intestinal bacteria are uncultivable, we developed Inter Spacer profiling (IS-pro); a rapid,highly reproducible method for high-throughput bacterial profiling. Aim To characterizethe intestinal microbiota adherent to duodenal mucosal biopsy specimens in individualsundergoing routine gastroscopy. Methods Duodenal biopsy specimens were harvested from21 consecutive individuals undergoing routine gastroscopy. In addition, four accompanyingcolonic mucosal biopsy specimens and 11 colonic mucosal biopsy specimens from otherindividuals were harvested during routine colonoscopy. Samples were analyzed with IS-pro,which is based on two features of the bacterial genome: species-specific length of the interspacer (IS) region between the 16S and 23S rDNA and phylum-specific labelled primersequences. With these specific labelled primers, species can be directly sorted into eitherFirmicutes/Actinobacteria or Bacteroidetes phylum. By color and size sorting of amplifiedfragments specific bacterial profiles are created. Intra- and inter-individual variation ofbacterial profiles was assessed with Pearson's correlation. Results Duodenal and colonicspecimens clustered in two distinct clusters. Colonic specimens were unique per individualwith low homology between individuals (6- 59%). In contrast, duodenal specimens showedvery high similarity up to 100% identity. Lowest similarity between duodenal specimens(64%) was higher than highest similarity in colonic cluster (59%). Matched duodenal andcolonic microbiota were distinct with clustering based upon region of sampling instead ofclustering per individual. ConclusionMicrobiota adherent to duodenal mucosa and to colonicmucosa were distinct. Whereas colonic microbiota was found to be host-specific, duodenalmicrobiota was extremely homogenous, with profiles from different individuals being almostidentical. Findings of the present study provide a solid basis for the analysis of potentialdisease-specific differences in microbiota adherent to duodenal mucosa.

W1821

Gut Commensal Microbiota Affects the Expression of Cannabinoid Receptors(Cb1 and Cb2) in the Intestine of RatsEvangelina Teran, Mónica Aguilera, Patri Vergara, Vicente Martinez

Background: Gut microbiota has been implicated in the pathophysiology of irritable bowelsyndrome (IBS). Changes in gut microbiota, associated to treatments with antibiotics or tothe use of probiotics, have been shown to modulate intestinal sensory mechanisms, leadingto the improvement of IBS symptoms in humans and the reduction of visceral pain responsesin animal models.Aim:To determine how spontaneous changes in gut commensal microbiotaaffect the expression of cannabinoid receptors (CB1 and CB2) in the cecum of rats. Methods:SD rats bred under controlled microbiological conditions (barrier group, n=12) or underconventional housing conditions (conventional group, n=15) were used. Cecal microbiotawas characterized using fluorescent in situ hybridization (FISH) and analysis of terminalrestriction fragment length polymorphism of the 16s rDNA (t-RFLP). In the same animals,the expression of CB1 and CB2 receptors was determined by RT-PCR. The same parameterswere also assessed in rats bred under controlled conditions and adapted for a 3-week periodto conventional conditions (adapted group, n=11). Results: Total number of cecal bacteriawas similar in the three experimental groups. Compared with the conventional group, thebarrier group had a higher number of strict anaerobic bacteria (Bacteroides spp and Clostridiumspp; P<0.05), while the number of Bifidobacterium spp was reduced (P<0.05). As expected, themicrobiota of the adapted group showed intermediate characteristics, suggesting spontaneousadaptation to the conventional environment. Expression levels of CB1 receptors were low,with a trend to be higher in the barrier group (0.15±0.02) compared with the conventionalor the adapted groups, which showed similar levels of expression (0.05±0.01 and 0.04±0.01,respectively). CB2 expression was 2-fold higher in the barrier (0.85±0.02) than in theconventional group (0.42±0.02, P<0.001), with the adapted group showing intermediatelevels of expression (0.59±0.02). Regardless the group considered, levels of expression ofCB receptors correlated positively with the counts of Bacteroides spp and Clostridium sppand negatively with the counts of Bifidobacterium spp. Conclusions: Expression of CB1 andCB2 receptors in the rat cecum correlated with spontaneous, environment-related, changesin commensal microbiota, particularly in the content of Bacteroides spp, Clostridium spp andBifidobacterium spp. Thus indicating that commensal microbiota modulates gut sensorymechanisms. These observations suggest that modulation of the intestinal expression of CBreceptors might be part of the mechanisms underlying the beneficial effects of probioticsin IBS.

W1822

Persistent Alterations of Gut Commensal Microbiota During the ChronicPhase of Inflammation in a Model of Indomethacin-Induced Enteritis in RatsEvangelina Teran, Manuel Góngora, Francisco Cabello, Gloria Nejar, Sandra Barbosa,Vicente Martinez, Jordi Cantó, Patri Vergara

Background: Gut microbiota plays a major role in the pathogenesis of inflammatory boweldisease. During acute inflammation gut microbiota suffers significant alterations; however,is not clear if these changes persist during chronic or latent inflammation of the gut. Aim:To determine changes in gut commensal microbiota associated to the chronic, sub-clinic,phase of inflammation in a model of indomethacin induced enteritis in rats. Methods:Intestinal inflammation was induced in SD rats by systemic administration of indomethacin.The progress of inflammation was followed through disease activity parameters (clinicalsymptoms, macroscopic/microscopic alterations and MPO levels). A group of rats wereeuthanized during the acute inflammatory response (day 4 post-indomethacin or vehicle,n=6-7). Another series was allowed to progress to chronicity and was euthanized 21 dayspost-indomethacin (n=11) or vehicle (n=5). In all cases disease activity parameters wereassessed and gut commensal microbiota characterized by fluorescent in situ hybridization(FISH). Results: After indomethacin animals developed acute ileitis, with significant clinicalsigns and macroscopic and microscopic alterations and increased MPO levels in the ileum(ng/100 μg protein; vehicle: 2.4±0.5; Indomethacin: 170±74; P<0.05). Total number of ilealbacteria was similar in control and inflamed animals. Compared to controls, in inflamedanimals there was a significant increase in the number of enterobacteriaceae, while lactobacil-lusspp and Bifidobacterium spp were unchanged. The Clostridium coccoides-Eubacterium rectalegroup, virtually absent in control conditions, increased during inflammation to levels similarto those of other bacterial groups. In animals allowed to progress to a chronic inflammatory

S-747 AGA Abstracts

state, the only signs of inflammation were a moderate increase in the ileal levels of MPO(ng/100 μg protein; vehicle: 4.2±1.4; indomethacin: 24.9±16.9; P>0.05) and the numberof mucosal NFκ-b-positive cells (vehicle: 4.6±0.7; indomethacin: 6.8±1.2; P=0.06), withoutother histological alterations. Nevertheless, there was a persistent alteration of gut microbiota,characterized by a significant increase in the numbers of enterobacteriaceae and Clostridiumcoccoides-Eubacterium rectale. Essentially the same changes observed during the acute phaseof inflammation. Conclusions: These results show that in rats, significant, and similar,alterations of gut commensal microbiota occur during states of active, acute inflammationor during chronic, subclinical, inflammation. These observations suggest that persistentbacterial changes might be a factor in the chronicity of intestinal inflammation.

W1823

Risk Factors for Small Intestinal Bacterial Overgrowth: A MultivariateAnalysisWalter W. Chan, Natan Feldman, Robert Burakoff

Background: Small intestinal bacterial overgrowth (SIBO) has been associated with numerousconditions andmedications. However, with few large clinical trials and potential confounding,their roles in clinically significant SIBO remain unclear. Identifying risk factors would helpin the understanding of the pathophysiology of SIBO and formulating effective preventionand treatment strategies. Aim: To identify independent risk factors for SIBO using lactulosebreath test (LBT). Methods: This was a retrospective cohort study of 273 symptomaticpatients referred to a tertiary care center for LBT. Conditions and use of medications thoughtto be associated with SIBO (gastrointestinal [GI] surgery, diabetes [DM], irritable bowelsyndrome [IBS] per Rome III criteria, pancreatic insufficiency [PI] diagnosed by stool trypsin,thyroid disorder, immunodeficiency, inflammatory bowel disease, narcotic/sedative/psychiat-ric medications, and proton pump inhibitors [PPI]) were reviewed. The patients' demographicdata and presenting symptoms (diarrhea, constipation, bloating, abdominal pain, cramps,nausea, and gas) were also recorded. Positive LBT required a rise >20ppm in breath hydrogenor methane concentration within 60 minutes. Univariate and multivariate analyses wereperformed using Chi-square or Fisher-exact test and logistic regression, respectively. Statist-ical significance was established using p<0.05. Results: 101 (37.0%) patients had positiveLBT. Multivariate analysis showed that PPI, GI surgery, and PI were independently associatedwith positive LBT (OR=2.44, p<0.01; OR=2.78, p=0.03; OR=5.89, p=0.04, respectively).DM was also found to be associated with positive LBT on univariate analysis (OR=2.33, p=0.04), but it was no longer significant on multivariate analysis. No significant associationwith LBT outcome was seen for other factors on both univariate and multivariate analyses.A secondary analysis was performed using PPI and presenting symptoms. PPI remainedan independent risk factor for positive LBT when presenting symptoms were controlled.Conclusions: PPI use, GI surgery, and pancreatic insufficiency were independent risk factorsassociated with SIBO. Given its widespread use, PPI may have been a confounder for somefactors previously linked to SIBO, such as DM and IBS. Discontinuation of PPI may beconsidered as part of SIBO treatment and its use should be judicious, especially in patientswith other SIBO risk factors. SIBO should be considered in symptomatic patients with priorsurgery or pancreatic insufficiency. Further prospective studies are needed to better delineatethe role of these risk factors on small intestinal bacterial flora.

W1824

Murine Guanylate Cyclase C Provides Resistance to Attaching/EffacingPathogensElizabeth A. Mann, Eleana Harmel-Laws, Mitchell B. Cohen, Kris A. Steinbrecher

Background: The transmembrane receptor Guanylate Cyclase C (GC-C) is activated by itssecreted ligand guanylin (Gn) to produce cGMP and stimulate intestinal secretion. In addition,work by us and others implicates this intestinal epithelial cell (IEC) signaling pathway insuppressing proliferation and apoptosis, as well as maintaining epithelial monolayer barrierfunction during cytokine stress. The murine attaching/effacing pathogen Citrobacter rodentiumcauses self-limiting colitis, IEC hyperplasia, and severe diarrhea in some susceptible strains.We hypothesized that mice lacking GC-C activity would be more prone to injury duringC. rodentium infection due to aberrant IEC proliferative/apoptotic homeostasis and barrierfunction. Methods: GC-C null, Gn null, and wildtype (WT) mice, all >10thgeneration C57BL/6J, were administered orogastric C. rodentium (10 9 cfu). Stool was collected every 3-4 daysup to day 20 to enumerate C. rodentium colonies. Mice were sacrificed at 4, 10, and 20days after infection. Proximal and distal colon, fecal contents, and liver were removedaseptically. Bacterial counts were obtained from samples of fecal content and liver andanalyzed by Chi-square. Histological analysis was performed by H&E staining. Results: Thepercent of mice with C. rodentium positive stool on Day 4 differed significantly by genotype.Only 63% (20/32) of WT mice were infected compared to 97% (28/29) of GCC-/- mice (p≤ 0.002) and 80% of Gn-/- mice (12/15, p≤0.3). Of those animals infected, the numberof bacteria recovered from stool was similar in all genotypes and reached a high at Day 7,and then fell to non-detectable levels in most mice by 20 days after infection. Importantly,infected WT, GC-C-/-, and Gn-/- mice did not display any significant luminal fluid handlingdefects, as measured by the lack of apparent diarrhea or weight loss. Histologic examinationrevealed crypt hyperplasia and inflammation throughout the large intestine in all genotypesby Day 10. In colonized mice at Day 10, 29% (2/7) of WT livers were positive for C.rodentium compared to 80% (8/10) of GC-C-/- livers (p ≤ 0.03) and 61% (8/13, p ≤ 0.2)of Gn-/- null livers. Liver pathology was evident in GC-C and Gn null mice in the form oflymphocyte infiltration and, in severely affected GC-C-/- mice, large areas of necrosis.Conclusions: We conclude that GC-C, and its colonic ligand Gn, confer substantial protec-tion against enteric colonization and systemic dissemination of attaching/effacing bacterialpathogens.

AG

AA

bst

ract

s