There is a need for models useful for predicting the efficacy of agents developed for treating human immunodeficiency virus (HIV) based on information obtained during the drug development process. A pharmacodynamic model that superimposes the pharmacokinetics of anti-HIV nucleoside reverse transcription (RT) and protease inhibitors over a previously published predator-prey model of HIV and CD4 dynamics was developed to address this need. This model was applied to in vitro measurements and patient-derived pharmacokinetics of the unbound antiviral drugs to simulate HIV-1 and CD4 counts versus time and dose. The primary mechanism for nucleoside RT inhibitors was assumed to be competitive inhibition of HIV-1-RT by the active nucleoside triphosphates (NTP). Cellular accumulation and breakdown rates of the NTP were estimated from previous in vivo pharmacokinetic studies. Median inhibition concentrations for the HIV-1 RT enzyme were estimated from previously published cell-free binding studies. The concentration of active protease inhibitor available for binding with HIV-1 protease was assumed equal to the unbound fraction in the plasma. The resulting simulations for mono- and dual nucleoside therapy with zidovudine and lamivudine single dose regimen with the protease inhibitor indinavir, produced similar HIV and CD4 response profiles to those reported in large Phase II and III clinical trials. Based on these findings this pharmacodynamic model can be applied to predict starting doses for a new agent based on simulated biological responses as a function of time for dosage regimens comprising one or two agents. However, the model overestimated the efficacy of highly effective drug combinations where all three agents are combined as in highly active anti-retroviral therapy.
Copyright 2002 Elsevier Science B.V.