of mortality. Odds ratios for death at 90, 180 and 365 days, comparing patients with low (<29) vs normal HDL Ch, were 8.8, 8.7 and 9.0, respectively (all p < 0.001). Conclusion: Low HDL Ch in cirrhosis is strongly indicative of poor prognosis. A prospective study to assess the~value of HDL as an independent predictor of pretransplant survival is in progress.

194

Regional Cerebral Blood Flow During Gastric Balloon Distention in Functional Dyspepsia Jons Vandenbergbe, Patrick Dupont, Philippe Persoons, Koen Demyttenaere, Jan Tack

Introduction: Recently, we demonstrated that proximal gastric balloon distention activates components of the lateral (left postcemral gyrus) and medial (right anterior cinguhate) pain system in healthy volunteers (Vandenberghe et al., DDW 2002). Increasing evidence suggests inv01vement of both pain systems in normal visceral sensitivity. Hypersensitivity to gastric distention is observed in a subset of Functional Dyspepsia (FD) patients. Imtable Bowel Syndrome (IBS) is another condition characterized by visceral hypersensitivity. In IBS, brain imaging during painful rectal distention showed aberrant activation, suggesting abnormal CNS processing of visceral stimuli. We hypothesized aberrant brain activation in FD with hypersensitivity to gastric distention. Methods: Brain l~O-water 3D Positron Emission Tomog- raphy (PET) was performed in 8 FD patients with symptoms of gastric hypersensitivity daring 3 conditions: no distention, sham distention and isobaric distention to unpleasant or painful sensation. Pain, discomfort, nausea and bloating during maximal distention were rated on Visual Analogue Scales (VAS). Data were analysed using Statistical Parametrical Mapping (SPM99, subtraction analysis, co-registered with MRI-scans). Results: Threshold for unpleasant sensation was 6 .3- -3 .3 mmHg above MDP. Pain (7.6 +- 2.1), bloating (7.6-+2.5) and discomfort (8.2+- 1.0) were predominant over nausea (5.6_+3.2). At the corrected p<.05 level, subtraction analysis ('mammal distention' minus 'no distention') showed activations in bilateral frontal infetior gyri (Brodmann area 47) and left postcentral gyms (primary somatosensory cortex). 'Sham distention' minus 'no distention' showed no activations. Conclusions: Similar to healthy volunteers, distention of the proximal stomach in FD activates components of the lateral pain system (left postcentral gyrus) and frontal inferior gyn (possibly involved in regulation of hunger, appetite, satiety). In FD, however, these activations occur at significantly lower distention pressures. In contrast to findings in normal visceral sensitivity, none of the components of the medial pain system were activated. Ths might reflect failure of activation in hypersensitive FD of descending antinociceptive pathways, originating in the anterior cingulate gyrus. Sham distentions showed no significant cerebral activation, questioning the role of anticipation and interpretation bias in FD.

195

kram of Achalasic Patients Decreases nNOS Expression in Myenteric Neurons of Normal Human Fundns Michel Neunlist, Stephanie Pimont, Jean Claude Le Neel, Jean Paul Galmiche, Stanisfas Bmley Des Varannes

In idiopathic achalasia, loss in nitrergic neurons as well as anti-myenteric neuronal antibodies have been reported. Although the disease primarily affects the esophagus, several studies have established that the proximal stomach is also involved. The a i m of this study was to determine whether the serum of achalasic patients can change the neurochemical coding of myenteric neurons in the normal human fundus. Methods: Fragments of fundus were obtained from patients undergoing surgery for esophageal cancer. Tissues were incubated with serum of control (n=5) or achalasic (n= 16) subjects for 16-18h in presence of c01chicine the last 12h of culture. The tissues were fixed in 4% paraformaldehyde and the myenteric plexus microdissected. Whole mounts of myenteri c plexus were stained with antibodies against neuron specific enolase (NSE), choline acetyltranferease (CHAT), nitric o~ade synthase (NOS), substance P (SP) and vasoactive intestinal polypeptide (VIP). Results: In control tissues, the proportion of the different populations of neurons (expressed as the % of NSE positive neurons) were as follows: CHAT = 57.2 + -5.6; NOS = 40.8 + -4.5; VIP = 19.6+-6.9; SP = 16.0+ -37. NOS and CHAT neurons preferentially contained a single neuromediator while, in contrast, V1 P and SP were usually colocalized with ChAT and N0S in the same neurons. When the serum of achalasic patients was added to the preparation, the proportions of nitrergic neurons significantly decreased and that of ChAT increased (by 33% and 20% as compared t o control, respectively). Further analysis of neurocberincally- identified subpopulations showed a significant decrease in the proportion of NOS only containing neurons (by 37%) and a significant increase in the proportion of ChAT/SP neurons (by 263 %) as compared to control. Conclusion: Serum of achalasic patients is able to ~uce neurochemical changes in normal human fundns which are reminiscent of those observed in patients with idiopathic achalasia. The identification of these presumably immune hctors are currently in progress.

196

^mplification and Immune Modulation in The Ac / cAMP Signaling Pathway Ltads to Hyperexcitability in Intrinsic Primary Afferent (AH) Neurons of Trichinella Spiralis Infected-Guinea Pigs J0tge Guzman, Jeffrey Palmer, Jun!Ge Yu, Zach Suntres, Jianjing Xue, Helen J. Cooke, Sheik Wisel, Robert E. Michler, Hamdy Hassanain, Fievos L. Christofi

Cyclic AMP is implicated in the hyperexcitability of jejunal myenteric Intrinsic Primary ~erent (IPAN) / AH neurons during acute inflammation with Trichinella Spiralis infection. SPECIFIC AIM. To elucidate the role of the adenylyl cycfase (AC) / cAMP signal transduction pathway in altered AH cell excitability during the acute phase of T. Spiralis infection. METHODS. Guinea pigs were orally inoculated with a 0.2 ml 0.85% saline bolus of 8 x 103 T spiralis muscle -stage larvae. Electrophysiofogical, immunochemical or cAMP analysis m isolated ganglia was done day 6-8 near peak AH hyperexcitability. RESULTS. Infected mmals (n=4) compared to age-matched controls (n=4) had 17 fold higher mast cell ~ptase (p= .016), 3.6 fold higher LTB4 (p = .0028), 2.6 fold higher PGE2 (p = .0001) and

78 fold higher TNFu (p =.0001 ). In AH cells, infection caused hyperexcitability and increases in action potentials (AP) / current pulse from 1.9 + _ 0.3 to 8 .7+0.7 (n=40, p<.0001), spontaneous APs in 67% and anodal breaks in 74% more neurons. Antagonists of both histamine H 1 and H2 receptors (lv, M pyrilamine / cimetidine,n = 6), a leukotriene antagonist (5~M WY48252, n = 6 ) or a cyclooxygenase inhibitor (lp.M meclofenamic acid, n = 4 ) all reduced AH hyperexcitability by more than 80% (p<.0001). Infection caused a 30 fold increase in the 147kDa AC protein band (n=4). Basal cAMP content was elevated 2 fold (p<.013), or 3 fold by Ro-20-1724 (10v, M) in the infected gut (p= .039). The cAMP response to histamine, substance P or forskolin (1-10p.M) was 2 to 3 fold higher in infected compared to control gut ( p=0 .03 ,n=4 each). In longitudinal muscle -myenteric plexus preparations, forskolin caused pCREB immunoreactivity (ir) in 90% of calbindin ir neurons / visual field compared to 6% in controls (n= 164 neurons, p<.0001); 70% more calbindin ir (AH) neurons had nuclear pCREB ir in infected gut than in control (n= 280, p<.0001). Forskolin increased a 47 kDa pCREB ir protein band by ~50 fold in normal gut (n = 1800 ganglia, p<.0001). CONCLUSION. In T. Spiralis infection, hyperexcitability in IPANS results from up-regulation and immune activation of AC/cAMP signaling leading to nuclear pCREB and alterations in cAMP-dependent gene expression. This signaling pathway likely has significant pathophysiological consequences in neural reflexes, motility and secretion during acute inflammation (NIH DK44179, DK37240, DK57016, DKr $1).

197

Strain-Dependent Effects of Probiotics on Intestinal Muscle Dysfunction in an Animal Model of Post-lnfective Irritable Bowel Syndrome Elena F. Verdu, Premysl Bercik, Patricia Blennerhassett, Xian Xi Huang, Gabriela Bergunzelli, Irene Corthesy-Theulaz, Stephen M. Collins

Background/Aim: Patients with Irritable Bowel Syndrome (IBS) have abnormal motor and sensory function. In this study we test the ability of probiotics to influence intestinal muscle dysfunction in a recently described mouse model of post-infective IBS, and explore underlying immunological mechanisms. Methods: Groups of N1H Swiss mice were garaged daily from day 10 to 21 post Trichinelha spiralis infection (Tspl) with 100# of bacteria growth medium (MRS) or 100pd of spent culture medium (SCM) containing 109 Lactobacillus paracasei (Lpa), Lactobacinus johnsonii (Ljo), Bif~obacter longum (BIo) or B. lactis (Bla). Additional mice received heat inactivated Lpa or bacteria free-Lpa SCM. Uninfected mice were garaged with MRS or Lpa. Contractility was measured on day 21 post infection (PI). Tissue was collected on day 21 PI for intestinal and mucosa-associated flora and gene array analysis. Longitudinal myen- teric plexus preparation (LMMP) was obtained on day 14 and 21 PI for cytokine protein (ELISA) and mRNA (RT-PCR) assays. Results: Lpa, but not Ljo, attenuated hypercontractility as assessed by a reduced area under the curve (AUC) after carbachol stimulation, compared to Tsp-MRS 21 days PI (p = 0.01). The two bifidobacteria strains tended to decrease the AUC compared to Tsp-MRS 21 days PI (p=0.09). Differential effects were recorded with live bacteria and free-Lpa SCM. The attenuation of muscle hypercontractility was paralleled by a 2-fold decrease in the secretion of IL-4 (p<0.0001), mRNA for TGF-IB (p =0.0001) and COX-2 (p = 0.001) in LMMP evident from day 14 P1, and by modulation of genes involved in innate defenses such as RANTES and cryptidin, as evaluated by germ army analysis. Normalization of PI contractility was independent of Lpa presence in the mucosa-associated flora. Conclusions: Probiotics exhibit strain dependent effects on muscle hypercontractility in an animal model of PI IBS. Improvement in intestinal muscle dysfunction by Lpa and free-Lpa SCM is likely due to attenuation of cytokine and inflammatory mediator production in the mnscufaris externa and modulation of innate defense genes in the small intestine. Supported by a grant by Nestec.

198

Do food elimination diets improve Irritable Bowel Syndrome? A double blind trial based on IgG antibodies to food W. Atkinson, R. Gumey, T. A. Sheldon, P. J. Whorwell

Introduction: Many patients with imtabha bowel syndrome (IBS) feel that they have some form of dietary allergy or intolerance. IgE mediated food reactions (classic allergies) are probably rare in 1BS but little attention has been paid to the potential role of IgG responses. This study aimed to assess the efficacy of an exclusion diet based on testing for the presence of igG food antibodies in patients with IBS. Methods: A double blind randomiscd controlled trial was undertaken in which 150 unselected out-patients with IBS (all subtypes) were randomised to receive either a diet excluding, for 3 months, all foods to which they had IgG antibodies (titre -> 3:1) or a sham diet excluding the same number o1 foods but not those to which they were sensitive. Symptom severity, non colonic symptoms, anxiety/ depression and quality of life were recorded at 0 and 3 months, and the primary outcome measure was a change in symptom severity score. Global outcome was also recorded on a seven point scale with only 'better' and 'excellent' regarded as an improvement. Patients who withdrew before the end of the study were also assessed after 3 months. Analysis was by 'intention to treat' using a generalised linear model for severity and ordinal regression for global outcomes in SPSS. Results: The true diet was significantly superior to the sham diet in reducing symptom severity scores (average reduction 34; 95% CI: 17.3, 68.6; p = 0.049). However, response to the diet was significantly affected by dietary adherence and the number of foods to which the patient was sensitive. When these factors were accounted for in the analysis, the difference in symptom scores rose to 89 (95% CI: 41, 137; p < 0.001). It is of interest that adherence to the diet affected the response observed in patients on the true diet, but not those on the sham diet (p=0.038). The analysis also revealed a significant difference in favour of the true diet with respect to global symptomatology (p = 0.007). All other outcome measures showed a trend towards benefit, but did not reach statistical significance. Conclusion: A clinically significant improvement can be achieved in some patients with IBS using a food elimination diet based on IgG food antibodies. The number needed to treat is 3-4.

A - 2 9 A G A A b s t r a c t s