Discrepancies in the P-glycoprotein-mediated transport of (18)F-MPPF: a pharmacokinetic study in mice and non-human primates

Several in vivo studies have found that the 5-HT(1A) PET radioligand (18)F-MPPF is a substrate of rodent P-glycoprotein (P-gp). However, in vitro assays suggest that MPPF is not a substrate of human P-gp. We have now tested the influence of inhibiting P-gp on the brain kinetics of (18)F-MPPF in mice and non-human primates. We measured the peripheral kinetics (arterial input function, metabolism, free fraction in plasma (f(P))) during (18)F-MPPF brain PET scanning in baboons with or without cyclosporine A (CsA) infusion. We measured (3)H-MPPF transport at the mouse BBB using in situ brain perfusion in P-gp/Bcrp deficient mice and after inhibiting P-gp with PSC833. There was an unexpected 1.9-fold increase in brain area under the curve in CsA-treated baboons (n = 4), with no change in radiometabolite-corrected arterial input. However, total volume of distribution corrected for f(P) (V(T)/f(P)) remained unchanged. In situ brain perfusion showed that P-gp restricted the permeability of the mouse BBB to (3)H-MPPF while Bcrp did not. These and previous in vitro results suggest that P-gp may not influence the permeability of human BBB to (18)F-MPPF. However, CsA treatment increased (18)F-MPPF free fraction, which is responsible for a misleading, P-gp unrelated enhanced brain uptake.