Introduction

Liver cirrhosis and its complications were reported to render a million deaths annually worldwide [1]. Infection, one of the leading causes of hospitalization and death in patients with cirrhosis [2], is considered due to increased intestinal permeability and gut microbial translocation [2,3,4]. Therefore, epithelial barrier dysfunction in cirrhosis has been attracted a lot of concerns, but studies focused on mucus barrier is still limited [5, 6].

Mucus, which covers the intestinal epithelia, is known as the first line of defense for mucosal surface [7]. Physiologically, intestinal mucus may trap bacteria and further hamper bacteria by antibacterial peptides and proteins within [8]. Some experiments have shown that animals developed spontaneous colitis after knocking out Muc2, one of the crucial components in mucus [9], and became susceptible to infection [10, 11], highlighting the role of mucus in maintaining normal function of intestine. Intestinal mucus is mainly secreted and stored by goblet cells, which are differentiated from intestinal progenitors. The balance between differentiation and proliferation of progenitor cells, which are important to intestinal epithelial homeostasis, is crucial for the development of goblet cell [12, 13] and is regulated by Notch1 signaling [14]. Notch1 is a transmembrane receptor. It may activate downstream target genes, such as Hes1 after binding to its ligand, such as Dll1 [15, 16] by shifting its intracellular domain, NICD to nucleus [17]. Hes1 has been taken as an inhibitor for ATOH1, SPDEF, and KLF4, which are required for the differentiation of goblet cells [14, 16]. There are some data showing that down-regulation of Notch1 activity and Hes1 may lead to secretory cell hyperplasia [14, 16, 18, 19]. When Notch1 is activated in the inflammatory situation such as ulcerative colitis, depletion of goblet cells with impaired mucus is observed [20,21,22].

This study was aimed to observe whether intestinal mucus barrier would be impaired in the cirrhotic rats induced with thioacetamide (TAA), to describe the imbalance between proliferation and differentiation of progenitor of goblet cells, and to evaluate the activation of Notch1 signaling in vivo. Notch1 signaling as a potential mechanism behind the poorly differentiated goblet cells was also addressed in vitro.

Methods

Animal Model of Liver Cirrhosis

Twenty male Sprague-Dawley rats weighing 200 to 250 g were obtained from Experimental Animal Center of Sichuan University (Chengdu, China). The animal experimentation was approved by institutional ethics committees and was conducted according to regulations of Sichuan University. Rats were randomly allocated to control and TAA (Sigma-Aldrich, St. Louis, MO, USA) group with 10 in each. Control group received peritoneal injection of normal saline. Liver cirrhosis was induced in TAA group by peritoneal injection of TAA for 16 weeks (200 mg/kg/3 days for initial 8 weeks and 100 mg/kg/3 days for the following 8 weeks).

Histopathological Evaluation of Tissues

Tissues were dissected immediately after kill. Livers were fixed by 4% polyformaldehyde, embedded by paraffin, sectioned (thickness of 5 μm), and stained with hematoxylin and eosin (H&E) or Sirius red. Ishak fibrosis stage (with the Ishak scale ranging from 0 = no fibrosis to 6 = cirrhosis) was evaluated for each section after Sirius red staining at × 100 magnifications. Five randomly selected fields were analyzed per rat.

Ilea were fixed by 4% polyformaldehyde or Carnoy’s fluid (ethanol–acetic acid, 3:1). Fixed ilea were then embedded by paraffin and sectioned (thickness of 4 μm). Then, ilea were stained with H&E or periodic acid–Schiff (PAS, G1280, Solarbio, Beijing, China). Villi length was measured. The ileal goblet cells were quantified by visualization and counting PAS-positive theca along with villus–crypt axis according to a previously reported method [23]. The perimeters of corresponding villi were measured, and goblet cell number per perimeter of villi was calculated for comparison. At least eight microscopic fields at × 400 magnifications were analyzed for each section of individual rat. Theca area of goblet cell was measured by ImagePro Plus 6.0 (Media Cybernetics, Rockville, MD, USA) after PAS staining according to a previously reported method [23]. The average optic density (AOD) of mucus staining in villus or crypt was measured by ImagePro Plus 6.0 (Media Cybernetics, Rockville, MD, USA) after PAS staining. At least three microscopic fields at × 400 magnifications of villus or crypt were selected per animal. AOD was calculated by dividing integrated optical density (sum) by area (sum). The results of AOD were shown as fold changes to the control group.

Immunohistochemistry (IHC) Staining

The sections were routinely deparaffinized in xylene and rehydrated with graded ethanol dilutions. Heat-induced antigen retrieval was performed under high pressure in sodium citrate buffer (10 mM, pH = 6.0) for 30 min. After blocked with hydrogen peroxide, the sections were incubated with 10% goat serum for 1 h at 37 °C and subsequent primary antibodies (Table 1) overnight at 4 °C followed by incubation with a detection kit (ZSGB-BIO, Beijing, China) at 37 °C for 30 min. Finally, the sections were stained with diaminobenzoate (DAB, ZSGB-BIO) and counterstained with hematoxylin. Negative control slides without adding primary antibody were included in all assays. The images of sections were captured by an optical microscope (CX41, Olympus, Tokyo, Japan) equipped with camera (DP72, Olympus). The AOD of Ki-67 staining in crypts of rat ileum was measured by ImagePro Plus 6.0 (Media Cybernetics, Rockville, MD, USA).

Table 1 Antibodies list for immunostain and western blot
Full size table

Fluorescence In Situ Hybridization (FISH)

After fixation by Carnoy’s fluid, paraffin-embedded sections of ileum were immersed in xylenes for 10 min (twice) and then in 100% ethanol for 5 min, 85% ethanol for 3 min, and 70% ethanol for 3 min. After washed with PBS with 0.1% DEPC for 10 min (twice), the slides were treated with lysozyme for 15 min at room temperature. The slides were washed with PBS with 0.1% DEPC for 10 min and blocked with 3% BSA for 2 h afterward. Then, hybridization was performed with bacterial 16S rRNA probe EUB 338 (TRITC-conjugated) at 40 °C overnight. The sections were washed with a wash solution (2 × SSC/0.1% Tween 20) for 15 min (three times). The slides were immersed in 75% ethanol for 2 min, 100% ethanol for 2 min, and air-dried for 20 min before counterstained by 4′,6-diamidino-2-phenylindole (DAPI) and covered with coverslip. The images of sections were captured by an optical microscope (CX41, Olympus, Tokyo, Japan) equipped with camera (DP72, Olympus).

Quantitative Real-Time PCR (qRT-PCR)

RNA was extracted from frozen ilea by TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Reverse transcription was performed by a cDNA kit (Invitrogen, Carlsbad, CA, USA). qRT-PCR was performed by using a master mix (Bio-Rad, Hercules, CA, USA). Primers are listed in Table 2. All reactions were run in triplicate by CFX96 real-time PCR detection system (Bio-Rad). The expressions of mRNA were normalized to GAPDH by using 2−ΔΔCt method and were shown as fold changes to the control group.

Table 2 List of primers for qRT-PCR
Full size table

ELISA

Blood from the portal vein was collected. Serum was obtained by centrifugation at 3000 rpm for 15 min. Lipopolysaccharide (LPS) in serum and protein level of MUC2 in rat ileum were determined by ELISA kits according to instruction (USCN Life Sciences, Wuhan, China).

Western Blot Analysis

The whole proteins from frozen ilea or cells were extracted by using RIPA buffer (Beyotime Biotechnology, Shanghai, China) on ice. Equal amount of proteins, 50 μg (except for detecting Notch1 intracellular domain, NICD) or 100 μg (only for detecting NICD) from tissues or 30 μg from cell samples, was resolved by 10% SDS-PAGE and transferred to PVDF membrane (Millipore, Billerica, MA, USA). Then, the membrane was blocked with 5% nonfat dry milk dissolved in TBST and was subsequently incubated with corresponding primary antibodies (Table 1) overnight at 4 °C. Blots were washed in TBST afterward and incubated with appropriate secondary antibodies (Santa Cruz Biotechnology, TX, USA) for 2 h with agitation at room temperature. Bands were visualized by using a western blot detection kit (Santa Cruz Biotechnology, TX, USA). After exposing the membranes to Kodak XAR film (Eastman Kodak, Rochester, NY, USA) for up to 2 h, the autoradiographs were scanned. Quantity One software 4.6.2 (Bio-Rad, Hercules, CA, USA) was used to measure intensity of the bands. Protein expression levels were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The results were shown as fold changes to the control group.

Cell Culture and Treatments

Human colon cancer cell line Ls174T with goblet cell phenotype [13] was obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were maintained in low-glucose Dulbecco’s modified Eagle’s medium (DMEM, HyClone, Logan, UT, USA) containing 10% fetal bovine serum (FBS, Biological Industries, Cromwell, CT, USA) with 100 U/mL penicillin and 100 μg/mL streptomycin (HyClone). Cells were incubated in a humidified atmosphere at 37 °C with 5% CO2 in air.

In order to extract protein, the cells were seeded in six-well plate to density of 60% confluence. For dibenzazepine (DBZ) treatment (inhibiting Notch1), DBZ dissolved in dimethyl sulfoxide (DMSO) of different concentrations (final concentration of DMSO was 0.1%) was added to serum-free culture medium for 48 h. For EDTA treatment, after removing culture medium and washing once in PBS, 5 mM EDTA dissolved in PBS (adjusting pH to 8.0) was used to treat cells for 15 min before protein extraction. For transfection of NICD plasmid (bought from Hanbio, Shanghai, China; for sequence information, see supplementary Fig. 1), Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA) was used according to instruction. 48 h after transfection, the protein was extracted.

Fig. 1
figure 1

Impaired mucus barrier in cirrhotic rats induced by thioacetamide (TAA). Rats in TAA group showed extensive liver cirrhosis in gross, which was presented with Sirius Red staining (a). Ishak fibrosis score was significantly increased in TAA group (b, n = 9 in each group). Mucus was visualized by periodic acid–Schiff (PAS) staining in Carnoy’s fluid-fixed ileal tissues (c). The average optic density (AOD) of PAS staining in villus (d, n = 6 in each group) and crypt (e, n = 8 in each group) was lower in TAA group. mRNA expressions of mucus components, Muc2 (f, n = 8 in control group, n = 7 in TAA group) and TFF3 (g, n = 8 in each group), were lower in TAA group. MUC2 protein of ileum was lower in TAA group (h, n = 8 in control group, n = 9 in TAA group). Lipopolysaccharide (LPS) of serum from portal vein was higher in TAA group (i, n = 6 in each group). Bacteria (red spot in green outlines) were located near to the intestinal crypts in TAA group by using fluorescence in situ hybridization (FISH), and counterstain of nuclei with 4′,6-diamidino-2-phenylindole (blue) was performed (j). *p < 0.05

Full size image

For PAS staining, cells were plated on glass chamber slides. DBZ treatment and transfection were the same as mentioned above. After EDTA treatment, EDTA solution was replaced with culture medium for additional 6 h before PAS staining. After brief washing by PBS solution for three times, Carnoy’s fluid was used for fixation for 15 min. PAS staining was performed afterward according to manufacturer’s instructions (G1280, Solarbio, Beijing, China).

CCK8 Assay

The cells were seeded in 96-well plate to density of 50% confluence. DBZ treatment and transfection was the same as mentioned above. Then the CCK8 assay was performed according to manufacturer’s instruction (Dojindo, Kumamoto, Japan). The final optical density was measured at 450 nm by a Thermo microplate reader (Thermo Fisher Scientific, Waltham, MA, USA).

Immunofluorescent (IF) Staining

Cells were plated on glass chamber slides. DBZ treatment and transfection was the same as mentioned above. After brief washing by PBS solution for three times, 4% PFA was used for fixation for 15 min. After permeabilization with 0.2% triton X100 in PBS for 15 min, the sections were incubated with 10% goat serum for 1 h at 37 °C and subsequent primary antibodies (Table 1) overnight at 4 °C followed by incubation with a goat antirabbit fluorescent secondary antibody (Abcam, Cambridge, UK) at 37 °C for 1 h. Finally, the sections were counterstained with DAPI. Negative control slides without adding primary antibody were included in all assays. The images of sections were captured by an optical microscope (CX41, Olympus, Tokyo, Japan) equipped with camera (DP72, Olympus).

Statistical Analysis

Data were depicted as mean ± standard deviation and were analyzed by GraphPad Prism 5 (San Diego, CA, USA). Student’s t test or one-way ANOVA with Tukey’s posttest was utilized for comparing quantitative data where applicable. A value of p < 0.05 was deemed significant.

Results

Ileal Mucus Was Impaired in Cirrhotic Rats

Rats in TAA group showed extensive liver cirrhosis grossly compared with control group, which was confirmed microscopically after Sirius Red staining (Fig. 1a). Ishak fibrosis score was also increased in TAA group compared with that in control (5.60 ± 0.57 vs 0.00 ± 0.00, Fig. 1b). As shown in Fig. 1c, ilea mucus in TAA group appeared to be thin, loose, and porous with small bubbles, while control group had thick, dense, and compact mucus. The AOD of mucus in villus (1.00 ± 0.15 vs 0.63 ± 0.04, p = 0.0384, Fig. 1d) and crypt (1.00 ± 0.07 vs 0.74 ± 0.07, p = 0.0174, Fig. 1e) was significantly less in TAA group, indicating less mucus protecting the villus–crypt axis. Consistently, the mRNA expressions of mucus components, Muc2 and TFF3 in the intestinal tract, were significantly reduced in TAA group, p < 0.05 (Fig. 1f, g). Moreover, the protein level of MUC2 in ileum was decreased in TAA group, 0.21 ± 0.06 versus 0.14 ± 0.05 ng/mg ileum; p = 0.0126 (Fig. 1h).

Bacteria Were Prone to Invading Ileal Crypts in Cirrhotic Rats

Serous LPS level in portal vein of TAA group was significantly higher than that in control, 99.58 ± 36.73 versus 52.37 ± 24.35 ng/ml; p = 0.0254 (Fig. 1i). Limited bacteria stained with FISH were detected in both groups. The bacteria in control rats were mainly located beyond the top of villi (Fig. 1j, upper panel). Differently, the bacteria in TAA rats (5 out of 10 rats, 50%) were visualized between the adjacent villi (Fig. 1j, lower panel), while no such phenomena could be observed in control rats.

Goblet Cells in Cirrhotic Rats Were Decreased and Smaller

Compared with control group, the theca of goblet cells in ilea of TAA group showed with H&E stain became less and small (Fig. 2a) and this was verified by PAS stain (Fig. 2b). Accordingly, the length of villi, number of goblet cells per perimeter of villi, and size of the theca area were compared quantitatively as 328.8 ± 52.1 versus 322.9 ± 43.9 μm, p > 0.05 (Fig. 2c), 0.068 ± 0.013 versus 0.054 ± 0.006, p = 0.0176 (Fig. 2d), and 110.30 ± 6.39 versus 96.26 ± 6.29 μm2, p = 0.0006 (Fig. 2e) in control and TAA group, respectively.

Fig. 2
figure 2

Imbalance of intestinal epithelial differentiation and proliferation in cirrhotic rats. Goblet cell was visualized by hematoxylin and eosin (H&E) or PAS stain in paraformaldehyde-fixed ileum tissues (a and b). The length of villus was not altered, but number of goblet cell per perimeter of villi, and theca area of goblet cells were reduced in TAA group (c–e, n = 8 in each group). The increase in Ki-67-positive epithelia and AOD in ileum was observed in TAA group (f–g, n = 6 in control group, n = 7 in TAA group). Expression of CDK4 (h, i, n = 10 in each group) was higher, and protein level of P21 (h, j, n = 8 in each group) was lower in TAA group. *p < 0.05

Full size image

Proliferation of Ileal Epithelia in Cirrhotic Rats Was Increased

More Ki-67-expressing epithelia could be observed in TAA group (Fig. 2f), and AOD of Ki-67 was increased in TAA group, 1.00 ± 0.25 versus 1.35 ± 0.28, p = 0.0365 (Fig. 2g). Proliferation-related protein CDK4 was significantly increased in TAA group, while P21, the protein commonly inhibiting proliferation, was reduced in cirrhotic rats, p < 0.05 (Fig. 2h–j).

Notch1–Hes1–KLF4 Signaling Was Activated in Cirrhotic Rats

Ileal Notch1 expression was obviously augmented in TAA group (Fig. 3a). Among the downstream factors of Notch1, only the mRNA expressions of KLF4 were remarkably inhibited in TAA group compared with control, p < 0.05 (Fig. 3b–e). Western blot data showed that Notch1 ligand Dll1, active form of Notch1 NICD, and Hes1 were significantly higher in TAA group, p < 0.05 (Fig. 3f–i). Consistent with the results of mRNA expression, SPDEF and ATOH1 was not altered, but KLF4 protein was significantly suppressed in cirrhotic rats, p < 0.05 (Fig. 3f, j–l).

Fig. 3
figure 3

Notch1 pathway was activated in cirrhotic rats. Notch1 expression was determined by IHC staining implying increased Notch1 level in TAA group (a). mRNA expressions of Hes1 (b, n = 8 in each group), ATOH1 (c, n = 9 in each group), and SPDEF (d, n = 8 in each group) were not different in two groups, but mRNA of KLF4 (e, n = 8 in each group) was significantly lower in TAA group. Protein levels of Notch1 intracellular domain (NICD; f, h, n = 5 in each group), Dll1 (f, g, n = 10 in each group), and Hes1 (f, i, n = 9 in each group) were higher in TAA group, expressions of ATOH1 (f, j, n = 6 in each group) and SPDEF (f, k, n = 9 in each group) were similar, and KLF4 (f, l, n = 10 in each group) expression was lower in TAA group. * p < 0.05

Full size image

Overexpression of Notch1 Reduced Mucus Production, but Increased Proliferation of Ls174T Cells

Induction of NICD expression by EDTA treatment led to depletion of mucus in cells (Fig. 4a). Overexpression of NICD and Hes1, and inhibition of KLF4 were confirmed in EDTA treated cells compared with PBS treated cells by western blot, p < 0.05 (Fig. 4b–e). Transfection of NICD plasmid to Ls174T cells also vanished mucus expression (Fig. 4f). CCK8 assay found significantly higher viability of cells transfected with NICD plasmid, 1.36 ± 0.08 versus 1.00 ± 0.19, p = 0.0013 (Fig. 4g). Then, more Ki-67-expressing cells in NICD plasmid-transfected cells were found compared with that in vector-transfected cells (Fig. 4h). Furthermore, the increased expressions of NICD and Hes1, and inhibition of KLF4 were confirmed in NICD plasmid-transfected cells by western blot, p < 0.05 (Fig. 4i–l).

Fig. 4
figure 4

Overexpression of Notch1 reduced mucus expression and increased proliferation of Ls174T cells. EDTA treatment reduced mucus expression in Ls174T cells (a). EDTA treatment increased expressions of NICD and Hes1, but decreased KLF4 (b–e). Transfection of NICD plasmid to Ls174T cells led to decreased mucus expression (f). Viability of cells was increased after transfection of NICD plasmid (g). More Ki-67-positive cells were observed after transfection of NICD plasmid (h). Transfection of NICD plasmid increased expressions of NICD and Hes1, but decreased KLF4 (i–l). Mucus expression was indicated by green arrows. n ≥ 3 in each group. *p < 0.05

Full size image

Suppressed Notch1 Expression Up-Regulated Mucus Expression and Reduced Proliferation of Ls174T Cells

The optimal mucus expression in Ls174T cells (Fig. 5a) was at the 100 μM DBZ, which inhibited Notch1/NICD protein dose-dependently at the concentration of 10–100 μM, p < 0.05 (Fig. 5b, c). Interestingly, DBZ at 100 μM could significantly suppress Hes1, but increase KLF4 expression, p < 0.05 (Fig. 5b, d, e). Additionally, CCK8 assay showed that all tested concentration of DBZ could significantly reduce the viability of cells, p < 0.05 (Fig. 5f). DBZ at 25 μM and 100 μM obviously reduces the number of proliferating cells showed as Ki-67 stain (Fig. 5g). All the results were summarized in a schema, as shown in Fig. 6.

Fig. 5
figure 5

Suppressed Notch1 expression up-regulated mucus expression and reduced proliferation of Ls174T cells. 100 μM of dibenzazepine (DBZ) treatment resulted in the increase in mucus expression (a). Expressions of NICD and Hes1 were reduced, but KLF4 was increased after DBZ treatment (b–e). Viability of cells was decreased after serial concentration of DBZ treatment (f). Few Ki-67-positive cells were observed after DBZ treatment (g). Mucus expression was indicated by green arrows. n ≥ 3 in each group. *indicates p < 0.05

Full size image
Fig. 6
figure 6

Schema

Full size image

Discussion

Very few experiments focused on mucus barrier of ileum in liver cirrhosis although some studies ever suggested that gastric mucus from antrum was significantly reduced in cirrhotic patients or rats when compared with normal control [24, 25]. It has been observed in a recent study that the thickness of ileal mucus in cirrhotic mice was reduced and GFP-E. coli could easily pass through ileal epithelium [6]. The mucus barrier may no longer restrain bacteria in liver cirrhosis. Our study also observed few and smaller goblet cells of ileum in cirrhotic rats in spite of increased epithelial proliferation, which indicate impaired epithelial homeostasis. Due to the poorly differentiated goblet cells, ileal mucus became thin, loose, and porous either above villus or in crypt. Besides the reduction in PAS staining, expressions of essential components of intestinal mucus Muc2 and TFF3, which covalently bound to each other to form netlike structure of mucus [26], were also greatly decreased in cirrhotic rats when compared with control ones.

Unlike colon mucus which is consisted with the inner and outer layers, the surface of the small intestine is usually covered with a loose layer of mucus. It is much more mobile and traps the bacteria moving forward to the distal gut. As long as the small intestine remains its motility, the amount of bacteria in the lumen is low [27]. The low density of microorganism in the ileal mucus made few positive spots of bacteria with EUB338 FISH staining in this experiment. Differently, the positive spots of ileal bacteria were always beyond the top of villi in control rats, while they were closer to the crypts in cirrhotic rats. As a result, LPS levels in the portal vein of cirrhotic rats were significantly enhanced. Bacteria were likely to invade ileal crypts, and the protection of the villus–crypt axis might be impaired.

The intestinal epithelium is proliferated and differentiated from stem cells located in the crypts, contributing to intestinal epithelial homeostasis. The present study observed that Notch1–Hes1–KLF4 signaling pathway was activated in cirrhotic rats and considered that it was also involved in the insufficient differentiation of goblet cells. The relationship between Notch1–Hes1–KLF4 signaling and goblet cell differentiation as well as its gene expression has been verified in many studies. After binding to its ligand, such as Dll1 [17], Notch1 is activated and its intracellular domain, NICD, enters nucleus to induce downstream target genes, such as Hes1 [15, 16]. The overexpressed NICD and Hes1 could result in increase in proliferation and loss of secretory lineage [22, 28]. On the other hand, the absence of Notch1 signaling or Hes1 leads to differentiation toward secretory lineage, including goblet cells, accompanied by loss of intestinal stem cells and decreased proliferation [14, 16, 18, 19, 29, 30]. Additionally, KLF4 is required for inducing goblet cell genes and its normal phenotype, which could be suppressed by Notch signaling [31,32,33,34].

Interestingly, by modulating Notch signaling in in vitro experiments with mucus secreting goblet cell-like Ls174T cells [13], we found similar results as observed in cirrhotic rats. By inducing Notch signaling with EDTA, a known NICD activator [35] and transfection of NICD plasmid, cells showed increased proliferation, scarce mucus with augmentation of NICD and Hes1, and inhibition of KLF4. In accordance, another study transfected the expression construct for NICD to HT29 cells and found KLF4 could be dramatically suppressed accompanied by the increase in cell proliferation [34]. Besides inhibiting Notch signaling via DBZ, Ls174T expressed more mucus, had fewer proliferating cells, and showed decreased NICD and Hes1, but increased KLF4 expressions. Similarly, when mice were treated with DBZ, goblet cells were increased alone with the enhanced KLF4 expression in small intestine [18]. Therefore, the reduction in mucus expression and increase in epithelial proliferation in cirrhotic rats could be explained by activated Notch1–Hes1–KLF4 pathway.

Taken together, ileal mucus was impaired, and epithelial homeostasis was compromised, manifesting as imbalance in differentiation and proliferation of ileal epithelia, in cirrhotic rats. Activated Notch1–Hes1–KLF4 signaling might account for these phenomena in vivo. By using Ls174T cells, modulation of Notch signaling could also affect the number of mucus-expressing and proliferating cells in vitro. The current study provides novel insights into our understanding of intestinal barrier in liver cirrhosis.