, Wageningen University, Wageningen, The Netherlands
    Search for more papers by this author
  • W. A. Buurman,

    1. School for Mental Health and Neurosciences, Maastricht University Medical Center+, Maastricht, The Netherlands
    Search for more papers by this author
  • A. A. M. Masclee

    1. Top Institute Food and Nutrition, Wageningen, The Netherlands
    2. Division of Gastroenterology-Hepatology, Department of Internal Medicine, Maastricht University Medical Centre+, Maastricht, The Netherlands
    Search for more papers by this author

  • This article was accepted for publication after full peer-review.

Summary

Background

Alterations in serotonergic (5-HT) metabolism and/or intestinal integrity have been associated with irritable bowel syndrome (IBS).

Aims

To assess the effects of the precursor of 5-HT, 5-hydroxytryptophan (5-HTP), on mucosal 5-HT availability and intestinal integrity, and to assess potential differences between healthy controls and IBS patients.

Methods

Fifteen IBS patients and 15 healthy volunteers participated in this randomised double-blind placebo-controlled study. Intestinal integrity was assessed by dual-sugar test and by determining the mucosal expression of tight junction proteins after ingestion of an oral bolus of 100 mg 5-HTP or placebo. Mucosal serotonergic metabolism was assessed in duodenal biopsy samples.

Results

5-HTP administration significantly increased mucosal levels of 5-HIAA, the main metabolite of 5-HT, in both healthy controls (7.1 ± 1.7 vs. 2.5 ± 0.7 pmol/mg, 5-HTP vs. placebo, = 0.02) and IBS patients (20.0 ± 4.8 vs. 8.1 ± 1.3 pmol/mg, 5-HTP vs. placebo, P = 0.02), with the latter group showing a significantly larger increase. Lactulose/L-rhamnose ratios were significantly lower after administration of 5-HTP (P < 0.05) in healthy controls and were accompanied by redistribution of zonula occludens-1 (ZO-1), pointing to reinforcement of the barrier. In IBS, expression of the tight junction proteins was significantly lower compared to healthy controls, and 5-HTP resulted in a further decrease in occludin expression.

Conclusions

Oral 5-HTP induced alterations in mucosal 5-HT metabolism. In healthy controls, a reinforcement of the intestinal barrier was seen whereas such reaction was absent in IBS patients. This could indicate the presence of a serotonin-mediated mechanism aimed to reinforce intestinal barrier function, which seems to dysfunction in IBS patients.

Introduction

Irritable bowel syndrome is a common functional gastrointestinal disorder affecting up to 15% of the Western population. It is characterised by lower abdominal discomfort or pain with disturbed defecation in the absence of apparent structural or biochemical abnormalities that might explain the symptoms.[1] Despite being very common, the pathophysiology of IBS is poorly understood, which may in part explain the current lack of effective therapeutic approaches.

Accumulating evidence suggests that the intestinal barrier function is impaired in IBS and that this impairment may be involved in the development of the disorder.[2, 3] A disturbance in barrier function has been documented in 12–50% of IBS patients by using sugar permeability tests; it appears to be most pronounced in patients with post-infectious IBS and diarrhoea-predominant IBS.[2, 4-6] More recent evidence suggests that the intestinal barrier is affected in IBS regardless of the predominant bowel habit.[2] Several recent reports have pointed to abnormalities in tight junction proteins of both colonic[7] and jejunal samples[8, 9] of IBS patients. To date, the majority of these studies demonstrated decreased expression of these tight junction proteins, which could be the result of either intracellular degradation (proteasome-mediated),[10] or intraluminal breakdown (mediated by extracellular proteases).[11]

Apart from intestinal permeability, alterations in serotonergic metabolism have been associated with IBS.[12, 13] Serotonin (5-hydroxytryptamine, 5-HT) has an important role in regulating human gastrointestinal function, including secretion, intestinal sensing and signalling.[14] Some reports have documented alterations in mucosal levels of 5-HT and the expression and activity of the serotonin transporter (SERT) in IBS,[15] however, these findings were not consistent.[16, 17] More recent evidence suggests an important role for the duodenum with respect to abnormal mucosal 5-HT metabolism.[18] The efficacy of serotonergic drugs in the treatment of IBS also supports the hypothesis that disturbances in serotonergic metabolism contribute to disease development.

The question therefore arises whether changes in serotonergic metabolism precede the pathogenetic changes of IBS or that they are merely an epiphenomenon of intestinal barrier dysfunction. Here, we hypothesised that changes in serotonergic signalling can initiate a cascade of events that may lead to alterations in epithelial integrity. A number of in vitro studies have shown that 5-HT affects tight junctions in Caco-2 intestinal epithelial[19] and mammary epithelial cells.[20] However, studies in healthy humans and in IBS patients on the role of serotonin in intestinal permeability are lacking. The present study was therefore designed to investigate the relation between 5-HT metabolism and intestinal permeability in healthy controls and IBS patients. More specifically, we hypothesised that the oral administration of the direct precursor of 5-HT, 5-hydroxytryptophan (5-HTP) will increase serotonergic metabolism, which will affect tight junctions and intestinal permeability. We aimed to ascertain differential effects of the response to 5-HTP between healthy controls and patients with IBS. To this purpose, we used a dual-sugar permeability test and assessed duodenal mucosal expression and transcription of tight junction proteins.

Methods

The study was approved by the Medical Ethics Committee of the Maastricht University Medical Centre (MUMC) and was conducted according to the revised version of the Declaration of Helsinki (October 2008, Seoul). All volunteers gave their written informed consent prior to participation. The study has been registered at the US National Library of Medicine (http://www.clinicaltrials.gov, NCT00731003).

Participants

Fifteen healthy volunteers and 15 patients with IBS diagnosed according to the Rome III criteria were included in this study. Patients fulfilling the Rome III criteria for functional dyspepsia were excluded from the study. All participants completed the study. All subjects were screened involving a standardised general physical examination. Investigation to exclude organic disease in IBS patients was performed during a minimal diagnostic work-up at the out-patient clinic prior to inclusion, including blood test (i.e. thyroid-stimulating hormone levels and antibodies to anti-tissue transglutaminase), stool analysis, and sigmoidoscopy or colonoscopy, when considered necessary by the primary physician. Exclusion criteria included history of gastrointestinal disorders (other than IBS) or psychiatric disorders including use of psychoactive medication, use of medication affecting gastrointestinal or psychoactive function and/or serotonergic metabolism within 14 days prior to testing, administration of investigational drugs in the 180 days prior to the study, major abdominal surgery interfering with gastrointestinal function, dieting, pregnancy, lactation, smoking or excessive alcohol (>2 units/day).

Sample size was determined using OpenEpi Sample Size Calculator (Emory University, Atlanta, GA). Considering the primary outcome of the present cross-over study, which is the plasma lactulose/L-rhamnose permeability test, the required number of included subjects was based on an estimated effect size of 40% (in the proportion of) alteration in small bowel permeability between the two interventions (5-HTP and placebo), a power of 80%, and a significance level of 5%, 15 study completers were required for each group (healthy controls and IBS patients). The effect size of 40% was based on the study of by Zeng et al.[21] In that study, the effect of a probiotic intervention was investigated on intestinal permeability in patients with IBS. We assumed that a comparable effect size could be observed using our intervention.

Study design

The study was designed in a randomised double-blind placebo-controlled cross-over fashion with a minimum wash-out period of 7 days between test days. Each test day, participants received either 5-HTP or placebo based on a random pre-selection. Randomisation was performed by the Pharmacy Department of our hospital.

All subjects were tested within 3 months to avoid possible seasonal variation. All women were tested in the follicular phase of the menstrual cycle or taking anti-conceptives. Participants were requested to abstain from heavy physical exercise and consumption of alcohol or caffeine-containing products and of tryptophan-rich food the day prior to their visit.

Study medication containing 100 mg 5-HTP was obtained from Tiofarma BV, Oud Beijerland, the Netherlands, and prepared by the Department of Pharmacy of our hospital. Identical tablets without 5-HTP were used as placebo medication. Previous studies indicated that 100 mg of 5-HTP was well tolerated without serious side effects and has an elimination half-life of 1.5 h.[22] Potential side effects were registered in the first three hours following intake of 5-HTP or placebo.

Intestinal permeability test

On each test day, participants received an intravenous cannula in the antecubital vein. After drawing blood samples at 8:00 AM, participants received either 100 mg 5-HTP or placebo orally based on a random pre-selection (see Figure S1). At 9:00 AM, participants received an oral sugar drink to assess intestinal permeability. The sugar drink consisted of 1 g lactulose (Centrafarm Services, Etten-Leur, the Netherlands) and 0.5 g L-rhamnose (Danisco Sweeteners, Thomson, IL, USA).

Blood samples were taken following 60 min of sugar ingestion. Directly hereafter, participants underwent a gastroduodenoscopy performed by an experienced gastroenterologist. Mucosal samples from the horizontal part of the duodenum were obtained by standard gastroduodenoscopy using standard forceps (diameter 2.8 mm) and immediately frozen in liquid nitrogen. Two of the samples were imbedded in Tissue-tek® (Sakura Finetek, Tokio, Japan) prior to freezing for immunohistochemical analyses.

Assessment of plasma sugar concentrations

In this study, we aimed to specifically assess the effect of a short-term modulation of the small intestinal barrier. Accordingly, we measured the one-hour plasma recovery of the sugars administered orally. Samples for plasma were collected using pre-chilled K2EDTA tubes prior to, and 60 after ingestion of the sugar drink. Plasma collection tubes were centrifuged at 3000g at 4 °C for 10 min. Supernatants were allocated and immediately frozen at −80 °C until analysis. Plasma sugar concentrations were determined in the t = 60 min samples with correction for initial sugar concentrations in the fasted blood samples by HPLC-MS, as described by van Wijck et al.[23] Ratios for lactulose/L-rhamnose were calculated to indicate small intestinal permeability.

Expression of tight junction proteins in mucosal samples

Gene transcription of tight junction proteins (occludin and ZO-1) in biopsy specimens was measured by qRT-PCR. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and 18S RNA were used as reference genes (see also Data S1).

Immunofluorescent staining for zonula occludens-1 (ZO-1) and occludin was performed on frozen sections of the biopsy specimens. Antibodies used were: primary antibodies monoclonal mouse anti-ZO-1 (1:750, Invitrogen, Camarillo, CA, USA) and polyclonal rabbit α-Occludin (1:800, Zymed, Carlsbad, CA, USA), secondary fluorescent antibodies Cy3 conjugated goat anti-rabbit (1:800, Jackson, West Grove, PA, USA) and Alexa-488 conjugated goat anti-mouse (1:150, Invitrogen, Carlsbad, CA, USA). Images were analysed by confocal microscopy (Leica, DMI 4000B). Immunohistochemical staining were performed with ‘blanco’ controls, in which samples only the secondary antibodies were added. Prior to confocal microscopic analysis, fluorescent signal intensity was adjusted to these ‘blanco’ samples to exclude effects of nonspecific antibody binding.

To quantify the fluorescent staining of ZO-1 and occludin in the TJ region, each coded tissue sample was subjected to confocal analysis of uniform Z sections perpendicular to the apical cell surface of the epithelium in three different areas per section (Figure 1). These areas were selected randomly within three different Z sections. Staining was analysed in the medium region of the villi. Plot profiles (e.g. insert in Figure 1) of the staining intensity along the perpendicular lines were generated using Image J software (National Institutes of Health, Bethesda, MD, USA), with the LSM toolbox plug-in. The fluorescent intensities were plotted as a function of the cell location using the peak fluorescence signal from the TJ region to align each intensity profile beginning from the apical side of the cell.

Figure 1.

Immunohistochemical staining for zonula occludens-1 (ZO-1) and occludin proteins. (a) depicts a representative microscopic image of a villus stained for tight junctions (green: ZO-1, red: occludin) and nuclei (blue). The arrow depicts the perpendicular line drawn to assess staining intensity. The insert demonstrates the generated plot profile of the staining intensity. (b) demonstrates staining intensity is plotted against location within the cell. Zero indicates the luminal side of the cell. Lines indicate the distribution of staining intensity as a function of the cellular location in the regions investigated (apical and cytosolic). The dotted line indicates the confidence interval (CI) of the minimal and maximal difference in staining intensity for 5-hydroxytryptopan (5-HTP) as compared to the placebo condition. The table insert demonstrates corresponding values for area under the curves.

Assessment of mucosal 5-HT metabolism

Biopsy specimens were weighed and homogenised prior to analysis. Mucosal concentrations of 5-HTP, 5-HT, 5-hydroxyindole acetic acid (5-HIAA, the main metabolic breakdown product of 5-HT) were determined by HPLC-MS as described previously.[24]

Statistical analyses

The descriptive statistical analyses were performed using SPSS 20 (SPSS Inc., Chicago, IL, USA). Data were tested for normality by the Kolmogorov–Smirnoff test. Normally distributed data were analysed by Student's test and two-way anova. Wilcoxon, Mann–Whitney U-test, Kruskal–Wallis test and Friedman test (for repeated measures) were used for nonparametric data. Coefficients for correlations were calculated according to Pearson and Spearman, respectively. Data were expressed as mean ± S.E.M. or as median (range), unless otherwise stated. PCR data are presented as expression normalised for GAPDH and 18S RNA.

Data analysis from the immunohistochemical staining were performed in the software application ‘R’[25] using the elliptic library[26] by a multivariate nonlinear Gaussian regression in order the integrate the plot profiles of the staining intensities and to ascertain the differences in staining intensities for ZO-1 and occludin between the two treatment conditions (see Data S1). Confidence intervals were calculated on basis of the maximum and minimum difference in staining intensities between the two treatments (5-HTP and placebo).

Results

Demographic characteristics

IBS patients did not differ significantly from healthy controls regarding age (33 ± 17 years vs. 44 ± 13 years, P = 0.06), gender (46% male vs. 33% male, P = 0.35) or BMI (24 ± 4 vs. 27 ± 5, P = 0.08, healthy vs. IBS, respectively). IBS subtypes were 47% diarrhoea predominant, 33% constipation predominant, 20% mixed subtype. No association was found between age, gender or IBS subtype and any of the parameters studied.

There were no abnormalities detected on endoscopy in either the healthy controls or IBS patients, and all of the biopsy samples were judged to be normal.

Side effects

Following 5-HTP intake, nausea and/or belching was experienced by three healthy volunteers and three IBS patients. In addition, two of these IBS patients complained of abdominal pain, one of bloating, one of diarrhoea following intake of 5-HTP. No side effects were experienced following placebo intake.

Mucosal 5-HT metabolism

Concentrations of 5-HT, as well as of the precursor 5-HTP and the principal breakdown product of 5-HT, 5-HIAA, were measured in duodenal mucosa. Effects of oral 5-HTP on these metabolites as well as potential differences between IBS patients and healthy controls were examined.

After administration of 5-HTP, mucosal 5-HTP concentrations in healthy volunteers were significantly higher compared to placebo (3.7 ± 1.8 vs. 0.09 ±0.03 pmol/mg, = 0.005), but not in IBS patients (1.1 ± 0.9 vs. 0.09 ± 0.03 pmol/mg, P = 0.26). Mucosal concentrations of 5-HT were neither affected by 5-HTP in healthy volunteers (60 ± 21 vs. 52 ± 18 pmol/mg, 5-HTP vs. placebo, P = 0.68), nor in IBS patients (13 ± 2 vs. 13 ± 3 pmol/mg, 5-HTP vs. placebo P = 0.90). However, 5-HT concentrations were significantly lower in IBS vs. healthy both under 5-HTP and placebo conditions (= 0.004 and P = 0.008, respectively). Mucosal concentrations of the breakdown product 5-HIAA were significantly increased in healthy volunteers after 5-HTP compared to placebo (7.1 ± 1.7 vs. 2.5 ± 0.7 pmol/mg, 5-HTP vs. placebo, P = 0.02). Similar changes were observed in IBS (20.0 ± 4.8 vs. 8.1 ± 1.3 pmol/mg, 5-HTP vs. placebo, P = 0.02). The mucosal 5-HIAA concentrations were significantly higher in IBS compared to healthy controls in both conditions: after 5-HTP (P = 0.01) and after placebo (P = 0.001), see Figure 2.

Figure 2.

Mucosal concentrations of 5-hydroxytryptophan (5-HTP), serotonin (5-HT) and 5-hydroxyindoleacetic acid (5-HIAA) in nmol/mg wet tissue in healthy controls and patients with irritable bowel syndrome (IBS) after placebo and 5-HTP. *P < 0.05, **P < 0.01.

Two-way anova analysis of the mucosal 5-HIAA concentrations revealed a significant effect for the condition (IBS or healthy, P < 0.001) and for the treatment (5-HTP or placebo, P = 0.03), but not for the interaction (= 0.5). This implies that the condition and the treatment both have a significant effect on mucosal 5-HIAA concentrations, but there is no additive effect when examining the concomitant presence of both factors. Indeed, levels of 5-HIAA observed in healthy controls after 5-HTP were comparable to those seen in IBS following placebo.

Furthermore, we assessed the 5-HIAA/5HT ratio, reflecting the mucosal turnover of 5-HT. The 5-HIAA/5-HT ratios remained unaltered following 5-HTP in both healthy controls (0.47 ± 0.3 vs. 0.31 ± 0.2, 5-HTP vs. placebo, > 0.05) and IBS patients (3.2 ± 1.1 vs. 2.0 ± 0.5, 5-HTP vs. placebo, > 0.05). However, 5-HIAA/5-HT ratio was significantly higher in IBS both following 5-HTP and placebo compared to healthy controls (P = 0.005, for the effect of the condition in two-way anova analysis), indicating increased mucosal turnover of 5-HT in IBS.

Intestinal sugar permeability

In healthy controls, L/R ratios were significantly lower following 5-HTP administration compared to placebo in controls (4.9 × 10−3 ± 0.0005 vs. 6.3 × 10−3 ± 0.0005, P = 0.01; Figure 4). This difference results from a decrease in lactulose concentration after 5-HTP vs placebo (1.3 ± 0.5 vs. 1.7 ± 0.6 μmol/L, = 0.035), as no changes were observed in the L-rhamnose concentration (91 ± 6 vs. 97 ± 5 μmol/L, P = 0.41). In IBS patients, no significant change was observed in L/R ratios following 5-HTP administration (20 × 10−3 ± 0.009 vs. 12 × 10−3 ± 0.003, 5-HTP vs. placebo, = 0.89), see Figure 3. When comparing the placebo conditions of IBS patients with healthy controls, the L/R ratios were not significantly different (P = 0.06). On the other hand, two-factor anova analysis showed a significant effect of the condition (IBS or healthy, P = 0.05), but not for the treatment or interaction, indicating that, as a group, IBS patients had increased ratios for lactulose/L-rhamnose (L/R) compared to controls.

Figure 3.

Plasma lactulose/L-rhamnose ratio 60 min after oral ingestion in healthy controls and patients with irritable bowel syndrome (IBS) after placebo and 5-hydroxytryptopan (5-HTP). *P < 0.05.

Transcription of tight junction proteins occludin and ZO-1

To further characterise intestinal integrity, transcription of the tight junction proteins ZO-1 and occludin was examined in duodenal mucosa by measuring RNA concentrations. After 5-HTP administration, the transcription of ZO-1 was higher in healthy volunteers (normalised expression ratios 1.27 ± 0.24 vs. 0.87 ±0.12, 5-HTP vs. placebo, P = 0.04). This increase might, however, not be statistically significant given the number of analyses performed in the study. In IBS patients, 5-HTP did not significantly affect ZO-1 transcription (0.75 ± 0.16 vs. 1.25 ± 0.23, 5-HTP vs. placebo, P = 0.06). No significant difference was observed in ZO-1 transcription between IBS and healthy controls in the placebo conditions (P = 0.24), see Figure 4. Two-way anova analysis showed a significant effect of the interaction (P = 0.03), but not for the treatment (5-HTP or placebo) or condition (IBS or healthy) alone (P = 0.81 and 0.70, respectively). Furthermore, mucosal levels of 5-HIAA correlated positively with the expression of ZO-1 mRNA in healthy controls (r = 0.70, P = 0.008) and in IBS patients (r = 0.60, = 0.04).

Figure 4.

Transcription of zonula occludens-1 (ZO-1) and occludin, expressed as normalised expression ratios in healthy controls and patients with irritable bowel syndrome (IBS) after placebo and 5-hydroxytryptopan (5-HTP). *P < 0.05, **P < 0.01.

5-HTP did not affect transcription of occludin in neither healthy controls nor IBS patients (normalised expression ratios 1.43 ± 0.14 vs. 1.39 ± 0.12, 5-HTP vs. placebo, P = 1.0; 0.57 ± 0.14 vs. 0.80 ± 0.15, 5-HTP vs. placebo, P = 0.20, respectively). However, occludin transcription was significantly lower in IBS than in healthy controls on both conditions (P = 0.006 for placebo and P = 0.001 for 5-HTP), see Figure 4.

Immunohistochemical staining for occludin and ZO-1

In order to assess tight junction expression on a protein level, ZO-1 and occludin were stained immunohistochemically in the duodenal mucosal biopsies. Data for immunohistochemical staining are presented in Figure 1. The area under the curves (AUCs) indicates the overall amount of protein in the cellular region investigated (i.e. the apical and cytosolic regions).

In the placebo condition, expression of both ZO-1 and occludin, expressed as AUC, was significantly lower in IBS patients compared to healthy controls. Administration of 5-HTP induced a decrease in the AUC of ZO-1 expression in healthy controls (see table insert in Figure 1). However, a significant shift in peak intensity of ZO-1 was observed in the direction of the apical surface of the cell, indicating increased staining intensity in this region, which corresponds to increased amounts of protein in the proximity of the tight junctions. In IBS, following 5-HTP, an apical shift in peak intensity for ZO-1 was observed similarly to that seen in healthy controls.

The expression of the transmembrane tight junction protein occludin was found to be statistically significantly lower following 5-HTP in healthy controls. However, given the minute magnitude of the change observed, the biological effect is most probably not essential. Also, a very similar curve shape was observed, indicating unaltered protein distribution.

In IBS, occludin was more diffusely distributed in the cytosol in the placebo condition, i.e. a substantial amount of the protein was located outside the direct proximity of the tight junction region. Remarkably, 5-HTP administration led to a decrease in overall occludin expression, with the majority of the protein localised to the tight junction region.

Taken together, the intestinal permeability data indicate an increased intestinal permeability in IBS compared to healthy controls. This appears to be the result of a primary defect of the intestinal barrier as reflected by decreased transcription and expression of tight junction proteins. Following 5-HTP, healthy controls demonstrate an increased barrier function, accompanied by increased transcription and redistribution of ZO-1. In IBS patients, 5-HTP administration led to a similar change in ZO-1 distribution, however, without enhanced barrier function. Furthermore, in IBS, altered distribution and decreased expression of occludin was observed following 5-HTP (see Figure S2).

Discussion

In this study, we demonstrate that oral administration of 5-HTP-induced significant alterations in mucosal 5-HT metabolism, which was associated with reinforcement of the intestinal barrier function in healthy controls but not in IBS patients. On the contrary, in IBS patients, more pronounced alterations were observed with respect to 5-HT metabolism, which was accompanied by a failure to reinforce intestinal barrier function.

In order to modulate mucosal 5-HT metabolism, we administrated the 5-HT precursor, 5-HTP, and specifically intended to assess changes in 5-HT metabolism in duodenal mucosa. Recent reports point to an important role for 5-HT metabolism in the duodenum as pathomechanistic factor in IBS.[18, 27] Based on early findings of Bülbring and colleagues,[28, 29] we expected that the administration of 5-HTP would increase mucosal serotonergic metabolism. Indeed, significant alterations were observed in mucosal concentrations of the 5-HT metabolite 5-HIAA, but not of 5-HT, upon oral intake of 5-HTP in both healthy controls and IBS. An increase in mucosal 5-HTP concentrations indicates that 5-HTP was taken up by the mucosa. Furthermore, the observed increase in mucosal 5-HIAA, the main breakdown product of 5-HT, suggests that 5-HTP was locally converted to 5-HIAA. Interestingly, 5-HT levels remained unaltered.

The changes induced by 5-HTP in 5-HT metabolites that were observed in healthy controls showed a pattern similar to those seen in IBS patients, but the magnitude of these alterations was different. More specifically, the increase in mucosal 5-HTP concentrations was significantly lower and the increase in 5-HIAA concentrations was significantly higher in IBS compared to healthy controls. This observation suggests that there is a more rapid conversion (turnover) of 5-HTP to 5-HIAA in IBS. Indeed, baseline 5-HT turnover, as determined by the 5-HIAA/5-HT ratio, was significantly higher in IBS patients. Furthermore, we also observed increased baseline 5-HIAA concentrations and decreased baseline mucosal concentrations of 5-HT in IBS. This observation may be explained by an increased 5-HT release into the mucosa from EC cells, which is rapidly converted to 5-HIAA. A recent study by Cremon et al.[30] showing increased mucosal release of 5-HT in IBS supports this hypothesis.

With regard to mucosal 5-HT concentrations, we made two important observations in this study: (i) 5-HT levels remained unchanged after 5-HTP and (ii) baseline 5-HT levels were significantly lower in IBS compared to controls. With respect to the unaltered 5-HT levels, we hypothesise that this can be explained by the action of the serotonin transporter (SERT). SERT is responsible for rapid clearance of 5-HT in the mucosa.[31] Turnover of 5-HT is a critical determinant of the efficacy of 5-HT-mediated signalling processes. We therefore postulate that following 5-HT release induced by 5-HTP, a fast uptake of 5-HT by SERT into epithelial cells occurs, which is followed by a rapid conversion into 5-HIAA. It is important to note however, that mucosal SERT activity was not directly assessed in this study.

With regard to the lower levels of 5-HT in IBS, a number of research groups have previously reported on mucosal 5-HT content in IBS. Wang et al.[32] and Coates et al.,[15] found decreased 5-HT in jejunal and colonic biopsies, respectively, which is in line with the findings of our study. Other groups, however, have not found decreased but elevated 5-HT[16] or unaltered mucosal 5-HT levels.[33-35] This discrepancy in findings is not easily explained but may have been influenced by differences in patient populations (IBS-D, IBS-C or mixed group), anatomical location of sampling (colonic or small intestinal) analytical techniques used (ELISA or HPLC-based measurements) and the heterogeneous nature of IBS should also be taken into account. Importantly, the alterations in 5-HT metabolism in this study were observed regardless of the predominant bowel habit. However, one should realise that our study was not powered to assess potential differences between IBS subgroups. Also, our IBS patient group consisted mainly of nonconstipated IBS patients.

In this study, we also examined the effects of modulation of 5-HT metabolism by 5-HTP on intestinal barrier function, as evidence for the role for an impairment of the intestinal barrier in IBS is rapidly increasing. In healthy controls, upon 5-HTP administration, an increase in the mucosal transcription of ZO-1 was observed, accompanied by a redistribution of the protein to the tight junction region of the epithelial cells. Such changes point to reinforcement of the tight junction complex as ZO-1 is a scaffold protein known to bind to both cytoskeletal actin and to the transmembrane tight junction proteins occludin and claudin.[36] Also, a decrease in plasma L/R ratio was observed indicative of an enhanced small intestinal barrier function. These changes may represent a protective mechanism related to increased serotonergic signalling in the mucosa, for instance, to prevent penetration of potentially noxious luminal substances into the lamina propria. Luminal actors or agents such as potentially noxious food allergens can stimulate 5-HT release from EC cells.[37] The 5-HT released in response to 5-HTP may trigger a response that eventually results in reinforcement of the intestinal barrier. The 5-HT receptors present on the basolateral side of neighbouring intestinal epithelial cells are potential candidates involved in mediating such responses. In support of this hypothesis, Stull et al.[38] demonstrated in an in vitro study that activation of the 5-HT7 receptor located on the basolateral membrane of epithelial cells led to an increase in the epithelial transmembrane resistance. These findings indicate that 5-HT may regulate the intestinal tight junction functionality locally through the stimulation of the 5-HT7 receptor. A potential role for serotonergic signalling in regulation of the intestinal barrier is supported further by our observation that the expression of ZO-1 mRNA positively correlates with mucosal 5-HIAA levels.

In IBS, although we did observe a similar shift in the distribution of ZO-1 protein as seen in healthy controls, this was not accompanied by either an increase in ZO-1 mRNA or a decrease in the sugar permeability. Furthermore, although more concentrated to the apical region, a significant decrease in overall expression of occludin protein expression was also observed following 5-HTP. Therefore, in IBS, 5-HTP appears to impair barrier function rather than to enhance it. This inability to reinforce barrier function may be related to the fact that the expression and transcription of tight junction proteins was also decreased in the placebo condition. This may be indicative of a primary defect in the intestinal barrier, which is in line with other recent findings.[7, 10] Also, in IBS following placebo, occludin protein had a wider distribution throughout the cell, with a substantial proportion located outside the direct proximity of the tight junction region. This could represent an internalised fraction of occludin in the cytosol. Occludin internalisation has been proposed as an important mechanism in the impairment of the barrier function.[39] This mechanism, although not directly investigated in our study, may also be of relevance with respect to increased permeability in IBS. Furthermore, the contribution of other tight junction proteins, such as claudins, cannot be excluded in this study, as these were not subject to analysis here.

The novelty of our finding is that we were able to relate the changes in intestinal barrier function to 5-HTP-induced alterations in serotonergic metabolism. The question still remains whether the alterations in serotonergic metabolism can in fact be considered as causal factors in the observed changes in barrier function. It appears however that the same 5-HT-mediated mechanism aimed to reinforce the intestinal barrier exists in both controls and in IBS patients, but that it malfunctions in the IBS group possibly due to a primary defect of the intestinal tight junctions. Alternatively, increased serotonergic metabolism in IBS may be reflective of an ineffective attempt to reinforce the intestinal barrier. It cannot be excluded, however, that primary defects in both intestinal barrier function and in serotonergic metabolism coexist in IBS and both contribute to the IBS pathogenesis. Furthermore, as we did not record symptoms of IBS patients in this study, the role of these alterations with respect to symptom generation could not be determined. There are several other limitations to this study. Given the relatively small sample size, more thorough insight into the factors regulating intestinal barrier function will be necessary to confirm whether the proposed mechanisms have a legitimate role in the development of IBS. In addition, due to the same reasons, differential responses between IBS subtypes could not be assessed in this study. Although impaired permeability appears to be present regardless of predominant bowel habit,[2] it cannot presently be ascertained whether the pathophysiological mechanisms presented in this study are uniformly present in all IBS subtypes. Moreover, we did not assess the potential role of intestinal microbiota with regard to the differential changes observed in IBS and healthy controls, which may be relevant given the increased appreciation of microbial changes in the pathogenesis of IBS,[40] as well as serotonin metabolism.[14]

With regard to the sugar permeability test, this study applied measurement of plasma sugar concentrations. An advantage of measuring plasma instead of urinary sugar concentrations, which is performed more often, is that plasma levels are less influenced by urinary excretion rate. On the other hand, measurements in plasma only allow investigation of permeability at a specific time point, in contrast to urine analysis, which can yield more cumulative data. Nevertheless, plasma sugar permeability has been recently been validated by van Wijck et al.[41]

In conclusion, we have demonstrated that increased serotonergic metabolism, caused by oral administration of the serotonin precursor 5-HTP, induces an increase in intestinal mucosal barrier function in healthy individuals. Our findings in healthy individuals suggest the presence of a potential protective mechanism involving serotonergic signalling that may protect against luminal insults. In IBS, there is an increased metabolic response to 5-HTP, accompanied by a failure to reinforce intestinal integrity, which may potentially contribute to disease development. Correction of such dysfunction theoretically offers an attractive entity to counteract pathophysiological processes resulting in impaired barrier function.

Authorship

Guarantor of the article: Prof. A. A. M. Masclee.

Author contributions: DK designed and executed study, collected and analysed data and wrote manuscript. FT, DM and JD were involved in experimental design and data interpretation. HvE performed HPLC-MS measurements. PL performed statistical analyses for immunohistochemistry data. WB critically reviewed manuscript. AM was responsible for overall supervision of the study. All authors approved the final version of the manuscript.

Acknowledgements

Declaration of personal interests: AM Masclee has served as a consultant for Grünenthal, and has received research funding from Grünenthal, Ipsen, Ferring Pharmaceuticals and DSM.

Declaration of funding interests: Top Institute Food and Nutrition.

Ancillary