In vitro perfusion of human placenta was evaluated for characterization of aromatase inhibitors. The results were compared with those in kinetic experiments in cell-free system. Inhibition constants (Ki) were determined by measuring the release of tritiated water during coincubation of human placenta microsomes with varying amounts of [1 beta,2 beta 3H]androstenedione and inhibitor in the presence of NADPH-generating system. Irreversible inactivation constants (Kinact) were determined in a similar manner following preincubation of the microsomes with different amounts of inhibitor for varying times. Lineweaver-Burk plots indicated a competitive type of inhibition with Ki values of 37 nM for 4-hydroxy-androstenedione, 3,700 nM for testolactone, 15 nM for 1-methyl-androsta-1,4-diene-3,17-dione, and 7.5 nM for 19-azido-androstenedione. Additionally, irreversible enzyme inactivation by all four substances could be demonstrated with Kinact values of 3.64 x 10(-3), 0.57 x 10(-3), 0.34 x 10(-3), and 0.69 x 10(-3)sec-1, respectively. Perfusion of a single cotyledon of human term placenta was performed by infusing medium through catheters placed in a fetal artery and in the maternal intervillous space. Perfused medium was collected from a cannulated fetal vein and from the maternal basal plate. The medium was supplemented with [3H]androstenedione (4.2 nM) and inhibitor. The perfusates were analyzed for their [3H]estrone and estradiol content following phenolic partition and Sephadex-LH 20 chromatography. The main results were, (1) the recovery of labelled steroids increased rapidly after perfusion started and reached a plateau within 60 min, when 55 and 30% (mean values) of the infused radioactivity were recovered in the fetal and maternal perfusates, respectively, (2) similar amounts of estrone and estradiol were found in both effluates, whereas androgens (mainly androstenedione and lower amounts of 5 alpha-androstane-3,17-dione) were found nearly exclusively in the fetal perfusate, (3) formation of estrogens (estrone + estradiol) reached a plateau within 20 min of perfusion. (4) The percentage of estrogens formed was not changed by increasing androstenedione concentration in the perfusion medium unless this concentration exceeded 3.5 microM indicating limited capacity of aromatase. (5) The four aromatase inhibitors reduced estrogen formation by 50% at concentrations about 100-fold of their Ki determined in the cell-free system, (6) irreversible aromatase inhibition could not be demonstrated in the perfusion model. It was concluded that the human placenta perfusion model can be successfully used to evaluate aromatase inhibitors.