To study mechanisms of aromatase inhibition in brain cells, a highly effective non-steroidal aromatase inhibitor (Fadrozole; 4-[5,6,7,8-tetra-hydroimidazo- (1,5-a)-pyridin-5-yl] benzonitrile HCl; CGS 16949A) was compared with endogenous C-19 steroids, known to be formed in the preoptic area, which inhibit oestrogen formation. Using a sensitive in vitro tritiated water assay for aromatase activity in avian (dove) preoptic tissue, the order of potency, with testosterone as substrate was: Fadrozole (Ki < 1 x 10(-9) M) > 4-androstenedione > 5 alpha-androstanedione > 5 alpha-dihydrotestosterone (Ki = 6 x 10(-8) M) > 5 beta-androstanedione > 5 beta-dihydrotestosterone (Ki = 3.5 x 10(-7) M) > 5 alpha-androstane-3 alpha,17 beta-diol (Ki = 5 x 10(-6) M) > 5 beta-androstane-3 beta,17 beta-diol. Five other steroids, 5 beta-androstane-3 alpha,17 beta-diol, 5 alpha-androstane-3 beta,17 beta-diol, progesterone, oestradiol and oestrone, showed no inhibition at 10(-4) M. The kinetics indicate that endogenous C-19 steroids show similar competitive inhibition of the aromatase as Fadrozole. Mouse (BALB/c) preoptic aromatase was also inhibited by Fadrozole. We conclude that endogenous C-19 metabolites of testosterone are effective inhibitors of the brain aromatase, and suggest that they bind competitively at the same active site as Fadrozole.