Purpose: To determine the role of P-glycoprotein (Pgp) on the CNS penetration of the HIV protease inhibitor (PI) amprenavir (141W94) and to test the hypothesis that co-administration of a second HIV PI (ritonavir) could enhance amprenavir's brain penetration in vivo.
Methods: Pgp-mediated efflux was investigated in vitro with Caco-2 cells and in vivo by whole-body autoradiography (WBA). "Genetic" mdr1a/1b double knockout mice, "chemical" Pgp knockout mice generated by administration of the Pgp inhibitor GF120918, and mice pretreated with ritonavir were used in WBA studies to investigate the effects of Pgp modulation on the CNS penetration of amprenavir.
Results: Amprenavir, indinavir, ritonavir, and saquinavir had 2- to 23-fold higher transport rates from the basolateral to apical direction than from the apical to basolateral direction across Caco-2 monolayers. Incubation with GF120918 negated this difference, suggesting that the efflux was Pgp-mediated. WBA studies demonstrated a 13- and 27-fold increase in the brain and a 3.3-fold increase in the CSF concentrations of amprenavir in mice pretreated with GF120918 and in mdr1a/1b double knockout mice. In contrast, pretreatment with ritonavir did not alter the CNS exposure of amprenavir.
Conclusions: These results provide evidence that amprenavir and other HIV PIs are Pgp substrates and that co-administration of a specific Pgp inhibitor will enhance amprenavir's CNS penetration in vivo. These results will have an important therapeutic impact in the treatment of AIDS dementia.