Integrated expression profiles analysis reveals novel predictive biomarker in pancreatic ductal adenocarcinoma

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is the most lethal human malignant tumor, with a dismal 5-year survival rate of less than 5%. The lack of specific symptoms at early tumor stages and the paucity of biomarkers contribute to the poor diagnosis of pancreatic ductal adenocarcinoma. To improve prognosis, a screening biomarker for early diagnosis of pancreatic cancer is in urgent need. We searched the databases of expression profiling by array on GEO, aiming at comparing gene expression profile of matched pairs of pancreatic tumor and adjacent non-tumor tissues, and we screen out 4 suitable series of gene expression microarray data (“GSE15471”, “GSE18670”, “GSE28735” and “GSE58561”). After carefully analyzing, 13 DEGs (MYOF, SLC6A6, S100P, HK2, IFI44L, OSBPL3, IGF2BP3, PDK4, IL1R2, ERO1A, EGLN3, PLAC8 and ACSL5) are significantly differentially expressed in four microarray databases in common. After analyzing mRNA expression data and clinical follow-up survey provided in the TCGA database and clinicopathological data of 137 pancreatic ductal adenocarcinoma patients, we carefully demonstrated that three of these differentially expressed genes (ERO1A, OSBPL3 and IFI44L) are correlated with poor prognosis of pancreatic ductal adenocarcinoma patients. In addition, we revealed that cell–matrix adhesion and extracellular matrix were top significantly regulated pathways in pancreatic ductal adenocarcinoma and depicted two protein-protein interactions networks of extracellular matrix related Genes which are dysregulated according to 4 gene expression microarray data mentioned above (“GSE15471”, “GSE18670”, “GSE28735” and “GSE58561”), hoping to shed light on the etiology of PDAC and mechanisms of drug resistance in PDAC in this study.