Testosterone lowers aromatase activity in cultured human genital skin fibroblasts

Abstract

The cytochrome P-450-dependent aromatase pathway utilizes the androgens testosterone (T) and androstenedione, as substrates for estrogen formation. In addition, androgens have been shown to influence the level of aromatase activity in various tissues. In cultured human skin fibroblasts, incubation with T for 14 h resulted in a dose-dependent decline in aromatase activity, the concentration of T producing a half-maximal decline being 6 nM. In the presence of T (50 nM), aromatase activity declined in a time-dependent fashion with maximal reduction occurring by 9 h. When aromatase kinetics were determined after preincubation of cells with T, there was a significant decline in the calculated V max with no significant change in the apparent K m, suggesting that incubation of cells with T reduced the number of active enzyme sites. Aromatase activity was unaffected by preincubation of cells with the synthetic androgen methyltrienolone. In addition, the decline in aromatase activity following preincubation with T was observed in cells derived from patients with complete androgen insensitivity demonstrating that the effect of T was not mediated by androgen receptors. Furthermore, new protein synthesis was not necessary for the T-mediated effect as the presence of cycloheximide (50 μg/ml) did not prevent it. When cells were incubated at low oxygen tension, the inhibition of aromatase activity by T was diminished. Testosterone is rapidly metabolized in genital skin fibroblasts to dihydrotestosterone, androstanedione, androsterone, 3α-androstanediol, 3β-androstanediol and estradiol. To determine if a metabolite of T might be responsible for the repression of aromatase activity, aromatase activity was determined in cells following preincubation with various metabolites of T. Preincubation of cells with androstenedione, androstanedione or 3α-androstanediol produced a small but significant decline in aromatase activity, whereas preincubation of cells with dihydrotestosterone, androsterone, or 3β-androstanediol did not have a significant effect. Aromatase activity was also unaffected by preincubation of cells with estradiol or diethylstilbestrol. When aromatase activity was assayed in microsomal preparations from cells preincubated with T, activity was reduced. Although cells preincubated with 50 nM [ 3H]T contained between 0.25 and 0.51 pmol of residual steroid/mg microsomal protein, the amount of [l- 3H]androstenedione and T was insufficient to account for the observed decline in aromatase activity on the basis of competitive inhibition. In summary, preincubation of genital skin fibroblasts with T results in a decline in aromatase activity through a mechanism which does not involve the androgen receptor but which appears to involve a local, oxygen-mediated effect of T on the aromatase complex within the endoplasmic reticulum. We propose that incubation of cells with T may permit the destruction of the aromatase by oxygen-derived free radicals formed during aromatization.