The disposition and metabolic fate of 3'-deoxy-2',3'-didehydrothymidine (D4T) were 3'-Deoxy-2',3'-didehydrothymidine (D4T) is a thymidine analog with potent anti-human immunodeficiency virus activity in vitro comparable to that of 3'-azido-3'-deoxythymidine (AZT) (23,24). Previous studies on D4T metabolism, by several groups including our own, have demonstrated that D4T is sequentially phosphorylated by cellular enzymes to D4T-TP, the active form which competitively inhibits TTr incorporation by human immunodeficiency virus reverse transcriptase (3,19,34,35). D4T has shown efficacy in patients with AIDS without substantial hematological toxicity (1, 14). The reduced bone marrow toxicity is probably due to the lower steady-state levels of D4T in nuclear DNA of human bone marrow cells compared with those of AZT under similar conditions (34). Peripheral neuropathy, also reported after treatment with 2',3'-dideoxycytidine or 2',3'-dideoxyinosine, was the dose-limiting toxicity and was reversible (6). Inhibition of mitochondrial DNA synthesis has been suggested as a potential mechanism of 2',3'-dideoxynucleoside-induced peripheral neuropathy (8, 21), but other mechanisms may also be involved (31).Understanding of the catabolic fate of D4T is still limited. Pharmacokinetic studies of D4T in mice, rats, dogs, and monkeys have suggested that this agent is not substantially * Corresponding author. degraded. Approximately 50 to 80% of the administered dose of D4T, in animals, is eliminated in urine unchanged (28,29). In a recent pharmacokinetic study of D4T in humans (13), about half of the administered dose was recovered unchanged. The metabolic fate of the remainder is unknown. The in vivo formation of a glucuronide conjugate is still controversial. Schinazi et al. (29) have suggested the presence of a D4T glucuronide in urine of monkeys, while others have not detected such a conjugate (28). Until recently no D4T metabolites, other than its 5'-phosphorylated derivatives, were reported either in vitro or in vivo (18). D4T was recently demonstrated by this laboratory to be a substrate for Escherichia coli thymidine phosphorylase in vitro, leading to the formation of thymine (30). Interestingly, we also detected the in vitro degradation of D4T to thymine, with subsequent formation of thymidine, in human bone marrow cells (34). This suggests that D4T is cleaved to thymine and that subsequent degradation and/or utilization of thymine in vivo accounts for the unrecovered D4T in clinical trials. Understanding the metabolic fate of D4T may be important for clinical trials design, as well as provide information on the mechanism(s) of action and the toxicity of this agent. The present study examines both the in vitro and the in vivo metabolic disposition of D4T in isolated hepatocytes and in nonhuman primates. Isolated rat hepatocytes were chosen, since we had previously demonstrated the usefulness of this in vitro system in examining the metabolic disposition of AZT and 3'-azido-2',3'-dideoxyuridine, which are structur-