AGA Abstracts phosphorylation and recruitment to the IP-10 promoter (ChIP analysis) as well as TNF- induced IP-10 mRNA and IP-10-specific reporter gene expression, suggesting post-transcrip- tional inhibitory mechanisms. Although L. casei triggered TNF-induced activation of the translational initiation machinery including the phosphorylation of EIF4G and EIF4E, the intracellular accumulation of IP-10 protein was only transient at early times. Co-immunoprec- ipitation analysis then showed that L.casei induces selective ubiquitination of TNF-induced IP-10, supporting subsequent proteasomal degradation of this T-cell chemokine in IEC. Consistent with the inability of L. casei to target the TNF signalling cascade, IP-10 overexpres- sion experiments revealed stimulus-independent mechanisms for the inhibition of IP-10 expression. Finally, VSL#3 and L. casei (1,3x10 9 cfu/day for 15 weeks) failed to attenuate IP-10 protein expression in primary ileal epithelial cells and the development of tissue pathology in the ileum of heterozygous TNFΔARE/WT mice (N = 5-7/group). Conclusion. We identified a novel mechanism of VSL#3 based on the post-translational blockade of the pro-inflammatory T-cell chemokine IP-10. L. casei was identified as active component of VSL#3 targeting ERK-dependent and ubiquitin-mediated inhibitory mechanisms of IP-10 secretion. Feeding experiments in TNFΔARE mice failed to demonstrate protective effects of VSL#3 and L. casei in an animal model of chronic ileitis, suggesting the colon as primary target of probiotic intervention. 1016 Probiotics Modulate the Distribution of Dendritic Cell Subsets in the Intestinal Mucosa Xiao Wang, Heng-Fu Bu, Vjola Koti, Maurice R. O'Gorman, Xiao-Di Tan BACKGROUND AND AIMS: Dendritic cells (DC) play a pivotal role in regulating immune responses. DCs are distributed in the intestinal mucosa. Phenotypic characteristics of the gut DC have been well defined. Probiotics are a group of beneficial bacteria. A growing body of evidence has shown that probiotics have a positive impact on the gut health and beyond. In the intestinal lumen, probiotics may communicate with DCs. However, the In Vivo effects of probiotics on mucosal DC's characters and behaviors have not been elucidated. In this study, we investigated whether probiotics modulate DC subset compositions in the intestinal mucosa. METHODS: C57BL/10J mice were gavaged with probiotic cocktail VSL#3 (100 mg/mouse) for 7 days. Mononuclear cells were isolated from MLN (mesenteric lymph node), PP (Peyer's patch), and LP (lamina propria) followed by staining with a cocktail containing appropriate antibodies. The stained cells were analyzed by flow cytometry using a lineage-cell-depletion strategy. RESULTS: In mice without exposure to probiotics, we identified two main mucosal DC populations: B220 high CD11c int IA/AE low Lin - defined as plasmacytoid DCs (pDCs) and B220 - CD11c high IA/IE + Lin - defined as myeloid DC (mDC). The pDCs were the dominant form in the LP and PP, whereas mDCs were the prevailing type in MLN. mDC in LP, PP and MLN were CX3CR1 - CD103 + DEC205 inter , suggesting that they displayed a similar pattern on functional associated DC markers. In contrast, pDC were found as CD103 inter DEC205 - subset in LP and PP. They heterogeneously expressed CX3CR1. However, pDC in MLN were negative for all these three markers. VSL#3 administration resulted in a marked decrease in the number of pDC in LP but increase in MLN. In contrast to pDC, mDC were dramatically increased in all lymphoid tissue compartments in the intestinal mucosa. In addition, we revealed that VSL#3 triggered an increase in the CX3CR1 + pDC subset in both LP and PP. CONCLUSIONS: (1) the distribution of DC subsets in the intestinal mucosa is in a lymphoid tissue compartment dependent manner in the steady state; (2) mucosal DC subsets contain distinct subpopulations and express various function- associated DC markers such as CX3CR1, CD103 and DEC205; and (3) exposure to probiotics causes a shifting DC distribution within intestinal mucosa in both lymphoid tissue compart- ment and DC subset dependent mechanisms. 1017 The Metabolic Activity of the Enteric Microbiota Influences the Fatty Acid Composition of Host Tissues Rebecca A. Wall, Reynolds P. Ross, Fergus Shanahan, Liam O'Mahony, Caitlin O'Mahony, Eamonn M. Quigley, Barry Kiely, Gerald Fitzgerald, Catherine Stanton Background: We previously reported that human gut-derived Bifidobacterium breve species can synthesize c9, t11 conjugated linoleic acid (CLA) from free linoleic acid (LA) In Vitro. This has anti-inflammatory and anti-carcinogenic properties. Aims: (a) to assess bacterial production of c9, t11 CLA in the gut In Vivo; and (b) to determine if it influences the fatty acid composition of host tissues. Methods: Three separate murine and porcine trials were conducted in which we determined whether oral administration of B. breve and LA results in production of c9, t11 CLA In Vivo in the gut. LA supplemented diets (1%) were fed in combination with B. breve (daily dose of 10 9 organisms) to healthy BALB/c mice (n=8), and to SCID (Severe Combined Immuno Deficient) mice, for 8-10 weeks. Feeding with LA without the B. breve strain was used as a control. In addition, LA and B. breve were fed to weanling pigs for 21 days. At the end of the trials, tissue fatty acid composition was determined by gas liquid chromatography (GLC). Results: In comparison with controls, there was a statistically significant increase in c9, t11 CLA in the livers of the BALB/c mice, SCID mice, and pigs after feeding with LA in combination with B. breve, providing evidence that B. breve is active In Vivo and produces c9, t11 CLA from LA in the gastrointestinal tract. In addition to quantitative differences in CLA, an altered profile of polyunsaturated fatty acid (PUFA) composition was observed in mice receiving B. breve compared with controls. This included a significantly higher level of the omega-3 fatty acids, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) in adipose tissue. An anti-inflammatory effect was noted in the SCID mice receiving either B. breve or pure c9, t11 CLA which was reflected by a significant reduction in the proinflammatory cytokines tumour necrosis factor α (TNF- α) (p less than 0.05) and interferon-γ (IFN-γ). Conclusion: These data indicate that c9, t11 CLA is produced In Vivo following ingestion of B. breve, resulting in modulation of the fatty acid composition of the host, including significantly elevated c9, t11 CLA in the liver. Furthermore, the results suggest that some of the physiological effects of Bifidobacterium e.g. immunomodulatory properties may be associated with physiological interactions between probiotics and PUFA. The study shows that the presence of bifidobacteria in the gut with A-154 AGA Abstracts ability to produce bioactive fatty acid metabolites can be directly linked to modulatory effects on host fatty acid metabolism. 1039 Natalizumab Reduces the Hospitalization Rate in Moderate to Severe Crohn's Disease Patients: A Pooled Analysis of the ENACT-1 and ENCORE Studies Bruce E. Sands, Corey A. Siegel, Mike Spencer, Steven Hass OBJECTIVE: To investigate the impact of natalizumab (NAT) treatment on the rate of hospitalization during induction. BACKGROUND: While clinical trials in Crohn's disease (CD) are generally designed to assess the benefit of treatments on disease markers, one of the major aims of treatment is to prevent adverse health outcomes such as hospitalization, a major contributor to the cost of CD care. METHODS: The ENACT-1 and ENCORE trials randomized patients with moderate to severe CD (Crohn's Disease Activity Index [CDAI] 220-450) to a 12-week induction period with intravenous infusions of NAT 300 mg or placebo (PBO) every 4 weeks. Data from both studies were combined and hospitalizations were identified from adverse event reports. Two independent experts, blinded to treatment allocation, identified the hospitalizations that were CD related. The rate of all-cause and CD-related hospitalizations (hospital admissions per 100 induction courses [/100IC]) over the induction period were analyzed using Poisson regression, correcting for study, potential confounders that were univariate predictors of hospitalization (CDAI, log C-reactive protein [CRP], body mass index), and time in study. Further analyses were performed in an a priori defined subpopulation (the “restricted population,” n=346) that had failed prior anti-TNF alpha therapy (ie, unresponsive, lost response, or intolerant) and had evidence of active inflammation (CRP >2.87 mg/L). RESULTS: A total of 1,373 patients (of 1,414 available) had data on all covariates and comprised the analysis population (mean age 38 years, 58% female, hospitalizations=134). NAT was associated with a 34% reduction in the rate of all- cause hospitalization compared to PBO (7.3/100IC vs 11/100IC, P=0.026) and a comparable reduction (30%) in CD-related hospitalization (5.7/100IC vs 8.1/100IC, P=0.08). Prior TNF alpha use and active inflammation (log CRP) were significant predictors of hospitalization (P<0.001, P=0.003). The PBO rate in the restricted population was 2.5 times that in the remainder of the population (20.8/100IC vs 7.9/100IC). NAT-treated patients in the restricted population had all-cause hospitalization rates 47% (9.7/100IC vs 20.8/100IC, P=0.014) and CD-related hospitalization rates 49% those of their PBO-treated counterparts (6.3/100IC vs 12.8/100IC, P=0.04). CONCLUSIONS: NAT was associated with a reduction in the rate of all-cause and CD-related hospitalizations compared to PBO-treated patients. This effect appeared to be greater in patients who had failed prior anti-TNF alpha therapy and had evidence of active inflammation. This patient subset was observed to have the highest hospitalization risk. 1040 Have the Phenotype and the Need for Colectomy in Paediatric Ulcerative Colitis Patients Changed Over the Last 44 Years? - a Population Based Dccd, Danish Crohn Colitis Database Study Christian Jakobsen, Anders Paerregaard, Pia S. Munkholm, Vibeke Wewer Aim: To describe the development in incidence, disease localisation, disease course and surgery in two Danish paediatric IBD population based cohorts from 1962-1987 (period I) and 1998-2006 (period II) in Copenhagen County. Material and Methods: Incident IBD patients < 15 years were included. The median background population (< 15 years) in period I was 135.235 and in period II 117.216. Incidence rates were calculated per 100.000 person-years under the age of 15. Disease localisation was classified according to the Montreal classification for UC and into small bowel only, large bowel only and small+large bowel for CD patients. Disease course and need of surgery (colectomy for UC and bowel resections for CD) were assessed during the first year after the year of diagnosis . Results: A total of 119 IBD patients (77 UC and 42 CD) were included in the two periods. Incidence rates were 2.0/100.000 and 0.2/100.000 for UC and CD in period I and 1.6 and 3.1 in period II, with incidence rate ratios of 0.81 (95% CI: 0.5-1.4) and 15.6 (95% CI: 7.5-32.7) for UC and CD. A significant increase of UC patients with extensive disease E3 (46.7 % and 94.1 %, p<0.001) and a decrease of patients with proctitis E1 (28.3 % and 0 %, p=0.017) was seen in period I and II, respectively. No significant difference was found in disease localisation in CD patients between the two periods. Disease activity in the two periods is shown in table 1.In 17 UC patients (28.3%) colectomy was performed in period I compared to none in period II (p=0.017). No significant difference in the use of surgery was found for CD patients between the two periods. During period II 6/7 (86%) UC patients with activity received azathioprine and none received anti-TNFα during the first year after dia- gnosis.Conclusion: A 15 fold increase in incidence of CD was seen over the last 45 years. Despite more extensive disease in UC patients in period II, we found a similar activity within the first year in both periods, but a lower colectomy rate in period II. The latter may be caused by the introduction of immune modulating drugs in period II thus leading to a change in disease course in childhood UC. Table 1. Disease activity during first year after the year of diagnosis of UC and CD in period I and II. * p=NS, comparing activity in UC and CD patients between period I and II.