Kong_2016_Oncol.Rep_36_1852

Gastric cancer (GC) is a common and deadly digestive tract tumor worldwide. Unfortunately, diagnosis of GC is usually confused and misleading because of atypical symptoms or incomplete complaints. Accordingly, exploring gene expression profile and identifying genes with analogical variance trend will bring new perspective into the diagnosis and treatment of GC. Herein, a RNASeq dataset from Caucasian GC and their matched noncancerous samples [Gene Expression Omnibus (GEO): SRP049809] and datasets from four microarrays constituted with tumor and nontumor tissues (GEO: GSE13911, GSE19826, GSE29272, GSE33335) were analyzed to explore the differentially expressed genes (DGEs). As a result, we identified a core set of 373 DGEs. Among these genes, we found that most downregulated genes were related to lipidmetabolic functions. Especially, the gastric lipase (LIPF) gene, which was connected with various lipid metabolism processes, was significantly decreased among all datasets. We then performed immunohistochemistry experiments using gastric tissue arrays to investigate the clinical effects, and the expression of a LIPF target gene, diacylglycerol kinase alpha (DGKA). Among the 90 samples of gastric adenocarcinoma, the LIPF and DGKA levels were both decreased in cancer tissues [LIPF, 59.1% (53/90); DGKA, 77.8% (70/90)] compared to normal tissues [LIPF, 94.4% (85/90); DGKA, 90% (81/90)]. The expression level of these two proteins in GC was associated with local invasion and disease stage. Cox regression identified high DGKA expression (HR, 0.49; 95% CI, 0.260.94; P=0.03) as a predictor of good prognosis and LNM status (HR, 4.63; 95% CI, 1.3915.51; P=0.01) as a predictor of poor prognosis. Thus we speculated that LIPFDGKA might serve as a potential possible biomarkers for diagnosis of GC, and their downregulation may bring new perspective into the investigation of GC prognosis.