intestinal villi in the small intestine. Carbohydrate absorptive and digestive capacities wereassessed by stable isotope methodology after administration of U-13C-glucose and 1-13C-lactose. Concentrations of the EFA linoleic acid (LA), and the enzyme activity and mRNAexpression of lactase, were measured in the mucosa of proximal, mid and distal smallintestine. Results: Mice fed the EFA-deficient diet were markedly EFA-deficient (triene/tetraene ratio 0.23±0.06 vs. 0.01±0.00 in controls, p<0.05) with a profound lower fatabsorption of dietary fat (81±3 vs. 99±0% in controls, of ingested amount, p<0.05). EFAdeficiency did not affect the histology or proliferative capacity of the small intestine asdemonstrated by the histological and proliferative staining, and by the morphometrcalmeasurements of the small intestinal villi. Blood 13C6-glucose appearance and disappearancewere similar in both groups, indicating unaffected monosaccharide absorption. In contrast,blood appearance of 13C-glucose, originating from 13C-lactose, was delayed in EFA-deficientmice. EFA deficiency profoundly reduced lactase activity (-58%, p<0.01) and mRNA expres-sion (-55%, p<0.01) in mid small intestine. Both lactase activity and its mRNA expressionstrongly correlated with mucosal LA concentrations (r=0.89 and 0.79, resp., p< 0.01).Conclusions: EFA deficiency in mice inhibits the capacity to digest lactose, but does notaffect small intestinal histology. These data underscore the observation that EFA deficiencyfunctionally impairs the small intestine, possibly mediated by low LA levels in the enterocytes.This research was funded by the Dutch Digestive Foundation (MWO 04-38).

M1731

Sphingosine-1-Phosphate Regulates Barrier Function in Cultured IntestinalEpithelial Cells and Murine Intestinal TissueRuiyun Li, Jose Greenspon, Emily Bellavance, Rex Sun, Lan Xiao, Rao N. Jaladanki, TerezShea-Donohue, Jian-Ying Wang, Douglas J. Turner

Introduction: Intestinal barrier dysfunction has been implicated in a wide variety of pathologicconditions such as infection, trauma, inflammation and malignancy. Disruption of intercellu-lar junctional proteins appears to be a key pathway to loss of barrier integrity. Preliminaryevidence from our lab has shown that Sphingosine-1-phophate (S1P) upregulates E-cadherinprotein and decreases paracellular permeability. We hypothesized that S1P regulated barrierfunction via direct interaction with the S1P-1 receptor in intestinal epithelial cells (IECs),and that S1P would be decrease paracellular permeability in murine intestinal mucosa aswell. Methods: Studies were performed upon cultured differentiated IECs (IEC-Cdx2L1 line).Western blotting was assessed and siRNA transfections were via Lipofectamine. Paracellularpermeability was assessed 14C-labelled mannitol, and transepithelial electrical resistancewas measured by Voltohmmeter. Muscle-free segments of small intestine taken from C57Bl/6 mice were mounted in microsnapwells to measure transepithelial electrical resistance(TEER), an index of mucosal permeability. S1P (0.5 µM - 5.0 µM) or vehicle was added tothe serosal side and TEER was measured for every 30 minutes for 120 minutes. Results:Treatment of IECs with S1P (0.5 µM) increased E-cadherin protein by Western blot, anddecreased paracellular permeability to mannitol by ~ 25%. IECs demonstrated the S1P-1receptor protein, but did not express the S1P-3 receptor. Consistent with results in non-transfected cells, S1P treatment (0.5 µM) reduced permeability versus control by ~20% incells transfected with control siRNA, whereas S1P did not reduce permeability when S1P-1 was silenced with siS1P-1. In vehicle-treated tissue, TEER dropped over the 120 minuteperiod to 74+/-5% of its original value. In contrast, TEER in segments exposed to S1P (5.0µM) was 92+/-56% of its initial value at 120 minutes. Lower concentrations of S1P wereeffective until 90 minutes. Conclusions: These results indicate that S1P promotes intestinalepithelial barrier function through a mechanism involving the S1P-1 receptor, but not the S1P-3 receptor. S1P promotes intestinal barrier function in IECs, and this effect is reproducible inmurine intestinal tissue.

M1732

Lipopolysaccharide (LPS) Sensitizes Bile Duct Epithelial Cells for TightJunction (TJ) Disruption By Hydrogen Peroxide and TNFα By Nitric Oxide-Dependent MechanismAnkur Seth, Suneet Jain, Nicholas F. LaRusso, Radhakrishna (RK) Rao

About 80% of patients with primary sclerosing cholangitis (PSC) also suffer from ulcerativecolitis. Elevated intestinal permeability and endotoxemia in ulcerative colitis suggest a rolefor the LPS-induced bile duct injury in the pathogenesis of PSC. In the present study, weinvestigated the role of LPS in priming the bile duct epithelium for TJ disruption by H2O2and TNFα. METHODS: NRC1 (normal rat cholangiocyte) cell monolayers were incubatedwith 1-10 ng/ml of LPS for 1-20 hrs, followed by incubation with H2O2 (1-250 µM) orTNFα (1-20 ng/ml). Barrier function was evaluated by measuring inulin permeability. Cellviability was determined by assay for LDH release, and caspase-3 activity or WST assay.Integrity of TJ was analyzed by confocal microscopy of occludin and ZO-1. Micro arrayanalysis was performed to determine the global changes in gene expression by LPS treatment.Individual gene expression was determined by RT-PCR. The role of nitric oxide (NO) inLPS-induced sensitization to H2O2 and TNFα was determined by pretreatment of cells withL-N-methyl-arginine (L-NMA) or L-N-nitro-arginine (L-NNA). NO-donors, NOC-5 andNOR-3, were used to mimic LPS-induced sensitization. RESULTS: Incubation of NRC1 cellmonolayers with LPS for 20 hrs produced no significant effect on inulin permeability. H2O2(1-30 µM) and TNFα (1-10 ng/ml) dose-dependently increased inulin permeability in LPS-treated cell monolayers. In the absence of LPS, H2O2 and TNFα at these doses showed nosignificant effect on inulin permeability. Increase in inulin permeability was associated witha redistribution of occludin and ZO-1 from the intercellular junction into the intracellularcompartment. Greater than 6 hrs of LPS treatment was required for the sensitization ofNRC1 cells for TJ disruption by H2O2 and TNFα. Micro array analysis demonstrated thatthe expression of iNOS2 was increased 20-fold by LPS, which was confirmed by RT-PCR.Pretreatment of cell monolayers with L-NMA or L-NNA dose-dependently prevented LPS-induced sensitization for TJ disruption by H2O2 and TNFα, suggesting that iNOS2 expressionand NO play an important role in LPS-induced sensitization. However, pretreatment of cellswith NO-donors (NOC-5 and NOR-3) did not sensitize cells for TJ disruption by eitherH2O2 or TNFα. RT-PCR showed increased expression of IFNγR2 and chemokines by LPS.IFNγ also failed to mimic sensitization of NRC1 cells, indicating that a combination of

T : 11501$$CH204-02-08 16:47:10 Page 407Layout: 11501B : o

A-407 AGA Abstracts

multiple gene functions may be involved in the effect of LPS. CONCLUSION: These resultsdemonstrate that LPS at low doses sensitizes bile duct epithelium for TJ disruption by H2O2and TNFα by a NO-dependent mechanism.

M1733

Protective Effect of Rebamipide On the Methotrexate-Induced IntestinalDysfunction in RatsKazuma Hamada, Yoshihisa Shitara, Shuichi Sekine, Toshiharu Horie

(Background and Aim) The intestinal mucosa plays a crucial role in the active transport ofa great number of endogenous and exogenous materials as well as a barrier function. It iswell known that various kinds of drugs such as NSAIDs and antineoplastic agents causesmall intestinal injury in clinical practice, possibly leading to life-threatening sepsis. Inthe present study, we investigated the effect of rebamipide, an antiulcer drug, on themethotrexate(MTX)-induced increase in paracellular permeability in rat small intestine.(Method) The permeation of fluorescein isothiocyanate-labeled dextran(molecular weight;4,400) (FD-4) was examined in the In Vitro everted intestinal sac. Luminol-enhanced chemilu-minescence from the intestinal mucosa was examined to evaluate the reactive oxygen species(ROS) generation. Proteins of small intestinal mucosa homogenate were tested for tightjunction (TJ) proteins occludin and zonula occludens-1 (ZO-1), inducible nitric oxidesynthase (iNOS) and phospho-(Ser/Thr) cAMP dependent protein kinase (PKA) substrateexpression by Western blot analysis. (Result) The permeation of FD-4 and chemiluminescencefrom the intestinal mucosa of MTX-treated rats were significantly higher than those inuntreated rats. Coadministration of rebamipide significantly reduced them. The expressionsof occludin and ZO-1 as well as iNOS in small intestine were not altered in MTX- and MTXplus rebamipide treated rats. However, the levels of phospho-PKA substrates were decreasedin MTX-treated rats. (Discussion) These data suggest that ROS may be involved in the MTX-induced increase in paracellular permeability. In addition, because paracellular permeabilityis regulated by TJs, we investigated the expression levels of TJ proteins, occludin and ZO-1, in rat small intestine by Western blot analysis. However, interestingly, there was noapparent difference in the expression of these TJ proteins between control and MTX-treatedrats. On the other hand, the levels of phospho-PKA substrates of MTX-treated rats weredecreased compared with control rats. Since alteration of phosphorylation status of TJproteins could affect TJ functions by regulation of protein-protein interactions, oxidativestress induced down-regulation of PKA activity may be associated with its disruption. Inconclusion, as rebamipide is a kind of antioxidants, it might prevent the increase in paracellu-lar permeability by regulation of redox status. Rebamipide might be useful in decreasingthe risk of MTX-induced small intestinal injury in human.

M1734

Intestinal Involvement in Autistic Disorder: Progress of ResearchValeria Familiari, Laura de Magistris, Anna Sapone, Gabriele Riegler, Maria Cartenì,Patrizia Iardino, Giancarlo Caravelli, Carmela Bravaccio, Alessandro Frolli, RobertoMiliterni, Antonio Pascotto

We already demonstrated that IP is altered in a consistent % of autistic patients and theirfirst-degree relatives. AIM of the present progress of research was to increase the numberof investigated subjects and to correlate IP values to gastrointestinal (GIs) and behaviouralsymptoms (ADI). MATERIALS AND METHODS: 90 consecutive children with autism (meanage±SD=7,4±5,1; F=14, M=76) diagnosed according to the DSM-IV criteria, most of themnaïve and without any dietary restrictions; 146 of their first degree relatives (mean age±SD=40,2±8,7; F=72, M=74); 160 normal adult subjects (mean age±SD=31,8±12,3; F=98, M=42) and 20 normal children (mean age±SD=10,1±3,9; F=11, M=9) were recruited. The IPwas evaluated by means of lactulose and mannitol test (LA/MA). Faecal calprotectin wasperformed to evaluate intestinal inflammation. Serum anti-trans-glutaminase (tTG) and HLA-DQ2/8 were determined to rule out celiac disease in all subjects. GIs were evaluated as:constipation(C), diarrhoea (D), alternating diarrhoea/constipation and abdominal pain (P).RESULTS: tTG resulted normal in all autistic patients. HLA-DQ2/8 distribution reflectedthat of general population. 36,7% of autistic patients showed LA/MA values higher than thecut off range (0,030). 21,7% of first-degree relatives had LA/MA higher than normal. LA/MA mean values resulted statistically different (p<0,05) among the groups: a) autistic,0,042±0,084; b) relatives, 0,028±0,050; c) adult 0,014±0,014 and children controls0,016±0,006. GIs resulted present in 48,1% of the autistics, being: C=46,2%, D= 34,6%,P= 19,2%. The presence of referred GIs was independent on IP alteration: no correlationwas found between LA/MA values and GIs (p=0,275). Moreover, no correlation was foundbetween LA/MA and typical autistic behavioural symptoms (ADI) such as: lack of communic-ative skills (p=0,163), deficits in social interaction (p=0,88) and restricted repetitive behaviourad interests (p=0,246), neither between GIs and ADI. Calprotectin values indicated thatinflammation was present in 24,6% of autistic patients and in 11,7% of their relatives. Asalready described, we found no correlation between IP and faecal calprotectin values (r=0,0177). CONCLUSION: A statistically significant increase of IP has been confirmed in largergroups of autistic patients and first degree relatives compared to controls, thus suggesting agenetically-determined defect of intestinal barrier function. GIs are moreover present inalmost 1:2 autistic children, thus confirming the importance of GI tract involvement inautistic disorder. The still unknown genetic intestinal factor seems anyway independentfrom behavioural symptoms.

AG

AA

bst

ract

s