The agonist activity of the antimineralocorticoid spironolactone was evaluated in various cell lines through the use of transfection experiments. The target promoters were derived from the deltaMTV promoter in which one or several glucocorticoid-responsive elements (GRE) were inserted in tandem. Spironolactone at 100 nM activated by 6-fold the GRE/deltaMTV promoter in the human hepatoma HepG2 cell line and only partially prevented the 10-fold activation of this promoter by 0.1 nM aldosterone. Both effects were completely dependent on the cotransfection of an expression vector for the mineralocorticoid receptor. The half-maximal agonist effect of spironolactone was similar to its half-maximal antagonist effect (approximately 10 nM). For the GRE-2/deltaMTV, GRE-4/deltaMTV, and wild-type MMTV promoters, the activation by aldosterone was much more potent (70-, 100-, and 110-fold, respectively), whereas spironolactone elicited a 10-, 24-, and 25-fold activation, respectively. Thus, the effect of both compounds and the relative efficiency of spironolactone, compared with that of aldosterone, were dependent on the number of GREs present in the regulatory region of the promoter. The agonist effect of spironolactone was cell specific. Indeed, although spironolactone agonist activity was observed in H5 kidney tubule cells, none could be detected at concentrations of < or = 1 microM in the CV1 monkey fibroblast cells. In contrast, the antagonist effect was observed in all cells. Furthermore, other antimineralocorticoids, such as RU 26752 and progesterone, also displayed mineralocorticoid receptor-dependent agonist activity in the HepG2 cells. The antiprogesterone RU 486 and the antiandrogen cyproterone acetate were ineffective at < or = 1 microM. In conclusion, we show that under certain experimental conditions, several antimineralocorticoids display significant agonist activity in a cell-specific and promoter-dependent manner.