Arsenic trioxide (ATO, Trisenox™) is currently being investigated as a potential drug to optimize imatinib based treatment of chronic myelogenous leukemia (CML). ATO has antileukemic activity against imatinib sensitive and resistant CML cell lines and primary cells derived from patients with CML. However, preliminary clinical data reveal only moderate activity of ATO in CML (imatinib sensitive or resistant). ATO-activity in acute myelogenous leukemia or multiple myeloma cells is inversely related to the intracellular Glutathione- (GSH-) content with cells bearing low GSH-levels being the most sensitive. To investigate whether this relationship plays a role in CML cell lines sensitive or resistant to the Bcr-Abl inhibitor imatinib, we determined the antiproliferative activity of ATO with regard to the cellular GSH-levels. Furthermore, we examined the ability of GSH-depletion as a therapeutic tool to sensitize CML cells to ATO. MTS-proliferation assays were performed to determine the concentration of ATO needed to induce 50% cellular growth inhibition (IC50). Cell lines used were the imatinib sensitive CML blast crisis lines AR230-s, KCL22-s, LAMA-s as well as their imatinib resistant derivatives AR230-r1, KCL22-r1, LAMA-r1. Known ATO-sensitive AML cell lines (NB4, HL60) with low GSH-content were also included in this study. Intracellular GSH-levels were measured biochemically using a commercially available kit. Protein content was analyzed using the Bradford method. ATO toxicity was analyzed using trypan blue exclusion and flow cytometric analysis of Annexin/PI-stained cells. MTS-proliferation assays indicate a cell type dependent activity of ATO with IC50-values ranging from 0.37±0.03 μM (NB4) up to 6.9±1.43 μM (AR230-r1). The most sensitive NB4 cells express low GSH-levels (8.74±2.9 nmol/mg), whereas highly resistant AR230-r1 cells reveal a threefold increased GSH-content (26.7±8nmol/mg). Imatinib resistance in AR230-r1 and LAMA-r1 cells is not associated with significant GSH-content modulation when compared to the imatinib naïve counterparts (23.6±5.4nmol/mg [AR230-s] vs 26.7±8nmol/mg [AR230-r1]; 8.3±1.8nmol/mg [LAMA-s] vs 6.9±0.2 [LAMA-r1]). Treatment of AR230-s cells with 100 μM of the GSH-depleting agent L-Buthionine-Sulfoximine (BSO) for 12 h leads to significant downregulation of cellular GSH (23.6±5.4 nmol/mg [control] vs 4,25±0,53 nmol/mg [100 μM BSO]). Treatment with BSO alone does not affect cellular viability nor induces apoptosis. Subsequent cotreatment of AR230-s cells with ATO (1 μM) and BSO (100 μM) for 24 h reduces viability to 31.6 % compared to untreated cells. In contrast, treatment with 1 μM ATO alone does not affect viability. Flow cytometric analysis of apoptosis reflects viability data with a 5.5 fold increase of apoptotic cells in the combined treated fraction. Similar data were generated using the KCL22-s cell line. Experiments using primary patient cells are currently in progress. Our data indicate that high intracellular GSH-content confers relative resistance to ATO in Bcr-Abl positive cell lines regardless whether they are imatinib sensitive or resistant. A promising tool to increase the antileukemic activity of ATO is the application of GSH-depleting agents. Therefore, GSH-dependent response to ATO treatment needs further investigation in individual CML patients.