Leading Off Increased Colonic Epithelial Permeability and Mucosal Eosinophilia in Ulcerative Colitis in Remission Compared With Irritable Bowel Syndrome and Health Georgios Katinios, MD,*,† Maite Casado-Bedmar, MSc,† Susanna A. Walter, MD, PhD,*,† Maria Vicario, PhD,‡ Ana M. González-Castro, PhD,‡ Olga Bednarska, MD, PhD,*,† Johan D. Söderholm, MD, PhD,†,§ Henrik Hjortswang, MD, PhD, *,† and Åsa V. Keita, PhD† Background: Barrier dysfunction is recognized as a pathogenic factor in ulcerative colitis (UC) and irritable bowel syndrome (IBS), but it is unclear to what extent the factors related to barrier dysfunction are disease-specific. The aim of this study was to compare these aspects in UC patients in remission, IBS patients, and healthy controls (HCs). Methods: Colonic biopsies were collected from 13 patients with UC in remission, 15 patients with IBS-mixed, and 15 HCs. Ulcerative colitis patients had recently been treated for relapse, and biopsies were taken from earlier inflamed areas. Biopsies were mounted in Ussing chambers for measurements of intestinal paracellular permeability to 51chromium (Cr)-ethylenediaminetetraacetic acid (EDTA). In addition, biopsies were analyzed for mast cells and eosinophils by histological procedures, and plasma tumor necrosis factor (TNF)-α was assessed by ELISA. Results: Ussing chamber experiments revealed an increased 51Cr-EDTA permeability in UC and IBS (P < 0.05). The 51Cr-EDTA permeability was higher in UC compared with IBS (P < 0.005). There were increased numbers of mucosal mast cells and eosinophils in UC and IBS and more eosinophils in UC compared with IBS (P < 0.05). Also, increased extracellular granule content was found in UC compared with HCs (P < 0.05). The 51Cr-EDTA permeability correlated significantly with eosinophils in all groups. Plasma TNF-α concentration was higher in UC compared with IBS and HCs (P < 0.0005). Conclusions: Results indicate a more permeable intestinal epithelium in inactive UC and IBS compared with HCs. Ulcerative colitis patients, even during remission, demonstrate a leakier barrier compared with IBS. Both eosinophil numbers and activation state might be involved in the increased barrier function seen in UC patients in remission. Key Words: eosinophils, intestinal permeability, irritable bowel syndrome, mast cells, ulcerative colitis INTRODUCTION active disease. Irritable bowel syndrome is one of the most Ulcerative colitis (UC) and irritable bowel syndrome common functional bowel disorders characterized by recurring (IBS) are 2 chronic gastrointestinal (GI) entities with incom- abdominal pain and associated with disturbed bowel habits.1 pletely understood pathophysiology. Ulcerative colitis is an in- In terms of mucosal barrier function studies, the majority flammatory bowel disease (IBD) characterized by intermittent it critically for important intellectual content, and approved the final version. ÅVK flares of bloody stools and increased stool frequency. Unlike contributed to the conception and design of the study, the acquisition, analysis, and IBS, UC features macroscopic mucosal inflammation during interpretation of data, and was mainly responsible for drafting the final version of the article before approval. Supported by: This work was supported by grants from Bengt Ihre research Received for publications July 30, 2019; Editorial Decision December 13, 2019. Scholarship (ÅVK), AFA research foundation (SAW); Bengt-Ihre fonden, County From the *Department of Gastroenterology, Linköping University, Linköping, Council of Östergötland (SAW), the Swedish Research Council (VR-Medicine Sweden; †Department of Biomedical and Clinical Sciences, Linköping University, and Health, 2017-02475) (JDS, ÅVK), and the Swedish Foundation For Strategic Linköping, Sweden; ‡Laboratory of Translational Mucosal Immunology, Digestive Research (RB13-016). It was upported in part by Instituto de Salud Carlos III, Diseases Research Unit. Vall d’Hebron Institut de Recerca, Hospital Universitari Fondo de Investigación Sanitaria (PI16/00583, MV), and Ministerio de Ciencia, Vall d’Hebron, Universitat Autònoma de Barcelona, Barcelona, Spain; §Department Innovación y Universidades, Spain (CIBER-EHD CB06/04/0021, MV). of Surgery, Linköping, Linköping University, Sweden Address correspondence to: Åsa V. Keita, PhD, Department of Biomedical Author Contribution: GK contributed to the inclusion of HC and UC patients, and Clinical Sciences, Linköping University, Medical Faculty, Campus US, 581 85 acquisition, analysis, and interpretation of data and the drafting of the article and Linköping, Sweden. E-mail: asa.keita@liu.se. approved the final version. MCB contributed to the acquisition and analysis of data, the drafting of the article, and approved the final version. SAW contributed to the © 2020 Crohn’s & Colitis Foundation. Published by Oxford University Press on conception and design of the study, interpretation of data, and the drafting of the behalf of Crohn’s & Colitis Foundation. article and approved the final version. MV contributed to the acquisition and anal- This is an Open Access article distributed under the terms of the Creative ysis of data and the drafting of the article and approved the final version. AMGC Commons Attribution Non-Commercial License (http://creativecommons.org/ contributed to the acquisition and analysis of data and the drafting of the article licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and repro- and approved the final version. OB contributed to the inclusion of HC and IBS pa- duction in any medium, provided the original work is properly cited. For commer- tients and the drafting of the article and approved the final version. JDS contributed cial re-use, please contact journals.permissions@oup.com to the interpretation of data, revised it critically for important intellectual content, doi: 10.1093/ibd/izz328 and approved the final version. HH contributed to the interpretation of data, revised Published online 16 January 2020 Inflamm Bowel Dis • Volume 26, Number 7, July 2020 974 Inflamm Bowel Dis • Volume 26, Number 7, July 2020 Increased Colonic Epithelial Permeability and Mucosal Eosinophilia of abnormalities found in UC and IBS are based on the com- contains moderate amounts of active eosinophils.22 However, parison of healthy controls (HCs) with either UC patients2–4 it is known that the numbers of activated eosinophils are in- or IBS patients.5–7 However, to evaluate if the barrier function creased in patients with active and inactive UC.23 Furthermore, disturbances demonstrated in UC and IBS are disease-specific the numbers have shown to be higher in the mucosa of inac- or generalizable for both disorders, control studies are needed. tive UC compared with the inflamed mucosa,23 suggesting that Today, there are only few such studies directly comparing bar- eosinophils may play diverse roles in the pathophysiology of rier function of UC and IBS.8, 9 It is well known that there also IBD—both pro-inflammatory and mediating tissue repair. exists a clinical overlap between UC and IBS because patients Moreover, a close interaction between eosinophils and mast with UC in remission often report IBS-like symptoms that do cells, potentially leading to a disrupted mucosal barrier and in- not necessarily correspond to a low-grade inflammation, even creased permeability, has been demonstrated in IBD.24, 25 For though increased levels of pro-inflammatory cytokines have IBS, colon studies have shown inconsistent results with both been reported.10 increased26, 27 and unaltered eosinophil counts.28 A low-grade mucosal inflammation is often present in the A better knowledge about the differences between IBS colon and ileum of IBS patients, characterized by mast cells, T and UC in remission are important for a deeper understanding cells, and pro-inflammatory cytokines.11, 12 Ahn et al13 showed of the disease-specific barrier function disturbances in these that patients with diarrhea-predominant IBS (IBS-D) had sig- disorders. The present study was designed to search for simi- nificantly more colonic mast cells and T cells in relation to HCs larities and differences in epithelial barrier function and related but had similar counts when compared with UC patients in re- immune cells between UC in remission and IBS. Colonic bi- mission. Nevertheless, colonic tissue from IBS-D patients had opsies from patients with inactive UC, IBS, and HC patients significantly fewer immune cells than the colon from patients were mounted in Ussing chambers to assess paracellular per- with mildly activated UC. meability. Mucosal mast cells and eosinophils were quantified; A barrier dysfunction is well recognized as an important intracellular and extracellular tryptase and eosinophil cationic pathogenic factor in UC,2–4 and an impaired barrier function protein (ECP), respectively, were analyzed; and plasma levels of has become evident also in IBS.5–7 There are few studies com- tumor necrosis factor (TNF)-α were measured. paring gut barrier function in IBS and UC in remission. Gecse et al8 found that intestinal permeability was elevated in patients MATERIALS AND METHODS with IBS-D and inactive UC compared with HCs by measuring 24-hour urine excretion of orally administered 51Chromium Participants (Cr)-ethylenediaminetetraacetic acid (EDTA). In another study The material included 13 patients with inactive by Vivinus-Nébot et al,9 both IBS (all subtypes) and quiescent UC, 15 patients with IBS, and 15 HCs, recruited from UC patients with IBS-like symptoms showed higher paracel- the Gastroenterology Department, University Hospital, lular permeability compared with HCs, as measured using fluo- Linköping, Sweden. Inclusion criteria for the UC group was rescein isothiocyanate-sulfonic acid in Ussing chambers, but clinical and endoscopic remission and treatment for relapse they found no difference between HCs and quiescent UC pa- within the last year. Biopsies were taken from the earlier in- tients with no IBS-like symptoms. To our knowledge, there is to flamed area. Inclusion criteria for the IBS group was that pa- date no study directly comparing the colonic barrier function tients were classified in the IBS-mixed (IBS-M) subgroup based between IBS, UC in remission, and HCs by using biopsies and on the predominant stool consistency according to the Rome permeability markers in Ussing chambers—a gold standard III questionnaire.29 The HCs were recruited by advertisement, technique for the assessment of intestinal permeability ex vivo. and inclusion criteria was that patients had no medical his- There is substantial evidence for an elevated number and tory of chronic GI symptoms or disorders and no medication increased degranulation of mast cells and secretion of medi- intake. Exclusion criteria for all groups included metabolic ators in UC14–16 and IBS.5, 17, 18 In addition, some studies have or neurological disorders. The Committee of Human Ethics, implicated mast cells in the regulation of intestinal permea- Linköping, approved the study, and all subjects gave their bility.19–21 Our group previously found an increased number written informed consent. of mast cells and their activation in the colonic mucosa in IBS compared with HCs and an increased permeability in colonic Collection of Biological Samples biopsies of IBS patients, which was significantly diminished A flexible sigmoidoscopy was performed in all partici- after stabilizing the mast cells with ketotifen.5 In addition, we pants, and colonic biopsies were taken from the sigmoideum, recently demonstrated higher amounts of mast cells in the co- rectosigmoidal junction, or rectum. All biopsies were taken lonic mucosa of UC and Crohn’s disease patients compared with forceps without a central lance and directly put in ice- with HCs.16 Together, this indicates that mast cells have an im- cold oxygenated Krebs buffer (115 mM NaCl, 1.25 mM CaCl2, portant role in both IBS and IBD, including UC, and are also 1.2 mM MgCl2, 2 mM KH2PO4, and 25 mM NaHCO3, pH 7.35). important regulators of barrier function. The intestinal mucosa Before sigmoidoscopy, venous blood samples were collected in 975 Katinios  et al Inflamm Bowel Dis • Volume 26, Number 7, July 2020 EDTA-treated tubes, blood was centrifuged at 3400g for 15 incubated in eosin (Casa Alvarez) for 1 minute and mounted minutes in 4°C, and plasma was redrawn and frozen in −80°C with Pertex mounting media (Histolab). until analysis of TNF-α concentration. Quantification of Mast Cells and Eosinophils Ussing Chamber Experiments The total number of mucosal mast cells and eosino- Colonic biopsies were mounted in Ussing chambers phils was quantified in a blinded fashion by 2 independent re- (Harvard Apparatus Inc., Holliston, MA, USA) as previously de- searchers in a Nikon E800 fluorescence microscope connected scribed.30 After 40 minutes of equilibration, 34 µCi/mL of the para- to software NIS elements (Nikon Instruments Inc. Tokyo, cellular probe 51Cr-EDTA (Perkin Elmer, MA, USA) was added to Japan). The number of cells positive for tryptase or eosin were the mucosal side of each chamber. Serosal samples were collected at quantified manually at 400x magnification, and negative con- 0, 60, and 120 minutes, and 51Cr-EDTA passage was measured by trols were included in all experiments. A minimum of 8 areas/ gamma-counting (1282 Compugamma, Sweden). The 51Cr-EDTA sections was counted. Every area had the same size, and only permeability was calculated during the 60 to 120 minute period, areas that were covered completely by the biopsy were used. and results are presented as Papp (apparent permeability coeffi- The mean value for each section was estimated for mast cell cient; cm/s ×10-6). The transepithelial potential difference (PD) and tryptase expression or eosin and a median value of the number the transepithelial resistance (TER) across the tissues were moni- of mast cells and eosinophils from patients with UC, IBS, and tored throughout the experiments to ensure tissue viability. HCs, respectively. Distribution of mast cell tryptase and eosinophil- Measurements of TNF-α Concentration in Plasma An ultrasensitive human TNF-α ELISA kit was used ac- ECP and cell ultrastructure To study the granule distribution and activation levels cording to the manufacturer’s instructions (Invitrogen, Thermo of eosinophils and mast cells, biopsies from randomly selected Fisher Scientific, Lund, Sweden). Undiluted plasma, positive subjects, 4 UC patients in remission, 4 IBS-M patients and 4 and negative controls, and standard points were added in du- HCs, were processed for immunofluorescence and evaluated plicates. Absorbance was measured at 450 nm in VERSAmax with confocal microscopy. Mast cells were identified using Tunable Microplate Reader (Molecular Devices, CA, USA) antihuman tryptase following the protocol described previously using Softmax pro 5 (Molecular Devices). for mast cell quantification. Similarly, eosinophils were iden- tified with antihuman ECP 1:200 (Diagnostic Development, Staining of Mast Cells and Eosinophils in the Uppsala, Sweden), an eosinophil activation marker. Slides were Colonic Mucosa stained using the same protocol described previously for mast Immediately after collection, colonic biopsies were fixed cells quantification, with an additional antigen retrieval step in 4% paraformaldehyde in phosphate buffered saline for 24 with citrate buffer at 60°C for 30 minutes. One transversal sec- hours at 4°C and further embedded in paraffin and sectioned tion/individual was used for each staining, and negative controls at 5 µm. Upon use, slides were hydrated according to standard were included in all experiments. From 8 to 14 images of single procedures. mast cells or eosinophils (single when possible) were acquired with LSM800 Zeiss Inverted Confocal with a 60X oil immersion Mast cells objective. Images were processed in ImageJ Fiji software (NIH, Sections were incubated with background sniper Bethseda, MD), followed by the analysis with Cell Profiler 2.2.0 (Histolab, Gothenburg, Sweden) for 5 minutes. Slides were software. Nuclei signal was used, together with tryptase or ECP rinsed and incubated over night at 4°C with mouse monoclonal- signal, to delimit and segment the cell limits of mast cell and antihuman mast cell tryptase 1:200 (Santa Cruz, CA, USA). eosinophils, respectively. All the granules detected inside those Slides were rinsed and incubated with 1:400 secondary an- limits were defined as intracellular; the granules localized out- tibody Alexa Fluor 594-conjugated rabbit-antimouse (Life side the cellular limits were defined as extracellular content Technologies, Stockholm, Sweden) for 1 hour at room tempera- and interpreted as an indicative of degranulation levels. After ture. Slides were rinsed and mounted with Prolong Gold-DAPI threshold, the average of integrated intensity per cell and the ex- (Life Technologies). tracellular content surrounding the cell were measured. Finally, a median value was calculated from all the averages to create a Eosinophils median of intensity expressed with arbitrary units (AUs). Slides were incubated in Harris hematoxylin (Casa To further identify and illustrate signs of mast cell and Alvarez, Madrid, Spain) for 4 minutes. After brief rinsing in tap eosinophil activation, biopsies were processed for transmission water, slides were immersed in 1% hydrochloric alcohol for 10 electron microscopy (TEM). Biopsies from all groups were fixed seconds, re-immersed in tap water, and immersed in tap water in 2.5% glutaraldehyde in 0.1 M phosphate buffer for 24 to 72 with 2 to 3 ammonia drops for 1 minute. Finally, slides were hours, postfixed, dehydrated, and further processed according 976 Inflamm Bowel Dis • Volume 26, Number 7, July 2020 Increased Colonic Epithelial Permeability and Mucosal Eosinophilia to standard protocols for TEM. Representative photographs were classified in the IBS-M subgroup and had moderate-severe of mast cells and eosinophils were achieved using a Hitachi symptoms with median symptom severity score (SSS)32 of 343 H-1400 electron microscope. (range 167–462), and none of the patients related the onset of their symptoms to infectious gastroenteritis. Measurements Statistical Analysis of plasma-C-reactive protein (p-CRP) showed that all UC pa- Data were analyzed using the GraphPad Prism 7 Software tients had p-CRP <10 mg/mL (median 2.1, range 0.2–7.1) when (GraphPad Software Inc., CA, USA). All data, except electro- biopsies were achieved. Analyses showed that biopsy location physiology data, were confirmed as normally distributed and (rectum/rectosigmoidal junction/sigmoideum) had no signifi- given as mean ±SEM. Comparisons between these groups were cant impact on any of the results. done with Student t or ANOVA test. Electrophysiology data (TER) was given as median (25th to 75th percentile), and com- Decreased TER in UC and IBS Compared parisons between groups were done with the Mann-Whitney U With HCs test when there were 2 groups or the Kruskal-Wallis test fol- Ussing chamber experiments showed a stable PD after lowed by the Dunn multiple comparisons test when there were equilibration in all biopsies, indicating a good tissue viability multiple groups. Differences with P < 0.05 were considered throughout the experiments (data not shown) and ensuring that significant. Correlation testing was done with 2-tailed Pearson differences observed were not due to differences in viability be- correlation test. Influence on the results of different parameters tween biopsies. related to patient’s characteristics (as indicated in Table 1) was Ussing chamber experiments revealed a lower TER in bi- analyzed with the Mann-Whitney U and Spearman correlation opsies from both UC and IBS patients compared with HCs al- tests (continuous variable and stratification). ready at time 0 (P < 0.05; Fig. 1A). Transepithelial resistance remained lower after both 60 and 120 minutes from start, both in UC (60 min, 22.1 [18.2–26.1] ohm × cm2, P < 0.05; 120 min, RESULTS 20.1 [16.9–23.0], P < 0.005) and IBS (60 min, 21.8 [18.2–27.2], P < 0.05; 120 min, 20.6 [17.7–26.5], P < 0.05) compared with HCs Patients and HCs (60 min, 27.0 [23.0–31.9]; 120 min, 25.4 [23.3–29.7]). Changes of The material included 13 patients with inactive UC, TER over time, 0 to 120 minutes, are shown in Figs. 1B–D. 15 patients with IBS-M, and 15 HCs (Table 1). All UC pa- tients were in clinical and endoscopic remission (Full Mayo score ≤2).31 At the day of endoscopy, UC patients graded the Increased Paracellular Permeability in UC presence of abdominal pain for the previous day, (none, mild, and IBS moderate, or severe). Four patients reported mild abdominal Permeability to 51Cr-EDTA over time (60–120 min) was pain, and 1 reported moderate abdominal pain. All IBS patients higher in both UC (P < 0.0005) and IBS (P < 0.05) compared TABLE 1. Patient Characteristics of the 13 Patients with UC in Remission and 15 Patients with IBS-mixed and 15 Healthy Controls Included in the Study UC IBS HCs Age, median (range) 36 (21–72) years 31 (19–52) years 27 (21–61) years Sex 7 women, 6 men 15 women 11 women, 4 men Biopsy location Sigmoideum (n = 7) Sigmoideum (n = 15) Sigmoideum (n = 10) Rectum (n = 6) Rectosigmoidal junction (n = 5) Endoscopic Mayo score 0 (n = 12) 0 (n = 15) 0 (n = 15) 1 (n = 1) Anti-inflammatory medication 5-ASA (n = 12) None None Steroids (n = 1) Infliximab (n = 1) Azathioprine (n = 3) Allergy/smoking Pollen (n = 2) None None Pollen & fur (n = 1) Smoking (n = 2) Biopsies were taken in October-December that is outside pollen season in Sweden. 977 Katinios  et al Inflamm Bowel Dis • Volume 26, Number 7, July 2020 FIGURE 1. A, Transepithelial resistance (TER) measured at start, 0 min, in colonic biopsies from patients with ulcerative colitis in remission, irritable bowel syndrome, and healthy controls. Bars represent median (25th–75th percentile). B–D, Changes of TER over time in UC (B), IBS (C), and HCs (D). Graphs show median TER values and the variability at each time point (25th–75th percentile). Comparisons between 2 groups were done with Mann-Whitney U test. *P < 0.05. FIGURE 3. Concentrations of TNF-α in plasma from patients with ulcer- FIGURE 2. Paracellular permeability to 51Chromium (Cr)-EDTA in colonic ative colitis in remission, irritable bowel syndrome, and healthy con- biopsies from patients with ulcerative colitis in remission, irritable trols. Bars represent median (25th–75th percentile), and comparisons bowel syndrome, and healthy controls. Biopsies were mounted in between 2 groups were done with Mann-Whitney U test. **P < 0.005, Ussing chambers, and permeability to 51Cr-EDTA was measured over ***P < 0.0005. the 60–120 min period after start. Bars represent median (25th–75th percentile), and comparisons between 2 groups were done with Mann- Whitney U test. *P < 0.05, **P < 0.005, ***P < 0.0005. (4.3 cm/s ×10-6) was recorded in the single UC patient who graded the presence of abdominal pain as moderate. In line with HCs. Moreover, paracellular permeability was signifi- with this, there was no correlation between 51Cr-EDTA per- cantly higher in UC biopsies compared with IBS (P < 0.005; meability and IBS SSS (r = 0.15, P = 0.31); however, the 2 Fig. 2). A higher paracellular permeability was not correlated IBS patients displaying the 2 highest SSS values displayed the to a higher grade of abdominal pain in UC patients (r = 0.34, 2 highest 51Cr-EDTA permeability measurements (1.75 and P = 0.27), but notably, the highest permeability of 51Cr-EDTA 1.95 cm/s ×10-6, respectively). 978 Inflamm Bowel Dis • Volume 26, Number 7, July 2020 Increased Colonic Epithelial Permeability and Mucosal Eosinophilia Increased TNF-α in Plasma of UC Patients in with HCs (Fig. 5A). In contrast to mast cells, the number of Remission eosinophils was significantly elevated in UC mucosa compared ELISA revealed higher TNF-α levels in plasma of UC with IBS (P < 0.0005). Representative photographs of cells compared with both IBS (P < 0.005) and HCs (P < 0.0005), identified as eosinophils are shown in Figure 5B–D. Eosinophils whereas there was no significant difference between IBS and were often found just beneath the epithelial cell border; in UC, HCs (Fig. 3). There was no significant correlation between they were often found in clusters, as illustrated in Figure 5B. plasma TNF-α concentrations and 51Cr-EDTA permeability. When comparing microscopy results patient by patient, it was shown that 100% of the UC patients and approximately 69% of the IBS patients had elevated eosinophil numbers compared Increased Number of Mucosal Mast Cells in UC with the mean of HCs. This indicates that a mucosal immune and IBS Compared With HCs activation seems to be present in almost all patients. Colonic biopsies from both UC and IBS patients showed increased number of mast cells compared with HCs (P < 0.05; Fig. 4A). However, there was no difference in the number Decreased intracellular tryptase in mast cells in UC of mast cells between UC (144.7 ± 19.2 cells/mm2) and IBS Intracellular and extracellular tryptase were quanti- (132.1 ± 12.7 cells/mm2). Representative photomicrographs of fied in single mast cells with confocal microscopy in biopsies tryptase-positive cells are shown in Figure 4B–D. from 4 UC patients (mean ±SD, 39.2 ± 9.2 years; 2 women), The majority of both UC (85%) and IBS (85%) had 4 patients with IBS-M (30.8 ± 6.2 years; 4 women) and 4 higher mast cell numbers compared with the mean of HCs. HCs (36 ± 7.9 years; 4 women). Intracellular measurements This indicates recruitment of mast cell progenitors from the cir- showed significant decreased tryptase content in mast cells culation into tissues of almost all IBS and UC patients. of UC patients compared with IBS (P < 0.05), but no differ- ence was found in extracellular tryptase surrounding the cells (Fig. 6A). Increased Number of Mucosal Eosinophils in UC Ultrastructural visual analysis identified signs of activa- Compared with IBS and HCs tion in mast cells from all samples, with higher heterogeneity Eosin staining revealed an increased number of eosino- in granular electrodensity and pseudopodes in the UC and IBS phils in both UC (P < 0.0005) and IBS (P < 0.005) compared groups. Representative images are shown in Figure 6A. FIGURE 4. A, Mucosal mast cells quantified by immunofluorescence FIGURE 5. A, Graph showing mucosal eosinophils quantified by he- in colonic biopsies from patients with ulcerative colitis in remission, matoxylin and eosin staining in colonic biopsies from patients with irritable bowel syndrome, and healthy controls. B–D, Representative ulcerative colitis in remission, irritable bowel syndrome, and healthy photographs of mast cell tryptase-positive cells (red, arrows) in the controls. B–D, Photographs illustrating cells positive for eosin (arrows) colonic mucosa of UC, IBS, and HC, respectively. Blue, nuclei staining in the colonic mucosa of UC, IBS, and HC, respectively. Magnification by DAPI. Magnification 600X. Bars represent median (25th–75th per- 600X. Bars represent median (25th–75th percentile), and comparisons centile), and comparisons between 2 groups were done with Mann- between 2 groups were done with Mann-Whitney U test. **P < 0.005, Whitney U test. **P < 0.005. ***P < 0.0005. 979 Katinios  et al Inflamm Bowel Dis • Volume 26, Number 7, July 2020 FIGURE 6. Mast cell tryptase and eosinophil cationic protein (ECP) distribution in colonic biopsies from patients with ulcerative colitis in re- mission, irritable bowel syndrome, and healthy controls. A, Quantification of intracellular/extracellular tryptase. Representative images of mast cells with identification of plasma cell membrane (white line), and tryptase located at intracellular (red) and extracellular (green) compartments. Ultrastructural analysis identifies partial degranulation of cytoplasmic granules (white arrow) in UC and IBS, indicative of activation, and a more stable phenotype in HC. B, Quantification of intracellular/extracellular ECP. Representative images of eosinophils with identification of plasma cell membrane (white line) and ECP located at intracellular (blue) and extracellular (pink) compartments. Ultrastructural analysis identifies partial de- granulation of some cytoplasmic granules (white arrow) in UC and IBS, indicative of activation, and a more stable phenotype in HC. Quantification was done in 600X confocal images and expressed in arbitrary units. Bars represent median (25th–75th percentile), and comparisons were done with Mann-Whitney U test. *P < 0.05. 980 Inflamm Bowel Dis • Volume 26, Number 7, July 2020 Increased Colonic Epithelial Permeability and Mucosal Eosinophilia Increased extracellular ECP in eosinophils in UC The electrophysiological measurements monitored throughout Intracellular and extracellular ECP were quantified in the Ussing chamber experiments demonstrated a lower TER single or eosinophils (a few clustered) in biopsies from patients in both UC and IBS biopsies as compared with HCs. This dif- with UC, IBS-M, and HC patients as indicated for mast cells. ference was detected from the start of experiments, suggesting Even though intracellular ECP was not significantly different structural changes in the integrity of the paracellular pathways. among groups, measurements showed increased extracellular Ussing experiments further showed an increased paracellular ECP content in UC compared with HCs (P < 0.05; Fig. 6B). permeability in both UC and IBS compared with HCs. This Ultrastructural visual analysis of eosinophils showed an is in line with a previous study,8 showing an increased colonic activated profile, with larger cytoplasmic granules in the UC permeability to orally administrated 51Cr-EDTA measured by group and partial degranulation of some cytoplasmic granules 24-hour urine excretion in both inactive UC and IBS-D com- in UC and IBS groups as compared with HCs. Representative pared with HCs. Interestingly, the increase in permeability was images are shown in Figure 6B. correlated to stool frequency in IBS patients—but not in UC patients. We previously showed25 an increased transcellular passage but an unaltered paracellular permeability in biopsies Increased Paracellular Permeability Correlated from quiescent UC patients. However, these patients were in to Number of Eosinophils deep remission and had not had a relapse in several years, in Correlation analysis of the measured parameters (TER, comparison with the UC patients of the present study. 51 Cr-EDTA permeability, mast cells, eosinophils, and plasma There was a higher number of mucosal mast cells in co- TNF-α concentrations) showed significant associations between lonic biopsies of both UC patients in remission and IBS-M paracellular permeability to 51Cr-EDTA and the number of eo- patients in remission compared with HCs, as revealed by immu- sinophils in all groups: UC (r = 0.60, P < 0.05), IBS (r = 0.60, nofluorescence. The fact that the amount of mast cells seems to P < 0.05), and HCs (r = 0.63, P < 0.05) (Fig. 7A–C). However, follow the same pattern both in inactive UC and IBS could in- no significant correlation was found between 51Cr-EDTA and dicate that mast cells are a contributing factor to the increased the number of mast cells in any of the groups. Interestingly in paracellular permeability and, therefore, to the pathogenesis UC patients, the number of mast cells was negatively correlated of these diseases, as is also shown previously.14 Though, there to the number of eosinophils (r = −0.73, P < 0.05; Fig. 7D). was neither a correlation between 51Cr-EDTA permeability In HCs, there was a correlation between 51Cr-EDTA permea- and mast cell quantity nor a difference in mast cell numbers bility and TNF-α concentrations in plasma (r = 0.66, P < 0.05), between UC and IBS, which indicates that mast cell amounts though no significant correlations could be found in UC or IBS may not be responsible for the increased paracellular perme- patients. Transepithelial resistance did not correlate with any of ability observed in UC compared with IBS. It might be that the other parameters (data not shown). despite an unaltered number, there could be a difference in the granular content or activation status of the mast cells. Even DISCUSSION though our results showed similar release of tryptase in both In this study, we showed that intestinal paracellular per- IBS and UC compared with HCs, it could still be a difference meability was increased in both patients with inactive UC and in granular content. There is a different mucosal environment IBS-M patients and that this increase was more evident in UC that is of higher inflammatory degree in UC compared with patients in remission in comparison with IBS-M patients. In ad- IBS and HCs.18 Mast cells adapt their granular content to ex- dition, both inactive UC and IBS-M patients had significantly ternal factors; therefore, it would be of interest in future studies more mucosal mast cells and eosinophils than HCs, but only the to identify which mediators are differentially produced in each number of eosinophils was significantly higher in UC in remis- entity. Previously, it was shown that the content of mast cells sion compared with IBS-M. Also, a significantly increased level differs not only between healthy controls and IBD but also be- of eosinophil degranulation, identified by higher amounts of tween UC and CD patients.34 Furthermore, Lloyd et al showed extracellular ECP, could only be observed in UC. Furthermore, already in 1975 that there was a marked degranulation of intes- there was a significant positive correlation between eosinophil tinal mast cells in CD compared with healthy controls.35 This numbers and paracellular permeability to 51Cr-EDTA in all was confirmed later on by Dvorak et al who described in more groups. detail, using transmission electron microscopy, an enhanced de- There is a realization that among other factors such as granulation of intestinal mast cells in patients with both CD36 low-grade inflammation and genetic susceptibility, increased in- and UC.37 Other studies showed similar results in both CD and testinal permeability also contributes to the pathophysiology of UC, using antibody specific immunohistochemistry against both IBS5–7 and UC.2–4 Moreover, it is known that a great per- human mast cell markers tryptase and chymase.38 centage of patients with quiescent UC present IBS symptoms.33 In the present study, histological staining with eosin re- Our findings suggest that increased paracellular permeability vealed significantly more eosinophils in both inactive UC and might be a pathophysiological feature typical for both diseases. IBS-M colonic biopsies compared with HCs. Interestingly, the 981 Katinios  et al Inflamm Bowel Dis • Volume 26, Number 7, July 2020 FIGURE 7. A–C, The correlation between eosinophil numbers and paracellular permeability to 51Chromium (Cr)-EDTA in colonic biopsies from pa- tients with inactive ulcerative colitis (A), patients with irritable bowel syndrome (B), and healthy controls (C). The correlation between the number of eosinophils and mast cells in colonic biopsies from ulcerative colitis patients is illustrated (D). Correlation testing was done with 2-tailed Pearson correlation test. number of eosinophils was significantly elevated in UC com- them to enhance their response to eotaxin.40 Pro-inflammatory pared with IBS, and degranulation analysis revealed signif- cytokines are known to be elevated during active disease com- icantly more degranulated eosinophils in UC compared with pared with inactive disease; 41 however, during both active and both IBS-M and HCs. Moreover, when looking at the data, it inactive phases of UC, significant levels of eotaxin has been was indicated that the number of eosinophils seemed to corre- reported in serum.42 It might be that, compared with IBS, there late to the extracellular ECP levels in UC patients. Even though is a different cytokine milieu in the mucosa of patients with there were too few subjects to draw any conclusion, this pre- UC—even during remission—with somewhat enhanced levels liminary observation might suggest that the function of ECP of cytokines promoting the recruitment of eosinophils. The is the key to the role of eosinophils. In line with this, Bystrom eotaxin CCL11 and IL-5 were increased in both active and et al39 previously reported that ECP levels in patients with other inactive UC, and CCL11 was positively correlated to eosino- conditions such as asthma, atopic dermatitis, lung disease, and phil numbers,43 supporting its contribution to eosinophil re- various autoimmune conditions robustly correlated with the cruitment. Moreover, Carlson et al44 showed increased release number of eosinophils present. of eosinophil granule proteins in the noninflamed mucosa of The reason for the increased number of eosinophils might patients with UC compared with controls, and Lampinen et al be due to a higher recruitment of eosinophils during UC com- showed more activated eosinophils in the colon of inactive UC pared with IBS. During active bowel inflammation, eosinophils patients compared with active disease patients.23 Further, the are known to migrate to the GI tract in response to chemokines early activation marker CD69 showed to be more expressed in (eg, RANTES and eotaxin, both constitutively expressed inactive UC compared with active,45 and in addition, monocyte throughout the gut). Eotaxin binds to the CCR-3 receptor on chemotactic factor‐4 and IL-33, known to activate human eo- eosinophils, which is required for their homing to the GI tract.40 sinophils, were increased compared with controls in both active It is, however, known that other cytokines are needed as well, and inactive UC.45 These findings indicate eosinophil activity in among them interleukin (IL)-4, IL-5, IL-13 and TNF-α, which the noninflamed colon, which could refer to the cytokines in the all increase the circulating number of eosinophils and prime microenvironment. In addition, there might be other stimulus 982 Inflamm Bowel Dis • Volume 26, Number 7, July 2020 Increased Colonic Epithelial Permeability and Mucosal Eosinophilia of eosinophil recruitment and activation present during remis- proportional increase of mucosal eosinophils is not known, al- sion; however, this needs to be further investigated. though several cases of 5-ASA–induced eosinophilic disorders An interesting observation was that eosinophils ap- have been described.53 The specific increase in eosinophils seen peared to concentrate just beneath the epithelium, and in UC, in UC is probably related to other factors, as well. For example, they were often found in clusters. In contrast, mast cells were it could be speculated that 5-ASA can have effects on cell mat- more seldom found this close to the epithelial cell lining—and uration and progenitors. In contrast to our findings, in asthma especially not in UC. One might speculate that the proximity and eosinophilic esophagitis, the levels of eosinophils and of clustered eosinophils and the epithelial barrier could con- mast cells generally mirror each other.54 However, 5-ASA treat- tribute to increased permeability in UC. Eosinophil quantity ment is generally not used in these conditions. In addition, our was positively correlated to paracellular permeability in all finding of a negative correlation between eosinophils and mast groups, which is noteworthy because no correlation could be cells in UC might seem contradictory to our previous data25 re- found between 51Cr-EDTA permeability and mast cells. This garding eosinophils secreting CRH, which activates mast cells might be interpreted at first that the eosinophil number could to secrete mediators affecting the barrier. Our degranulation have a higher impact on paracellular permeability than mast analyses showed no difference in extracellular tryptase in UC. cells in the colonic mucosa. However, it could also be inferred This could either indicate that eosinophils have a more promi- that the increased amount and/or activation of eosinophils nent role compared with mast cells in UC barrier dysfunction give rise to an enhanced secretion of corticotropin-releasing or that the 5-ASA treatment drastically downregulates mast cell hormone (CRH), for example.24 As we previously described,25 activation levels. Further studies are needed to understand the CRH can activate mast cells, and upon activation, mast cells mechanism behind this. start secreting a wide variety of bioactive mediators such as Our findings of significantly higher TNF-α levels in UC TNF-α, IL-3, IL-4, IL-5, IL-16, and IFNγ. Several of these me- plasma compared with IBS and HC—and comparable TNF-α diators have effects on mucosal barrier function (eg, chymase, levels in IBS and HCs—are consistent with previous studies55 tryptase,46 IFNγ, and TNF-α).47 In this study, TNF-α levels and in line with common knowledge that IBS does not seem were higher in plasma of UC patients compared with both IBS to be a TNF-driven disease, in contrast to UC. There was no and HC patients. This, together with the correlation observed correlation between TNF-α levels in plasma and paracellular between paracellular permeability and eosinophil numbers, permeability in UC or IBS patients; however, a correlation was suggests an indirect involvement of mast cells, even though the found in HCs. This might simply be explained by the fact that effect on the barrier might be due to activation state rather than the levels of 51Cr-EDTA permeability, and especially TNF-α, amount of mast cells. are overall low in healthy controls which might give rise to the One of the HCs had a very high number of eosinophils, observed correlation between them. 146 cells/mm2, in comparison with the mean of HCs that was In conclusion, our findings indicate a more permeable 25.8 cells/mm2. This HC did not report any allergies, use of mucosa in patients with inactive UC and patients with mod- nonsteroidal anti-inflammatory drugs, or any other known erate to severe IBS compared with HCs. Ulcerative colitis pa- diseases that could otherwise explain the enhanced eosinophil tients, even during remission, display a leakier mucosal barrier count.48 However, there could be a nonsymptomatic parasite compared with IBS, where eosinophils might be involved. infection49 or disease (eg, vasculitis),50 where eosinophils are in- Results further suggest a generalizable association between eo- volved, of which the HC was not aware. Colonic eosinophilia sinophils and paracellular permeability; however, because of is known to be caused by conditions such as allergies, parasitic the cross-sectional design of the study, we were not able to dem- infections, or drugs, and it is also a hallmark of IBD and eosin- onstrate causality. This was also true for patients with IBS and ophilic colitis.50, 51 In the present study, the presence or absence HCs. The present results contribute to a better understanding of allergies did not have any impact on any of the results; how- of colonic paracellular permeability in patients with UC in re- ever, we do not have enough data to confidently exclude other mission and IBS, though more studies are needed to confirm conditions, as mentioned previously, that could influence the this in a larger material. results as far as colon eosinophils are concerned. Surprisingly, we found that the number of mast cells ACKNOWLEDGEMENTS was negatively correlated to the number of eosinophils in UC The authors thank all the patients for participating in in remission. This could be partly explained by ongoing treat- this study. They also thank Mr. Martin Winberg, Linköping ment with aminosalicylate (5-ASA) in the UC patients, as it has University, for assistance in the Ussing laboratory and Mrs. been shown (at least in IBS patients) that this treatment causes Lena Svensson, Linköping University, for the embedding and downregulation of mast cells.52 To our knowledge, data on the sectioning of samples. Further, they acknowledge the contribu- effect of 5-ASA on mast cells in UC are lacking; however, it tion of Dr. Alejandro Sánchez-Chardi and Francisca Cardoso, seems reasonable to extrapolate the results of the previous study Universitat Autònoma de Barcelona, for technical assistance also to UC patients in remission. Whether 5-ASA could cause and excellent performance in electron microscopy procedures. 983 Katinios  et al Inflamm Bowel Dis • Volume 26, Number 7, July 2020 REFERENCES 27. De Silva AP, Nandasiri SD, Hewavisenthi J, et al. Subclinical mucosal inflamma- tion in diarrhea-predominant irritable bowel syndrome (IBS) in a tropical setting. 1. Mearin F, Lacy BE, Chang L, et al. Bowel disorders. Gastroenterology. Scand J Gastroenterol. 2012;47:619–624. 2016;150:1393–1407. 28. O’Sullivan M, Clayton N, Breslin NP, et al. Increased mast cells in the irritable 2. Jenkins RT, Ramage JK, Jones DB, et al. Small bowel and colonic permeability to bowel syndrome. Neurogastroenterol Motil. 2000;12:449–457. 51Cr-EDTA in patients with active inflammatory bowel disease. Clin Invest Med. 29. Longstreth GF, Thompson WG, Chey WD, et al. Functional bowel disorders. 1988;11:151–155. Gastroenterology. 2006;130:1480–1491. 3. Arslan G, Atasever T, Cindoruk M, et al. (51)CrEDTA colonic permeability 30. Keita AV, Gullberg E, Ericson AC, et al. Characterization of antigen and bacte- and therapy response in patients with ulcerative colitis. Nucl Med Commun. rial transport in the follicle-associated epithelium of human ileum. Lab Invest. 2001;22:997–1001. 2006;86:504–516. 4. Teahon K, Somasundaram S, Smith T, et al. Assessing the site of increased 31. Schroeder KW, Tremaine WJ, Ilstrup DM. Coated oral 5-aminosalicylic acid intestinal permeability in coeliac and inflammatory bowel disease. Gut. therapy for mildly to moderately active ulcerative colitis. A randomized study. N 1996;38:864–869. Engl J Med. 1987;317:1625–1629. 5. Bednarska O, Walter SA, Casado-Bedmar M, et al. Vasoactive intestinal polypep- 32. Francis CY, Morris J, Whorwell PJ. The irritable bowel severity scoring system: a tide and mast cells regulate increased passage of colonic bacteria in patients with simple method of monitoring irritable bowel syndrome and its progress. Aliment irritable bowel syndrome. Gastroenterology. 2017;153:948–960.e3. Pharmacol Ther. 1997;11:395–402. 6. Dunlop SP, Hebden J, Campbell E, et al. Abnormal intestinal permeability in sub- 33. Fukuba N, Ishihara S, Tada Y, et al. Prevalence of irritable bowel syndrome-like groups of diarrhea-predominant irritable bowel syndromes. Am J Gastroenterol. symptoms in ulcerative colitis patients with clinical and endoscopic evidence of 2006;101:1288–1294. remission: prospective multicenter study. Scand J Gastroenterol. 2014;49:674–680. 7. Piche T, Barbara G, Aubert P, et al. Impaired intestinal barrier integrity in the 34. Sasaki Y, Tanaka M, Kudo H. Differentiation between ulcerative colitis and colon of patients with irritable bowel syndrome: involvement of soluble medi- Crohn’s disease by a quantitative immunohistochemical evaluation of T lympho- ators. Gut. 2009;58:196–201. cytes, neutrophils, histiocytes and mast cells. Pathol Int. 2002;52:277–285. 8. Gecse K, Róka R, Séra T, et al. Leaky gut in patients with diarrhea- 35. Lloyd G, Green FH, Fox H, et al. Mast cells and immunoglobulin E in inflamma- predominant irritable bowel syndrome and inactive ulcerative colitis. Digestion. tory bowel disease. Gut. 1975;16:861–865. 2012;85:40–46. 36. Dvorak AM, Monahan RA, Osage JE, et al. Crohn’s disease: transmission electron 9. Vivinus-Nébot M, Frin-Mathy G, Bzioueche H, et al. Functional bowel symp- microscopic studies. II. Immunologic inflammatory response. Alterations of mast toms in quiescent inflammatory bowel diseases: role of epithelial barrier disrup- cells, basophils, eosinophils, and the microvasculature. Hum Pathol. 1980;11:606–619. tion and low-grade inflammation. Gut. 2014;63:744–752. 37. Dvorak AM, McLeod RS, Onderdonk A, et al. Ultrastructural evidence for 10. Jonefjäll B, Öhman L, Simrén M, et al. IBS-like symptoms in patients with piecemeal and anaphylactic degranulation of human gut mucosal mast cells in ulcerative colitis in deep remission are associated with increased levels of vivo. Int Arch Allergy Immunol. 1992;99:74–83. serum cytokines and poor psychological well-being. Inflamm Bowel Dis. 38. Bischoff SC, Wedemeyer J, Herrmann A, et al. Quantitative assessment of intes- 2016;22:2630–2640. tinal eosinophils and mast cells in inflammatory bowel disease. Histopathology. 11. Ohman L, Isaksson S, Lindmark AC, et al. T-cell activation in patients with irri- 1996;28:1–13. table bowel syndrome. Am J Gastroenterol. 2009;104:1205–1212. 39. Bystrom J, Amin K, Bishop-Bailey D. Analysing the eosinophil cationic protein–a 12. An S, Zong G, Wang Z, et al. Expression of inducible nitric oxide synthase in clue to the function of the eosinophil granulocyte. Respir Res. 2011;12:10. mast cells contributes to the regulation of inflammatory cytokines in irritable 40. Filippone RT, Sahakian L, Apostolopoulos V, et al. Eosinophils in inflammatory bowel syndrome with diarrhea. Neurogastroenterol Motil. 2016;28:1083–1093. bowel disease. Inflamm Bowel Dis. 2019;25:1140–1151. 13. Ahn JY, Lee KH, Choi CH, et al. Colonic mucosal immune activity in irritable 41. Inoue S, Matsumoto T, Iida M, et al. Characterization of cytokine expression bowel syndrome: comparison with healthy controls and patients with ulcerative in the rectal mucosa of ulcerative colitis: correlation with disease activity. Am J colitis. Dig Dis Sci. 2014;59:1001–1011. Gastroenterol. 1999;94:2441–2446. 14. De Winter BY, van den Wijngaard RM, de Jonge WJ. Intestinal mast cells in gut 42. Mir A, Minguez M, Tatay J, et al. Elevated serum eotaxin levels in patients with inflammation and motility disturbances. Biochim Biophys Acta. 2012;1822:66–73. inflammatory bowel disease. Am J Gastroenterol. 2002;97:1452–1457. 15. He SH. Key role of mast cells and their major secretory products in inflammatory 43. Lampinen M, Waddell A, Ahrens R, et al. CD14+CD33+ myeloid cell-CCL11- bowel disease. World J Gastroenterol. 2004;10:309–318. eosinophil signature in ulcerative colitis. J Leukoc Biol. 2013;94:1061–1070. 16. Casado-Bedmar M, Heil SDS, Myrelid P, et al. Upregulation of intestinal mu- 44. Carlson M, Raab Y, Peterson C, et al. Increased intraluminal release of eosin- cosal mast cells expressing VPAC1 in close proximity to vasoactive intestinal pol- ophil granule proteins EPO, ECP, EPX, and cytokines in ulcerative colitis and ypeptide in inflammatory bowel disease and murine colitis. Neurogastroenterol proctitis in segmental perfusion. Am J Gastroenterol. 1999;94:1876–1883. Motil. 2019;31:e13503. 45. Lampinen M, Fredricsson A, Vessby J, et al. Downregulated eosinophil activity in 17. Guilarte M, Santos J, de Torres I, et al. Diarrhoea-predominant IBS patients ulcerative colitis with concomitant primary sclerosing cholangitis. J Leukoc Biol. show mast cell activation and hyperplasia in the jejunum. Gut. 2007;56:203–209. 2018;104:173–183. 18. Cremon C, Gargano L, Morselli-Labate AM, et al. Mucosal immune activation 46. Wilcz-Villega EM, McClean S, et al. Mast cell tryptase reduces junctional adhe- in irritable bowel syndrome: gender-dependence and association with digestive sion molecule-A (JAM-A) expression in intestinal epithelial cells: implications symptoms. Am J Gastroenterol. 2009;104:392–400. for the mechanisms of barrier dysfunction in irritable bowel syndrome. Am J 19. Neunlist M, Toumi F, Oreschkova T, et al. Human ENS regulates the in- Gastroenterol. 2013;108:1140–1151. testinal epithelial barrier permeability and a tight junction-associated pro- 47. Li Q, Zhang Q, Wang M, et al. Interferon-gamma and tumor necrosis factor- tein ZO-1 via VIPergic pathways. Am J Physiol Gastrointest Liver Physiol. alpha disrupt epithelial barrier function by altering lipid composition in mem- 2003;285:G1028–G1036. brane microdomains of tight junction. Clin Immunol. 2008;126:67–80. 20. Bakker R, Dekker K, De Jonge HR, et al. VIP, serotonin, and epinephrine mod- 48. Jiménez-Sáenz M, González-Cámpora R, Linares-Santiago E, et al. Bleeding co- ulate the ion selectivity of tight junctions of goldfish intestine. Am J Physiol. lonic ulcer and eosinophilic colitis: a rare complication of nonsteroidal anti-in- 1993;264:R362–R368. flammatory drugs. J Clin Gastroenterol. 2006;40:84–85. 21. Conlin VS, Wu X, Nguyen C, et al. Vasoactive intestinal peptide ameliorates in- 49. Al Samman M, Haque S, Long JD. Strongyloidiasis colitis: a case report and re- testinal barrier disruption associated with Citrobacter rodentium-induced colitis. view of the literature. J Clin Gastroenterol. 1999;28:77–80. Am J Physiol Gastrointest Liver Physiol. 2009;297:G735–G750. 50. Alfadda AA, Storr MA, Shaffer EA. Eosinophilic colitis: epidemiology, clinical 22. Winterkamp S, Raithel M, Hahn EG. Secretion and tissue content of eosinophil features, and current management. Therap Adv Gastroenterol. 2011;4:301–309. cationic protein in Crohn’s disease. J Clin Gastroenterol. 2000;30:170–175. 51. Lucendo AJ. Cellular and molecular immunological mechanisms in eosino- 23. Lampinen M, Rönnblom A, Amin K, et al. Eosinophil granulocytes are activated philic esophagitis: an updated overview of their clinical implications. Expert Rev during the remission phase of ulcerative colitis. Gut. 2005;54:1714–1720. Gastroenterol Hepatol. 2014;8:669–685. 24. Zheng PY, Feng BS, Oluwole C, et al. Psychological stress induces eosino- 52. Corinaldesi R, Stanghellini V, Cremon C, et al. Effect of mesalazine on mucosal phils to produce corticotrophin releasing hormone in the intestine. Gut. immune biomarkers in irritable bowel syndrome: a randomized controlled proof- 2009;58:1473–1479. of-concept study. Aliment Pharmacol Ther. 2009;30:245–252. 25. Wallon C, Persborn M, Jönsson M, et al. Eosinophils express muscarinic recep- 53. Bartal C, Sagy I, Barski L. Drug-induced eosinophilic pneumonia: a review of tors and corticotropin-releasing factor to disrupt the mucosal barrier in ulcerative 196 case reports. Medicine (Baltimore). 2018;97:e9688. colitis. Gastroenterology. 2011;140:1597–1607. 54. Galdiero MR, Varricchi G, Seaf M, et al. Bidirectional mast cell-eosinophil inter- 26. Park KS, Ahn SH, Hwang JS, et al. A survey about irritable bowel syndrome in actions in inflammatory disorders and cancer. Front Med (Lausanne). 2017;4:103. South Korea: prevalence and observable organic abnormalities in IBS patients. 55. Chang L, Adeyemo M, Karagiannides I, et al. Serum and colonic mucosal im- Dig Dis Sci. 2008;53:704–711. mune markers in irritable bowel syndrome. Am J Gastroenterol. 2012;107:262–272. 984