AG

AA

bst

ract

s

775

Influence of the Cannabinoid Type 1 Receptor Antagonist Rimonabant OnEsophageal Function in ManEmidio Scarpellini, Rita Vos, Pieter Vanden Berghe, Jan F. Tack

Background: Cannabinoid type 1 (CB1) receptors have been implicated in the control oftransient lower esophageal sphincter (LES) relaxations (TLESRs) in animals. In man, non-scpecific cannabinoid receptor agonism was found to inhibit TLESRS, but it is unclearwhether CB1 receptors are involved in the control of esophageal function. Aim: To studythe effects of the CB1 receptor antagonist rimonabant on fasting and postprandial LESfunction in healthy subjects. Methods: Eleven healthy volunteers underwent two esophagealsleeve manometry studies after 3 days premedication with placebo or rimonabant 20 mg/day. Drug was administered after 30 minutes and a standardized meal was administeredafter 90 minutes, with measurements continuing to 120 minutes postprandially. Throughoutthe study, 10 wet swallows were administered at 30-minute intervals. Results: Intragastricpressure was significantly enhanced by pretreatment with rimonabant (24.5±1.5 vs. 20.4±1.7mmHg, p=0.01). Rimonabant did not significantly alter preprandial LES pressure (20.3±4.0vs. 17.3±3.0 mmHg, NS), but postprandial LES pressures were significantly enhanced(ANOVA p<0.01), with post-hoc significant differences at 30, 60, 90, and 120 minutes afterthe end of the meal (respectively 17.1±2.7 vs. 10.0±1.7; 17.8±3.3 vs. 10.2±1.6; 19.3±3.6vs. 10.4±1.3 mmHg and 19.6±3.1 vs. 8.6±3.0; all p<0.05). Swallow-induced relaxations(97±1 vs. 97±1%, NS) and amplitude of peristaltic contractions were not altered, butrimonabant significantly increased the duration of peristaltic contraction waves at all timepoints (e.g. 5.0±0.3 vs. 8.0±0.3 sec preprandially and 5.0±0.2 vs. 8.2±0.3 sec at 30 minpostprandially, both p<0.01). The tendency of a decreased number of TLESRs after rimonab-ant did not reach statistical significance (TLESRs total 2.3±0.8 vs. 3.3±0.5, p=0.08; prepran-dially 0.6±0.2 vs.1.3±0.3, p=0.14; postprandially 1.6±0.8 vs. 2.0±0.4, p=0.10). Conclusion:The CB1 receptor antagonist rimonabant enhances postprandial LES pressure and tends todecrease TLESRs in healthy subjects.

776

Novel Fluorescence-Labeled Peptide for Targeting Intra-Epithelial Neoplasia inBarrett's EsophagusMeng Li, Chris M. Komarck, Yong M. Kwon, Thomas D. Wang

Background: Peptides are promising for use as molecular probes to target neoplasia in thedigestive tract. We have previously demonstrated the use of fluorescence-labeled peptidesto identify colonic dysplasia In Vivo on confocal microendoscopy. Aims: We aim to selectan affinity peptide that binds preferentially to cell surface targets for future use in surveillanceof intra-epithelial neoplasia in Barrett's esophagus. Methods: We perform peptide selectionusing a 7mer phage display library with a diversity of 2x109. First, non-specific binderswere cleared by biopanning against cultured QhTERT (intestinal metaplasia) cells. Thetarget phage were identified by biopanning the supernatant against cultured OE33 (humanesophageal adenocarcinoma) cells for 5 rounds. Binding specificity was validated on fluores-cence microscopy, phage binding counts, and ELISA for OE33 (target) compared to thatfor QhTERT (control) cells. On fluorescence microscopy, the target phage was incubatedwith the cells using a FITC-labeled antibody, and binding to the plasma membrane wasquantified using NIH Image J. The antibody used in ELISA was horseradish peroxidase-labeled anti-phage monoclonal antibody. Results: A total of 60 randomly selected candidateswere sequenced from the enriched pool of candidate phage, and we found that 40 of theseclones revealed the peptide sequence SNFYMPL. On fluorescence microscopy, binding ofthis target peptide was localized to the plasma membrane of OE33 cells, and the meanintensity for OE33 and QhTERT cells was 69±18 and 26±3, respectively, resulting in atarget-to-background ratio of 2.8. On phage binding counts, the number of the target peptidespecific for OE33 cells was greater by a factor of 250 compared to that for wild type, andgreater by a factor of 25 compared to that for QhTERT cells. On ELISA, the optical density(OD) for binding of the target peptide to OE33 cells is greater by a factor of 2 comparedto that of wild type phage. A much lower OD was observed for binding to QhTERT cells.Conclusions: We have identified and demonstrated preferential binding of the affinity peptideSNFYMPL for esophageal adenocarcinoma cells that can be used for future In Vivo surveillanceof intra-epithelial neoplasia in Barrett's esophagus.

A-122AGA Abstracts

A) Binding of target peptide to OE33 esophageal adenocarcinoma cells. B) DAPI stainidentifies cell nuclei. C) Overlay image registers binding to cell surface targets.

777

Label-Free Nuclear Morphology Measurements of Dysplasia in the Egda RatModel Using Angle-Resolved Low Coherence InterferometryAdam Wax, Yizheng Zhu, Neil G. Terry, Xiaoxin L. Chen, Steven C. Gebhart, William J.Brown

BACKGROUND: Tissue characterization is still performed primarily by physical biopsy withmicroscopic analysis of fixed tissue. Angle-resolved low coherence interferometry (a/LCI)obtains real time nuclear morphology measurements of epithelial tissues by combining thedepth resolution of optical coherence tomography with the sensitivity to nuclear morphologyof light scattering spectroscopy. We report the first use of a/LCI to measure nuclear morpho-logy for detection of dysplasia in the esophagogastro-duodenal anastomosis (EGDA) rat modelof BE. METHODS: In the rat EGDA model, the gastroesophageal junction is anastomosised tothe duodenum to create a condition similar to Barrett's esophagus in humans. The surgeryalong with iron supplementation leads to esophageal adenocarcinoma (EAC) with an incid-ence of ~80% at 40 weeks after surgery. Forty male Sprague-Dawley rats (6-wks-old) weighing~200 g were purchased from Harlan Inc. After two weeks of acclimatization, 31 rats receivedEGDA surgery. Thirteen animals (10 treated, 3 control) were sacrificed at each of three timepoints (12 weeks, 20 weeks, and 30 weeks), with one lost to surgery. Immediately aftersacrificing, the esophagus was removed, opened longitudinally, and a/LCI nuclear morpho-logy data acquired from 1 mm2 points beginning at the gastric junction and moving upwardsin 5 mm increments. A tissue sample was taken at each imaged point, fixed in 10% bufferedformalin, and later processed for histological examination. The primary endpoints were thesensitivity and specificity of a/LCI for identifying dysplasia. RESULTS: Based on the firsttwo time points, 143 paired tissue specimens and a/LCI images were taken in 26 rats.Measured nuclear sizes ranged from 6 μm to 15.5 μm. By setting a decision line of 8.9 μmfor the week 12 animals and 9.0 μm for the week 20 animals, we achieved 85% sensitivity(63/74 samples) and 97% specificity (67/69) for distinguishing dysplasia from non-dysplastictissue. Results from the third time point will be processed and added to the study. CONCLU-SION: In this first pilot study, a/LCI demonstrated good sensitivity and specificity fordetecting dysplasia based on nuclear morphology in EGDA rat model of BE.

778

Cytokeratin Assisted Laser Scanning Cytometry (LSC) Permits FullyAutomated Detection of Aneuploidy in Fixed Barrett's BiopsiesAjay Bansal, Joyce G. Slusser, Sharad C. Mathur, Amit Rastogi, Sachin B. Wani, VikasSingh, Mandeep Singh, Prateek Sharma

Background: DNA aneuploidy as measured by flow cytometry has been shown to be apredictor of progression in Barrett's Esophagus (BE) in prospective, phase 4 biomarkerstudies. However, flow cytometry requires biopsies in special media and is technicallydemanding. To date, LSC performed on fixed tissues for the detection of aneuploidy hasnot yet been fully automated [requires calculation of the ratio of DNA content betweenmanually selected epithelial and non-epithelial (stromal) cells to calculate the diploid index(DI)] and suffers from selection bias of manually selected cells. Aim : To evaluate thefeasibility of fully automated detection of aneuploidy by cytokeratin assisted LSC in fixedBarrett's biopsies. Methods: In this pilot study, biopsies of BE patients collected in Bouin'sfixative were retrieved. 5 μm thin sections were deparaffinized and rehydrated followed byDAPI nuclear staining. Cytokeratin staining was performed (Pancytokeratin Ab, BD Pharma,1:200 dilution) to distinguish epithelial from stromal cells (to avoid the need for manualselection). The cytometer was configured using 40X objective and 5 mW of Argon laserpower to calculate the DI for cytokeratin positive and negative cells for the entire slide inan automated fashion. Deviation from the ratio of 1.0 indicated the degree of aneuploidy[mild (DI 1.1-1.3),moderate (DI 1.3-1.8) and aneuploidy (DI >1.8)]. Results: Twenty-fourbiopsies obtained from 12 BE patients (mean age 64 yrs) were evaluated - 6 patients withintestinal metaplasia (IM) and 6 with high-grade dysplasia (HGD). All patients were malesand Caucasian with a mean BE length of 3 cm. Cytokeratin LSC could be accomplished in10/12 (83.3%) patients whereas sections from 2 patients were dry and folded making theresults uninterpretable. The mean DI in the biopsies with IM was 1.48±0.07 compared to2.28±0.30 in those with HGD (p=0.007). Two patients with IM had mild, three had moderateand none had severe aneuploidy. One patient with HGD had moderate while four had severeaneuploidy. Conclusion: The results of our pilot project demonstrate that cytokeratin assistedLSC allows for fully automated detection of aneuploidy in fixed BE tissues. Furthermore,cytokeratin staining of epithelial cells for LCM a) obviates the need formanually distinguishingepithelial from stromal cells b) avoids selection bias as the entire cellular population on theslide is used to generate an overall diploid index. This fully automated technique needs tobe further tested and validated as an easy and effective method of aneuploidy analysis inBE biomarker studies with the potential for wide scale clinical applicability.