Microarray-based Prediction of Cytotoxicity of Tumor Cells to Arsenic Trioxide

Arsenic has been used since ancient times as a medicinal agent. Currently, arsenic trioxide is experiencing a thriving revival in modern oncology. The aim of this study was to identify the molecular predictors of sensitivity and resistance to arsenic trioxide. We mined the microarray database of the National Cancer Institute (NCI), USA, for genes whose expression correlated with the IC₅₀ values for arsenic trioxide of 60 cell lines of different tumor types. By COMPARE analysis, Kendall's τ test, and false discovery rate (FDR) analyses, 47 out of 9706 genes or expressed sequence tags (ESTs) were identified. If the mRNA expression of the 47 genes or ESTs was subjected to hierarchical cluster analysis and cluster image mapping, sensitivity or resistance of the 60 cell lines to arsenic trioxide was predictable with statistical significance (p=1.01 × 10^-5). While the proteins encoded by the 47 genes identified differ in their specific functions (signal transducers, transcription factors, proteasome degradation proteins, proliferation-related proteins, regulators of oxidative stress etc.), it is intriguing that many of them are in one way or another involved in the apoptotic machinery, indicating that apoptosis is an important mechanism of arsenic trioxide's cytotoxicity.