Abstract

Using human liver microsomes (HLMs) and recombinant human cytochrome P450 (CYP450) isoforms, we identified the major route of pimozide metabolism, the CYP450 isoforms involved, and documented the inhibitory effect of pimozide on CYP450 isoforms. Pimozide was predominantly N-dealkylated to 1,3-dihydro-1-(4-piperidinyl)-2H-benzimidazol-2-one (DHPBI). The formation rate of DHPBI showed biphasic kinetics in HLMs, which suggests the participation of at least two activities. These were characterized as high-affinity (K_m1 andV_max1) and low-affinity (K_m2 and V_max2) components. The ratio of V_max1 (14 pmol/min/mg protein)/K_m1 (0.73 μM) was 5.2 times higher than the ratio of V_max2 (244 pmol/min/mg protein)/K_m2 (34 μM).K_m2 was 91 times higher thanK_m1. The formation rate of DHPBI from 25 μM pimozide in nine human livers correlated significantly with the catalytic activity of CYP3A (Spearman r = 0.79, P = .028), but not with other isoforms. Potent inhibition of DHPBI formation from 10 μM pimozide was observed with ketoconazole (88%), troleandomycin (79%), furafylline (48%) and a combination of furafylline and ketoconazole (96%). Recombinant human CYP3A4 catalyzed DHPBI formation from 10 μM pimozide at the highest rate (V = 2.2 ± 0.89 pmol/min/pmol P450) followed by CYP1A2 (V = 0.23 ± 0.08 pmol/min/pmol P450), but other isoforms tested did not. TheK_m values derived with recombinant CYP3A4 and CYP1A2 were 5.7 μM and 36.1 μM, respectively. Pimozide itself was a potent inhibitor of CYP2D6 in HLMs when preincubated for 15 min (K_i = 0.75 ± 0.98 μM) and a moderate inhibitor of CYP3A (K_i = 76.7 ± 34.5 μM), with no significant effect on other isoforms tested. Our results suggest that pimozide metabolism is catalyzed mainly by CYP3A, but CYP1A2 also contributes. Pimozide metabolism is likely to be subject to interindividual variability in CYP3A and CYP1A2 expression and to drug interactions involving these isoforms. Pimozide itself may inhibit the metabolism of drugs that are substrates of CYP2D6.