17β-Hydroxysteroid Dehydrogenase-2 Deficiency and Progesterone Resistance in Endometriosis

Existing evidence points to a differentiation defect in endometriotic stromal cells due to resistance to selective actions of progesterone. Progesterone resistance in endometriosis seems to be subtle and manifested by specific molecular abnormalities. It is likely that progesterone resistance is primarily due to significantly lower PR-B and total PR in endometriotic stromal cells compared with endometrial stromal cells. PR deficiency is likely responsible for increased levels of E2 in endometriosis because progesterone fails to induce the E2-metabolizing enzyme HSD17B2 in endometriotic tissue ( Fig. 3 ). We localized deficient progesterone action to a stromal cell defect in endometriosis. Our conclusion resonates with gene expression microarray studies performed on eutopic endometrium of women with endometriosis compared with that from disease-free women. 31 , 32 These studies on eutopic endometrium identified distinct molecular defects that are consistent with the progesterone resistance hypothesis.

Progesterone induces the differentiation of both endometrial stromal and epithelial cells as exemplified by decidual and secretory changes and reduction in epithelial cell proliferation. The histological markers of differentiation, stromal decidualization and epithelial secretion, are associated with the presence of nuclear PRs and increased levels of circulating progesterone during the luteal phase. Molecular markers of progesterone action with respect to differentiation include increased production of lactoferrin and glycodelin in epithelial cells and prolactin and insulinlike growth factor binding protein-1 (IGFBP-1) in stromal cells of the endometrium. Progesterone inhibits and even reverses estrogen-induced endometrial growth, hyperplasia, or adenocarcinoma of the endometrium. This effect was found to be mediated by PR in stromal cells in tissue recombination studies of PR knockout and wild-type mice. 11 These findings reinforced the concept that progesterone-dependent stromal cell-derived factors act on epithelial cells to offset the effects of estrogen. The antiestrogenic effect of progesterone in endometrial tissue is in part mediated by its stimulation of 17β-hydroxysteroid dehydrogenase type 2 activity, which catalyzes the conversion of biologically active E2 to the inactive steroid estrone (E1). 12 – 14 17β-hydroxysteroid dehydrogenase type 2 catalyzes the conversions of E2→E1 and testosterone→androstenedione in the endometrium as well as in the placenta and liver. 15 This enzyme is encoded by the HSD17B2 gene in endometrial epithelial cells 16 , 17 ( Fig. 2 ). Very high levels of HSD17B2 mRNA have been demonstrated in the epithelial cell component of whole endometrial tissue upon exposure to progesterone both in vivo and in vitro. 16 , 17 Direct treatment of isolated endometrial epithelial cells with progesterone only slightly induced HSD17B2 expression. We found, however, that endometrial epithelial HSD17B2 expression was regulated robustly by soluble factors in conditioned medium from endometrial stromal cells incubated with progesterone. 18 Stromal PRs mediated progesterone-dependent production of a cocktail that stimulated epithelial HSD17B2 promoter activity, mRNA, and enzyme activity 18 ( Fig. 2 ). Serial deletion analysis of the HSD17B2 promoter showed that the 200 bp region immediately upstream of the transcription start site was essential for the maximum stromal PR-dependent induction of this epithelial promoter. 18 Site-directed mutants of the HSD17B2 promoter demonstrated that two overlapping specificity protein 1 (Sp1) motifs within this critical regulatory region are essential for induction of promoter activity by stromal cell-conditioned medium or overexpression of Sp1/Sp3. 19 Conditioned medium markedly enhanced binding of Sp1/Sp3 to this region of the HSD17B2 promoter. In vivo, we demonstrated a significant spatiotemporal association between epithelial Sp1/Sp3 and HSD17B2 protein levels in human endometrial biopsies. 19 Taken together, these data suggest that HSD17B2 expression in endometrial epithelial cells, and therefore E2 inactivation, is regulated by Sp1 and Sp3, which are downstream targets of progesterone-dependent paracrine signals originating from endometrial stromal cells ( Fig. 2 ). 19 We also found that 9-cis or all-trans retinoic acid (RA) stimulated HSD17B2 expression in endometrial epithelial cells. 20 The liganded RA receptors RARα/RXRα tethered Sp1/Sp3 at the HSD17B2 promoter to induce its activity 20 ( Fig. 2 ). The Sp1-binding cis-regulatory site at HSD17B2 promoter, Sp1, Sp3, RARα, and RXRα were all necessary for RA-dependent stimulation of HSD17B2 expression in endometrial epithelial cells. 20 Taken together, these findings suggest that progesterone acts on endometrial stromal cell PR to stimulate formation of paracrine factors including RA. This in turn activates RA- and Sp1/Sp3-dependent signaling to up-regulate HSD17B2 expression in endometrial epithelial cells leading to conversion of E2 to E1. 20 We characterized primary endometriotic stromal cells by demonstrating prolactin mRNA expression in response to treatment with medroxyprogesterone acetate (MPA) and dibutyryl cyclic adenosine monophosphate (Bt 2 cAMP). 21 Prolactin is a molecular marker of decidualization of endometrial stromal cells. Prolactin mRNA was detected in cultured stromal cells from eutopic endometrium and endometriosis but not in ovarian granulosa and theca cells. 21 Progesterone-induced prolactin mRNA levels in endometriotic cells were much lower compared with eutopic endometrial cells. 21 These findings further support the concept of impaired stromal differentiation owing to a relative resistance to progesterone action in endometriosis. Previous immunohistochemical studies revealed that PR levels are lower in endometriotic tissue compared with eutopic endometrium. 22 – 24 The molecular evidence of progesterone resistance in endometrial tissues and cells, together with the clinical observation of resistance of endometriosis to treatment with progestins, prompted us to determine the protein and mRNA levels of PR-B and total PR in matched eutopic endometrial and endometriotic tissues exposed to the same circulating hormones. In eutopic endometrium, PR-B protein and mRNA levels progressively increased during the proliferative phase, peaked immediately before ovulation, and diminished after ovulation, suggesting that E2 stimulates PR-B levels. 9 In contrast, PR-B mRNA and protein levels were not detectable in any of the matched extraovarian eutopic endometrial samples ( Fig. 1 ). In a separate matched tissue set of eutopic endometrial samples and ovarian endometrioma cyst walls, Affymetrix gene microarray (confirmed by real-time polymerase chain reaction [PCR]) showed severely decreased PR-B (11-fold) and total PR (8-fold) mRNA levels. These data suggest that the absence of PR-B protein in endometriotic tissue is transcription related (our unpublished observations). These in vivo observations in PR expression were accompanied by comparable differences in primary endometrial and endometriotic stromal cells in that both PR-B and PR-A levels are significantly lower in endometriotic stromal cells (our unpublished observations). Thus progesterone resistance in endometriotic tissue may be due to an overall reduction in PR and the absence of PR-B. 9 It was previously reported that progesterone did not induce the conversion of E2→E1 in endometriotic tissue, in contrast to eutopic endometrium. 26 Interconversions of E2 ↔ E1 are catalyzed by several enzymes encoded by separate genes. 15 , 27 , 28 For example, HSD17B1 favors the formation of E2, whereas HSD17B2 encodes the most efficient enzyme for inactivation of E2 by conversion to E1. 12 , 13 , 16 HSD17B2 enzyme activity and mRNA were found to be present in extremely high levels in eutopic endometrial tissues and epithelial cells (from disease-free women) during the secretory phase. 12 , 13 , 16 We demonstrated that HSD17B2 mRNA and protein are readily detectable in secretory eutopic endometrium but not in paired extraovarian endometriotic samples exposed to the same circulating levels of progesterone. 25 Immunoreactive HSD17B2 was localized to secretory epithelial cells. The lack of HSD17B2 in endometriotic epithelial cells of matched tissues despite histologically detectable secretory changes in response to progesterone suggested that endometriotic tissue exhibits selective resistance to certain actions of progesterone. We also observed that inactivation of E2 is impaired in endometriotic tissues due to deficient expression of HSD17B2, which is normally expressed in eutopic endometrium in response to progesterone. 25 Furthermore, aberrant aromatase expression in endometriotic stromal cells favors the conversion of androstenedione to E1; aromatase expression is normally absent in the eutopic endometrium. 29 We also demonstrated that transcripts of the reductive HSD17B1, which catalyzes the conversion of E 1 →E 2 , were present in both eutopic endometrium and endometriotic tissue. 25 Thus the weakly estrogenic product of the aromatase reaction, E1, is readily converted to the potent estrogen, E2, in endometriotic tissues. Local concentrations of E2 in endometriotic tissues are further amplified by the lack of HSD17B2 expression and subsequent metabolism of E 2 →E 1 . Taken together, multiple molecular mechanisms contribute to the increased estrogenic activity in endometriosis as compared with eutopic endometrial tissue. Physiologically, a stromal PR-mediated paracrine mechanism is responsible for induction of epithelial HSD17B2 expression. Paracrine factors secreted from stromal cells induce epithelial cells to produce the transcription factor Sp1 that binds to the promoter of the 17βHSD2 gene to regulate its expression. 19 In contrast to eutopic endometrium, epithelial cells of endometriosis do not express HSD17B2 in vivo in response to progesterone, resulting in accumulation of E2 in ectopic tissue. 25 We attempted to determine the underlying cellular mechanism for this lack of HSD17B2 expression in endometriotic epithelial cells. The addition of conditioned medium from progestin-pretreated endometrial stromal cells significantly stimulated Sp1 protein and HSD17B2 mRNA and promoter activity in endometrial epithelial cells. 30 In contrast, conditioned medium from endometriotic stromal cells failed to induce Sp1 or HSD17B2 in epithelial cells ( Fig. 3 ). Moreover, conditioned medium from progestin-pretreated endometrial stromal but not from endometriotic stromal cells enhanced binding of Sp1 to the HSD17B2 promoter region. 30 Taken together, a stromal cell defect in endometriosis appears to be responsible for the lack of progesterone-mediated secretion of factors that induce Sp1 production and its binding activity to the HSD17B2 promoter for expression of this gene in epithelial cells. 30 This may be a key molecular consequence clinically observed progesterone resistance in endometriosis ( Fig. 3 ).