Aromatase and Endometriosis
Posted 05/13/2004
Original story at

Serdar E. Bulun, M.D.; Zongjuan Fang, M.D.; Gonca Imir, M.D.; Bilgin Gurates, M.D.; Mitsutoshi Tamura, M.D.; Bertan Yilmaz, M.S.; David Langoi, D.V.M.; Sanober Amin, B.S.; Sijun Yang, M.D.; Santanu Deb, Ph.D.

Abstract and Introduction


Aromatase P450 (P450arom) is the key enzyme for biosynthesis of estrogen, which is an essential hormone for the establishment and growth of endometriosis. There is no detectable aromatase enzyme activity in normal endometrium; therefore, estrogen is not locally produced in endometrium. Endometriosis tissue, however, contains very high levels of aromatase enzyme, which leads to production of significant quantities of estrogen. Moreover, one of the best-known mediators of inflammation and pain, prostaglandin E2, strikingly induces aromatase enzyme activity and formation of local estrogen in this tissue. Additionally, estrogen itself stimulates cyclo-oxygenase-2 and therefore increases the formation of prostaglandin E2 in endometriosis. We were able to target this positive feedback cycle in endometriosis using aromatase inhibitors. In fact, pilot trials showed that aromatase inhibitors could decrease pelvic pain associated with endometriosis.


Endometriosis is one of the most prominent public health problems in the United States.[1,2] It is characterized by the presence of endometrial glands and stroma within the pelvic peritoneum and other extrauterine sites and is linked to pelvic pain and infertility. It is estimated to affect 5% of women in the reproductive age group.[1,2] Endometriosis is a polygenically inherited disease of complex multifactorial etiology.[3] Sampson's theory of transplantation of endometrial tissue on the pelvic peritoneum via retrograde menstruation is the most widely accepted explanation for the development of pelvic endometriosis because of convincing circumstantial and experimental evidence.[4] Given that retrograde menstruation is observed in almost all cycling women, endometriosis is postulated to develop as a result of the coexistence of a defect in clearance of the menstrual efflux from pelvic peritoneal surfaces, possibly involving the immune system.[5] Alternatively, intrinsic molecular aberrations in pelvic endometriotic implants were proposed to significantly contribute to development of endometriosis. Aberrant expression of aromatase, certain cytokines and tissue metalloproteinases, deficiency of 17β-hydroxysteroid dehydrogenase (17β-HSD) type 2 and resistance to the protective action of progesterone are some of these molecular abnormalities.[6-12] Because endometriosis is an estrogen-dependent disorder, aromatase expression and 17β-HSD type 2 deficiency are of paramount importance in the pathophysiology of endometriosis. Aromatase causes accumulation of the biologically active estrogen estradiol (E2) in this tissue. 17β-HSD type 2 that metabolizes E2 to estrone (E1) is deficient in endometriosis. The combination of these two abnormalities serves to maintain high levels of E2 in endometriosis tissue.

Mechanisms of Estrogen Biosynthesis in Human Tissues The enzyme aromatase catalyzes the conversion of androstenedione and testosterone to estrone and estradiol. The gene that encodes this enzyme is expressed in several human tissues and cells such as ovarian granulosa cells, placental syncytiotrophoblast, adipose tissue and skin fibroblasts, and the brain. In the reproductive-age woman, the ovary is the most important site of estrogen biosynthesis, and this takes place in a cyclic fashion. Upon binding of follicle-stimulating hormone (FSH) to its G-protein-coupled receptor in the granulosa cell membrane, intracellular cyclic adenosine monophosphate (cAMP) levels rise and enhance binding of two critical transcription factors [i.e., steroidogenic factor-1 (SF-1) and cAMP response element binding protein (CREB)], to the classically located proximal promoter II of the aromatase gene.[13,14] This, in turn, activates aromatase expression and consequently estrogen secretion from the preovulatory follicle.[14,15]

On the other hand, in postmenopausal women, estrogen formation takes place in extraglandular tissues such as the adipose tissue and skin[16-18] (Fig. 1). In contrast to cAMP regulation of aromatase expression in the ovary, this is controlled primarily by cytokines [interleukin (IL)-6, IL-11, tumor necrosis factor alpha (TNFα)] and glucocorticoids via the alternative use of promoter I.4 in adipose tissue and skin fibroblasts.[15] The major substrate for aromatase in adipose tissue and skin is androstenedione of adrenal origin. In postmenopausal women, ~2% of circulating androstenedione is converted to estrone, which is further converted to estradiol in these peripheral tissues. This may give rise to significant serum levels of estradiol capable of causing endometrial hyperplasia or even carcinoma.[17,18]

Click to zoom
Figure 1. (click image to zoom) Mechanisms of estrogen production. Estradiol (E2) is either directly secreted by the ovary or produced in peripheral sites (adipose tissue and skin). The principal substrate for extraglandular aromatase activity in ovulatory women is androstenedione (A) of adrenal and ovarian origins. In women receiving gonadotropin-releasing hormone (GnRH) agonists or postmenopausal women, however, the adrenal remains as the primary source of androstenedione. Androstenedione is converted by aromatase to estrone (E1) in adipose tissue and skin fibroblasts. Estrone is further converted to estradiol by 17β hydroxysteroid dehydrogenase (17β-HSD) (reductase) activity in these peripheral tissues. Thus peripheral aromatization is the major source for circulating estradiol in the postmenopausal period or during ovarian suppression.

Aromatase in Endometriosis

Endometrium and myometrium contain extremely high levels of estrogen receptors and, thus, are prime targets of estrogen. Until recently, estrogen action has been classically viewed to occur only via an "endocrine" mechanism: in other words, it was thought that only circulating estradiol, whether secreted by the ovary or formed in the adipose tissue, could exert an estrogenic effect after delivery to target tissues via the bloodstream. Studies on aromatase expression in breast cancer demonstrated that paracrine mechanisms play an important role in estrogen action in this tissue.[19] Estrogen produced by aromatase activity in breast adipose tissue fibroblasts was demonstrated to promote the growth of adjacent malignant breast epithelial cells.[20] Finally, we demonstrated an "intracrine" effect of estrogen in uterine leiomyomas and endometriosis: Estrogen produced by aromatase activity in the cytoplasm of leiomyoma smooth muscle cells or endometriotic stromal cells can exert its effects by readily binding to its nuclear receptor within the same cell.[6,21,22] Disease-free endometrium and myometrium, on the other hand, lack aromatase expression.[21,22]

Among estrogen-responsive pelvic disorders, aromatase expression was studied in greatest detail in endometriosis.[6,7,22,23] First, extremely high levels of aromatase mRNA were found in extraovarian endometriotic implants and endometriomas. Second, endometriosis-derived stromal cells in culture incubated with a cAMP analog displayed extraordinarily high levels of aromatase activity comparable to that in placental syncytiotrophoblast.[22] These exciting findings led us to test a battery of growth factors, cytokines, and other substances that might induce aromatase activity via a cAMP-dependent pathway in endometriosis. Prostaglandin E2 (PGE2) was found to be the most potent known inducer of aromatase activity in endometriotic stromal cells.[22] In fact, this PGE2 effect was found to be mediated via the cAMP-inducing EP2 receptor subtype (our unpublished observations). Moreover, estrogen was reported to increase PGE2 formation by stimulating cyclooxygenase type 2 (COX-2) enzyme in endometrial stromal cells in culture.[24] Thus, a positive feedback loop for continuous local production of estrogen and prostaglandins (PGs) is established, favoring the proliferative and inflammatory characteristics of endometriosis (Fig. 2). Additionally, aromatase mRNA was also detected in the eutopic endometrial samples of women with moderate to severe endometriosis (but not in those of disease-free women) albeit in much smaller quantities compared with endometriotic implants.[6] This may be suggestive of a genetic defect in women with endometriosis, which is manifested by this subtle finding in the eutopic endometrium. We propose that, when defective endometrium with low levels of aberrant aromatase expression reaches the pelvic peritoneum by retrograde menstruation, it causes an inflammatory reaction that exponentially increases local aromatase activity (i.e., estrogen formation) induced directly or indirectly by PGs and cytokines.[22] It would be rather naive to propose that aberrant aromatase expression is the only important molecular mechanism in the development and growth of pelvic endometriosis. There may be many other molecular mechanisms that favor the development of endometriosis: abnormal expression of proteinase type enzymes that remodel tissues or their inhibitors (matrix metalloproteinases, tissue inhibitor of metalloproteinase-1), certain cytokines (IL-6, RANTES [regulated on activation, normal T cell expressed and secreted]), and growth factors (epidermal growth factor) represent some of the mechanisms.[8-11] Alternatively, a defective immune system that fails to clear peritoneal surfaces of the retrograde menstrual efflux has been proposed in the development of endometriosis.[5,25] The development of endometriosis in an individual woman probably requires the coexistence of a threshold number of these aberrations. Nonetheless, the clinical importance of aromatase expression pertains because we could treat endometriosis using aromatase inhibitors.[26,27]

Click to zoom
Figure 2. (click image to zoom) A positive feedback cycle for estrogen and prostaglandin formation. Estradiol (E2) in an endometriotic lesion arises from several body sites. In an ovulatory woman, estradiol is secreted directly from the ovary in a cyclic fashion. In the early follicular phase and after menopause, peripheral tissues (adipose and skin) are the most important sources to account for the circulating estradiol. Estradiol is also produced locally in the endometriotic implant itself in both ovulatory and postmenopausal women. The most important precursor, androstenedione (A) of adrenal origin, becomes converted to estrone (E1), which is in turn reduced to estradiol in the peripheral tissues and endometriotic implants. We demonstrated significant levels of 17β-hydroxysteroid dehydrogenase type 1 (reductase) expression in endometriosis, which catalyzes the conversion of estrone to estradiol.[12] Estradiol both directly and indirectly (through cytokines) induces cyclo-oxygenase-2 (COX-2), which gives rise to elevated concentrations of prostaglandin E2 (PGE2) in endometriosis.[24] PGE2 in turn, is the most potent known stimulator of aromatase in endometriotic stromal cells.[22] This establishes a positive feedback loop in favor of continuous estrogen formation in endometriosis.

Regulation of Aromatase Expression in Endometriotic Stromal Cells

PGE2 induces aromatase activity strikingly by increasing cAMP levels in endometriotic stromal cells.[22] On the other hand, neither cAMP analogs nor PGE2 was capable of stimulating any detectable aromatase activity in eutopic endometrial stromal cells in culture. The obvious question then addresses the molecular differences that give rise to aromatase expression in endometriosis and its inhibition in eutopic endometrium. To explore this, we first determined that the cAMP-inducible promoter II was used for in vivo aromatase expression in endometriotic tissue.[7] Then a stimulatory transcription factor, SF-1, and an inhibitory factor, chicken ovalbumin upstream promoter transcription factor (COUP-TF), were found to compete for the same binding site in aromatase promoter II. COUP-TF was ubiquitously expressed in both eutopic endometrium and endometriosis, whereas SF-1 was expressed, specifically in endometriosis but not in eutopic endometrium and binds to aromatase promoter more avidly than COUP-TF.[7] Thus SF-1 and other transcription factors (e.g., CREB) activate transcription in endometriosis whereas COUP-TF, which occupies the same DNA site in eutopic endometrium, inhibits this process[7] (Fig. 3). In summary, one of the molecular alterations leading to local aromatase expression in endometriosis but not in normal endometrium is the aberrant production of SF-1 in endometriotic stromal cells, which overcomes the protective inhibition maintained normally by COUP-TF in the eutopic endometrium.

Click to zoom
Figure 3. (click image to zoom) Molecular mechanism of aberrant aromatase expression in endometriosis. Normally, aromatase is not expressed in endometrium. Thus there is no local estrogen formation. This is maintained by inhibitory proteins that bind to DNA in the promoter region of the P450arom gene. These inhibitory factors may directly bind to DNA in the promoter region (inhibitors 1 and 2) or may bind to transcription factors and repress their activities through protein-protein interactions (corepressors 1 and 2). In endometriosis, however, the inhibitors and corepressors are decreased or absent. Instead, they are replaced by aberrantly present proteins that bind to the P450arom promoter and activate its transcription. Again, these proteins either directly bind to DNA as classical transcription factors (stimulators 1 and 2) or interact with DNA-binding proteins and enhance their transcriptional activity (coactivators 1 and 2). PGE2, prostaglandin E2; EP2-R, type 2 receptor for PGE2; cAMP, cyclic adenosine monophosphate; P450arom, aromatase P450.

Rationale for Treatment of Endometriosis With Aromatase Inhibitors

Endometriosis is successfully suppressed by estrogen deprivation with GnRH analogs or the induction of surgical menopause. Control of pelvic pain with gonadotropin-releasing hormone (GnRH) agonists is usually successful during and immediately after the treatment, whereas pain associated with endometriosis returns in up to 75% of these women.[28,29] There may be multiple reasons for the failure of GnRH agonist treatment of endometriosis. One likely explanation is the presence of significant estradiol production that continues in the adipose tissue, skin, and endometriotic implant per se during the GnRH agonist treatment (Fig. 4). Therefore, blockage of aromatase activity in these extraovarian sites with an aromatase inhibitor may keep a larger number of patients in remission for longer periods of time. The most striking evidence for the significance of extraovarian estrogen production is the recurrence of endometriosis after successfully completed hysterectomy and bilateral salpingo-oophorectomy in several women.[26,30] Endometriotic tissue in one such aggressive case was found to express much higher levels of aromatase mRNA compared with premenopausal endometriosis.[26] We reported the treatment of a 57-year-old overweight woman who had recurrence of severe endometriosis after hysterectomy and bilateral salpingo-oophorectomy. Two additional laparotomies were performed owing to persistent severe pelvic pain and bilateral ureteral obstruction leading to left renal atrophy and right hydronephrosis. Treatment with megestrol acetate was ineffective. A large (3 cm) vaginal endometriotic lesion contained unusually high levels of aromatase mRNA. The patient was given anastrozole (an aromatase inhibitor) for 9 months. Despite the addition of calcium and alendronate (a nonsteroidal inhibitor of bone resorption), bone density in the lumbar spine decreased by 6.2%. The occurrence of significant bone loss in this particular case should be studied further. Dramatic relief of the pain and regression of the vaginal endometriotic lesion were observed within the first month of treatment. At the same time, circulating estradiol levels were reduced to 50% of the baseline value. Markedly high pretreatment levels of aromatase mRNA in the endometriotic tissue became undetectable in a repeat biopsy 6 months later, and the lesion nearly disappeared after 9 months of therapy. Two potential mechanisms may have accounted for this strikingly successful result. First, there was evidence of suppression of peripheral (i.e., skin and adipose tissue) aromatase activity, giving rise to a significant decrease in serum estradiol level. Second, unusually high levels of aromatase expression in the endometriotic lesion disappeared after treatment with the aromatase inhibitor, anastrozole. Besides the expected direct inhibition of aromatase activity in endometriosis by anastrozole, the disappearance of aromatase mRNA expression in the lesion may be explicable by denial of estrogen that is known to stimulate local biosynthesis of PGE2, which, in turn, stimulates aromatase expression (Fig. 2).

Click to zoom
Figure 4. (click image to zoom) The role of aromatase inhibition in the treatment of endometriosis. Estrogen formation takes place via various mechanisms in endometriosis patients: (1) delivery from the ovary and peripheral tissues via circulation and (2) local biosynthesis in endometriosis. Gonadotropin-releasing hormone (GnRH) analogs will eliminate estradiol secreted by the ovary by downregulating the hypothalamic-pituitary unit. In cases resistant to treatment with GnRH agonists or in postmenopausal endometriosis, the use of aromatase inhibitors to block estrogen formation in the peripheral tissues as well as in endometriotic stromal cells may be critical in controlling the growth of endometriosis.

A recent publication now shows that premenopausal women with endometriosis can be treated with a combination of an aromatase inhibitor and norethindrone acetate. These women previously had not responded to the existing surgical or medical treatments exemplified by multiple recurrences or persistence of pain during the treatment. The majority of them had pain relief and decreased laparoscopically detectable endometriosis after a 6-month treatment with letrozole and norethindrone acetate.[27]

In conclusion, aromatase inhibitors may represent a new generation of medications for the treatment of endometriosis in the near future. Larger clinical trials are needed to address this question.

Funding Information

The preparation of this article was supported, in part, by NIH grants HD38691 and TW01339.

Reprint Address

Serdar E. Bulun, M.D., Department of Obstetrics and Gynecology, Northwestern University, 333 E. Superior Street, Suite 484, Chicago, IL 60611.

Commentary by Steph: