Based on acute cytotoxicity studies, selenosulfate (SeSO3-) has been suggested to possess a generally higher toxic activity in tumor cells than selenite. The reason for this difference in cytotoxic activity remained unclear. In the present study, cytotoxicity tests with human hepatoma (HepG2), malignant melanoma (A375), and urinary bladder carcinoma cells (T24) showed that the selenosulfate toxicity was very similar between all three tested cell lines (IC50 6.6-7.1 µM after 24 h). It was largely independent of exposure time and presence or absence of amino acids. What changed, however, was the toxicity of selenite, which was lower than that of selenosulfate only for HepG2 cells (IC50 > 15 µM), but similar to and higher than that of selenosulfate for A375 (IC50 4.7 µM) and T24 cells (IC50 3.5 µM), respectively. Addition of amino acids to T24 cell growth medium down-regulated short-term selenite uptake (1.5 versus 12.9 ng Se /106 cells) and decreased its cytotoxicity (IC50 8.4 µM), rendering it less toxic than selenosulfate. The suggested mechanism is a stronger expression of the xc-¬ transport system in the more sensitive T24 compared to HepG2 cells which creates a reductive extracellular microenvironment and facilitates selenite uptake by reduction. Selenosulfate is already reduced and so less affected. The cytotoxic activity of selenosulfate and selenite to tumor cells therefore depends on the sensitivity of each cell line, supplements like amino acids as well as the reductive state of the extracellular environment.