It is well documented that estrogens have atheroprotective effects in humans. Peripheral aromatization of circulating androgens has been demonstrated to exert estrogenic actions in many human tissues, especially in men and post-menopausal women. Recently, production of estrogens mediated by aromatase was detected in cultured smooth muscle cells and aortic endothelial cells and it has been proposed that this in situ produced estrogen may influence the development of atherosclerosis. In this study, we first examined aromatase expression by immunohistochemistry in human aortic tissues obtained from 85 autopsy cases (50 males, 35 females, 49.6 +/- 2.9-year-old) and by mRNA in situ hybridization in 10 cases. We then semi-quantified the level of aromatase mRNA in aortic tissues of 12 men and 12 post-menopausal women by reverse transcriptase-polymerase chain reaction (RT-PCR) to examine whether or not and in which cell types aromatase was expressed. We also studied alternative use of multiple exon 1 of its gene and immunolocalization of 17beta-hydroxysteroid dehydrogenase type I (17beta-HSD I), which converts estrone produced by aromatase to estradiol, a biologically active estrogen. Aromatase immunoreactivity and mRNA hybridization signals and 17beta-HSD I immunoreactivity were all detected in smooth muscle cell (SMC) of the media and thickened intima, especially in SMC adjacent to an atheromatous plaque. The levels of aromatase mRNA were significantly higher in female cases than in male cases (P<0.05). The amount of aromatase mRNA was significantly higher in the specimens with fibroatheroma (P<0.05) than other lesions, and was also significantly higher in the cases utilizing 1c (I.3) or 1d (PII) of exon 1, i.e. gonadal types than those utilizing 1b (I.4), i.e. fibroblasts type as the promoter (P<0.01). These results suggest that estrone and estradiol are produced in SMC of the human aortic wall and that their production is mediated by aromatase and 17beta-HSD I, respectively. Moreover, it was suggested that aromatase overexpression, possibly as a result of alternative splicing, may play some roles in the development of atherosclerosis.