l~d:LoA. Mutating 4 or 2 bases through a C-rich element (-79 GCCCCTCCC -70) reduced the responses by ->65% and ->40%, respectively. Mutations in other regions of the promoter caused a less marked reduction of the responses, in EMSAs, a nnnrber of complexes bound a wild type (wt) labeled probe, spanning -86 m -64bp, and binding was prevented by inchision at excess unlabeled wt probe. For one complex, mutations of two or more bases through the region -79 to -70 prevented unlabeled probe competing Ibr binding to the complex. Tbe intensity of this complex was not altered by gastrin. De-incuhation of nuclear extract with antibodies to two members of the Sp family of transcription factors, Sp-3 and Sp-4, disrupted binding to the wt probe. Conclusions: 1. Gastrin stimulates Reg expression via activation of RhoA. 2 A C-rich redan (-78 to -71) is critical to the gastrin stimulation of Reg expression 3 Sp family transcription factors bind to this region of the Peg promoter.

T1026

Differential Expression of Two Alternative Spliced Variants of Fibroblast Growth Factor Receptor 2 (FGFR2) in Gastric Fundns of Rats Pu-Qing Yuam S V. Wu, David R. Scott, George Sachs, Gordon V Ohning

Background: The effects of keratinocyte growth factor (KGF) on proliferation, differentiation, wound repair and tumor growth in stomach are mediated by" KGE receptor (KGFR), a member of FGFR tamdy. ~ k gene encodes two spliced variants of FGFR2: FGFR21Ilb (lIIb), also known as KGFR, binds KGF, and FGFR2Illc (lIlc) binds FGF2 and FGF1, but not KGF Pew data are currently available regarding the tissue-specific distribution of these FGER2 spliced variants in stomach. Aims: To ct~racterize the differential expression of two FGPR2 spliced variants m rat tundus and determine if KGFR is present in proliferating zone of oxyntic mucosa. Methods: Total RNA was extracted separately" from whole fnndic tissue strips, oD'ntic mucosa or submucosal/musele layers. RT-PCR was carried out using specific primers ior lllb and lIlc of the knc~'n rat KGER2 sequences. Wild type and transformed MDCK cells were used as a positive control for lIlb and lIlc. To further determine the tissue- specific expression of lllb and lIIc in oxyntic mucosa, the surface epithdia (SE), the neck zone (NZ), the glandular middle portion (GM) and the glandular basal portion (GB) were captured with Laser Capture Microdtssection (LCM) System from 6 ~m cryostat hindus sections stained with H&E. Total RNA (1000 shots/sample) was extracted and processed for eDNA synthesis, tallowed by PCR for lllb and Illc. PCR for acidic ribosomal protein was used as a control for RNA quality and nora'mlization. Bek immunohistochemistry was performed with poly- and monoclonal anti-Bek antibody to demonstrate the Bek localition in oxyntic mucosa. Pesuhs Both lllb and lIlc were detected in whole fundic tissue and the kvef of lIIb Is significantly higher than that of Ilk. IIlb was expressed in oxyntic mucosa, but not in submucosal/muscle layers. RT-PCR analysis on the LCM-dissected tissues showed that llib was present in all layers and distributed in a distinct pattern as SE>GB>NZ>GM. IIIc was only detected in GM. Bek immunomactivity was localized in the cells of all layers. Particularly, the strot:gest Bek immunoreactivity was found in the neck zone. Conclusion: Two FGFR2 spliced variants were differentially expressed in fundns at which KGE, FGPI and PGF2 may generate specific tunctions at their respective target sites. Variations of KGFR distribution in oxyntic ntucosa are also consistent with nmltifunctional nature of KGF in different cell types Our data suggest that KGP may act directly on oxyntic mucosal prolit}ra- tion in the neck zone

T1027

Synergistic Cyclooxygenase-2 Expression by Gastrin and EGF bee W Slice, Elena Zhukova

Background. Over-expression of Cyclooxygenase-2 (COX-2), which is induced by multiple growth lactors, has been implicated in numerous human cancers. In turn, gastrin and EGF are growth factors tar multiple cell types. The aim of this study was to determine the signaling pafhways used by gastrin and EGF/or COX-2 expression. Methods CCKjgastnn receptor was expressed m Swiss 3T3 (S3T3/CCK_,R) cells by retroviral transduciion. Western and Northern blot analysis was used to measure COX-2 protein and mRNA expression. Enzyme immutmassays were used to measure PGEz secretion. Nuclear run-on assays were used to measure gene transcription rates. RNA stability assays were done using dichloroben- zinnde riboside (DRB) to bhx:k de nova RNA synthesis. Reanlts. Gastrin and EGF induced PGE~ secretion in S3T3/CCK2R ceils. Basal COX-2 expression in S3T3/CCK~R cells was very, low. Gasmn dose-dependently induced COX-2 protein, reaching a maximum level by 8 hr. The CCK, receptor antagonist, I_365,260, inhibited gasmn-dependent COX-2 expression. Gasmn also dose-dependently induced COX-2 mRNA, reaching a maximum between 2 and 4 hr and then decreasing. 1.365, 260, inhibited gastrin-induced COX-2 mRNA levels. Inhibi- tion at de nova protein synthesis by cyclobeximide did not block COX-2 mRNA induction by gastrin. Inhibition of PKC activity, ERK activation by MEK and EGF receptor activation did not block induction of either COX-2 protein or COX-2 mPvNA by gastnn Nuclear run- on assays showed 12-told, 2-laid and 5-fold increases in the COX-2 gene transcription rate relative to the GAPDH gene in cells treated with gastrin, EGF, and gastrin & EGF, respectivdy. COX-2 mRNA stabilization assays showed half-lives of 18 min, 91 min and 100 rain for COX-2 mRNA in DRB treated cells that were stimulated with gastrin, EGF, and gastrin & EGF, respectiveb;. Co-st~mulation with gastrin and EGF resulted in a synergistic increase in COX-2 mRNA, COX-2 protein, and PGE~ secretion Conclusions. Gastrin induces COX-2 expression in S3T3/CCK~R cells as an immediate/early gene response that results in PGE,~ ~cretion. Gasmn and EGF synergasticaUy induce PGE~ secretkm by induction of COX-2 expression through distinct molecular mechanisms, namely an increase in COX-2 gene transcription by gastrin and enhanced COX-2 mRNA stabitity by EGF DK35740, DK17294

T1028

Genome-wide Detection of Novel Transcriptional Target Genes of Gastrin in a Gastric Cell Line Shigeo Takaishi, John Fleming, Dana Frederick, Timothy Wang

Backgrom{d & Aim: Gastrin regulates both acid secretion and gastric epithelial proliferation / morphogenesis and has been implicated m gastric carcinogenesis. While a number of down- stream targets of gastrin, such as TGF-alpha, HB-EGF, Peg I, IL-8, Cox-2, PAl-2, MMP-9 and cyclin D1, have been identified, recent high-throughput m~cmarray technology' has allowed the possibility of genome-wide survey of potential transcriptional targets. Methods: Stable clrares of a gastric cell line AGS cell transfected with human gastrin/CCKB receptor were stimulated with or without gastrin (100riM), and 24 hours later total RNAs were extracted Fluorescence-labelled cRNA~ were hybridized to GeneChip human U133A and B array (Affymet>,:, C~.), on which a total of around 40,000 cDNAs were spotted. Experiments were repeated twice, and data were analyzed by Affymetrix Microarray Suite 5.0. Results: 1,445 cDNAs were up-regulated more than 2-fold and 1,558 cDNAs were down-regulated less than 0.5-ibid. Among these candidate target genes, we confirmed previous gasmn targets were included. The following 12 new target genes were selected, and all the changes of these genes were confirmed by" addtional methods such as not'them hybridization and/or realtime quarttitative RT-PCR: Up-regulated genes: MMP-10 26-fold, MMP-3 1 l-fold, TFF1 8-fold, Claudm-4 4-fold, IL6 receptor alpha 4-fold, gp130 2.5-fold, GKLF (KLF4) 2.5-fbld, IKLF(KLF5) 3.5-told; Down-lvgulated genes: Claudin-2 0.2-fold, PTTG(securin) 0.4-fold, STATi 0 4-1bid, p53 0.4-tbld; Treatment vAth gastrin/CCKB receptor antagonist YF476 (llxM) completely blocked these changes in gene expreasion by gastrin. Pharmacologacal signal inhibitors such as PD98059(MEK1 inhibitor), Go6976(PKC-alpha, beta & gamma inhibitor) and Rottlerin(PKC-deha & theta inhibitor) suppressed these regulations in a differential manner. For example, up-mgnlation of GKLF and IKLF were inhibited with Rottterm but not with PD98059 & G~6976, whereas up-regulation of IL6 receptor alpha and down-regulation of STAT1 were partially" inhibited with PD98059 & Go6976 but not with Rottlerin. Conclusions: We have identified possible candidate genes transcriptionally tvgnlated by gastrin in the almost whole human genome. Naval target genes such as GKLF, IKLF, IL6 receptor alpha and STAT1 seem to be regulated through distinct signal pathways.

T1029

Microarray Analysis of Gastrin-Regulated Gene Expression in Parietal Cells Penu N. Jain, Raren L Hinkle, Cynthia S. Bmnkan, Gina C. Bane, Catherine S Chew, Linda C. Samuelson

BACKGROUND & AIMS: Gastrtn-deficiem (GasKO) mice have a severe impairment in acid secretion, which can be partially repaired by gastrin replacement. The ann of this study was to evahiate changes in gene expression m parietal cells resulting fi'om the loss of gasmn METHODS: Parietal cells were isolated from 8 week-old male mice to > 90% punty by light-scatter ceil sorting from dispersed gasmc nrucosal cells. Parietal ceils fiom 10-14 GasKO or wild-type mice were pooled and three separate pooled RNA samples were used for gene expression microarray anal}sis using the Afl),metrix U74AV2 chip. The gene expressinn results obtained from microarray were validated by quantitative RT-PCR (Q-PCR), and protein expression was analyzed by western blot. RESULTS: Significant changes (p<0.01) were detected in 234 probe sets out of a total of 12,487 probes on the chip. Most of the changes in expression were small with only 26 genes showing a greater than twofold change. Selected genes were chosen, for further analysis, including those that may participate in gastric acid secretion (W/K+-ATPase, lasp-1 and H2 receptor), growth/proliferation of parietal ceils (hedgehog lamdy, amphiregulin, trefoil lactor), cellular cytoskeleton (ezrin, vilhil and radixin), and others (ADAMs family and annexins). Several genes were shown to have reduced expression in GasKO rnice. Expression of the alpha subunit of H+/K *-ATPase was markedly" reduced in GasKO mice as determined by Q-PCR and western blot analysis. I.asp-1 was also down regnlated in GasKO mice, as demonstrated by both microarray and western blot analysis CONCLUSION: Gastfin appears to be responslbk ~br up regnhtion of expression of genes involved in acid secretion and growth regulation. Reduction in the levels of H +/K + -ATPase and lasp- 1 could contribute to reduced acid secretion in GasKO mice

T1030

Cloning and Expression Profiling of Canine NK~, NK_, and NK3 Receptors Pieter Peeters, Benoit Moreaux, Ronald De Hoogt, Joke Van Den Berg Peter Verhasseh, .ann Meu/emans, Bernard Coulie

BACKGROUND: Studies on the role of NK, NK2 and NK receptors (NK~R, NKzR and NK~R) m visceral pain are limited to rodents At a pharmacological and molecular level human NK~R, NKeR and NK3R differ from their rodent counterparts. We hypothesized that canine neurokinm receptm~ might constitute closer homologues to the human neurokinin receptors. ALMS: 1) To identify the eDNA sequence of canine NK~R, NK2R and NK~R and to do sequence comparison acmas species. 2) To assess expression levels of neurokinin receptors in canine peripheral nervous and cola-rectal tissues. METHODS: Dog genomie libraries were screened with eDNA probes derived from human NK~R, NKeR and NK3R in order to isolate canine NK receptor related sequences. Based upon these sequences RACE was performed on RNA fi'om ascending colon and dorsal root ganglia resulting in full length eDNA encoding the canine NKjR, NK,R and NK3R. Real time quantitative RT-PCR was performed on RNA derived from dog peripheral nervons and cola-rectal tissues to quantify expression levels of neurokinin receptors. RESULTS: The amino acid sequences of dog NK/R, NK2R and NK3R resemble more closely" human NKtR, NK~R and NK~R (98%, 88 .%, and 91% homology, respectively) than those of mouse (96%, 86% and 74% homologous to human, respectively) and mt (96.%, 88.% and 88% homologona to human, respectively). High expression levels of NK~R were found in lumbar and sacral parts of the spinal cord and their respectively dorsal root ganglia. In addition lower expression levels were detected in cola-rectal tissues In contrast to NK~R, expression of NK~R was confined to cola-rectal tissues and urinary' bladder. Expression of NK3R was complementary to that of NK~R since it was only detected in central and peripheral nervous tissues. Centrally expression was

AGA Abstracts A-470