Endometriosis is the most prevalent gynecological disease in women of reproductive age (Taylor et al., 2021), and it is an important worldwide public health problem (Olive and Schwartz, 1993; Signorile and Baldi, 2010). Endometriosis is characterized by endometrial tissue localized outside the uterus (ectopic lesions), usually at several sites, that causes a chronic inflammatory process predisposing to adhesion formation (Jackson and Telner, 2006; Vermeulen et al., 2021). The current estimation of endometriosis prevalence is around 10–15% of women of childbearing age (Eskenazi and Warner, 1997; Giudice and Kao, 2004) of whom 20–50% present infertility-related symptoms (Smolarz et al., 2021).

Intracrinology of endometriosis debates on the balance between estrogen biosynthesis and metabolism that determines local estrogen availability in the tissues (Piccinato et al., 2018a). Evidence of this mechanism was revealed by data showing that the levels of estradiol and estrone remained constant in the lesions, despite the menstrual cyclical changes in circulating levels and the eutopic endometrium (Huhtinen et al., 2012). Such features indicate that, within the lesions, there are intrinsic regulatory pathways, which maintain high tissue availability of estrogen necessary for disease progression. As reported in previous studies, there are controversial data on gene regulation and protein levels of estradiol-synthesizing enzymes. Some studies have shown that, in endometriotic lesions, there are enzymes that synthesize estrogenfrom androgenic precursors as well as for de novo synthesis from cholesterol (Cornillie et al., 1990; Nap et al., 2004; Attar and Bulun, 2006; Kamergorodsky et al., 2009; Gibson et al., 2018; Mori et al., 2018; Mori et al., 2019). Conversely, estrogen-inactivating enzymes are clearly dysregulated in endometrial and/or endometriotic tissues evidencing a putative compensatory mechanism in view of the increased local estrogen concentrations (Piccinato et al., 2016b, 2017; 2018b).

Overall, two estrogen synthesizing pathways stand out: the aromatase (ARO)-dependent pathway and the steroid sulfatase (STS)-dependent pathway (Piccinato et al., 2018a). Briefly, ARO catalyzes the aromatization of androstenedione and testosterone to E1 or E2, respectively. STS mediates the local desulphonation of inactive circulating steroids, called estrogen sulfates, to their unconjugated, biologically active forms (Miller and Fraser, 2015; Foster, 2021). Interestingly, both pathways share a common final step through the action of hydroxysteroid 17-beta dehydrogenase (HSD17B1). This reducing enzyme belongs to the HSD17B family with high specificity for steroid substrates and high catalytic activity for the conversion of E1 to E2; therefore, HSD17B1 is an important regulator of intracellular steroid hormone concentration at the final stages of biosynthesis (Konings et al., 2018; Foster et al., 2019).

Although no clear association between lesion sites and enzymes function has been described in endometriosis, different types of lesions, as well as different localizations, seem to directly determine the dynamics of the disease (Taylor et al., 2021). Considering that endometriosis is an estrogen-dependent disease, the expression regulation of estrogen-synthesizing enzymes may have an even more pre-eminent impact on endometriosis dynamics. In this context, although sometimes with contradicting results, some studies have focused on site-specific regulation of aromatase. A couple of studies have described ovarian lesions/endometriomas (OMA) express high levels of aromatase gene (CYP19A1) and protein (ARO) (Heilier et al., 2006; Szaflik et al., 2020), whereas deep infiltrating endometriotic lesions (DIE) either fail to express ARO or show reduced expression (Bulun et al., 2004). Another example of site-specific regulation occurs with STS, which is highly expressed in DIE and OMA (Šmuc et al., 2007; Šmuc et al., 2009; Piccinato et al., 2016b) but is minimally present or even absent (Colette et al., 2009) in the eutopic endometrium of endometriosis-affected women (Dassen et al., 2007). The unbalanced redox metabolism of enzymes belonging to the HSD17B1 family is the putative cause of the higher E2 synthesis that has been detected in ectopic and eutopic endometrium. Expression of reducing type- HSD17B1 mRNA was significantly higher in the ectopic tissue (superficial peritoneal endometriosis: SUP, DIE, and OMA), whereas no difference was seen in the expression of HSD17B1-oxidizing types 2 and 4 (Šmuc et al., 2007; Bulun et al., 2010; Delvoux et al., 2014). Altogether, these observations strengthen the idea that the production of estrogen, which takes place through complex enzymatic pathways, is site-specific.

There is no consensus among published studies regarding the expression of estradiol-synthesizing enzymes in endometriotic lesions. Furthermore, there are no published studies simultaneously investigating the expression of these enzymes, ARO, STS, and HSD17B1, in endometriotic tissue samples obtained from a large cohort of patients. The present study explores previous significant reports that demonstrate high and constant levels of estrogens (estradiol and estrone) in different endometriotic tissues (Huhtinen et al., 2012). We, thus, hypothesize that the observed site-specific variations in estrogen concentrations are a result of differential expression regulation of estrogen-synthesizing enzymes, namely, ARO and STS, and also HSD17B1.

Although intracrine estrogen concentration depends on the complex interaction of synthesizing and metabolizing enzymes, our aim was to evaluate gene and protein expression of the synthesizing part (ARO, STS, and HSD17B1) in different types of endometriotic lesions (SUP, OMA, and DIE) as well as in samples of eutopic endometrium (EE) obtained from the same patients with endometriosis. Patients without endometriosis were donors of nonendometriosis eutopic endometrial tissue, here called the control group. In addition, we investigated by laser-capture microdissection and immunohistochemistry the cellular expression of the chosen enzymes.

Our hypothesis in this study was that previously reported estrogen levels detected in endometriosis sites (Huhtinen et al., 2012) are different in function of the distinct expression of estrogen-synthesizing enzymes in each lesion site. Our data show that the relative expression of ARO, STS, and HSD17B1 vary in endometriotic tissue obtained from deep infiltrating lesions (DIE), superficial (SUP), ovarian (OMA), and eutopic endometrium from endometriosis patients (EE) lend support to the hypothesis. The expression of those enzymes in endometriosis was also compared to endometrium samples of nonendometriosis patients (control). It should be mentioned that we analyzed material from a large number of patients taking into account the patients’ menstrual cycle phases.

The gene encoding ARO, CYP19A1, was detected in all endometriosis tissue at levels all greater than found in control; STS and HSD17B1 presented consistently higher mRNA expression levels in DIE. In addition, analysis of gene expression levels applied to the dataset indicated that STS and HSD17B11 genes were the most closely regulated. Overall, for protein expression, EE presented higher levels of ARO and endometriotic lesions had superior levels of STS and HSD17B1. Particularly, unique patterns of mRNA expression levels of the STS and HSD17B1 were described in DIE when compared to OMA and SUP lesion sites; EE gene and protein expression regulations of ARO, STS, and HSD17B1 were distinct from those of the endometriotic sites. Taken together, the findings of different levels of the estrogen-synthesizing enzymes in distinct lesion sites support the hypothesis that the local tissue microenvironment can influence and be influenced by the regulation of different estrogenic pathways. Ultimately, this determines the local availability of estrogen, previously reported in other studies (Huhtinen et al., 2012), which in turn directly affects the local growth of ectopic endometrium.

Gene expression (mRNA) and tissue protein expression (as measured in IHC stained sections) of ARO, STS, and HSD17B1, as well as their relative ratios, did not always match. STS mRNA was more expressed in relation to the others enzymes, in all sample groups analyzed, followed by HSD17B1 and finally by CYP19A1. ARO protein expression was higher among groups, followed by STS and HSD17B1. We detected differences between CYP19A1:STS ratios, CYP19A1:HSD17B1 (gene expression), ARO:STS, and ARO:HSD17B1 (protein expression) in different sites of endometriosis. This would reinforce the idea that there is a regulation of enzymes in a site-specific manner, as the proportions are variable among sites. On the other hand, the lack of difference between gene and protein expression for the ratio HSD17B1:STS, in different endometriosis sites indicates that these enzymes are more closely regulated, as they show less expression variation. Supporting this observation, the Spearman correlation analysis of gene expression between STS and HSD17B1 revealed a positive, moderate correlation between these same enzymes, not seen between them and CYP19A1. Correlation among protein levels was, in general, weak and, although significant, might not be clinically relevant.

It seems that endometriotic lesion sites’ estrogen synthesis is more closely regulated by the enzymes STS and HSD17B1, whereas in the EE, ARO is the dominant regulator. Although STS expressed higher mRNA but not protein in comparison with the other enzymes, the stromal areas of OMA/SUP and DIE lesions showed STS elevated levels of protein. Conversely, ARO was significantly more expressed at the protein level, and it was prominent in stromal areas of EE.

The existence of tissues similar to the endometrium in sites other than the uterine cavity implies that there are local and specific mechanisms that support the survival and growth of the ectopic tissue in an otherwise alien environment. Comparative to the other sites, DIE presented the most distinct and elevated STS and HSD17B1 mRNA expression levels. This suggests that a unique hormonal environment developed in this lesion type favoring high estrogen-synthesis enzymes. Indeed, estradiol measurements by HPLC-MSMS showed its high and persistent concentrations in endometriotic lesions, including deep lesions (Huhtinen et al., 2012). Because vascularization is limited in the DIE (Liu et al., 2008) and the antagonizing enzyme, sulfotransferase is upregulated (Piccinato et al., 2016a), and it seems reasonable to propose that regulatory mechanisms might exist to maintain estradiol levels and disease progression despite poor blood irrigation. In OMA and largely in SUP lesions, blood supply is more pronounced favoring the mRNA expression of CYP19A1, allowing the use of the precursor’s testosterone/androstenedione from circulation to synthesize estradiol.

Either higher levels, or no difference, or absence of CYP19A1 mRNA was previously described in endometriotic lesions in comparison with the endometrium of disease-free, women-control, or of endometriosis patients, EE (Noble et al., 1997; Matsuzaki et al., 2006; Dassen et al., 2007; Colette et al., 2009; Mori et al., 2018). Despite having verified that there is CYP19A1 expression in control endometrium (although at lower levels compared to EE), this study corroborates previous findings showing increased gene and protein expression of ARO in EE (Hudelist et al., 2007; Aghajanova et al., 2009).

The enzymes STS and HSD17B1 have been considered important in the pathogenesis of endometriosis, as indicated by the use of their inhibitors in the treatment of endometriosis (Purohit et al., 2008; Salah et al., 2017; Maltais et al., 2018; Barra et al., 2019). Our data evidenced a parallel regulation between them suggesting a local combined relationship. Thus, STS and HSD17B1 together can play an important role in estrogen production by desulfating sulfated steroidogenic compounds (DHEAS, E2S, and E1S) into active compounds that reach endometriotic tissues in the bloodstream (Purohit et al., 2008). Expression of STS (gene and protein), being greater than of HSD17B1, may provide more E1 or E2, but the final balance between these two estrogens depends on HSD17B1 regulating the formation of the more active estradiol, namely, E2.

In consonance with other authors (Konings et al., 2018), we found in a larger cohort of patients that the expression of STS and HSD17B1 genes was much higher in DIE compared to all other endometriotic lesion types and EE or endometrium from nonaffected women. Increased STS mRNA expression has also been demonstrated in OMA, whereas HSD17B1 expression was higher in all types of endometriotic lesions when compared to control (Šmuc et al., 2007) or EE (Mori et al., 2015). In fact, only low levels of the HSD17B1 gene and protein (Dassen et al., 2007) or total absence of this enzyme (Utsunomiya et al., 2001) were reported in the endometrium of nonaffected women. Colette et al. (2013) described similar levels of 1 HSD17B1 gene and protein expressions among different types of lesions.

The elevated expression of STS and HSD17B1 genes in DIE lesions in comparison to other lesion types was not paralleled by the corresponding protein expression of these enzymes. Although staining by IHC was present for both enzymes, there were no significant quantitative differences between the analyzed groups (EE, SUP, OMA, and DIE). Other studies have also failed to find significant differences in STS protein expression among different endometriotic lesions and EE (Dassen et al., 2007).

The lack of association between the results of gene and protein expressions of the analyzed estradiol-producing enzymes can be explained by posttranscriptional regulatory events, which can prevent the formation of functional proteins from the mRNA transcript. Estrogens are consistently present in endometriotic tissue, regardless of the menstrual cycle phases (Huhtinen et al., 2012). Interestingly, the stability of several mRNAs is regulated by estrogens (Ing, 2005). Several microRNAs (miRNA) control gene expression at the posttranscriptional level and can be induced or inhibited by estrogen (Klinge, 2012; Kolanska et al., 2021). In addition, E2 canonical signaling pathways mainly mediated by nuclear estrogen receptors can perform as transcription factors to stabilize or destabilize mRNAs via miRNAs modifying gene expression (Kim et al., 2021). These controls promote the expression of genes that are critical to either strengthen or diminish the effects of steroid hormones that depend on the feedback of the same enzymes they regulate. Furthermore, estrogen-regulated expression of miRNA is both cell- and tissue-specific (Cohen et al., 2008).

Regarding the participation of the glandular versus stromal areas of endometriotic lesions in estrogen synthesis, we looked at the expression of ARO, STS, and HSD17B1 in the various types of endometriotic lesions and of eutopic endometrium, EE. Molecular and histophysiological differences between stromal and glandular cells from EE or endometriotic lesions have already been described such as alterations in prosurvival enzymes, cytoskeletal proteins, proteasomes, and cell repair mechanisms (Bastos Wolff et al., 2012). It was suggested that endometrial stromal cells responsive to progesterone could induce HSD17B2 (i.e., inactivation of E2) in epithelial cells (Yang et al., 2001) via paracrine factors. However, primary cultures of OMA stromal cells stimulated with progesterone resulted in the suppression of HSD17B2 expression in epithelial Ishikawa cells (Cheng et al., 2007).

We found higher protein expression of HSD17B1 in the stromal cell regions from endometriotic lesions. This result lends support to the idea that the regulation of estrogen secretion in endometriotic lesions is cell-specific. A pilot study detected, by immunofluorescence, the presence of stromal cells (but not epithelial cells) in the bloodstream of women with endometriosis, suggesting that endometrial stromal cells may migrate from the EE to distant sites via the bloodstream (Lin et al., 1995). It may also suggest that stromal cells could participate in the maintenance of endometriotic lesions. For the STS pathway, both stromal and epithelial cells seem to act at maintaining high levels of estrogen-synthesizing enzymes, as observed in our results. It appears that stromal cells are more susceptible to deficient steroid hormonal secretion.

Because steroid hormones regulate the expression of some estrogen metabolizing and synthesizing enzymes, which could be present at different levels throughout the menstrual cycle, the experimental design of our study included samples from progesterone-secretory and nonsecretory phases. Although the menstrual cycle phase was taken into account in the statistical analysis, it was not always possible to separate the main findings of the study considering this variable, especially for the gene analysis. Interestingly, stromal STS secretion was higher during the secretory phase, when compared to glandular STS expression in the nonsecretory phase. Furthermore, DIE lesions expressed more STS protein in the secretory phase than in the nonsecretory phase. These results agree with previously published results (Simpson et al., 1994) and with those from our group (Piccinato et al., 2016b), in which increased STS expression levels were seen in the secretory phase of the cycle. In contrast, other authors have reported that STS expression levels are not altered in the different phases of the menstrual cycle (Šmuc et al., 2007; Colette et al., 2013).

The higher protein expression of ARO was found in stromal EE in relation to DIE, in secretory phase samples, which agrees with the results of ARO tissue protein expression, suggesting that the stromal cells in the endometrium may be the main site of estrogen production. However, some authors have found ARO mRNA and protein expressed preferentially in glandular cells (Aghajanova et al., 2009; Wang et al., 2010). It is known that the endometrial tissue undergoes cyclical changes according to systemic hormonal variation (Foti et al., 2018). However, our findings suggest that the EE may be autonomous with regard to regulation of the ARO pathway, maintaining its expression, and the increased local estrogen production, regardless of systemic hormonal variations. Considering the basic characterization of endometriosis, such as the presence of endometrial tissue in other sites, it is important to understand not only the ectopic behavior of the endometrial-like tissue but also the behavior of that tissue in its place of origin (uterus) during the onset of the disease.

We found high variability among samples in the gene expression of all enzymes in stromal and glandular cell areas. Of note, most samples did not express detectable levels of CYP19A1 mRNA. The absence of CYP19A1 in some microdissected biopsies may signify no expression or expression below the detection levels of RT-PCR. In fact, LCM yields small amounts of microdissected material (Matsuzaki et al., 2004) and, consequently, amplification of genes with a low expression such as CYP19A1 might not reach detectable levels. It is important to consider that the number of microdissected samples for the experiments on gene expression was rather small. Despite having generated interesting primary results, we have to consider them preliminary and subject to confirmation.

Although the level of the single enzymes in the intracrine machinery varies with apparently no clear association with the localization of endometriotic lesions, a combination of enzymes and site intracrinology may explain site-specific characteristics of the disease. The results of the present study support this idea, as endometriotic sites present, at both gene and protein levels, higher STS and HSD17B1, whereas in EE, there was increased ARO expression. The cellular compartment of in situ protein expression seems to emphasize the importance of cell regulation in a paracrine manner. Finally, even though STS was the most expressed gene, ARO was the most expressed protein, revealing a complex site-specific system of transcript activation. Although enzyme activity studies are still required to fully understand how estrogen biosynthesis is regulated in the distinct sites where endometriosis occurs, it was possible to substantiate the pronounced gene and protein regulation of STS and HSD17B1, in endometriotic sites (particularly, in DIE) and of ARO in EE. It is apparent that endometriotic lesions at different sites correspond to different facets of the same disease and develop site-dependent intrinsic regulatory mechanisms.