(HT29,SW 1116 and Colo 320) and p21WAF1 expression has been lost, as well as APC,p53, p73, c-myc, c-Ki-ras, survivin genes expressed in SWl116 and Colo-320 cells before treat- ment. After lmm but not 10mm of 5-aza-dC for 24hours, the methylation status of the promoter of the p161NK4A gene significantly decreased and mRNA expression increased in HT29 ceils. In SW1116 and Colo-320 cells, the expression of p161NK4A and APC genes were markedly increased after treatment of 5-aza-dC, although 5-aza-dC treatment did not activate the expression of p21WAF1 gene. Interestingly, the treatment of TSA or sodium butyrate resulted in the siguificantly over-expression of p21WAF1 in these two cell lines. Most genes including p53, p73, c-myc, c-Ki-ras and survivin genes did not change following 5-aza-dC or TSA or sodium butyrate treatment. In addition, 5-aza-dC treatment failed to affect cell cycle distribution, but TSA or sodium butyl'ate blocked cells mainly in the G1 phases. Conclusion: The expression of pl61NK4A and APC genes were regulated by DNA methylation, and histone hyperacetylation is the major mechanism of up-regulation of p21WAF1 gene, and the transcription level of p53, p73, c-myc, c-Ki-ras, survivin genes were not regulated by either methylation or/and histone acetylation in two (or three) human colonic cancer cells. In these two cell lines, histone hyperacetylation causes a G1 cell cycle arrest. $987 in Vivo Activation of The P53-Mdm2 System in Microsate|lite Unstable-Hmlhl Defective Colorectal Cancers Luigi Ricciardiello, Graziella Angiolini, Claudio Ceccarefli, DonateUa Sammi, Enrico Roda, Franco Bazzoli MicrosateUite instability (MS1) is characterized by insertions, deletions or mutations of single or multiple nucleotides at highly repetitive-short sequences. This phenomenon is due to lack of function of the DNA mismatch repair (MMR) system. MSI tumors represent 15% of all colorectal cancers while the majority display chromosomal instability (CIN) characterized by gross chromosomal abnormalities and loss of beterozygosity. CIN tumors often lack p53 function, which is mediated by the MDM2 co-factor: when MDM2 is activated the protein promotes p53 degradation. AIM: to evaluate p53 and MDM2 proteins in colorectal cancers with and without microsatellite instability. Methods: Thirty one consecutive colorectalcancers were evaluated for MSI by PCR with BAT25 and BAT26 mononucleotide markers using an automated sequencer for fragment analysis. MSI-H tumors were defined by having both markers mutated (N=4). In this case those having one marker mutated were defined as MS1-L (N = 10) and those negative were MSS (N = 17). Nuclear immunostaining for hMLH1, hMSH2, p53 and MDM2 proteins was quantified by image cytometry with Cytometrica software. A labelling index was expressed as the percentage of the labelled nuclear area over the total neopastic nuclear area in the section. Results: In four cases both MSI markers were mutated (12.9%). In all cases there was loss of hMLH1 expression, while hMSH2 was normally expressed. MSI-H tumors showed reduced p53 expression (Labelling index 51.3%), while in the MSI-L and MSS groups the protein showed a higher expression (75,2% and 65 2%. respectively). MDM2 expression was higher in MS1-MMR in the MS1-H group compared to MSI-L (1,9% vs. 0.5%, P=0.023), while no difference was found in between those groups and MSS cases. Conclusions: in MMR defective-MSl positive colorectal cancers our results suggest a higher activation of the complex p53-MDM2. This indicates an attempt to repair DNA errors and preserve the cell from genomic derangements. $988 No-Donating NSAIDs Are More Potent than Traditional NSAIDs in Inhibiting The Growth of Cultured Human Colon Cancer Cells Raymond Yeh, Jennie L. Williams, Srinivas Kalala, Thomas R. Hundley, Jie Cben, Khosrow Kashfi, Basil Rigas INTRODUCTION: NO-donating NSAiDs (NO-NSAIDs) are novel drugs consisting of a traditional NSAID linked to a NO-donating moiety. They have enhanced safety and efficacy compared to their traditional counterparts. In animal models, NO-aspirin (NO-ASA) pre- vented the formation of aberrant crypt loci, a precursor lesion of colon cancer. Here, we examine the effect of NO-NSAIDs from various classes, on human colon cancer cells. METHODS: HT-29 (COX positive) and HCT 15 (COX negative) human colon cancer cells were treated for 48 hours with NO-(salicylic acid, ASA, sulindac, ibuprofen, piroxicam, flurbiprofen, indomethacin) and compared to their traditional counterparts in a concentra- tion-dependent manner. IC50 values were calculated. We determined cell proliferation (PCNA expression), apoptosis (DAP1 staining, sub-G0/G1 peak), and cell cycle (flow cytome- try of PI stained cells). Imracellular NO levels were monitored using DAF 2 and measuring fluorescence. We evaluated morphology by light and electron microscopy. For comparison, we also evaluated the effect of NO-ASA and ASA on HUV-EC cells. RESULTS: All NO- NSAIDs were more potent than their parent compounds in inhibiting the growth of these cell lines. The 1C50s of the NO-NSAIDs were lower than those of the corresponding NSAIDs by a factor of up to 2,000. NO-NSAIDs inhibited cell proliferation, induced apoptusis and akered their cell cycle phase distribution. DAPI stained cells showed a significant population of atypical cells that were prominent in NO-ASA treated cells. We have previously reported these atypical cells with NO-ASA in colon cancer cells; these cells are unique to NO-ASA and independent of cell type, as they were observed in HUV-EC cells. Light microscopy showed pronounced alterations in cellular morphology; closer examination using electron microscopy revealed nuclear disintegration. Treatment of cells with NO-NSAIDs was associ- ated with NO release. CONCLUSION: All NO-NSAIDs tested were significantly more potent and effective than their corresponding parent compounds in inhibiting human colon cancer cells, indicating a generalized property. These effects were seen in both HT-29 and HCT 15 cells, suggesting a COX independent effect. These agents exercise pleiotropic effects involving cell renewal and cell death as well as cell cycle phase transitions. Our findings suggest that NO-NSAIDs possess siguthcant chemopreventive and/or chemotherapeutic activ- ity agaifist human colon cancer. $989 Egf-Receptor Related Protein (ERRP): A Potential Therapeutic Agent for Coloreceal and Prostate Cancers Dorota Marcmiak, Ramzi Mohammad, Lathika Moraguda, Yingjie Yu, Amro Aboukameel, Volkan Adsay, Arun K. Rishi, Fazlul H. Sarkar, Adhip P. N. Majumdar EGF-receptor (EGFR) is frequently implicated in epithelial cancers, and is, therefore, being considered as a potential target for therapy. Recently, we reported the isolation and character- ization of a negative regulator of EGFR, termed ERRP (EGFR Related Protein; Am.J.Phy- siol.Cen Physiol.280: C1083,2001). To determine whether ERRP could be an effective therapeutic agent for co[orecta[ and prostate cancers, we generated ERRP fusion protein in Drosophila expression system and studied its effect on the growth of colon and prostate cancer cells in vivo and in vitro. We also determined expression of ERRP in colorectal and prostate cancers using rabbbit polyclona[ antibodies against ERRP. ERRP was found to be high in benign human colonic and prostate epithelium but low in invasive adenocarcinomas. Exposure of colon (HCT-116 and Caco-2) or prostate (PC-3) cancer cells to purified recombi- nant ERRP caused a marked inhibition of proliferation as well as attenuation of basal and ligand-induced stimulation of EGFR phosphorylation. ERRP-induced inhibition of prolifera- tion could be prevented by ERRP antibodies. Reduced EGFR phosphorylation was partly due to sequestration of EGFR ligands by ERRP, resulting in the formation of inactive heterodimers with EGFR. Intratumoral or subcutaneous (away from the tumor site) injections of purified ERRP caused regression of palpable tumors/n SCID mice xenografts of HCT- 116 cells in some animals and arrested tumor growth in others. We propose that ERRP inhibits EGFR signaling processes and is an effective therapeutic agent for co[orectal and prostate cancers. $990 Intra-vilious Adenomas in the Duodenum of Patients with Familial Adenomatous Polyposis (FAP): Implications for the Origins of Adenomas and for the Relevance of Mouse Models for Intestinal Neoplasia Marco NoveUi, Scan L. Preston, Evangelns Kyriakides, Dab.mane Oukrif, fan Tomhnson. Richard Poulson, Nicholas A. Wright lntroduction:Careinomas in the duodenum of patients with FAP are well-recogmsed, and am held to follow the adenoma/careinoma sequence, but little is known of their natural history. Many well-known genetic mouse models of intestinal neoplasia, such as the Min mouse, show lesions which are predominantly in the small bowel; as the small bowel is rarely involved in humans, such models have been criticised on their relevance to human disease. Methods:We examined pancreatico-duodenal resections from three FAP patieots. Serial sections of multiple blocks were stained with H&E, by immunohistocbemistry for [3- catenin and gene products which constitute downstream targets of 13-catenin-mediated transcription (CD44, c-MYC and p21CIP1/WAF1), Cell 2002; 111: 241-250) and for lyso- zyme for Paneth cells. Results:We found that (i) microadenomas are extremely common in the duodenum of FAP patients; (if) in routine section these adenomas grow and ramify in the core of duodenal villi, forming intra-villous lesions not dissimilar from those seen in the Min mouse; (iii) many lesions appear to arise as monocryptal adenomas - lesions occupying only one crypt; (iv) these lesions grow by a process of crypt fission and branching; (v) the neoplastic tubules open out onto both the surfaces of villi and into the inter-villous basins, and neoplastic cells are seen migrating onto multiple villi from a single adenomatous tubule; (vi) migrating cells appear to lose the overt neoplastic phenotype during the migration process, and mature (vii) Paneth cells lose positional information, and are found throughout the tubules. Conclusions:These findings suggest that (i) duodenal adenomas are very common in FAP, and originate in monocryptal adenomas and thus follow the 'bottom-up' rather than the 'top-down' model of morphogenesis, growing by crypt fission and branching; (ii) the presence of such intra-villous adenomas, which resemble the Min mouse vindicates it as a model for FAP; (iii) adenomatous cells can migrate unto normal villi, displacing the inthgenous cell populations and apparently differentiate and (iv) the defects in the positioning of Paneth cells suggests disruption of the Ephrin-B: EphB3 receptor system (Cell 2002; 111: 251- 263). The availability of such duodenal lesions in FAP now form an important model for looking at the molecular pathology of intestinal neoplasia. $991 Dendritic Cell-Tumor Fusion Vaccine is Superior in Protecting Mice Against Colon Cancer Compared to Temor-Lysate Pulsed Dendritic Cell Vaccine John Y. Kao, Chnan-Min Chen, Ynsong Gong, Jian-Jun Chen Background: Dendritic-cell antitumor vaccine is a promising cancer immunotherapy that is highly tumor specific. For cancers without known tumor-specific antigens, DC vaccine can be generated by stimulating DCs with whole tumor antigens in the form of tumor lysate or by fusion of dendritic cells and live tumor cells. However, it is still unknown which type of DC vaccine is more effective to indcuce anti-tumor immunity. While tumor-lysate pulsed DCs stimulate CD8 + T cells, fusion vaccine has the theoretic advantage of priming both CD4+ and CD8+ T cells utilizing MHC class I and If molecules. Aim: our goal of the study is to compare the relative protective immunity against colon cancer in mice vaccinated with DC-tumor fusion cells and with tumor lysated-pulsed DCs. Methods: Bone marrow- derived DCs were enriched on day 7 (90% purity by flow cytometry) and were either fused (polyethylene glycoI method) to mouse colon adenocarcinoma cells, CT26, or pulsed with CT26 tumor lysate (freeze-thawed method) overnight. For the fusion cells, the fusion effi- ciency was approximately 50% and the fusion cells were irradiated (16K rad) before use. Animals (n = 27 per group) were immunized on day 0 and 14 followed by tumor challenge (2x106 cells) on day 21. Tumor incidence and animal survival were recorded. Results: Dendritic-CT26 fusion vaccine was superior in protecting mice against subsequent tumor challenge compared to CT26-1ysate pulsed DC vaccine (tumor incidence: 23% vs 70%, p<O.05, and survival: 77% vs 30%, p<O.05). The protective immunity is durable in 9 out of 12 animals rechallenged at 1 year post immunization. Conclusions: Dendritic cell-tumor AGA Abstracts A- 132