ABSTRACT
PURPOSE: Long-acting formulations of the potent antiretroviral prodrug tenofovir alafenamide (TAF) hold potential as biomedical HIV prevention modalities. Here, we present a rigorous comparison of three animal models, C57BL/6 J mice, beagle dogs, and merino sheep for evaluating TAF implant pharmacokinetics (PKs). METHODS: Implants delivering TAF over a wide range of controlled release rates were tested in vitro and in mice and dogs. Our existing PK model, supported by an intravenous (IV) dosing dog study, was adapted to analyze mechanistic aspects underlying implant TAF delivery. RESULTS: TAF in vitro release in the 0.13 to 9.8 mg d-1 range with zero order kinetics were attained. Implants with equivalent fabrication parameters released TAF in mice and sheep at rates that were not statistically different, but were 3 times higher in dogs. When two implants were placed in the same subcutaneous pocket, a two-week creep to Cmax was observed in dogs for systemic drug and metabolite concentrations, but not in mice. Co-modeling IV and TAF implant PK data in dogs led to an apparent TAF bioavailability of 9.6 in the single implant groups (compared to the IV group), but only 1.5 when two implants were placed in the same subcutaneous pocket. CONCLUSIONS: Based on the current results, we recommend using mice and sheep, with macaques as a complementary species, for preclinical TAF implant evaluation with the caveat that our observations may be specific to the implant technology used here. Our report provides fundamental, translatable insights into multispecies TAF delivery via long-acting implants.
