{"paper_id":"3c4aa619-ef4f-4ec4-9ef8-3522b56f18a8","body_text":"Endocrine-Related Cancer  (1999) 6 293-301\nEndocrine-Related Cancer (1999) 6 293-301 Online version via http://www.endocrinology.org\n1351-0088/99/006-293  © 1999 Society for Endocrinology  Printed in Great Britain                                                                 \nIntroduction\nEndometriosis is a complex disorder that is characterized\nby the presence of endometrial tissue in ectopic sites\noutside the uterus and is linked to pelvic pain and infer-\ntility. The prevalence of endometriosis in women of\nreproductive age is estimated to be as high as 10% (Aral\n& Cates 1983, Wheeler 1989). Endometriosis is diagnosed\nin approximately 25% of women who undergo lapar-\noscopy because of pelvic pain and in 20% of infertile\nwomen (Hasson 1976, Goldstein et al. 1980, Eskenazi &\nWarner 1997). It is a chronic and progressive disease that\nmay give rise to a variety of severe and disabling\nsymptoms including painful menses, painful intercourse,\nchronic pelvic pain, and infertility.\nEndometriosis is probably inherited in a polygenic\nmanner with an etiology of complex and multifactorial\nnature (Olive & Schwartz 1993). The most widely\naccepted mechanism for the pelvic disease is implantation\nof endometrial tissue on the peritoneum through retro-\ngrade menstruation, which was first proposed by Sampson\n(1927). Since retrograde menstruation occurs in at least\n90% of all women, the presence of immunologic defects\nin women with endometriosis were hypothesized (Syrop\n& Halme 1987, Hill & Anderson 1989, Hill 1992, Olive &\nEstrogen production in endometriosis    \nand use of aromatase inhibitors to          \ntreat endometriosis\nS E Bulun1, K Zeitoun, K T akayama, L Noble, \nD Michael, E Simpson, A Johns, M Putman               \nand H Sasano 2\nDepartment of Obstetrics and Gynecology, University of T exas Southwestern Medical Center at Dallas, Texas, \nUSA\n1Department of Obstetrics-Gynecology, University of Illinois at Chicago, 820 S. Wood St, M/C 808 Chicago, \nIllinois 60612, USA\n2Department of Pathology, Tohoku University School of Medicine, Sendai, Japan\n(Requests for offprints should be addressed to S E Bulun)\nAbstract\nEstrogen is the most important known factor that stimulates the growth of endometriosis. Estrogen\ndelivery to endometriotic implants was classically viewed to be only via the circulating blood in an\nendocrine fashion. We recently uncovered an autocrine positive feedback mechanism, which favored\nthe continuous production of estrogen and prostaglandin (PG)E 2 in the endometriotic stromal cells.\nThe enzyme, aromatase, is aberrantly expressed in endometriotic stromal cells and catalyzes the\nconversion of C\n19 steroids to estrogens, which then stimulate cyclooxygenase-2 to increase the levels\nof PGE 2. PGE 2, in turn, is a potent inducer of aromatase activity in endometriotic stromal cells.\nAromatase is not expressed in the eutopic endometrium. Aromatase expression in endometriosis and\nits inhibition in eutopic endometrium are controlled by the competitive binding of a stimulatory tran-\nscription factor, steroidogenic factor-1, and an inhibitory factor, chicken ovalbumin upstream promoter-\ntranscription factor to a regulatory element in the aromatase P450 gene promoter. In addition, we find\nthat endometriotic tissue is deficient in 17 β-hydroxysteroid dehydrogenase type 2, which is normally\nexpressed in eutopic endometrial glandular cells and inactivates estradiol-17 β to estrone. This defi-\nciency is another aberration that favors higher levels of estradiol-17 β in endometriotic tissues in\ncomparison with the eutopic endometrium. The clinical relevance of local aromatase expression in\nendometriosis was exemplified by the successful treatment of an unusually aggressive form of recur-\nEndocrine-Related Cancer (1999) 6 293-301\nrent endometriosis in a postmenopausal woman using an aromatase inhibitor.\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access\n\n\nBulun et al.: Aromatase in endometriosis\n294\nSchwartz 1993). These defects would presumably lead to\nimpaired clearance of the menstrual debris on the\nperitoneal surfaces. On the other hand, data from other\nlaboratories suggested that intrinsic molecular aberrations\nin the endometrium of women with endometriosis\nfacilitated implantation of the endometrium on the pelvic\nperitoneum. The proposed intrinsic aberrations that were\nimportant in this hypothesis included deficient expression\nof an integrin (Lessey et al. 1994) and overexpression of\ncomplement 3 (Isaacson et al. 1990) and certain cytokines\n(Ryan & Taylor 1997). Moreover, certain molecules such\nas tissue metalloproteinase inhibitor type 1 were shown to\nbe expressed in endometriosis but not in the endometrium\n(Sharp et al.  1993, Sharpe-Timms et al.  1995). We\nrecently demonstrated significant levels of aromatase\nactivity and mRNA in the stromal cell component of\nendometriosis, whereas aromatase expression was either\nabsent or barely detectable in the eutopic endometrium\n(Noble et al.  1996, 1997). We would like to clarify the\nterminology to be used here in reference to the tissues and\ncells that were studied: the terms ‘endometriotic tissue’\nand ‘endometriosis’ will refer to the pathological ectopic\nendometrium-like tissues in the pelvic peritoneum or\novaries. Extremely high levels of aromatase expression\nwere found in the stromal cell component of endometriotic\ntissues. The term ‘endometrium’ refers to the eutopic or\nintrauterine endometrial tissue in its normal location.\nAromatase expression is absent in the eutopic endo-\nmetrium of women without endometriosis or any other\nuterine pathology and is barely detectable (only by reverse\ntranscription (RT)-PCR) in the eutopic endometrium of\nwomen with endometriosis.\nConsiderable circumstantial and laboratory evidence\nsuggests that endometriosis is an estrogen-dependent\ndisease (Dizerga et al. 1980). For example, the usefulness\nof gonadatropin-releasing hormone agonists in supp-\nressing ovarian steroidogenesis and progestins (which act\nto inhibit estrogen action) in the management of\nendometriosis is well recognized. Moreover, we recently\nreported the successful treatment of an unusually\naggressive type of recurrent postmenopausal endometri-\nosis using an aromatase inhibitor (Takayama et al. 1998).\nThe responsiveness of endometriosis to estrogen and\nprogesterone is also evident from hormone-dependent\nhistological changes in this tissue similar to those in\neutopic endometrium. In addition, the expression of\nestrogen and progesterone receptors has been demon-\nstrated in endometriotic tissue (Lessey et al. 1989).\nThe delivery of estrogen to endometriotic implants has\nbeen assumed by many to be only via the circulating blood\nin an endocrine fashion. We, and others, however, have\nrecently demonstrated markedly high levels of aromatase\nP450 mRNA and activity in pelvic endometriotic implants\n(Noble et al. 1996, 1997, Kitawaki et al. 1997). Moreover,\nprostaglandin (PG)E\n2, which is produced in very high\nlevels in endometriotic tissues, was found to be the most\npotent inducer of aromatase activity in endometriosis-\nderived stromal cells (Badawy et al. 1984, De Leon et al.\n1988, Karck et al.  1996, Noble et al. 1996, 1997). The\nproduction of PGE\n2 in eutopic endometrial stromal cells,\nin turn, was demonstrated to be greatly stimulated by\ncytokines and estradiol-17 β via enhancement of\ncyclooxygenase-2 (COX-2) expression (Ishihara et\nal.1995, Kennard et al. 1995, Huang et al. 1996). Finally,\nthe expression of 17 β-hydroxysteroid dehydrogenase\n(17β-HSD), the enzyme that is induced by progesterone\nand inactivates estradiol-17β (by conversion to estrone) in\neutopic endometrium, was recently shown to be deficient\nin endometriotic tissues biopsied during the mid-secretory\nphase of the cycle (Zeitoun et al. 1998). Collectively, these\ndata support the model in which alterations in the\nexpression of aromatase, COX-2, and 17 β-HSD type 2 in\nendometriosis may lead to increased local concentrations\nof estradiol-17 β by enhancing its production and\ndiminishing its metabolism (Fig. 1). In fact, higher\nconcentrations of estradiol-17 β have been detected in the\nperitoneal fluid of women with endometriosis than normal\ncontrols (DeLeon et al. 1986).\nMechanisms of estrogen biosynthesis and \nmetabolism in endometriosis\nEstrogen biosynthesis and metabolism                     \nin humans\nAromatase P450 (P450arom) catalyzes the conversion of\nandrostenedione to estrone, and testosterone to\nestradiol-17β in a number of human cells, including\nplacental syncytiotrophoblast, ovarian granulosa cells,\nand adipose and skin fibroblasts (Simpson et al. 1994). In\nthe human, aromatase expression is regulated by usage of\nalternative and partially tissue-specific promoters in the\nplacenta (promoter I.1), adipose tissue (promoters I.4, I.3\nand II), and ovary (promoter II). Activation of these\npromoters, and thus aromatase expression, in these tissues\nis controlled by various hormones. In ovarian granulosa\ncells, follicle-stimulating hormone stimulates the\nactivation of promoter II via a cAMP-dependent signaling\npathway. In adipose fibroblasts, glucocorticoids and\nmembers of the interleukin (IL)-6 cytokine family give\nrise to activation of promoter I.4, whereas treatment with\ncAMP analogs or PGE\n2 switches the promoter use to I.3\nand II in these cells. Estrogen biosynthesis in peripheral\ntissues (adipose tissue, skin, and endometriosis) is\ndependent for substrate on circulating androstenedione,\nwhich is produced by the adrenal cortex. Importantly, the\nproduct of aromatase activity in these tissues, namely\nestrone, is only very weakly estrogenic, and must\ntherefore be converted to estradiol-17 β in tissue sites of\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access\n\n\nEndocrine-Related Cancer  (1999) 6 293-301\n295\naction. Evidence from several laboratories indicates that\n17β-HSD type 1, which is present in these peripheral\ntissues, catalyzes this conversion. We recently demon-\nstrated the expression of 17β-HSD type 1 in endometriotic\ntissues (Zeitoun et al. 1998). Another 17 β-HSD isozyme,\n17β-HSD type 2, catalyzes the conversions of estradiol-\n17β to estrone, and testosterone to androstenedione, in a\nnumber of human tissues, including the placenta and liver\n(Andersson & Moghrabi 1997). In addition, very high\nlevels of 17 β-HSD type 2 transcripts have been demon-\nstrated in the glandular epithelial cell fraction of the\nhuman endometrium during the secretory phase, suggest-\ning that progesterone stimulates this enzyme (Casey et al.\n1994, Mustonen et al.  1998). In fact, estradiol\ndehydrogenase activity (oxidation of estradiol-17 β to\nestrone) in endometrial tissues and isolated glandular\nepithelial component has been shown to be stimulated by\nprogesterone in earlier reports (Tseng & Gurpide 1974,\n1975, Satyaswaroop et al. 1979). The inactivation of\nestradiol-17β to estrone by the secretory phase endo-\nmetrium has been viewed as an important protective\nmechanism in this estrogen-responsive tissue.\nAromatase expression in Müllerian-               \nderived tissues\nMüllerian-derived tissues are targets of estrogen action.\nBecause aromatase is expressed in extraglandular tissues,\nwe have investigated the regulation of expression of this\ngene in estrogen-dependent neoplasia or disorders that\ninvolve müllerian-derived tissues. First, using an\n[\n3H]water assay and quantitative RT-PCR, we were\nunable to detect aromatase activity or mRNA in disease-\nfree endometrium, myometrium, or endometrial stromal\ncells in culture (derived from eutopic endometrium from\ndisease-free women) (Bulun et al.  1993). On the other\nhand, aromatase expression was demonstrable in the\ndisease states of these tissues. For example, in endometrial\ncancer, aromatase transcripts are readily demonstrable by\nRT-PCR, and aromatase expression was found to be\nregulated by promoter II in this malignant tissue (Bulun et\nal. 1994). Next, extremely high levels of aromatase\ntranscripts were found in uterine leiomyoma tissues from\n32 of 35 women and in apparently normal myometrial\ntissues adjacent to leiomyomata (18 of 24 evaluated) but\nFigure 1 Estrogen biosynthesis and metabolism in endometriotic lesions. Estradiol-17ß (E 2) reaches the endometriotic \nlesion via the bloodstream (and possibly peritoneal fluid). Aromatase P450 (P450arom) in the stromal cell catalyzes the \nconversion of androstenedione (A) to estrone (E1), which is further reduced to E2 by 17β-HSD type 1 in the endometriotic \ntissue. (At this time, the cell type that expresses 17 β-HSD type 1 in endometriotic lesions is not known.) E 2 is normally \ninactivated by conversion to E1 by 17β-HSD type 2 in epithelial cells of the eutopic endometrium. In endometriotic tissue, \nhowever, E2 is not metabolized because of the lack of 17β-HSD type 2, giving rise to increased local concentration of this \npotent estrogen. Elevated E2, in turn, will promote the growth of endometriotic tissue and, also, local PGE 2 formation in \nstromal cells. Since PGE2 is the most potent known inducer of aromatase in endometriosis, this will complete the positive \nfeedback cycle that favors increased levels of E2 in endometriosis through enhanced biosynthesis and deficient \nmetabolism.\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access\n\n\nBulun et al.: Aromatase in endometriosis\n296\nnot in normal myometrial tissues from disease-free uteri\n(Bulun et al.  1994 a). In leiomyoma-derived smooth\nmuscle cells maintained in primary culture, treatment with\ndibutyryl (Bt2)cAMP acted to increase aromatase activity.\nAddition of phorbol diacetate potentiated this stimulatory\neffect of Bt\n2cAMP . Again, promoter II was found to be\nprimarily responsible for aromatase expression in\nleiomyoma tissues and cells. These findings led us to\ninvestigate the expression of aromatase in endometriosis,\nanother estrogen-dependent disorder of a Müllerian tissue.\nIn an initial study, we found high levels of P450arom\ntranscripts in all 17 endometriotic tissues from\nextraovarian pelvic sites evaluated (Noble et al. 1996,\n1997). The levels of P450arom transcripts (normalized to\ntotal RNA) in endometriotic tissues were 3.2 times those\nin adipose tissue. Eutopic endometrium (obtained by\nendometrial curettage) from these patients also contained\nP450arom transcripts, albeit in quantities barely\ndetectable by RT-PCR (Noble et al. 1996, 1997).\nP450arom transcripts could not be detected in the disease-\nfree pelvic peritoneum proximal to endometriotic implants\nor in the intrauterine endometrial curettings from disease-\nfree women. Thus, we hypothesize that estrogen-\nresponsive müllerian-derived neoplasia and endometriosis\nare disorders with aberrant aromatase expression that may\ngive rise to an increase in the concentration of bioactive\nestrogen in situ (Bulun et al. 1994b). Moreover, a common\ncAMP-dependent signaling pathway seems to be\nresponsible for activating P450arom promoter II in these\ndisorders (Bulun et al. 1997).\nRegulation of aromatase expression in \nendometriotic stromal cells (Noble \net al.1997)\nUpon demonstration of relatively high quantities of\nP450arom transcripts in endometriosis (much higher than\nthose found in the adipose tissue), we next used\nendometriotic stromal cells in monolayer culture as a\nmodel system to study the regulation of aromatase\nexpression (Noble et al.1997). Glands and stromal cells of\novarian endometriomas and eutopic endometrium were\nseparated by the method of Satyaswaroop et al. (1979) and\nthe stromal cells were cultured using a previously reported\nprotocol (Satyaswaroop et al.  1979, Ryan et al. 1994).\nThese cultured stromal cells were reported to retain\nestrogen receptors and estrogen responsiveness (Ryan et\nal. 1994). The endometriotic stromal cells cultured by this\nmethod were also characterized in terms of vimentin and\ncytokeratin expression (Ryan et al. 1994). Baseline\naromatase activity in endometriotic stromal cells ranged\nfrom 0.65 to 6 pmol/4 h per mg protein. No significant\nstimulation of aromatase activity was observed by various\ncytokines (IL-1β, IL-2, IL-6, IL-11, oncostatin M, IL-15,\ntumour necrosis factor) or steroids (estradiol-17 β,\nprogesterone agonist R5020, dexamethasone). Bt\n2cAMP\ninduced aromatase activity in these cells by 26 to 60 times\nthe baseline values (Fig. 2), whereas the addition of\nphorbol acetate neither potentiated nor diminished this\nresponse. Because of the inflammatory nature of\nendometriosis, we treated these stromal cells with various\nprostanoids. Whereas treatments with PGI\n2, PGF 2α, or\nPGJ2 failed to elicit a response, PGE 2 treatment gave rise\nto a dose-dependent induction of aromatase activity by up\nto 19- to 44-fold in endometriosis-derived cells from\ndifferent patients (Fig. 2) (Noble et al. 1997). These\nchanges in aromatase activity were accompanied by\ncomparable changes in the levels of P450arom mRNA. A\nmodified rapid amplification of 5'-cDNA ends (5'-RACE)/\nSouthern hybridization of the promoter-specific\nsequences in P450arom transcripts revealed almost\nexclusive use of promoter II for aromatase expression in\nPGE\n2- or Bt2cAMP-treated endometriotic cells.\nThe summary of our findings thus far is as follows.\nPGE2 induction of aromatase activity in endometriotic\nstromal cells is mediated possibly through increased\nintracellular levels of cAMP. The basis for markedly high\nlevels of aromatase expression in endometriosis in\ncontrast with absent or barely detectable quantities in the\neutopic endometrium may be due to the transformation of\nendometrial stromal cells after implantation in the pelvic\nperitoneum and ovary in response to locally produced\nFigure 2 Aromatase activity in endometriosis-derived \nstromal cells. Confluent stromal cells in primary culture \nwere maintained for 24 h in serum-free medium. \nTreatments consisted of (1) dexamethasone (DEX; \n250 nmol/l) in serum-free medium plus one of the \nfollowing cytokines: IL-1β (1 ng/ml), IL-2 (2 ng/ml), or \nIL-15 (2 ng/ml); (2) Bt\n2cAMP (0.5 mmol/) in serum-free \nmedium; and (3) PGE2 (10–8 mol/l). All treatments \nwere continued for 24 h. Note that Bt2cAMP and PGE2 \ntreatments gave rise to extremely high activity levels \ncomparable with those in the placental syncytio-\ntrophoblast or ovarian granulosa cells.\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access\n\n\nEndocrine-Related Cancer  (1999) 6 293-301\n297\nparacrine factors. The potential aromatization capability\nof eutopic endometrial cells from women with the genetic\npredisposition to develop endometriosis may facilitate the\nimplantation process and growth in pelvic peritoneum by\nincreasing local estradiol-17 β concentrations by the\nactivities of aromatase and 17 β-HSD type 1 (Noble et\nal.1996, 1997, Zeitoun et al. 1998). Estradiol-17β, in turn,\nwill induce the activity of COX-2, the rate-limiting\nenzyme for PGE\n2 biosynthesis (Huang et al. 1996). The\ninflammatory process in endometriotic tissues giving rise\nto increased production of cytokines (e.g. IL-1 β, tumour\nnecrosis factor α) by monocytes and macrophages will\nalso promote PGE 2 production in this tissue (Guan et\nal.1997). Thus a positive feedback cycle is established,\nwhereby local production of estrogen and PGE 2 is\nenhanced by complex molecular interactions (Fig. 3).\nStimulation of aromatase P450 promoter (II) \nactivity in endometriosis and its inhibition in \nendometrium are regulated by competitive \nbinding of steroidogenic factor-1 (SF-1) and \nCOUP-TF to the same \ncis-acting element \n(Zeitoun et al. 1999)\nAn intriguing observation made during the previous\nstudies was the lack of aromatase expression in eutopic\nendometrial stromal cells in contrast with significant\nlevels of aromatase mRNA and activity in endometriotic\nstromal cells, which can be strikingly induced by cAMP\nanalogs. Thus, we sought to determine whether\ndifferential binding of transcription factors to the\nP450arom promoter in response to cAMP is a mechanism\ninvolved in this process. First, we demonstrated by\n5'-RACE that P450arom expression in pelvic endo-\nmetriotic lesions is regulated almost exclusively via the\nalternative promoter II. Then, luciferase reporter plasmids\ncontaining deletion mutations of the 5'-flanking region of\npromoter II were transfected into endometriotic stromal\ncells. We identified two critical regulatory regions for\ncAMP induction of promoter II activity: (i) –214/–100 bp\nproximal region responsible for a 3.7-fold induction, and\n(ii) –517/–214 bp distal region responsible for potentiation\nof cAMP response up to 13-fold. In the –214/–100 bp\nregion, we studied eutopic endometrial and endometriotic\nnuclear protein binding to a nuclear receptor half-site\n(NRHS) (AGGTCA) and an imperfect cAMP-responsive\nelement (CRE) (TGCACGTCA). Using an electro-\nphoretic mobility-shift assay, CRE-binding activity in\nnuclear proteins from both endometriotic and eutopic\nendometrial cells was found to give rise to formation of\nidentical DNA-protein complexes, which led us to\nconclude that CRE did not account for differential\naromatase expression. The NRHS probe, on the other\nhand, formed a distinct complex with nuclear proteins\nfrom endometriotic cells, which migrated at a much faster\nrate than the complex formed with nuclear proteins from\neutopic endometrial cells. Employing recombinant\nproteins and antibodies against SF-1 and COUP-TF, we\ndemonstrated that COUP-TF but not SF-1 bound to NRHS\nin eutopic endometrial cells, whereas SF-1 was the\nprimary NRHS-binding protein in endometriotic cells. In\nfact, COUP-TF transcripts were present in both eutopic\nendometrial ( n=12) and endometriotic tissues ( n=8),\nwhereas SF-1 transcripts were detected in all\nendometriotic tissues ( n=12), but in only three out of 15\neutopic endometrial tissues. Moreover, we demonstrated a\ndose-dependent direct competition between SF-1 and\nCOUP-TF for occupancy of the NRHS, to which SF-1\nbound with a higher affinity. Finally, overexpression of\nSF-1 in endometriotic cells strikingly potentiated baseline\nand cAMP-induced activities of the –517 promoter II\nFigure 3 Local estrogen biosynthesis in endometriotic \ntissue. This model indicates the origin of estadiol-17β \nin a postmenopausal woman or a woman in her \nreproductive years, who is treated with a \ngonadotropin-releasing hormone agonist and thus has \ninactive ovaries. Therefore, the body sites of estrogen \nbiosynthesis are peripheral tissues (adipose and skin) \nand the endometriotic implant itself. The most \nimportant precursor, androstenedione, of adrenal \norigin is converted to estrone which is, in turn, reduced \nto estradiol-17β in the peripheral tissues and \nendometriotic implants. We demonstrated significant \nlevels of 17β-HSD type 1 expression in endometriosis, \nwhich catalyzes the conversion of estrone to \nestradiol-17β. Estradiol-17β induces prostaglandin \nsynthase-2 (COX-2), which gives rise to elevated \nconcentrations of PGE\n2 in endometriotic tissues. \nPGE2 in turn, is the most potent known inducer of \naromatase in endometriotic stromal cells. Therefore, a \npositive feedback loop in favor of continuous estrogen \nformation is established in endometriosis.\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access\n\n\nBulun et al.: Aromatase in endometriosis\n298\nconstruct, whereas overexpression of COUP-TF almost\ncompletely abolished these activities. In conclusion,\nCOUP-TF is responsible for the inhibition of P450arom\nexpression in eutopic endometrial stromal cells, which\nlack SF-1 expression in the majority (80%) of the samples,\nwhereas aberrant SF-1 expression in endometriotic stro-\nmal cells overrides this inhibition by competing for the\nsame DNA-binding site, which is likely to account for\nhigh levels of baseline and cAMP-induced aromatase\nactivity (Fig. 4).\nDeficient expression of 17 β-HSD type 2 in \nendometriosis in contrast with eutopic \nendometrium (Zeitoun \net al. 1998)\nInterconversions of estradiol-17β ↔ estrone are catalyzed\nby two enzymes encoded by two separate genes (Penning\nFigure 4 Proposed mechanism for the regulation of aromatase P450 expression by SF-1 and COUP-TF in eutopic \nendometrium and endometriosis. (A) Binding of COUP-TF readily to the nuclear receptor half-site in aromatase P450 \npromoter II in the absence of SF-1 in eutopic endometrial stromal cells. Thus, COUP-TF exerts its inhibitory effect on the \ncomplex of general transcription factors (GTFs) that bind to TATA box. (B) In endometriotic stromal cells that contain both \nSF-1 and COUP-TF , however, SF-1 binds to the nuclear receptor half-site with a higher affinity than COUP-TF and \nsynergizes with CRE-binding protein (CREB) and other transcription factors to activate the transcription of the CYP19 \n(P450arom) gene in response to cAMP .\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access\n\n\nEndocrine-Related Cancer  (1999) 6 293-301\n299\n1997). 17β-HSD type 1 favors the formation of estradiol-\n17β, whereas type 2 inactivates estradiol-17 β by convert-\ning it to estrone. We recently demonstrated by Northern-\nblot analysis the presence of transcripts of 17 β-HSD type\n1, which catalyzes the conversion of estrone to\nestradiol-17β, in both eutopic endometrium and endo-\nmetriosis. Thus, it follows that the product of the aroma-\ntase reaction, namely estrone, which is weakly estrogenic\ncan be converted to the potent estrogen, estradiol-17 β, in\nendometriotic tissues.\nIt was previously demonstrated that progesterone\nstimulates the inactivation of estradiol-17 β through con-\nversion to estrone in eutopic endometrial epithelial cells.\nSubsequently, 17β-HSD type 2 was shown to catalyze this\nreaction, and its transcripts were detected in the epithelial\ncell component of eutopic endometrium in secretory\nphase. Because estradiol-17 β plays a critical role in the\ndevelopment and growth of endometriosis, we studied\n17β-HSD type 2 expression in endometriotic tissues and\neutopic endometrium. We demonstrated by Northern-blot\nanalysis the presence of 17β-HSD type 2 transcripts in all\nRNA samples of secretory eutopic endometrium ( n=12)\nbut not in samples of secretory endometriotic lesions\n(n=10), including paired samples of secretory eutopic\nendometrium and endometriosis ob-tained simultaneously\nfrom four patients. These transcripts were not detectable\nin any paired samples of proliferative eutopic endo-\nmetrium or endometriosis ( n=4), as expected. Next, we\nconfirmed these findings by demonstration of immuno-\nreactive 17 β-HSD type 2 in epithelial cells of secretory\neutopic endometrium in 11 out of 13 samples employing\na monoclonal antibody against 17 β-HSD type 2, whereas\n17β-HSD type 2 was absent from paired secretory\nendometriotic tissues ( n=4). Proliferative eutopic endo-\nmetrial (n=8) and endometriotic ( n=4) tissues were both\nnegative for immunoreactive 17 β-HSD type 2 except for\nbarely detectable levels in one eutopic endometrial\nsample. Finally, we sought to determine whether deficient\n17β-HSD type 2 expression in endometriotic tissues is due\nto impaired progesterone action in endometriosis. We\ndetermined by immuno-histochemistry the expression of\nprogesterone and estrogen receptors in these paired\nsamples of secretory (n=4) and proliferative (n=4) eutopic\nendometrium and endometriosis, and no differences could\nbe demonstrated. In conclusion, inactivation of estradiol-\n17β is impaired in endometriotic tissues as the result of\ndeficient expression of 17 β-HSD type 2, which is\nnormally expressed in eutopic endometrium in response to\nprogesterone. The lack of 17 β-HSD type 2 expression in\nendometriosis is not due to alterations in the levels of\nimmunoreactive progesterone or estrogen receptors in this\ntissue and may be related to an inhibitory aberration in the\nsignaling pathway that regulates 17 β-HSD type 2\nexpression.\nThe first reported use of an aromatase inhibitor \nto treat endometriosis (Takayama 1998)\nAromatase inhibitors have been widely used to treat breast\ncancer (Brodie 1991). We recently evaluated a\n57-year-old woman, who presented with recurrent severe\nendo-metriosis after hysterectomy and bilateral salpingo-\noophorectomy. Two additional laparotomies were per-\nformed because of severe pelvic pain and bilateral ureteral\nobstruction giving rise to left renal atrophy and right\nhydronephrosis. Recently, recurrent pelvic endometriosis,\nevident from a 30 mm vaginal lesion visible on speculum\nexamination, did not respond to oral megestrol acetate\ntreatment for 4 months. We administered anastrozole (an\naromatase inhibitor) orally, 1 mg/day, and elemental\ncalcium, 1.5 g/day, for 9 months. Alendronate (a non-\nestrogenic inhibitor of bone resorption), 10 mg/day, was\nadded to this regimen. The vaginal lesion was biopsied\nbefore and 6 months after the onset of treatment. The\ncirculating levels of estradiol-17 β were reduced to\napproximately 50% of the baseline value after treatment\nwith anastrozole. Pain rapidly decreased and completely\ndisappeared after the second month of treatment. The\n30×30×20 mm bright-red polypoid vaginal lesion was\nreduced to a 3 mm area of gray tissue by the end of 9\nmonths of treatment. Markedly high pretreatment levels of\nP450arom mRNA in the endometriotic tissue became\nundetectable in a rebiopsy specimen after 6 months of\ntreatment. Bone density of the lumbar spine had decreased\nby 6.2% after 9 months of treatment. No other side effects\nwere noted. This is the first description of the use of an\naromatase inhibitor in the treatment of endometriosis. The\nshort-term results were extraordinarily successful, with\nelimination of pain and near-complete eradication of\nimplants associated with severe endometriosis not\nresponsive to other therapy. The occurrence of significant\nbone loss despite the addition of alendronate to the\ntreatment regimen in this particular case should be studied\nfurther in large clinical trials. Besides the expected\ninhibition of aromatase enzyme activity by anastrozole,\nthe disappearance of aromatase mRNA expression in the\nlesion may be explained by denial of estrogen which is\nknown to stimulate local biosynthesis of PGE\n2, which in\nturn, stimulates aromatase expression (Fig. 5). We\nconclude that the recently developed potent aromatase\ninhibitors are candidate drugs in the treatment of\nendometriosis that is resistant to standard regimens.\nConclusions\nThe development and growth of endometriosis is\nestrogen-dependent. Several molecular aberrations were\nfound to be present in endometriotic tissues (in contrast\nwith the eutopic endometrium), which favor increased\nlocal levels of estradiol-17 β. In fact, we uncovered a\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access\n\n\nBulun et al.: Aromatase in endometriosis\n300\npositive feedback mechanism that is responsible for\ncontinuous formation of estradiol-17 β and PGE2 through\nupregulation of aromatase and COX-2 in endometriotic\nstromal cells. Levels of estradiol-17 β in endometriotic\ntissue are further increased by impaired inactivation of this\nsteroid because of deficient 17β-HSD type 2 expression in\nendometriotic epithelial cells. Aberrant regulation of\nsteroidogenic enzymes in endometriotic tissues giving rise\nto elevated estradiol-17 β levels is possibly one of many\nmetabolic abnormalities that promote the development\nand growth of this tissue. These studies have already led\nus to successfully use an aromatase inhibitor to treat\nendometriosis. We believe that determination of such\nmolecular aberrations in endometriosis will give rise to\nidentification of other molecular targets for potential\ntreatments.\nAcknowledgements\nThis work was supported by an unrestricted grant from the\nAmerican Society for Reproductive Medicine-Organon\n(to SEB) and an American Association of Obstetricians\nand Gynecologists Foundation Fellowship Award (to KZ).\nThe authors thank Rosemary Bell for expert editorial\nassistance.\nReferences\nAndersson S & Moghrabi N 1997 Physiology and molecular \ngenetics of 17β-hydroxysteroid dehydrogenases. Steroids 62 \n143-147.\nAral SO & Cates W, Jr 1983 The increasing concern with \ninfertility: why now? Journal of the American Medical \nAssociation 250 2327-2331.\nBadawy S, Cuenca V , Marshall L, Munchback R, Rinas A & \nCoble D 1984 Cellular components in peritoneal fluid in \ninfertile patients with and without endometriosis. Fertility and \nSterility 42 704-708.\nBrodie A 1991 Aromatase and its inhibitors: an overview. \nJournal of Steroid Biochemistry and Molecular Biology 40 \n255-261.\nBulun SE, Mahendroo MS & Simpson ER 1993 Polymerase \nchain reaction amplification fails to detect aromatase \ncytochrome P450 transcripts in normal human endometrium \nor decidua. Journal of Clinical Endocrinology and \nMetabolism 76 1458-1463.\nBulun SE, Economos K, Miller D & Simpson ER 1994a CYP19 \n(aromatase cytochrome P450) gene expression in human \nmalignant endometrial tumors. Journal of Clinical \nEndocrinology and Metabolism 79 1831-1834.\nBulun SE, Simpson ER & Word RA 1994b Expression of the \nCYP19 gene and its product aromatase cytochrome P450 in \nhuman leiomyoma tissues and cells in culture. Journal of \nClinical Endocrinology and Metabolism 78 736-743.\nBulun SE, Noble LS, Takayama K, Michael MD, Agarwal V , \nFisher C, Zhao Y , Hinshelwood MM, Ito Y & Simpson ER \n1997 Endocrine disorders associated with inappropriately \nhigh aromatase expression. Journal of Steroid Biochemistry \nand Molecular Biology 61 133-139.\nCasey ML, MacDonald PC & Andersson S 1994 17 β-Hydroxy-\nsteroid dehydrogenase type 2: chromosomal assignment and \nprogestin regulation of gene expression in human endo-\nmetrium. Journal of Clinical Investigation 94 2135-2141.\nDe Leon FD, Vijayakumar R, Brown M, Rao CV , Y ussman MA \n& Schultz G 1986 Peritoneal fluid volume, estrogen, \nprogesterone, prostaglandin, epidermal growth factor \nconcentrations in patients with and without endometriosis. \nObstetrics and Gynecology 68 189-194.\nDe Leon FD, Vijayakumar R, Rao CV & Yussman M 1988 \nProstaglandin F\n2α and E2 release by peritoneum with and \nwithout endometriosis. International Journal of Fertility 33 \n48-51.\nDizerga GS, Barber DL & Hodgen GD 1980 Endometriosis: role \nof ovarian steroids in initiation, maintenance and suppression. \nFertility and Sterility 33 649-653.\nEskenazi B & Warner M 1997 Epidemiology of endometriosis. \nObstetrics and Gynecology Clinics of North America  24 235-\n258.\nGoldstein DP, deCholnoky C, Emans SJ & Leventhal JM 1980 \nLaparoscopy on the diagnosis and management of pelvic pain \nin adolescents. Journal of Reproductive Medicine 24 251-256.\nGuan Z, Baier LD & Morrison AR 1997 p38 Mitogen-activated \nprotein kinase down-regulates nitric oxide and up-regulates \nFigure 5 The molecular basis for the treatment of \nendometriosis using an aromatase inhibitor. After \nadministration of a non-steroidal competitive \naromatase inhibitor such as anastrozole, aromatase \nactivity is significantly reduced in the peripheral tissues \nand in endometriotic implants, giving rise to markedly \ndiminished estradiol-17β availability for endometriosis \nfrom both sources. Moreover, the positive feedback \nloop involving PGE\n2 stimulation of local aromatase \nexpression is also interrupted. The end result is \nsignificantly lower concentrations of estradiol-17 β in \nendometriotic tissues.\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access\n\n\nEndocrine-Related Cancer  (1999) 6 293-301\n301\nprostaglandin E2 biosynthesis stimulated by interleukin-1 β. \nJournal of Biological Chemistry 272 8083-8089.\nHasson HM 1976 Incidence of endometriosis in diagnostic \nlaparoscopy. Journal of Reproductive Medicine 16 135-138.\nHill JA 1992 Immunology and endometriosis. Fertility and \nSterility 58 262-264.\nHill J & Anderson D 1989 Lymphocyte activity in the presence \nof peritoneal fluid from fertile women and infertile women \nwith and without endometriosis . American Journal of \nObstetrics and Gynecology 161 861-864.\nHuang JC, Dawood MY & Wu KK 1996 Regulation of \ncyclooxygenase-2 gene in cultured endometrial stromal cells \nby sex steroids. American Society for Reproductive Medicine \n5 (Abstract).\nIsaacson KB, Galman M, Coutifaris C & Lyttle CR 1990 \nEndometrial synthesis and secretion of complement \ncomponent-3 by patients with and without endometriosis. \nFertility and Sterility 53 836-840.\nIshihara O, Matsuoka K, Kinoshita K, Sullivan M & Elder M \n1995 Interleukin-1β-stimulated PGE\n2 production from early \nfirst trimester human decidual cells is inhibited by \ndexamethasone and progesterone. Prostaglandins 49 15-26.\nKarck U, Reister F, Schafer W, Zahradnik H & Breckwoldt 1996 \nM. PGE2 and PGF2α release by human peritoneal \nmacrophages in endometriosis. Prostaglandins 51 49-60.\nKennard E, Zimmerman P, Friedman C & Kniss D 1995 \nInterleukin-1β induces cyclooxygenase-2 in cultured human \ndecidual cells. American Journal of Reproductive \nImmunology 34 65-71.\nKitawaki J, Noguchi T, Amatsu T, Maeda K, Tsukamoto K, \nYamamoto T, Fushiki S, Osawa Y & Honjo H 1997 \nExpression of aromatase cytochrome P450 protein and \nmessenger ribonucleic acid in human endometriotic and \nadenomyotic tissues but not in normal endometrium. Biology \nof Reproduction 57 514-519.\nLessey BA, Metzger DA, Haney AF & McCarty KS, Jr 1989 \nImmunohistochemical analysis of estrogen and progesterone \nreceptors in endometriosis: comparison with normal \nendometrium during the menstrual cycle and the effect of \nmedical therapy. Fertility and Sterility 51 409-415.\nLessey BA, Castelbaum AJ, Sawin SW, Buck CA, Schinnar R, \nBilker W & Strom BL 1994 Aberrant integrin expression in \nthe endometrium of women with endometriosis. Journal of \nClinical Endocrinology and Metabolism 79 643-649.\nMustonen MVJ, Isomaa VV , V askivuo T, Tapanainen J, Poutanen \nMH, Stenback F, Vihko RK & Vihko PT 1998 Human \n17β-hydroxysteroid dehydrogenase type 2 messenger \nribonucleic acid expression and localization in term placenta \nand in endometrium during the menstrual cycle. Journal of \nClinical Endocrinology and Metabolism 83 1319-1324.\nNoble LS, Simpson ER, Johns A & Bulun SE 1996 Aromatase \nexpression in endometriosis. Journal of Clinical \nEndocrinology and Metabolism 81 174-179.\nNoble LS, Takayama K, Putman JM, Johns DA, Hinshelwood \nMM, Agarwal VR, Zhao Y , Carr BR & Sulun SE 1997 \nProstaglandin E\n2 stimulates aromatase expression in \nendometriosis-derived stromal cells. Journal of Clinical \nEndocrinology and Metabolism 82 600-606.\nOlive DL & Schwartz LB 1993 Endometriosis. New England \nJournal of Medicine 328 1759-1769.\nPenning TM 1997 Molecular endocrinology of hydroxysteroid \ndehydrogenases. Endocrine Reviews 18 281-305.\nRyan IP & Taylor RN 1997 Endometriosis and infertility: new \nconcepts. Obstetrical and Gynecological Survey  52 365-371.\nRyan I, Schriock ED & Taylor R 1994 Isolation, characterization, \nand comparison of human endometrial and endometriosis \ncells in vitro. Journal of Clinical Endocrinology and \nMetabolism 78 642-649.\nSampson JA 1927 Peritoneal endometriosis due to the menstrual \ndissemination of endometrial tissue into the peritoneal cavity. \nAmerican Journal of Obstetrics and Gynecology 14 422-425.\nSatyaswaroop PG, Bressler RS, de la Pena MM & Gurpide E \n1979 Isolation and culture of human endometrial glands. \nJournal of Clinical Endocrinology and Metabolism 48 639-\n641.\nSharpe KL, Zimmer RL, Griffin WT & Penney LL 1993 \nPolypeptides synthesized and released by human endo-\nmetriosis differ from those of the uterine endometrium in cell \nand tissue explant culture. Fertility and Sterility 60 839-851.\nSharpe-Timms KL, Penney LL, Zimmer RL, Wright JA, Zhang \nY & Surewicz K 1995 Partial purification and amino acid \nsequence analysis of endometriosis protein-II (ENDO-II) \nreveals homology with tissue inhibitor of metallo-\nproteinases-1 (TIMP-1). Journal of Clinical Endocrinology \nand Metabolism 80 3784-3787.\nSimpson ER, Mahendroo MS, Means GD, Kilgore MW, \nHinshelwood MM, Graham-Lorence S, Amarney B, Ito Y , \nFisher CR, Michael MD, Mendelson CR & Bulun SE 1994 \nAromatase cytochrome P450, the enzyme responsible for \nestrogen biosynthesis. Endocrine Reviews 15 342-355.\nSyrop CH & Halme J 1987 Peritoneal fluid environment and \ninfertility. Fertility and Sterility  48 1-9.\nTakayama K, Zeitoun K, Gunby RT, Sasano H, Carr BR & Bulun \nSE 1998 Treatment of severe postmenopausal endometriosis \nwith an aromatase inhibitor. Fertility and Sterility 69 709-713.\nTseng L & Gurpide E 1974 Estradiol and 20 α-dihydro-\nprogesterone dehydrogenase activities in human endometrium \nduring the menstrual cycle. Endocrinology 94 419-423.\nTseng L & Gurpide E 1975 Induction of human endometrial \nestradiol dehydrogenase by progestins. Endocrinology 97 \n825-833.\nWheeler JM 1989 Epidemiology of endometriosis-associated \ninfertility. Journal of Reproductive Medicine 34 41-46.\nZeitoun K, Takayama K, Sasano H, Suzuki T, Moghrabi N, \nAndersson S, Johns A, Meng L, Putman M, Carr B & Bulun \nSE 1998 Deficient 17β-hydroxysteroid dehydrogenase type 2 \nexpression in endometriosis: failure to metabolize \nestradiol-17β. Journal of Clinical Endocrinology and \nMetabiolism 83 4474-4480.\nZeitoun K, Takayama K, Michael MD & Bulun SE 1999 \nStimulation of aromatase P450 promoter (II) activity in \nendometriosis and its inhibition in endometrium are regulated \nby competitive binding of SF-1 and COUP-TF to the same \ncis-acting element. Molecular Endocrinology  13 239-253.\nDownloaded from Bioscientifica.com at 06/12/2026 10:46:11AM\nvia free access","source_license":"CC0","license_restricted":false}