In human endometriotic stromal cells, markedly high levels of aromatase P450 (P450arom) mRNA and promoter II activity are present and can be vigorously stimulated by PGE(2) via a cAMP-dependent pathway to give rise to physiologically significant estrogen biosynthesis. Stromal cells of eutopic endometrium, on the other hand, do not express sufficient levels of P450arom for detectable enzyme activity. Because P450arom is up-regulated in the ovaries of CCAAT/enhancer binding protein (C/EBP) beta knockout mice and activation of the ovarian-type P450arom promoter (II) is responsible for aberrant P450arom expression in endometriosis, we sought here to evaluate the possible roles of C/EBP isoforms in the regulation of P450arom expression in endometriotic vs. eutopic endometrial stromal cells. We previously found that the -517-bp flanking region of promoter II contained the critical cis-acting elements for baseline and cAMP (analog)-induced activity. In this study, we disrupted several potential sequences and found that mutations of a -211/-197-bp cAMP-response element (CRE) and a -317/-304-bp C/EBP binding site abolished both baseline and cAMP-induced promoter II activity. Ectopic expression of C/EBPalpha increased both baseline and cAMP-dependent promoter II activity significantly in endometriotic cells, whereas ectopic expression of C/EBPbeta or C/EBPdelta abolished promoter II activity in both untreated and cAMP-treated endometriotic stromal cells. Comparable changes in promoter II activity were observed using endometrial stromal cells, which showed, however, seemingly diminished levels of baseline and cAMP-induced promoter II activity in comparison with endometriotic cells. EMSA using a probe containing the critical -317/-304-bp C/EBP site upstream of promoter II demonstrated a distinct DNA-protein complex in endometriotic, but not in endometrial stromal cells. This specific complex, however, could not be altered using antibodies against C/EBPalpha, -beta, or -delta. Because CRE is another potential DNA motif that can bind C/EBP isoforms, we next used EMSA using a probe containing the -211/-197-bp CRE and demonstrated that specific DNA-protein complexes contained C/EBPalpha but not C/EBPbeta or C/EBPdelta in endometriotic stromal cells. In contrast, C/EBPbeta and C/EBPdelta but not C/EBPalpha were detected in DNA-protein complexes using nuclear extracts from endometrial stromal cells. Western blotting and immunohistochemistry demonstrated expression of C/EBPalpha, -beta, and -delta in human endometriotic and endometrial stroma and epithelium. Intriguingly, C/EBPbeta was expressed at increased levels in stromal cells of human eutopic endometrium compared with simultaneously biopsied endometriotic tissues. We conclude that both -317/-304 and -211/-197-bp elements in promoter II are critical for the robust cAMP-dependent induction in endometriosis. C/EBPalpha up-regulates, whereas C/EBPbeta and C/EBPdelta inhibit P450arom promoter activity via binding primarily to the -211/-197-bp CRE under in vitro conditions. In vivo down-regulation of C/EBPbeta in endometriotic stromal cells and its up-regulation in endometrial stromal cells may in part account for the induction of P450arom expression in endometriosis and its inhibition in endometrium.