Previous in vitro investigations demonstrated that human leukemia cells, when incubated with hematopoietic growth factors such as granulocyte-colony-stimulating factor (G-CSF), augment the accumulation of the triphosphate 1-beta-D-arabinofuranosylcytosine (ara-C cytarabine). To test whether G-CSF infusion prior to ara-C infusion would biologically modulate the accumulation of ara-9-beta-D-arabinofuranosylcytosine 5'-triphosphate (ara-CTP) and other ara nucleotides in the leukemia blasts during therapy, protocols were designed to infuse G-CSF prior to fludarabine (9-beta-D-arabinofuranosyl-2-fluoroadenine monophosphate) and ara-C to increase the accumulation of the active triphosphates [9-beta-D-arabinofuranosyl-2-fluoroadenine 5'-triphosphate (F-ara-ATP) and ara-CTP] in acute myelogenous leukemia (AML) blasts during therapy. To complement these in vivo studies, ex vivo accumulation of ara-CTP was also investigated before and after G-CSF infusion. Patients (n = 5) treated on the fludarabine/ara-C/G-CSF regimen received a 30 mg/m2 dose of fludarabine followed by a 2 g/m2 dose of ara-C infused i.v. for 4 h. Beginning at 24 h, and every day, patients received a 6-h infusion of 400 microgram/m2 G-CSF. At 48 h, the fludarabine and ara-C couplet was repeated. Comparison of F-ara-ATP pharmacokinetics in circulating AML cells of patients on the fludarabine/ara- C/G-CSF regimen demonstrated that the area under concentration time curve (AUC) of F-ara-ATP increased significantly (median, 1.4-fold; range, 0.9-1.5; P = 0.045) after G-CSF infusion. This was due to an increased rate of F-ara-ATP accumulation by AML cells. The AUC of ara-CTP, on the other hand, was not affected (median, 1.0-fold; range, 1.0-1.2; P = 0.571) after G-CSF infusion. Because fludarabine potentiates the accumulation of ara-CTP, the effect of G-CSF on ara-CTP metabolism may not be evident in the AML blasts of patients on the fludarabine/ara-C/G-CSF regimen. To determine the effect of G-CSF when ara-C was infused alone, four additional patients were treated on a pilot protocol in which ara-C (2 g/m2) was infused on days 1 and 3 and G-CSF on day 2. The AUC of ara-CTP accumulation in these patients decreased by a median of 48% after G-CSF infusion. Consistent with these in vivo investigations, ex vivo ara-CTP accumulation was decreased in the AML blasts after G-CSF infusion. Based on these data it could be concluded that (a) infusion of G-CSF before fludarabine augmented the rate of F-ara-ATP synthesis in circulating AML blasts during therapy, suggesting that G-CSF may benefit fludarabine therapy by biological modulation; (b) G-CSF did not increase ara-CTP accumulation, rather it may have caused it to decrease; and (c) these data imply that when G-CSF and ara-C are used in combination, administration of fludarabine prior to ara-C may maintain the ara-CTP AUC.