Population pharmacokinetic/pharmacodynamic analysis of CCR5 receptor occupancy by maraviroc in healthy subjects and HIV-positive patients.

BACKGROUND: Maraviroc, a noncompetitive antagonist of the CCR5 coreceptor, was recently approved in the USA as a treatment of HIV infection. For antiretroviral agents that target the virus, antiviral effect can be related to some extent to plasma drug concentrations. For CCR5 antagonists that target the host cells, receptor occupancy in vivo might be a better predictor of efficacy. AIMS To develop a population pharmacokinetic (PK)-pharmacodynamic (PD) model that describes CCR5 receptor occupancy by maraviroc after oral administration at different doses in healthy volunteers and HIV-positive patients and to assess the relevance of receptor occupancy in predicting the decrease in viral load (HIV-1 RNA copies ml(-1)) in HIV-positive patients. METHODS: Receptor occupancy data from 88 individuals enrolled in two multiple dose trials were included in the population PK-receptor binding model. Out of the 88 individuals, 25 were HIV-1-infected patients and had viral load measurements, whereas the remaining 63 were healthy volunteers. Doses ranged from 3 mg b.i.d. to 600 mg q.d. A previously published PK-PD disease model describing the effect of maraviroc on the viral load was updated by replacing its PD module by the receptor occupancy model. Simulated viral load-time profiles with the updated model were compared with the profiles observed in patients. RESULTS: The majority of measured plasma concentrations were associated with receptor occupancy > or = 50% even at the lowest dose of 3 mg b.i.d. A simple direct E(max) model appeared to describe satisfactorily the PK-receptor occupancy relationship. The estimated K(D) was around 0.0894 ng ml(-1), far below the operational in vivo antiviral IC(50) of 8 ng ml(-1). Accordingly, simulations led to marked overprediction of the decrease in viral load-time profiles. CONCLUSIONS: Maraviroc receptor occupancy close to the maximum is required to induce a significant decrease in viral load, indicating that in vivo CCR5 receptor occupancy by maraviroc is not a direct measure of drug inhibitory activity. Considering the imprecision of the measurement in the upper flat part of the maraviroc concentration vs. percent CCR5 occupancy curve, it can reasonably be concluded that routine monitoring of receptor occupancy as a biomarker for maraviroc efficacy will not be helpful. Based on this analysis, it was decided not to use receptor occupancy as a biomarker of viral load inhibition during the development of CCR5 antagonist compounds.