Alterations of nitric oxide homeostasis as trigger of intestinal barrier dysfunction in non‐alcoholic fatty liver disease

Abstract Changes in intestinal nitric oxide metabolism are discussed to contribute for the development of intestinal barrier dysfunction in non‐alcoholic fatty liver disease (NAFLD). To induce steatosis, female C57BL/6J mice were pair‐fed with a liquid control diet (C) or a fat‐, fructose‐ and cholesterol‐rich diet (FFC) for 8 weeks. Mice received the diets ± 2.49 g L‐arginine/kg bw/day for additional 5 weeks. Furthermore, mice fed C or FFC ± L‐arginine/kg bw/day for 8 weeks were concomitantly treated with the arginase inhibitor Nω‐hydroxy‐nor‐L‐arginine (nor‐NOHA, 0.01 g/kg bw). Liver damage, intestinal barrier function, nitric oxide levels and arginase activity in small intestine were assessed. Also, arginase activity was measured in serum from 13 patients with steatosis (NAFL) and 14 controls. The development of steatosis with beginning inflammation was associated with impaired intestinal barrier function, increased nitric oxide levels and a loss of arginase activity in small intestine in mice. L‐arginine supplementation abolished the latter along with an improvement of intestinal barrier dysfunction; nor‐NOHA attenuated these effects. In patients with NAFL, arginase activity in serum was significantly lower than in healthy controls. Our data suggest that increased formation of nitric oxide and a loss of intestinal arginase activity is critical in NAFLD‐associated intestinal barrier dysfunction.

besides diet and physical inactivity, dysfunction of intestinal barrier and subsequently, an increased translocation of bacterial endotoxin may also be among the key factors in the onset and the progression of NAFLD. [3][4][5] Indeed, it has been shown that plasma bacterial endotoxin levels in portal vein are elevated and expression of toll-like receptor 4 (Tlr4), as well as dependent signalling cascades in liver, are induced in patients with NAFLD. [6][7][8] Additionally, results of animal models of NAFLD targeting endotoxin-dependent signalling cascades, and even more so, intestinal barrier function, 3,5,9 suggest that intestinal barrier dysfunction is critical in the development of the disease.
Alterations of nitric oxide bioavailability in intestinal epithelial cells have been discussed to be critical in maintaining the intestinal epithelial barrier structure. 10 Indeed, while moderate amounts of nitric oxide are continuously produced by the neuronal and endothelial nitric oxide synthase (NOS1, NOS3) in the intestinal mucosa, being critical in maintaining intestinal homeostasis and integrity, an induction of inducible nitric oxide synthase (iNOS, NOS2) has been proposed to be associated with intestinal barrier dysfunction in various settings including NAFLD in humans. 11,12 In support of these findings, it has been shown that supplementation of the amino acids L-citrulline and L-arginine, shown to restore intracellularly the nitric oxide formation, suppress iNOS activity in the intestine in the presence of endotoxaemia in rodents. 13 However, the underlying molecular mechanism of the diet-induced disruption of intestinal barrier function in mice with diet-induced NAFLD and the role of nitric oxide, herein, have not yet been fully clarified. Based on this background, the aim of the study was to assess whether changes in intestinal nitric oxide and L-arginine metabolism are critical in the development of diet-induced intestinal barrier dysfunction in mice with NAFLD and if targeting these changes through an oral supplementation of L-arginine or the arginase inhibitor N ω -hydroxy-nor-L-arginine (nor-NOHA) alters the progression of steatosis to later disease stages, for example NASH.

| Animals and intervention trials
Six-eight weeks old female C57BL/6J mice (Janvier SAS, Le Genest-Saint-Isle, France) were housed in controlled conditions in a specific pathogen-free barrier facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care. All procedures and treatments were approved by the local institutional animal care and use committee (Landesamt für Verbraucherschutz, Thuringia, Germany and Federal Ministry Republic of Austria Education, Science and Research, Vienna, Austria). Mice had free access to water at all times and were housed in groups. It has been shown that female C57BL/6J mice were more susceptible to fructose-induced NAFLD 14 and results of a meta-analysis suggest that variability of traits and parameters is similar between male and female mice. 15 For the ex-vivo-everted gut sac experiments, naïve female C57BL/6J mice were sacrificed via cervical dislocation. For feeding experiments, animals were randomly assigned to the following feeding groups (the study designs are summarized in Figure S1).
The calculation of sample size was based on previous findings. 3,16 Intervention trial 1: Mice were fed a liquid control diet (C; 15.7 MJ/ kg diet: 69E% carbohydrates, 12E% fat, 19E% protein; Ssniff) or a liquid fat-, fructose-and cholesterol-rich diet (FFC; 17.8 MJ/kg diet: 60E% carbohydrates, 25E% fat, 15E% protein with 50% wt/ wt fructose and 0.16% wt/wt cholesterol; Ssniff) as detailed previously. 3 After an adaption phase to the liquid diet, mice were pair-fed C or FFC for 8 weeks. In week 8, tissue and blood were collected from some mice fed C or FFC as detailed below. Remaining animals were randomized to the following groups (n = 6-8/group): mice fed liquid C ± 2.49 g L-arginine/kg bw/day (C, C + Arg) and mice fed FFC ± 2.49 g L-arginine/kg bw/day (FFC, FFC + Arg) for additional 5 weeks. Intervention trial 2: Furthermore, 6-8 weeks old female C57BL/6J mice (n = 6/group) were fed drinking water enriched with 30% (w/v) fructose (F) in addition to standard chow for 16 weeks or plain water. For details regarding feeding and liver damage and markers of intestinal permeability see also Sellmann et al. 17 Intervention trial 3: After the adaption to the liquid diets, mice were pair-fed C or FFC supplemented with a mixture of non-resorbable antibiotics (AB) for 8 weeks resulting in the following experimental groups: C, FFC or FFC + AB. AB mixtures consisted of polymyxin (92 mg/kg bw/day) and neomycin (216 mg/kg bw/day) as detailed before. 18 Intervention trial 4: Once adaption to the liquid diet, animals were pair-fed a liquid C or FFC ± 2.49 g L-arginine/kg bw/day and were treated i.p. with the arginase inhibitor nor-NOHA (0.01 g/kg bw; Bachem AG) or vehicle 3 times per week for 8 weeks resulting in the following experimental groups: C, FFC, FFC + NOHA, FFC + Arg and FFC + Arg + NOHA.
At the end of the trials, mice were anaesthetized with 100 mg ketamine/kg bw and 16 mg xylazine/kg bw. Blood from portal vein was collected just prior to cervical dislocation. Blood, livers and intestinal tissue were collected to determine markers of liver damage and intestinal barrier function. Therefore, tissue was fixed in neutralbuffered formalin or snap-frozen for further analyses.

| Everted sac model of mice ex vivo
Small intestines (n = 7/treatment) from naïve female C57BL/6J mice were collected and everted with a rod as previously described, 19,20 cut into equal sections, ligated at both ends and filled with 1X Krebs-Henseleit-bicarbonate buffer supplemented with 0.2% bovine serum albumin (KRH buffer). Everted gut sacs were preincubated in gassed (95% O 2 /5% CO 2 ) KRH buffer (Ctr) ± 0.04 mM L-arginine at 37°C for 10 min and then further incubated with 5 mM fructose (F) ± 0.04 mM L-arginine (Arg + F), respectively, at 37°C for 1 h. To determine tissue permeability, everted small intestinal tissue sacs were incubated in 0.1% D-xylose (Sigma-Aldrich Chemie) for 5 min in above-mentioned incubation solutions, subsequently. After incubation, liquids inside the everted tissue sacs were collected, and intestinal tissue was snap-frozen until further analysis.

| Human study
Serum was collected from 13 patients with steatosis (NAFL patients) as diagnosed by ultrasound and/or liver biopsies and 14 agematched healthy controls. All procedures were approved by the ethics committee of the Medical University of Vienna and informed consent was obtained from all subjects before the study (747/2011).
Characteristics of the NAFL patients and controls and liver histology of NAFL patients are summarized in Table S1.

| Histological evaluation of liver and hepatic triglyceride accumulation
Paraffin-embedded liver sections (4 µm) of mice were stained with haematoxylin and eosin (Sigma-Aldrich Chemie) and evaluated using NAFLD activity score (NAS) adapted from Kleiner et al. 21 Neutrophilic granulocytes were stained using a commercially available Naphthol AS-D Chloroacetate Specific Esterase Kit (Sigma-Aldrich Chemie) as described before. 14 Concentration of hepatic triglycerides was measured in liver homogenates using a commercially available kit (Randox Laboratories).

| Blood parameters of liver damage
Activities of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) in mouse plasma were measured in a routine laboratory (Friedrich-Alexander University).

− ) measurement
Arginase activity was measured in proximal intestinal tissue of mice and in serum of NAFL patients and healthy controls as detailed previously. 22 To determine arginase activity in tissue samples, proximal small intestine was homogenized in 10 mM Tris-HCl containing 0.4% (w/v) Triton X-100 and protease inhibitor cocktail. Levels of NO 2 − in proximal small intestine were detected using Griess assay (Promega).

| D-xylose assay
To determine tissue permeability, xylose concentration in collected liquids of everted gut sacs was measured using a commercially available kit (Megazyme, Bray).

| Endotoxin, TLR4 ligand measurement, myeloperoxidase (MPO) assay and ELISA
Bacterial endotoxin levels in mouse portal plasma were determined using a limulus amebocyte lysate assay (Charles River) as reported in detail previously. 3 TLR4 ligands in mouse portal plasma and human serum were measured as previously described 23

| RNA isolation, real-time RT-PCR
Total RNA from liver was extracted with Trizol (peqGOLD TriFast, Peqlab). cDNA was synthetized with a reverse transcription system (Promega) and real-time PCR was performed using iTaq TM Universal SYBR Green Supermix (Bio-Rad Laboratories). Table S2 summarizes the primer sequences that were used.

| Western Blot analysis
Intestinal tissue was homogenized in RIPA buffer containing pro-

| Statistics
Data are presented as means ± standard error of the means (SEM).
Statistical analyses were performed with GraphPad Prism Version 7.0. Grubbs test was used to identify outliers. A Student t test was used to analyse differences between C-and FFC-fed animals fed for 8 weeks, C and F-fed animals fed for 16 weeks, or for analysis between NAFL patients and healthy controls. A one-way ANOVA and two-way ANOVA, respectively, were applied to determine statistical differences between three and more different treatment groups. For analysing parameters obtained from ex-vivo-everted gut sac experiments, one-way ANOVA was used. A p-value <0.05 was defined as significant.  Figure 1K). Furthermore, arginase activity, which is suggested to be the opponent of iNOS in the balance of nitric oxide availability, and subsequently, pro-and anti-inflammatory processes, 24,25 was significantly lower in small intestine of FFC-fed mice than in controls ( Figure 1L). Interestingly, protein levels of ARG-2 found to be the predominant arginase isotype in small intestine, 26 were similar between groups while ARG-1 protein was not detectable in small intestinal tissue ( Figure S2). In line with these findings, To elucidate if alterations alike are also associated with the development of NAFLD in humans, markers of intestinal permeability and arginase activity were determined in serum of NAFL patients and controls. In line with the findings in mice, marker of intestinal permeability in serum, like TLR4 ligands were higher in NAFL patients than in healthy probands, while arginase activity was significantly lower (p < 0.05, Figure 2A,B).

| Effect of antibiotics on intestinal permeability and arginase activity in small intestine in FFC-fed mice
To determine whether the alterations found in small intestinal tissue of FFC-fed mice were related to intestinal microbiota, mice pair-fed the FFC were concomitantly treated with the non-resorbable antibiotics polymyxin B and neomycin. C-fed mice are shown for comparison. As expected, the development of NAFLD was significantly attenuated in mice fed the FFC + AB as determined by NAS ( Figure 3A,B). Despite the similar caloric intake, absolute body-and liver weight and liver to body weight ratio, AST activities in plasma were significantly lower in FFC + AB-fed mice compared to FFCfed mice (Table S3). In contrast, protein concentrations of the tight junction ZO-1 in small intestinal tissue were similar between the two FFC-fed groups and markedly lower than in C-fed mice ( Figure 3C; Figure S4). Furthermore, arginase activity in small intestinal tissue was also similar in both FFC-fed groups. However, arginase activity was markedly lower than in controls ( Figure 3D).

| Effect of fructose on arginase activity and markers of intestinal permeability in everted gut sacs of mice
To assess if a loss of arginase activity plays an important role in the development of intestinal barrier dysfunction in NAFLD and if fructose found in diet is critical; herein, everted gut sacs of naïve mice were incubated with fructose. The ex-vivo-everted sac technique is summarized in Figure 3E.

| Effect of L-arginine on markers of liver damage and intestinal barrier function in FFCfed mice
To determine whether targeting intestinal arginase activity has a beneficial effect on intestinal barrier dysfunction in vivo, mice with diet-induced NAFLD were either pair-fed the FFC or an FFC supplemented with L-arginine for 5 weeks. In accordance with previous findings, 16 Table 1). However, liver weight and liver to body weight ratio were still significantly higher in both FFC-fed groups compared to controls (Table 1) Furthermore, as data varied considerably, ALT and AST activity levels were similar between groups ( Table 1).
The progression of NAFLD was associated with a significant  Figure 4G,H). Again, protein levels of ARG-2 were not different between groups ( Figure S5).

| Effect of inhibiting arginase activity on the development of liver damage and markers of intestinal barrier function in FFC-fed mice
To further delineate if a loss of arginase activity is an important factor in the development of NAFLD, mice were additionally treated with the arginase inhibitor nor-NOHA while being fed the C-and Despite similar caloric intake, markers of liver damage, for example NAS, liver weight and liver to body weight ratio were significantly higher in FFC-and FFC + NOHA-fed mice when compared to FFC + Arg-and FFC + Arg + NOHA-fed mice ( Figure 5A,B; Table 2).
Similar to the NAS, number of neutrophils was higher in all FFC-fed groups than in controls. Number of neutrophils was also significantly higher in livers of FFC + NOHA-fed mice than in all other groups, whereas number of neutrophils were similar between FFC-and FFC + NOHA + Arg-fed mice. In FFC + Arg-fed mice, number of neutrophils was significantly lower than in FFC-fed mice ( Figure 5C). The activity of MPO in liver was also higher in FFC-fed groups than in Cfed animals but did not differ between groups as data varied considerably within some groups (Table 2). In contrast, mRNA expression of F4/80 in liver was similar between C-and FFC-fed groups, while mRNA expression of Tnfα was higher in FFC-fed groups regardless of additional treatments compared to controls; however, as data varied considerably in some groups, no differences were found between FFC groups (Table 2)  Data are shown as means ± SEM, n = 5-8.

ACK N OWLED G EM ENTS
This work was funded by the German Research Foundation (DFG): BE2376/6-1 and BE2376/6-3 (both I.B.). Open access funding provided by University of Vienna.

CO N FLI C T O F I NTE R E S T S
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.