3'-Fluoro-3'-deoxythymidine (FLT), a candidate anti-AIDS compound in clinical trials, showed anti-human immunodeficiency virus type 1 (HIV-1) potency (50% effective concentration, 0.0052 microM) slightly better than or equal to that of 3'-azido-3'-deoxythymidine (AZT) in MT4 cells and was threefold more potent in H9 cells. There was no FLT resistance demonstrable in the AZT-resistant HIV-1 strains. Both FLT and AZT showed low cytotoxicity for MT4 cells, with selectivity indices (efficacy/toxicity ratio) of greater than 47,000 and greater than 33,000, respectively. Cellular permeation of FLT and thymidine (dThd) was greater than that of AZT, and FLT and dThd permeated the cell membranes by a carrier-mediated mechanism as well as by simple diffusion, as indicated by the existence of nitrobenzylthioinosine-5'-monophosphate-sensitive and -insensitive components. By contrast, transport of AZT into cells was by simple diffusion. The intracellular level of the triphosphate of FLT (FLTTP) in MT4 cells was two- to threefold higher than that of AZT (AZTTP) after exposure to 1.8 microM each compound for 12 h. The elimination kinetics of FLTTP and AZTTP in HIV-1-infected MT4 cells in fresh medium showed biphasic patterns, with initial half-lives of 1.03 and 1.09 h, respectively. In phytohemagglutinin-stimulated human peripheral blood lymphocytes, the FLTTP level was increased 59-fold compared with that in unstimulated cells at 12 h, was four- to sixfold higher than the level of AZTTP in stimulated cells at 12 h, and remained four- to fivefold higher during a 4-h elimination period in fresh medium and twofold higher at the end of a 12-h elimination period. Two- to eightfold more [3H]AZT than [3H]FLT was incorporated into the host cell DNA, and both [3H]AZT and [3H]FLT remained persistently incorporated for over 24 h. The incorporated [3H]AZT and [3H]FLT were alkali labile, whereas incorporated [3H]dThd was alkali stable. Pharmacokinetics of FLT in plasma of monkeys after intravenous (i.v.) administration showed that the FLT concentration in plasma declined, with a half-life of 1.19 +/- 0.1 h; the steady-state volume of distribution was 0.93 +/- 0.2 liter/kg of body weight, and total clearance was 0.56 +/- 0.15 liter/kg. Oral bioavailability of FLT was excellent and comparable to i.v. bioavailability in terms of areas under the concentration-time curves for three monkeys. Of the total dose, 41 to 61% was excreted in urine as unchanged FLT, and only 3.2 to 7.4% of the total dose was identified as glucuronide-conjugated FLT in urine 48 h after i.v. administration to monkeys. We conclude that FLT exhibits an anti-HIV-1 potency similar to that of AZT but with slightly better selectivity of effects and with higher intracellular active metabolite levels.