major determinant of cost-effectiveness in treating arthritis. We therefore performed a meta-
`analysis to compare rates of dyspeptic symptoms for two commonly used therapies in high-
`risk patients with arthritis: 1) Coxib alone, and 2) NSAIDⳭPPI combination.
`Methods: We performed a structured search of MEDLINE and published abstracts to
`identify English-language randomized trials from 1990-2004 comparing either a Coxib
`vs NSAID or NSAIDⴐPPI combination vs NSAID alonein chronic arthritis. Two review-
`ers independently selected studies that report incident dyspeptic symptoms, defined
`a priori as ‘‘epigastric pain,’’ ‘‘dyspepsia,’’ and ‘‘nausea.’’ The reviewers independently
`abstracted data and assigned a quality score for each study. We performed meta-
`analysis with a fixed effects model to compare the relative risk reduction (RRR) and
`Absolute Risk Reduction (ARR) of dyspeptic symptoms for Coxib vs NSAID and NSAID-
`ⴐPPI vs NSAID, and performed an Egger’s test to assess for publication bias.
`Results: We identified 840 titles, of which 37 were selected for final review (kappa⬎0.9
`for agreement). Meta-analysis of 32 studies (N⳱60,163 patients) comparing dyspeptic
`symptoms between Coxibs and NSAIDs revealed a 12% RRR for Coxibs (RR⳱0.88; 95%
`CI⳱0.85-0.90) with an ARR of 3.7%. Meta-analysis of 5 studies comparing dyspeptic
`symptom between the NSAIDⳭPPI combination and NSAIDs alone revealed a 66% RRR
`for NSAIDⳭPPI (RR⳱0.34; CI⳱0.22-0.54) with an ARR of 9%. There was no evidence
`of heterogeneity (p⬎0.05) or publication bias (p⬎0.05) in either analysis. Compared to the
`NSAID strategy, the number needed to treat in order to prevent a dyspeptic symptom was
`27 for Coxibs and 11 for NSAIDⳭPPI.
`Conclusions: The NSAIDⳭPPI strategy affords a greater risk reduction for dyspepsia than
`Coxibs alone when compared to the common baseline of NSAIDs alone. Because there are
`limited head-to-head data comparing Coxibs vs NSAIDⳭPPI, these meta-analytic data pro-
`vide the best indirect evidence that the NSAIDⳭPPI strategy may be superior to Coxibs in
`minimizing incident dyspeptic symptoms during the treatment of chronic arthritis.
`
`864
`
`Do Proton Pump Inhibitors (ppi) Infer Additional Gastrointestinal Protection in
`Patients Given Celecoxib? a Retrospective Cohort Study.
`Elham Rahme, Alan Barkun, Youssef Toubouti, Sophie Rochon, Jacques LeLorier
`
`Introduction: Proton pump inhibitors are prescribed with non-selective NSAIDs to prevent
`and treat NSAID-associated gastropathy. It is unclear whether the utilization of a PPI with
`celecoxib confers additional gastrointestinal (GI) protection in elderly patients.
`Objectives: To assess the association between GI hospitalizations and the use of celecoxib
`& PPI, celecoxib alone, NSAID& PPI or NSAID alone and to identify patient subgroups in
`whom the addition of a PPI to celecoxib is beneficial.
`Methods: We conducted a population-based retrospective cohort study using Quebec govern-
`ment administrative databases. Patients 66 years of age or older were included at the
`dispensing date of their first filled prescription (index date) for celecoxib or an NSAID
`between April 1999 and December 2002. They were followed from the index-date until the
`occurrence of a GI hospitalization, death or the last day of supplied medication for either
`celecoxib or an NSAID in the study period. Cox regression models with timedependent
`exposure were used to compare the hazard rates of GI hospitalization between the four
`groups: celecoxib & PPI, celecoxib alone, NSAID& PPI or NSAID alone, adjusting for patient
`characteristics at the index date.
`Results: A total of 332,491 patients were included. The adjusted GI hospitalization hazard
`rate was significantly lower among patients given celecoxib & PPI compared to those given
`celecoxib alone (hazard ratio (HR) 0.69, 95% CI 0.52-0.93). The adjusted hazard rate among
`patients given NSAID & PPI was similar to that of patients given celecoxib alone (0.98,
`0.67-1.45) while the rate among patients given NSAID alone was about twice as high as
`that of patients given celecoxib (2.18, 1.82-2.61). Stratified analyses showed that celecoxib
`alone was the GI-safest treatment option in patients 66-74 years of age, not taking aspirin
`and who did not have other GI risk factors. In all other groups the GI-risk associated with
`celecoxib seemed similar to that associated with NSAIDs & PPI. The results also showed
`that the use of a PPI with celecoxib may be beneficial in high-risk patients aged 75 or older
`and in patients using aspirin.
`Conclusions: Celecoxib alone seemed as GI-safe as NSAIDs combined with a PPI in most
`patients. PPI conferred additional protection to celecoxib for older patients and for patients
`taking aspirin. The addition of a PPI to celecoxib did not seem beneficial for patients without
`these GI-risk factors.
`
`865
`
`Identification of a Gastric Progenitor Cell
`Xiaotan Qiao, Joshua Ziel, Blair Madison, Andrea Todisco, Linda C. Samuelson, Juanita
`Merchant, Deborah Gumucio
`
`The stomach and intestine develop from a tube of endoderm surrounded by mesoderm that
`becomes regionally pattered during ontogeny. To trace the development of intestinal identity,
`we placed a ␤-galactosidase cDNA within an intestine-specific gene locus (villin) by homolo-
`gous recombination. At E15.5, ␤-gal is expressed throughout the intestine and in the distal
`half of the stomach. At E16.6, a sharp epithelial border forms, and ␤-gal is robustly expressed
`in epithelial cells on the intestinal side of this border, but not in the adjacent stomach
`epithelial cells (Braunstein et al., Dev. Dyn. 2002). However, closer examination of the
`antrum of these mice reveals rare ␤-gal (Ⳮ) cells within the stem cell zone. These cells are
`small and granule free with a prominent nucleolus; they are first seen in E16.5 animals,
`and still present in adults. Treatment of adult mice with interferon ␥ (IFN␥), which induces
`pseudopyloric metaplasia, results in a 5-10 fold increase in the number of these cells. We
`were intrigued by the fact that the number of ␤-gal (Ⳮ) cells per gastric pit was always ⱕ
`1. Since they were located in the stem cell zone, and induced by INF␥, a cytokine known
`to expand stem cells in liver and pancreas, we speculated that the ␤-gal (Ⳮ) cells may be
`progenitors whose daughter cells differentiate into gastric lineages, ceasing to express ␤-gal.
`To test this, we traced the progeny of the ␤-gal (Ⳮ) cells. Villin-Cre mice were mated to
`the Rosa indicator strain, R26R (in the offspring, Cre expression results in ␤-gal activation).
`We expected that if ␤-gal (Ⳮ) cells divide to give rise to gastric cells, then entire antral pits
`should be ␤-gal positive. This outcome was observed and cells within ␤-gal (Ⳮ) pits
`
`84476$CH21
`
`03-31-05 13:11:51
`
`expressed markers of gastric pit and neck cells. We are currently isolating and characterizing
`these progenitors from transgenic mice expressing Villin-EGFP, generated in our laboratory.
`These cells may represent the first prospectively identified gastric progenitor cells; it will
`be important to establish their possible links to intestinal metaplasia and gastric cancer.
`
`866
`
`Gli3 Mediates Sonic Hedgehog Dependent Development of the Glandular Stomach
`Jae H. Kim, Zhen Huang, Rong Mo
`
`The role of the Hedgehog signaling pathway in various aspects of gut development is still
`poorly understood. In the developing stomach, secreted Sonic (Shh) and Indian hedgehog
`are expressed in both distinct and overlapping regions. The loss of function of Sonic hedgehog
`in the stomach results in a dramatic phenotype of intestinal transformation and glandular
`overgrowth. These changes are reminiscent of the pre-malignant lesion, intestinal metaplasia.
`Consistent with these findings, in adult forms of intestinal metaplasia, the expression of
`Sonic hedgehog is lost. In an effort to determine if the embryonic stomach would be a
`suitable model of intestinal metaplasia, we conducted a mutant analysis of glandular stomach
`from mice mutant in Hedgehog pathway molecules. METHODS: We studied mice mutant
`in Shh and the transcription factors, Gli2 and Gli3. Paraffin section slides generated from
`dissected stomachs at E18.5 days post coitum were subjected to immunohistochemical and
`transmission electron microscopic analysis. Antibodies and biotin-conjugated lectins studied
`were against gastric markers (HⳭ/KⳭ-ATPase, intrinsic factor, lectins AAA and GSII), tissue
`markers (alpha-smooth muscle actin, PECAM, beta-tubulin III), and proliferation markers
`(Ki67, phospho-histone H3). RESULTS: We discovered that striking glandular expansion
`occurred in both the Shh and Gli3 mutant stomachs but not the Gli2 mutant stomach. In
`the first two cases, such glandular expansion was mostly epithelial and gastric in nature
`and these changes arose from increased glandular branching. Increases in proliferation were
`not seen, however. All gastric cell markers tested were present in all wild-type and mutant
`stomachs but in the cases of Shh and Gli3 mutant mice, they co-existed with early features
`of intestinal metaplasia. CONCLUSIONS: These data strongly suggest that Gli3 is the primary
`mediator through which Shh negatively directs proper specification and control of growth
`of the glandular stomach. We suggest that this defines an uncommon example whereby full
`length Gli3 activator plays a dominant physiologic role. To our knowledge, this is the first
`instance where the full length Gli3 activator functions in gastrointestinal development. This
`study furthers our understanding of the role of Hedgehog pathway molecules in this model
`of intestinal metaplasia. The embryonic glandular stomach may assist in further illuminating
`the mechanisms of gastric carcinogenesis.
`
`867
`
`Signal Transduction Pathways Regulating the Actions of Bmp-4 in Isolated Parietal
`Cells.
`Andrea Todisco, Saravanan Ramamoorthy, Yinghua Xiao, Colin Delaney
`
`BACKGROUND: BMP-4 is an important regulator of cellular growth and differentiation.
`Expression of BMP-4 has been documented in the gastric mucosa. We previously reported
`that incubation of purified (⬎95%)canine parietal cells in primary culture with BMP-4 (20
`ng/ml)induces HⳭ/KⳭ-ATP-ase alfa subunit gene expression. In addition, we demonstrated
`that protein kinase B/Akt plays an important role in the regulation of parietal cell differentia-
`tion. AIM: To explore the signal transduction pathways mediating the actions of BMP-4 in
`the parietal cells. METHODS: HⳭ/KⳭ-ATP-ase alfa subunit gene expression was examined
`by Northern blot analysis. Akt activation was measured by immunoprecipitation and in-
`vitro kinase assay using GSK3 as substrate. Inhibition of Cdcd42 and Ras function was
`achieved by transduction of the parietal cells with 100 moi of adenoviral vectors expressing
`either dominant negative Ras or Green Fluorescent Protein-tagged dominant negative Cdc42.
`Expression of the proteins was documented by either western blots with anti-Ras antibodies
`or by examination of the parietal cells by fluorescent microscopy. Control experiments were
`performed with the adenoviral vector expressing beta-galactosidase. Inhibition of Akt was
`achieved by treatment of the cells with the specific Akt inhibitor 1L-6-Hydroxymethyl-chiro-
`inositol 2-(R)-2-O-methyl-3-O-octadecylcarbonate (5 micromolar). Cdc42 activation was
`measured by affinity precipitation and western blots with anti-Cdc42 antibodies. RESULTS:
`BMP-4 induced Akt after 48 h of incubation and it stimulated the activation of Cdc42. The
`stimulatory effects of BMP-4 on Akt activation and HⳭ/KⳭ-ATP-ase alfa subunit gene
`expression were blocked by either treatment of the cells with the Akt inhibitor or by
`expression of dominant negative Cdc42. The specificity of this effect was confirmed by the
`observation that dominant negative Ras failed to block BMP-4 induction of HⳭ/KⳭ-ATP-
`ase gene expression, while it reversed the inhibitory effect on the expression of the HⳭ/
`KⳭ-ATP-ase gene, seen after prolonged exposure (48 to 72h) of the parietal cells to EGF
`(10 nM). CONCLUSIONS: BMP-4 induces HⳭ/KⳭ-ATP-ase gene expression through a
`signal transduction pathway that requires the sequential activation of Cdc42 and Akt. These
`findings provide new clues for a better understanding of the mechanisms that regulate gastric
`epithelial cell differentiation.
`
`868
`
`A Potential Interplay Between Gastrin and Hedgehog in Regulating Apoptosis in the
`Gastric Mucosa
`Mo El-Zaatari, Andrew McKenzie, Susan A. Watson
`
`Introduction: High gastrin levels have been implicated in regulating apoptotic activity and
`the development of cancer. Sonic Hedgehog (Shh) has been implicated in the maintenance
`of gastric cancer. In addition, Gastrin and Shh have been implicated in regulating proliferation
`and differentiation of stem cells within gastric glands. The aim of this study was to investigate
`the interplay between gastrin and Shh by using mice treated with lansoprazole (a proton
`pump inhibitor that increases gastrin levels) and cyclopamine (Shh pathway inhibitor).
`Methods: Animals were treated for 28 days with lansoprazole, cyclopamine, a combination
`of lansoprazole and cyclopamine, and untreated (n⳱3-4 per group). Serum was collected
`
`A-139
`
`AGA Abstracts
`
`Page 1 of 1
`
`Patent Owner Ex. 2018
`CFAD v. Pozen
`IPR2015-01718