The novel compound N-benzoxazol-2-yl-N'-1-(isoquinolin-3-yl-ethylidene)-hydrazine (EPH136) has been shown to exhibit antitumor activity in vitro and in vivo. A COMPARE analysis showed that the patterns of cellular effects of EPH136 are not related to any of 175 standard antitumor agents with a known mechanism of action. In order to help identify the mechanism of action we employed a bioinformatics approach called partial least squares modelling in latent variables in which the expression levels of approximately 8,000 genes in each of 56 untreated NCI panel cell lines were correlated with the respective IC(50) values of each cell line following treatment with EPH136. The 60 genes found to be most important for the antiproliferative effect of EPH136 are involved in nucleoside, nucleotide, nucleic acid binding and metabolism, developmental processes, protein modification and metabolism. In addition, using a DNA microarray we measured the expression of approximately 5,000 known genes following treatment of HT-29 colon carcinoma cells with a two-fold IC(50) concentration of EPH136. The genes that were up-regulated more than two-fold compared to untreated controls belong to the same classes as found by the bioinformatic approach. Many of these proteins are regulated by oxidation/reduction and so we concluded that formation of radicals may be involved in the mechanism of action. We show here that EPH136 leads to generation of oxygen radicals, swelling of mitochondria and dissipation of the mitochondrial membrane potential. The antiproliferative activity of EPH136 was prevented by the radical scavenger N-acetylcysteine. Cells with elevated glutathione exhibited resistance to EPH136. In summary, the mechanism of the novel experimental anticancer drug EPH136 is generation of radicals and dissipation of the mitochondrial membrane potential.