1086 the clonal origins of dysplasia from metaplasia in the human stomach

1
1078 Pediatric Intervention Study with a Probiotic Mixture (Bacilac Forte) in Acute Gastroenteritis Gigi Veereman, Sofie Staelens, Jan Wijffels The aim of this study was to evaluate the acceptability, safety & effect on diarrhea and accompanying symptoms of a probiotic mixture (Bacilac Forte) compared to placebo in children with a clinical diagnosis of mild acute gastroenteritis. The active product contained 3, 3 10 9CFU pro capsule of a probiotic mixture:20% Lactobacilus rhamnosus, 20% L acidophilus, 20% L casei, 20% L plantarum, 20% Bifidobacterium infantis. Study design was prospective DBPC in a primary care setting. IRB approval was obtained and subjects were included after parental consent. Otherwise healthy eutrophic children ages 1 mth-5 yrs, consulting their GP b/o liquid stools more than 3 tid for less than 48 h with dehydration less than 9% were included. Exclusion criteria were exposure to antibiotics or probiotics within 4 wks, bloody diarrhea & fever. Subjects received active or placebo tid. Parents kept a diary until D9 rating acceptability of the product, various parameters and any adverse event. Subjects were re-examined between D7-9. In total 48 subjects enrolled, 47 were analyzed (22 in the intervention group), 53% were girls, 91% Caucasian, age was 24+-15 mths, with no differences between groups. Average stool frequency was 5/d, 45% had cramps, 21% regurgitation, 45% vomiting and 23% upper respiratory symptoms. Statistical analysis used GLMM and non-parametric tests. Based on diary there was no difference in duration of diarrhea and stool frequency (although a sharper decrease in the active group) but stool consistency returned to normal more rapidly (D3) in the active group (p=0.0205). Also stools were more soft and formed over the observation period in the active group whereas subjects in the placebo group showed more variation (more hard and watery stools). During the course of the study there were no differences in cramps, regurgitation, vomiting, upper respiratory symptoms and appetite between groups. Acceptability of the capsules was rated easy by over 70% on all occasions for both groups, Tolererance was excellent: crying, flatulence, sleep and general behavior were rated similarly. Two adverse events (AE) were recorded in the intervention group: 1 otitis media and 1 gastroenteritis with dehydration necessitating hospitalization (frequency of AE not different for both groups by fisher exact p=0.203). In conclusion: tid administration of the probiotic mixture Bacilac Forte to young children with mild gastro-enteritis was safe, well tolerated and led to more rapid normalization of stool consistency with overall more formed and soft stools. 1079 A Probiotic Treatment (Lactobacillus farciminis) Attenuates the Hypothalamic- Pituitary-Adrenal (HPA) Axis Response to Acute Stress in Rats Afifa Ait-Belgnaoui, Helene Eutamene, Christel Salvador-Cartier, Henri Durand, Eric Houdeau, Jean Fioramonti, Lionel Bueno, Vassilia Theodorou Aims: We have previously shown that Lactobacillus farciminis (L. farciminis) treatment reduced stress-hypersensitivity to colorectal distension (CRD) in rats. Quantitative assessment of Fos expression in relevant spinal structure 1 indicated a decrease in the stress-induced activation/ sensitization of sensory neurons in response to CRD. Since L. farciminis reduces neuronal activation, we hypothesize that this probiotic may influence immediate HPA response to stress. This study aimed to evaluate the effect of L. farciminis on PRS-induced HPA response by assessing adrenocorticotropic hormone (ACTH) and corticosterone (CS) plasma levels and the number of corticotropin-releasing factor (CRF) positive neurons in the hypothalamic paraventricular nucleus (PVN). Materiel and Methods: In a first series of experiments, kinetics of plasma ACTH and CS were determined in 14 groups of 5 female Wistar rats receiving orally during 15 days either saline (groups 1-7) or 10 11 cfu/day of L. farciminis (groups 8- 14). All animals were submitted to PRS excepted groups 1 and 8 (control, no CRD and no stress). After 15, 30, 45, 60, 90 and 120 min of stress, under general anaesthesia blood was drawn from the abdominal aorta and ACTH and CS levels were determinated using a Lincoplex kit. In a second series of experiments, 4 groups of 4 female rats received orally during 15 days either saline (groups 1, 2) or 10 11 cfu/day of L. farciminis (groups 3, 4). Rats were submitted to a 2 hour PRS session (groups 1, 3) or control (groups 2, 4). At the end of the PRS session, the brain was removed and free-floating sections (35 μm) were immunostained for CRF positive neurons evaluation. Results: PRS increased (P<0.05) plasma ACTH levels at 15 and 30 min of stress exposure (532.3 ± 90.8 and 437.9 ± 61.2 vs 136.7 ± 4.7 pg/mL in control) whereas CS levels were increased (P<0.05) at 90 and 120 min of stress exposure (552. 2 ± 11.9 and 784.4 ± 28.3 vs 421.3 ± 33.7 ng/mL in control). L. farciminis treatment decreased (P<0.05) basal CS levels (226.3 ± 41.2 vs 421.3 ± 33.7 ng/ mL in control) and reduced the effect of PRS on ACTH at 15 and 30 min of stress exposure (156.0 ± 6.3 and 199.5 ± 25.4 ng/mL) and CS levels at 90 and 120 min of stress exposure (233.6 ± 15.1 and 345.6 ± 43.0 ng/mL). PRS increased the number of CRF positive neurons in the parvocellular regions of the PVN and L. farciminis treatment significantly reduced by 31% this effect. Conclusion: This study shows that a 15 day treatment by L. farciminis reduces basal corticosteronemia and attenuates the activation of the HPA axis triggered by an acute stress. (1) Ait-Belgnaoui et al., Gut 2005;54 (Suppl VII) A16. 1086 The Clonal Origins of Dysplasia from Metaplasia in the Human Stomach Lydia Gutierrez, Manuel Rodriguez-Justo, Simon Leedham, David Stoker, Ian Mitchell, Marco Novelli, Janusz A. Jankowski, Nicholas A. Wright, Stuart A. McDonald Background: Intestinal metaplasia of the stomach has been heavily implicated as a premalig- nant condition that can lead to gastric adenocarcinoma (GA). However, while there have been many studies describing the mutation load within GA, there is little evidence to elucidate its genetic origins from metaplasia. Furthermore previous studies from our laboratory have shown that the development of a field metaplasia depends on the division of metaplastic crypts to form a patch. However there is some controversy about the clonality of such crypts. Here we describe experiments where we define the clonal nature of metaplastic crypts within the human stomach and demonstrate the genetic relationship these have with associated dysplasia. Methods: Patients undergoing gastrectomy for gastric adenocarcinoma A-167 AGA Abstracts with dysplasia-associated metaplasia were used in this study. To determine clonality of metaplastic crypts, we stained frozen sections with a dual enzyme assay for cytochrome c oxidase (CCO) and succinate dehydrogenase (SDH). It has been established that mitochon- drial DNA (mtDNA) mutations that result in CCO-deficiency, are a reliable marker of clonality. We laser-captured as many individual cells as possible from a single CCO-deficient crypt. To determine the genetic relationship between metaplasia and dysplasia we laser captured crypts along a strip of gastric mucosa that contained patches of metaplastic and dysplastic crypts. We then screened these for APC mutations by nested PCR sequencing. Results: Every cell within a CCO-deficient metaplastic crypt contained the same mtDNA mutation (G7588A transition in mtCoII, whereas all cells from neighbouring CCO-positive crypts were wild type. Therefore such crypts are clonal. Initial screening of one of our patients revealed a truncating APC mutation (G4682 ins, leading to a stop codon at position 1566). All crypts within this section were laser captured and genotyped for this mutation. We demonstrated that 50% of dysplastic crypts were APC-mutated, interestingly, there were a few metaplastic crypts that also contained the same mutation, indicating the dysplasia did arise from a field of metaplasia. Conclusion: We present evidence to refute the claim that crypts from gastric mucosa showing metaplasia are polyclonal, but are in fact clonal. We also demonstrate the clonal expansion into a large field of dysplasia of an APC mutation seen in metaplasia, confirming the metaplastic origins of gastric adenocarcinoma. Although this is clearly not the founder mutation of the dysplasia, it does confirm that dysplasia arises from metaplasia in the human stomach. 1087 Hypermethylation of Let-7 Promoter Is Involved in cagA-Induced RAS Upregulation in Gastric Epithelial Cells Yoshito Hayashi, Masahiko Tsujii, Shingo Tsuji, Ying Jin, Takahiro Inoue, Jumpei Kondo, Sachiko Nakajima, Toshiyuki Yoshio, Shuji Ishii, Shinichiro Shinzaki, Satoshi Egawa, Tsutomu Nishida, Kenji Watabe, Hideki Iijima, Shusaku Tsutsui, Sunao Kawano, Tetsuo Takehara, Norio Hayashi <Background&Aim>Infection with CagA positive H. pylori is considered to be a risk factor for the development of gastric cancer. We have already reported that CagA expression induced ERK activation and cell growth, however, its precise mechanism remains unclear. Recently, microRNAs have been reported to be participating in the regulation of multiple carcinogenic processes. We investigated the effect of CagA expression on microRNA expres- sion in gastric epithelial cells by microRNA array analyzing 238 microRNAs. Let-7 was detected as one of the microRNAs whose expression was most significantly changed between before and after CagA expression. The aim of this study was to determine the effect of CagA on expression of let-7 target molecules in gastric epithelial cells and mechanism by which CagA suppressed let-7 expression. <Method>Western or East Asian type Cag (ABCCC or ABD, respectively) was expressed in a non-transformed rat gastric mucosal cultured cell line (RGM-1) using stable transfection of expression vectors containing tetracycline-off system. The effects of CagA on Ras, Myc and HMGA2 expression were determined by immunoblotting assay. The effect of let-7 induction in RGM-1 on target protein was investigated by let-7 precursor transfection. Many microRNA genes, including let-7, are located in CpG islands, suggesting possible epigenetic regulation of their expression. The effects of demethylating agent on let-7 expression and Ras expression were investigated by immunoblotting assay. Promoter CpG island methylation of rat let-7 was ascertained by a PCR based HpaII methyl- ation sensitive restriction enzyme assay. Genomic DNA prepared from RGM-1 control and CagA expressing RGM-1 was digested by HpaII, followed by each digested DNA, was analyzed by PCR using primers for rat let-7 promoter. <Result>Using Real time RT-PCR, it was confirmed that CagA significantly reduced let-7 microRNA expression. Immunoblotting assay revealed CagA induced Ras expression but not Myc or HMGA2 in rat gastric epithelial cells. Let-7 precursor suppressed Ras expression in CagA expressing cells. Adding demethylating agent led to recovery of let-7 expression and suppress of Ras expression. CpG methylation analyze showed that CagA expression induced hypermethylation of rat let-7 promoter region. <Conclusion>These results indicate that CagA is involved in let-7 downregulation through hypermethylation, leading to its downstream Ras upregulation. Furthermore, let-7 downreg- ulation and Ras upregulation might account for CagA-related ERK activation, resulting in accelerated cell proliferation. 1088 Role of MicroRNA-29c in the Chemoprevention of Gastric Cancer By a Selective COX-2 Inhibitor Yoshimasa Saito, Hidekazu Suzuki, Izumi Nakagawa, Juntaro Matsuzaki, Eisuke Iwasaki, Hitoshi Tsugawa, Hiroyuki Imaeda, Toshifumi Hibi microRNAs (miRNAs) are small non-coding RNAs that can function as endogenous silencers of target genes and play critical roles in human malignancies. We have recently proposed that the regulation of tumor suppressor miRNAs could be a novel therapeutic approach for human cancers (Cancer Cell 9: 435, 2006). The selective COX-2 inhibitor, celecoxib, has been proposed to be a potential drug for the chemoprevention of gastrointestinal tumors (Gut 53: 195, 2004). The aim of this study is to investigate roles of miRNAs in the chemoprevention of gastric cancer by selective COX-2 inhibitors. Methods. Forty-two patients of early gastric cancers and gastric adenomas who underwent endoscopic submucosal dissection (ESD) were enrolled for the study. This study was approved by the ethics committee of Keio University and registered with the UMIN Clinical Trials Registry (UMIN 000001057). Tissue specimens from early gastric cancers and gastric adenomas and surrounding non- tumor gastric mucosae were biopsied under the endoscope. miRNA expression profile was analyzed by miRNA microarray and quantitative RT-PCR. Separately, human gastric cancer cells (AGS, MKN-45) were exposed to celecoxib and then miRNA expression profile was analyzed by miRNA microarray. Expression of miRNA target genes was analyzed by Western blotting, and apoptosis was evaluated. Results. miRNA expression profile reveals that miR- 29c is significantly downregulated in early gastric cancers and gastric adenomas compared to the corresponding non-tumor mucosae (p<0.005). Downregulation of miR-29c is especially observed in early gastric cancers and gastric adenomas located at lower corpus and antrum. On the other hand, microarray analysis shows that miR-29c is activated by celecoxib in AGS AGA Abstracts

Upload: lydia-gutierrez

Post on 01-Dec-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: 1086 The Clonal Origins of Dysplasia from Metaplasia in the Human Stomach

1078

Pediatric Intervention Study with a Probiotic Mixture (Bacilac Forte) in AcuteGastroenteritisGigi Veereman, Sofie Staelens, Jan Wijffels

The aim of this study was to evaluate the acceptability, safety & effect on diarrhea andaccompanying symptoms of a probiotic mixture (Bacilac Forte) compared to placebo inchildren with a clinical diagnosis of mild acute gastroenteritis. The active product contained3, 3 10 9CFU pro capsule of a probiotic mixture:20% Lactobacilus rhamnosus, 20% Lacidophilus, 20% L casei, 20% L plantarum, 20% Bifidobacterium infantis. Study designwas prospective DBPC in a primary care setting. IRB approval was obtained and subjectswere included after parental consent. Otherwise healthy eutrophic children ages 1 mth-5yrs, consulting their GP b/o liquid stools more than 3 tid for less than 48 h with dehydrationless than 9% were included. Exclusion criteria were exposure to antibiotics or probioticswithin 4 wks, bloody diarrhea & fever. Subjects received active or placebo tid. Parents kepta diary until D9 rating acceptability of the product, various parameters and any adverseevent. Subjects were re-examined between D7-9. In total 48 subjects enrolled, 47 wereanalyzed (22 in the intervention group), 53% were girls, 91% Caucasian, age was 24+-15mths, with no differences between groups. Average stool frequency was 5/d, 45% hadcramps, 21% regurgitation, 45% vomiting and 23% upper respiratory symptoms. Statisticalanalysis used GLMM and non-parametric tests. Based on diary there was no difference induration of diarrhea and stool frequency (although a sharper decrease in the active group)but stool consistency returned to normal more rapidly (D3) in the active group (p=0.0205).Also stools were more soft and formed over the observation period in the active groupwhereas subjects in the placebo group showed more variation (more hard and watery stools).During the course of the study there were no differences in cramps, regurgitation, vomiting,upper respiratory symptoms and appetite between groups. Acceptability of the capsules wasrated easy by over 70% on all occasions for both groups, Tolererance was excellent: crying,flatulence, sleep and general behavior were rated similarly. Two adverse events (AE) wererecorded in the intervention group: 1 otitis media and 1 gastroenteritis with dehydrationnecessitating hospitalization (frequency of AE not different for both groups by fisher exactp=0.203). In conclusion: tid administration of the probiotic mixture Bacilac Forte to youngchildrenwithmild gastro-enteritis was safe, well tolerated and led tomore rapid normalizationof stool consistency with overall more formed and soft stools.

1079

A Probiotic Treatment (Lactobacillus farciminis) Attenuates the Hypothalamic-Pituitary-Adrenal (HPA) Axis Response to Acute Stress in RatsAfifa Ait-Belgnaoui, Helene Eutamene, Christel Salvador-Cartier, Henri Durand, EricHoudeau, Jean Fioramonti, Lionel Bueno, Vassilia Theodorou

Aims: We have previously shown that Lactobacillus farciminis (L. farciminis) treatment reducedstress-hypersensitivity to colorectal distension (CRD) in rats. Quantitative assessment of Fosexpression in relevant spinal structure1 indicated a decrease in the stress-induced activation/sensitization of sensory neurons in response to CRD. Since L. farciminis reduces neuronalactivation, we hypothesize that this probiotic may influence immediate HPA response tostress. This study aimed to evaluate the effect of L. farciminis on PRS-induced HPA responseby assessing adrenocorticotropic hormone (ACTH) and corticosterone (CS) plasma levelsand the number of corticotropin-releasing factor (CRF) positive neurons in the hypothalamicparaventricular nucleus (PVN). Materiel and Methods: In a first series of experiments, kineticsof plasma ACTH and CS were determined in 14 groups of 5 female Wistar rats receivingorally during 15 days either saline (groups 1-7) or 1011 cfu/day of L. farciminis (groups 8-14). All animals were submitted to PRS excepted groups 1 and 8 (control, no CRD and nostress). After 15, 30, 45, 60, 90 and 120 min of stress, under general anaesthesia blood wasdrawn from the abdominal aorta and ACTH and CS levels were determinated using aLincoplex kit. In a second series of experiments, 4 groups of 4 female rats received orallyduring 15 days either saline (groups 1, 2) or 1011 cfu/day of L. farciminis (groups 3, 4).Rats were submitted to a 2 hour PRS session (groups 1, 3) or control (groups 2, 4). At theend of the PRS session, the brain was removed and free-floating sections (35 μm) wereimmunostained for CRF positive neurons evaluation. Results: PRS increased (P<0.05) plasmaACTH levels at 15 and 30 min of stress exposure (532.3 ± 90.8 and 437.9 ± 61.2 vs 136.7± 4.7 pg/mL in control) whereas CS levels were increased (P<0.05) at 90 and 120 min ofstress exposure (552. 2 ± 11.9 and 784.4 ± 28.3 vs 421.3 ± 33.7 ng/mL in control). L.farciminis treatment decreased (P<0.05) basal CS levels (226.3 ± 41.2 vs 421.3 ± 33.7 ng/mL in control) and reduced the effect of PRS on ACTH at 15 and 30 min of stress exposure(156.0 ± 6.3 and 199.5 ± 25.4 ng/mL) and CS levels at 90 and 120 min of stress exposure(233.6 ± 15.1 and 345.6 ± 43.0 ng/mL). PRS increased the number of CRF positive neuronsin the parvocellular regions of the PVN and L. farciminis treatment significantly reduced by31% this effect. Conclusion: This study shows that a 15 day treatment by L. farciminisreduces basal corticosteronemia and attenuates the activation of the HPA axis triggered byan acute stress. (1) Ait-Belgnaoui et al., Gut 2005;54 (Suppl VII) A16.

1086

The Clonal Origins of Dysplasia from Metaplasia in the Human StomachLydia Gutierrez, Manuel Rodriguez-Justo, Simon Leedham, David Stoker, Ian Mitchell,Marco Novelli, Janusz A. Jankowski, Nicholas A. Wright, Stuart A. McDonald

Background: Intestinal metaplasia of the stomach has been heavily implicated as a premalig-nant condition that can lead to gastric adenocarcinoma (GA). However, while there havebeen many studies describing the mutation load within GA, there is little evidence to elucidateits genetic origins from metaplasia. Furthermore previous studies from our laboratory haveshown that the development of a field metaplasia depends on the division of metaplasticcrypts to form a patch. However there is some controversy about the clonality of suchcrypts. Here we describe experiments where we define the clonal nature of metaplasticcrypts within the human stomach and demonstrate the genetic relationship these have withassociated dysplasia. Methods: Patients undergoing gastrectomy for gastric adenocarcinoma

A-167 AGA Abstracts

with dysplasia-associated metaplasia were used in this study. To determine clonality ofmetaplastic crypts, we stained frozen sections with a dual enzyme assay for cytochrome coxidase (CCO) and succinate dehydrogenase (SDH). It has been established that mitochon-drial DNA (mtDNA) mutations that result in CCO-deficiency, are a reliable marker ofclonality. We laser-captured as many individual cells as possible from a single CCO-deficientcrypt. To determine the genetic relationship between metaplasia and dysplasia we lasercaptured crypts along a strip of gastric mucosa that contained patches of metaplastic anddysplastic crypts. We then screened these for APC mutations by nested PCR sequencing.Results: Every cell within a CCO-deficient metaplastic crypt contained the same mtDNAmutation (G7588A transition in mtCoII, whereas all cells from neighbouring CCO-positivecrypts were wild type. Therefore such crypts are clonal. Initial screening of one of ourpatients revealed a truncating APC mutation (G4682 ins, leading to a stop codon at position1566). All crypts within this section were laser captured and genotyped for this mutation.We demonstrated that 50% of dysplastic crypts were APC-mutated, interestingly, there werea few metaplastic crypts that also contained the same mutation, indicating the dysplasia didarise from a field of metaplasia. Conclusion: We present evidence to refute the claim thatcrypts from gastric mucosa showing metaplasia are polyclonal, but are in fact clonal. Wealso demonstrate the clonal expansion into a large field of dysplasia of an APC mutationseen in metaplasia, confirming the metaplastic origins of gastric adenocarcinoma. Althoughthis is clearly not the founder mutation of the dysplasia, it does confirm that dysplasia arisesfrom metaplasia in the human stomach.

1087

Hypermethylation of Let-7 Promoter Is Involved in cagA-Induced RASUpregulation in Gastric Epithelial CellsYoshito Hayashi, Masahiko Tsujii, Shingo Tsuji, Ying Jin, Takahiro Inoue, Jumpei Kondo,Sachiko Nakajima, Toshiyuki Yoshio, Shuji Ishii, Shinichiro Shinzaki, Satoshi Egawa,Tsutomu Nishida, Kenji Watabe, Hideki Iijima, Shusaku Tsutsui, Sunao Kawano, TetsuoTakehara, Norio Hayashi

<Background&Aim>Infection with CagA positive H. pylori is considered to be a risk factorfor the development of gastric cancer. We have already reported that CagA expressioninduced ERK activation and cell growth, however, its precise mechanism remains unclear.Recently, microRNAs have been reported to be participating in the regulation of multiplecarcinogenic processes. We investigated the effect of CagA expression on microRNA expres-sion in gastric epithelial cells by microRNA array analyzing 238 microRNAs. Let-7 wasdetected as one of the microRNAs whose expression was most significantly changed betweenbefore and after CagA expression. The aim of this study was to determine the effect of CagAon expression of let-7 target molecules in gastric epithelial cells and mechanism by whichCagA suppressed let-7 expression. <Method>Western or East Asian type Cag (ABCCC orABD, respectively) was expressed in a non-transformed rat gastric mucosal cultured cell line(RGM-1) using stable transfection of expression vectors containing tetracycline-off system.The effects of CagA on Ras, Myc andHMGA2 expression were determined by immunoblottingassay. The effect of let-7 induction in RGM-1 on target protein was investigated by let-7precursor transfection. Many microRNA genes, including let-7, are located in CpG islands,suggesting possible epigenetic regulation of their expression. The effects of demethylatingagent on let-7 expression and Ras expression were investigated by immunoblotting assay.Promoter CpG island methylation of rat let-7 was ascertained by a PCR based HpaII methyl-ation sensitive restriction enzyme assay. Genomic DNA prepared from RGM-1 control andCagA expressing RGM-1 was digested by HpaII, followed by each digested DNA, was analyzedby PCR using primers for rat let-7 promoter. <Result>Using Real time RT-PCR, it wasconfirmed that CagA significantly reduced let-7 microRNA expression. Immunoblotting assayrevealed CagA induced Ras expression but not Myc or HMGA2 in rat gastric epithelial cells.Let-7 precursor suppressed Ras expression in CagA expressing cells. Adding demethylatingagent led to recovery of let-7 expression and suppress of Ras expression. CpG methylationanalyze showed that CagA expression induced hypermethylation of rat let-7 promoter region.<Conclusion>These results indicate that CagA is involved in let-7 downregulation throughhypermethylation, leading to its downstream Ras upregulation. Furthermore, let-7 downreg-ulation and Ras upregulation might account for CagA-related ERK activation, resulting inaccelerated cell proliferation.

1088

Role of MicroRNA-29c in the Chemoprevention of Gastric Cancer By aSelective COX-2 InhibitorYoshimasa Saito, Hidekazu Suzuki, Izumi Nakagawa, Juntaro Matsuzaki, Eisuke Iwasaki,Hitoshi Tsugawa, Hiroyuki Imaeda, Toshifumi Hibi

microRNAs (miRNAs) are small non-coding RNAs that can function as endogenous silencersof target genes and play critical roles in human malignancies. We have recently proposedthat the regulation of tumor suppressor miRNAs could be a novel therapeutic approach forhuman cancers (Cancer Cell 9: 435, 2006). The selective COX-2 inhibitor, celecoxib, hasbeen proposed to be a potential drug for the chemoprevention of gastrointestinal tumors(Gut 53: 195, 2004). The aim of this study is to investigate roles of miRNAs in thechemoprevention of gastric cancer by selective COX-2 inhibitors. Methods. Forty-twopatients of early gastric cancers and gastric adenomas who underwent endoscopic submucosaldissection (ESD)were enrolled for the study. This study was approved by the ethics committeeof Keio University and registered with the UMIN Clinical Trials Registry (UMIN 000001057).Tissue specimens from early gastric cancers and gastric adenomas and surrounding non-tumor gastric mucosae were biopsied under the endoscope. miRNA expression profile wasanalyzed by miRNA microarray and quantitative RT-PCR. Separately, human gastric cancercells (AGS, MKN-45) were exposed to celecoxib and then miRNA expression profile wasanalyzed by miRNA microarray. Expression of miRNA target genes was analyzed by Westernblotting, and apoptosis was evaluated. Results. miRNA expression profile reveals that miR-29c is significantly downregulated in early gastric cancers and gastric adenomas comparedto the corresponding non-tumor mucosae (p<0.005). Downregulation ofmiR-29c is especiallyobserved in early gastric cancers and gastric adenomas located at lower corpus and antrum.On the other hand, microarray analysis shows that miR-29c is activated by celecoxib in AGS

AG

AA

bst

ract

s