BACKGROUND Over the last decade, several drugs that inhibit class I and/or class II histone deacetylases (HDACs) have been identified, including trichostatin A, the cyclic depsipeptide FR901228 and the antibiotic apicidin. These compounds have had immediate application in cancer research because of their ability to reactivate aberrantly silenced tumour suppressor genes and/or block tumour cell growth. Although a number of HDAC inhibitors are being evaluated in preclinical cancer models and in clinical trials, little is known about the differences in their specific mechanism of action and about the unique determinants of cancer cell sensitivity to each of these inhibitors. METHODS Using a combination of cell viability assays, HDAC enzyme activity measurements, western blots for histone modifications, microarray gene expression analysis and qRT-PCR, we have characterised differences in trichostatin A vs depsipeptide-induced phenotypes in lung cancer, breast cancer and skin cancer cells and in normal cells and have then expanded these studies to other HDAC inhibitors. RESULTS Cell viability profiles across panels of lung cancer, breast cancer and melanoma cell lines showed distinct sensitivities to the pan-inhibitor TSA compared with the class 1 selective inhibitor depsipeptide. In several instances, the cell lines most sensitive to one inhibitor were most resistant to the other inhibitor, demonstrating these drugs act on at least some non-overlapping cellular targets. These differences were not explained by the HDAC selectivity of these inhibitors alone since apicidin, which is a class 1 selective compound similar to depsipeptide, also showed a unique drug sensitivity profile of its own. TSA had greater specificity for cancer vs normal cells compared with other HDAC inhibitors. In addition, at concentrations that blocked cancer cell viability, TSA effectively inhibited purified recombinant HDACs 1, 2 and 5 and moderately inhibited HDAC8, while depsipeptide did not inhibit the activity of purified HDACs in vitro but did in cellular extracts, suggesting a potentially indirect action of this drug. Although both depsipeptide and TSA increased levels of histone acetylation in cancer cells, only depsipeptide decreased global levels of transcriptionally repressive histone methylation marks. Analysis of gene expression profiles of an isogenic cell line pair that showed discrepant sensitivity to depsipeptide, suggested that resistance to this inhibitor may be mediated by increased expression of multidrug resistance genes triggered by exposure to chemotherapy as was confirmed by verapamil studies. CONCLUSIONS Although generally thought to have similar activities, the HDAC modulators trichostatin A and depsipeptide demonstrated distinct phenotypes in the inhibition of cancer cell viability and of HDAC activity, in their selectivity for cancer vs normal cells, and in their effects on histone modifications. These differences in mode of action may bear on the future therapeutic and research application of these inhibitors.