AG

AA

bst

ract

sSa1914

Prognostic Value of Tumoral Expression of Galectin-9 in Gastric CancerKi-woo Seo, Soo-Jeong Cho, Myeong-Cherl Kook, Keun Won Ryu, Young-Woo Kim,Chan Gyoo Kim, Il Ju Choi

BACKGROUND & AIMS: Galectin-9 (Gal-9) is a member of the β-galactoside-bindinggalectin family. Our previous study revealed that Gal-9 suppresses cell migration, invasion,and epithelial-mesenchymal transition in gastric cancer cells. Gal-9 was reported to havean antimetastatic activity in patients with hepatocellular carcinoma. Therefore, we aimed toevaluate the prognostic significance of Gal-9 in gastric cancer patients. METHODS: Theclinical significance of Gal-9 was explored using tissue microarray methods and immunohisto-chemical staining of specimens from 619 gastric cancer patients. Median follow-up durationwas 65.7 months (range 0 ~ 79 months). The Kaplan-Meier method was used to calculatesurvival curves, and differences between the survival curves were assessed using the log-rank test. RESULTS: According to the Gal-9 tumoral expression, 619 patients with gastriccancer were classified into two groups; 292 negative patients (47.2%), 327 positive patients(52.8%). The Gal-9 positive patients had significantly lower overall (P = 0.001) and gastriccancer-specific mortalities (P < 0.001) than did those with Gal-9-negative tumors. CONCLU-SIONS: Tumor expression of Gal-9 might be a suppressor of tumor progression. Keywords:Gastric cancer, Galectin-9 (Gal-9), Tumor progression suppressor, Survival

Sa1915

How Solid Are Iron Indices for Risk Stratification in Suspected IronDeficiency Anemia?Toufic Kachaamy, Jason D. Adam, Jianfeng Cheng, Puneet Puri

Background: Endoscopic evaluation for iron deficiency anemia (IDA) is the standard of caregiven an increased risk of gastrointestinal (GI) malignancy. Bone marrow biopsy is the goldstandard for diagnosing IDA and serum ferritin is the best acceptable surrogate marker.Serum ferritin however is an acute phase reactant and can be falsely elevated in patientswith GI malignancy. On the other hand, there have been studies suggesting that lower valuesof ferritin predict the risk of malignancy and can be used for risk stratification in determiningwho should undergo endoscopic evaluation. Additional iron indices are often used clinicallyto diagnose IDA and patients are then referred for endoscopic evaluation. Objective: Toinvestigate the utility of serum iron and ferritin in predicting GI malignancy in the subjectsreferred for IDA. Methods: All subjects referred for endoscopic evaluation of IDA in a tertiarycare academic center during 2007 to 2012 were screened retrospectively. Subjects that hadesophagogastroduodenoscopy (EGD), colonoscopy and iron indices were included. Dataincluding demographics, serum iron, ferritin, transferrin, hemoglobin, creatinine and erythro-cyte sedimentation rate were recorded when available. Continuous variables were analyzedby t-tests or ANOVA, categorical data were analyzed via Pearson chi-square. Logistic regres-sion models were fit to find the possible significant predictors. All analysis was performedin SAS software(Cary, NC, V9.3) Results: A total of 370 patients underwent both upper andlower endoscopies for IDA and had iron indices available. The mean age was 57.2 years (range18-88 years) and 56.7% were female. The prevalence of pathologically proven malignancy was3.8% (2.7% colon and 1.1% gastric cancers). Patients were further stratified into 3 groups:G1: ferritin ≤50ng/ML, G2: ferritin of 51-100ng/mL and G3 with ferritin>100ng/mL. Ferritinlevels were not significant of malignancy (p=.05). Serum iron level was significantly lowerin G1 vs G3(47.1 vs 64.6, p=0.016). However, iron level were not significant predictor forcolon malignancy (p=0.19, adjusting for ferritin level and p=0.23 without adjusting ferritin)or gastric cancer (p=0.63 adjusting for ferritin level and p=0.53 without adjusting ferritin).There was no statistically significant difference between the serum iron or ferritin concentra-tions between patients with and without malignancy (colonic and gastric combined). Conclu-sions: Our data demonstrate that subjects referred for endoscopic evaluation of IDA have~4% prevalence of GI malignancy. Iron indices are not predictive of malignancy risk andthus endoscopic evaluation is an important test in the work up of all patients with Irondeficiency anemia.

Sa1916

Increased Expression of ROCK I and Met in Colorectal Cancer Is AssociatedWith Decreased Patient SurvivalJian Li, Sarah C. Glover, Altin Gjymishka

Introduction and hypothesis: Despite the advancements in early detection and treatment,colorectal cancer (CRC) remains a leading cause of cancer-related deaths. There is an ongoingsearch for novel tumor markers that can serve as predictors of response to therapy, recurrenceand/or overall survival in CRC. In this retrospective study, we hypothesized that increasedexpression of Rho-associated protein kinase I (ROCK I), - II (ROCK II) and hepatocytegrowth factor receptor (MET), is correlated with poor survival of CRC patients. Methods:Atissue microarray (TMA) was constructed from CRC specimens obtained from 110 patients.The TMA slides and ROCK I, -II and MET protein expression levels categorized as low,medium and high were independently analyzed by two pathologists that were blind withrespect to clinical and histopathological results. Clinical information was collected from thereview of electronic charts and it was organized in a database for further analysis. Results:Tu-mor marker expression was positive in 107 patients for ROCK I (97.3%), 102 (92.7%) forROCK II and 92 (83.6%) for MET.CRC patients with high expression levels of ROCK I andMET were identified only in more advanced tumor stages (2, 3 and 4) but not in stage 1.In addition, the comparison of ROCK I and MET mean expression levels between stage 1(n=16, ROCKI 1.38±0.09; MET 1.03±0.17) and stage 4 (n=25, ROCKI 1.65±0.09; MET1.55±0.12) revealed a significant increase (p =0.05 and p=0.01, respectively). Whereas, the5-year overall survival of CRC patients with high ROCK I expression was significantly lowerthan that of patients with low ROCK I expression (p=0.037). Also, a significant decrease inROCK I expression levels was observed between non-survival (n=64, 1.62±0.07) and survivalgroups (n=46, 1.41±0.08) after a 5-year follow up period (p=0.048). In addition, there wasa decrease in overall survival of CRC patients with high MET expression relative to thosewith low MET expression during the 5-year (p=0.08) and 6-year follow up periods (p=

S-328AGA Abstracts

0.049). Whereas, ROCK II expression levels that did not change significantly among differenttumor stages, showed no association with survival outcomes. Conclusion:The data demon-strate that the high expression of ROCK I and MET, but not that of ROCK II, is correlatedwith poor survival in CRC patients.

Sa1917

Detection of Mutations in KRAS, BRAF and PIK3CA Genes in ColorectalCancer Patients Using Allele Specific Primer Extension on Luminex PlatformMiao Cui, Zhiqing Wang, Weihua Tong, David Zhang, Fei Ye

BACKGROUND: Colorectal cancer (CRC) is one of the leading causes of cancer-relateddeaths in the USA, which can be developed through different pathways. It is now widelyaccepted that cancer progression in human CRC through the activation of oncogenic signalingpathways as well as the inactivation of tumor suppressor genes and mismatch repair genes.EGFR is a transmembrane tyrosine kinase receptor that, on ligand binding, triggers two mainsignaling pathways: RAS-RAF-MAPK pathway and PIK3-PTEN-AKT pathway. Mutations inKRAS, BRAF or PIK3CA genes result in continuous activation of the downstream RAS orPIK3 pathways in many human cancers in the absence of EGFR ligand. Therefore, detectionof the activation mutations in these genes is critical for personalized treatment of thesecancers. In this study, we developed a multiplex assay to detect 17 hotspot mutations inKRAS, BRAF and PIK3CA genes simultaneously in FFPE cancer tissues samples. METHODS:Six cell lines including HCT-116, HT-29, MCF-7, SW-620, YULAC and K562 were usedto determine analytical sensitivity. 125 colon polypes and cancer samples collected at MountSinai Medical Center, New York from 2012 to 2013 were included in this study. DNAsfrom FFPE tissues are extracted using Maxwell® CSC instrument using the Maxwell® FFPEPlus LEV DNA Purification kit. Multiplex PCR was carried out to generate 4 ampliconscovering 17 hotspot mutations in KRAS, BRAF and PIK3CA genes (Table 1). Allele specificprimer extension (ASPE) assay was performed to detect the mutations and the biotin-labeledPCR-ASPE products were detected using Luminex 200 xMAP platform. RESULTS: Theanalytical sensitivity for the is 2.5% of mutant allele in a background of wild-type allele,except for PIK3CA exon 9 (E545K) mutations as determined using DNAs from five mutantcell lines (HCT-116, HT-29, MCF-7, SW-620 and YULAC) mixed serially with non-mutantDNA. Among 125 samples tested, 60 (48%) samples had KRAS codon 12 mutations and20 (16%) samples had KRAS codon 13 mutations. BRAF codon V600E mutation was detectedin 14 cases (11.2%). PIK3CA mutations were detected in 9 (7.2%) cases and of these 9cases, two were positive for PIK3CA exon 9 mutation, one had mutations in both PIK3CAexon 9 and KRAS codon 13, two had mutations in both PIK3CA exon 9 and KRAS codon12, three had mutations in both PIK3CA exon 20 and KRAS codon 12, and one hadmutations in both PIK3CA exon 20 and KRAS codon 13 (Table 2). The results were 100%consistent with those obtained with the reference methods. CONCLUSION: Our resultsdemonstrated that PCR- ASPE on Luminex platform is a sensitive, specific and high-through-put method for multiplex detection of hotspot mutations in KRAS, BRAF and PIK3CA genesin FFPE samples. Mutations in KRAS and BRAF genes are mutually exclusive. However,PIK3CA mutations often co-exist with KRAS mutations.

Sa1918

Effects and Possible Mechanism of Inhibition of PI3K/AKT/mTOR SignalingPathway on Glycolysis of Gastric Adenocarcinoma Cell LinesMin Chen, Zhu Zhu, Ying Lv, Xiaoqi Zhang, Xiaoping Zou

Background and Aims: The PI3K/Akt/mTOR signaling pathway has been linked to tumormalignant phenotype, including glycolysis. Pyruvate Kinase 2(PKM2) is one of the rate-limiting enzyme, mainly regulating tumor glycolysis metabolism, while the underlying molec-ular mechanisms remain largely elusive. Consequently, our study aimed to investigate theinhibition of PI3K/Akt/mTOR signaling pathway on glycolysis in gastric adenocarcinomacell lines and its possible mechanism. Methods: Gastric adenocarcinoma cells BGC-823 weretreated with Ly294002 at different concentrations. CCK-8 assay was used to asses the cellrelative proliferation rate. Western blotting was used to determinate the expressions of p-Akt, p-m TOR, HIF-1α and PKM2. The intracellular distribution of PKM2 was assessed byimmunofluorescence. Cell apoptosis rate was analyzed by flow cytometry .The intracellularlactic dehydrogenase and the extracellular lactic acid also were detected by related kits.SiRNA was used to inhibit HIF-1α in BGC-823, and PKM2 expression was detected. Afterthe use of pShRNA-PKM2 in BGC-823/SGC-7901, the levels of proliferation and invasionalso were assessed. Results: Ly294002 could inhibit the proliferation of BGC-823 in a dose-and-time dependent manner. Also, the inhibition of p-Akt, p-mTOR and HIF-1α showed