The clinical effects of cytosine arabinoside (AraC) are highly dependent on schedule and dose. Many regimens administered to patients are derived from artificial model systems involving permanent leukemic cell lines. The differences in pharmacokinetics between the in vivo situation and such cell lines are largely neglected. However, cytidine deaminase activity in particular has a major impact on AraC pharmacokinetics by degrading AraC to its inactive metabolite AraU, and it has been shown to be of prognostic relevance in the treatment of acute myeloid leukemia. This study therefore investigated cytidine deaminase activities and AraC deamination in a variety of the most commonly used leukemic cell lines and fresh blasts and their impact on the results of an in vitro model system. It was found that cells from different cell lines (BLIN, CEM, HL60, K562, RAJI, REH, U937) vary greatly in cytidine deaminase activity (e.g., 1.89 nmol per min/mg in K562 versus 0.01 in BLIN cells) and degrade between 18.5 (BLIN) and 96.5% (REH) of AraC to AraU in the incubation medium. This degradation results in highly different AraC exposures for different cells (e.g., AUC of 960 ng per h/ml in REH versus 4048 ng per h/ml in BLIN cells) in spite of identical starting concentrations of the drug. Formation of AraCTP as the main cytotoxic metabolite of AraC is significantly influenced by the differences in cell type-dependent cytidine deaminase activity (e.g., 35.6 ng/10(7) cells in REH versus 180.2 ng/10(7) cells in BLIN cells). In contrast to permanent cell lines, fresh leukemic blasts and normal bone marrow mononuclear cells featured low AraC degradation in the model system.