The aim of the present study was to evaluate the use of recombinant human cytochrome P-450 1A2 (rH-CYP1A2) in studies performed in vitro in order to predict metabolic drug-drug interactions occurring in man. In vitro metabolism of tacrine (a CYP1A2 probe) in the presence and absence of fluvoxamine, a CYP1A2 inhibitor, was investigated in human liver mircrosomes and with different rH-CYP. Vmax, Km and Ki determined with human liver microsomes were compared with those observed using rH-CYP1A2, assuming that 1 mg of liver microsomes contains, on average, 69 pmol of CYP1A2. The extent of tacrine metabolism inhibition procured by fluvoxamine with rH-CYP1A2, was compared with previous results observed in man. The Vax and Km for 1-hydroxytacrine formation rates obtained with rH-CYP1A2 were in good agreement with those observed in human liver microsomes (175+/-9 versus 140+/-60 pmol/min/mg for Vmax and 14+/-2 versus 16+/-2 microM for Km, respectively. The Ki of fluvoxamine on 1-hydroxytacrine formation rate observed with rH-CYP1A2 was similar to that observed with human liver microsome (0.35+/-0.05 versus 0.20+/-0.20 microM, respectively). Using the Km, Vmax and Ki determined with rH-CYP1A2, we calculated that fluvoxamine produced an inhibition of 1-, 2- and 4-hydroxytacrine formation rate of 91, 87 and 88%, respectively, in the range of tacrine and fluvoxamine concentrations observed in man. These percentages of inhibition calculated in vitro were in agreement with the percentage of fluvoxamine-dependent decrease in tacrine apparent oral clearance previously observed in man (83+/-13%). We conclude that human CYP1A2 expressed in yeast is a powerful tool to predict and to quantify drug-drug interactions in man.