Microarray analysis of gene expression profile of multidrug resistance in pancreatic cancer

Abstract

Background: Chemotherapy is the most frequently adopted adjuvant therapy of pancreatic ductal adenocarcinoma (PDAC), but the development of drug resistance reduces its effectiveness. Clarification of the mechanism of multidrug resistance (MDR) development in PDAC is needed to improve the therapeutic effect of chemotherapy. This study was aimed to investigate the molecular mechanism of MDR of PDAC and to identify genes associated with MDR development.

Methods: The gene expression profiles of cell line SW1990 and three drug-selected pancreatic chemoresistant sub-lines, SW1990/5-Fu, SW1990/ADM and SW1990/GEM, were obtained using an oligonucleotide microarray (Affymetrix HG U133 2.0 plus) that contained approximately 38,000 human genes. The microarray results were validated by real-time quantitative polymerase chain reaction and Western blot analysis.

Results: There were 165 genes and expressed sequence tags, some of which have never been linked to drug resistance, that were up- or down-regulated at least 2-fold in all resistant sub-lines when compared with SW1990. According to Gene Ontology annotation, differentially expressed genes related to MDR in pancreatic cancer belong to many functional families and with diverse biological processes. Genes related to antioxidant activity, apoptosis, the cell cycle, signal transduction and intracellular adhesion may undergo epigenetic changes preceding MDR development. A hierarchical clustering was conducted and several interesting clusters were discovered that may be primarily related to cell cycle and developmental regulation. A prediction rule was built from the expression profiles of 117 genes after support vector machine (SVM) analysis, and the prediction result was examined by cytotoxic testing. As a result, a differential gene expression pattern was constructed in multidrug resistant pancreatic cancer cells.

Conclusions: The findings of this study prove that construction of a chemoresistance prediction rule, based on gene expression patterns, is practical. These data provide new insights into the molecular mechanism of pancreatic cancer MDR development and may be useful for the detection and treatment of MDR in pancreatic cancer patients.