Epithelial–mesenchymal transition (EMT) is an important process of cell remodeling characterized by the gradual loss of the epithelial phenotype and progressive gain of a mesenchymal phenotype. EMT is not an all-or-nothing process, but instead a transition of epithelial to mesenchymal cells with intermediate cell states. Recently, EMT was described in endometriosis, and many EMT-specific pathways like Twist, Snail, Slug, Zinc finger E-box-binding homeobox 1/2 (ZEB1/2), E/N-cadherin, keratins, and claudins are involved. However, as pointed out in this review, a comparison of the eutopic endometrium of women with and without endometriosis yielded only subtle changes of these EMT markers. Furthermore, only very few alterations in cell–cell contacts could be found but without changes in the epithelial phenotype. This suggests only a partial EMT which is not a prerequisite for the detachment of endometrial cells and, thus, not critical for the first step(s) in the pathogenesis of endometriosis. In contrast, the majority of changes in the EMT-related marker expression were found in the ectopic endometrium, especially in the three endometriotic entities, ovarian, peritoneal, and deep infiltrating endometriosis (DIE), compared with the eutopic endometrium. In this review, we examine the most important EMT pathways described in endometriosis and propose that partial EMT might result from the interaction of endometrial implants with their surrounding microenvironment.

A diagnosis of endometriosis is based upon the histological identification of endometrial tissue at ectopic sites which are commonly located on the pelvic organs, the peritoneum and ovary. In rare cases, ectopic lesions can be found in other organs, such as kidney, bladder, lung or brain. Diagnosis is achieved by laparoscopic intervention followed by histological confirmation of endometriotic tissue. Prevalence is estimated at approximately 10% in the general female population with many patients experiencing pain and/or infertility. Currently, the implantation hypothesis by Sampson is the most accepted hypothesis about the pathogenesis of endometriosis. However, the occurrence of endometriosis in patients with Mayer–Rokitansky–Küster–Hauser (MRKH) syndrome who sometimes lack a uterus or endometrium seems to suggest metaplasia as a cause of endometriosis. A critical reevaluation of the literature about MRKH does not reveal conclusive evidence of an association of uterus/endometrium agenesis and endometriosis. Most often only MRI diagnoses of uterus/endometrium agenesis and only very rarely conclusive histological evidence of the endometriotic lesions are presented. In contrast, whenever biopsies were performed endometriosis always appeared together with uterus/endometrium remnants. Taken together, we suggest that MRKH patients only develop endometriosis if a uterus/endometrium is present which underscores and not contradicts the implantation hypothesis of Sampson.

Claudins are the major components of tight junctions and are often deregulated in human cancer, permitting escape of cancer cells along with the acquisition of invasive properties. Similarly, endometrial cells also show invasive capabilities; however, the role of tight junctions in endometriosis has only rarely been examined. In this study, we analyzed the protein expression and localization of claudin-7 and claudin-11 in human eutopic and ectopic endometrium and endometrial cell lines. We identified claudin-7 primarily at the basolateral junctions of the glandular epithelial cells in eutopic endometrium as well as in the ectopic lesions in nearly all glands and cysts. Quantification of claudin-7 localization by HSCORE showed a slight increase in peritoneal and deep infiltrating endometriosis (DIE) compared to eutopic endometrium. In contrast, claudin-11 was localized mainly in the apicolateral junctions in nearly all glandular epithelial cells of the eutopic endometrium. Interestingly, we observed a deregulation of claudin-11 localization to a basal or basolateral localization in ovarian ( P < .001), peritoneal ( P < .01), and DIE ( P < .05) and a moderately decreased abundance in ovarian endometriosis. In endometrial cell lines, claudin-7 was only present in epithelial Ishikawa cells, and silencing by small-interfering RNA increased cell invasiveness. In contrast, claudin-11 could be demonstrated in Ishikawa and endometriotic 12Z and 49Z cells. Silencing of claudin-11 decreased invasiveness of 12Z slightly but significantly in 49Z. We suggest that although claudin-7 and claudin-11 can be found in nearly all eutopic and ectopic epithelial cells, the impaired localization of claudin-11 in ectopic endometrium might contribute to the pathogenesis of endometriosis.

Epithelial-mesenchymal transition (EMT) is characterized by the loss of epithelial and acquisition of mesenchymal cell characteristics. Our aim was to assess the epithelial phenotype in the pathogenesis of endometriosis with epithelial and mesenchymal markers. We used 2 structural (keratin-18, -19 [K18, K19]), 1 membrane-associated (mucin-1 [MUC1]), and 2 mesenchymal proteins (vimentin; zinc finger E-box-binding homeobox 1, [ZEB1]) to compare epithelial and mesenchymal characteristics in eutopic endometrium with the 3 endometriotic entities, peritoneal, ovarian, and deep infiltrating endometriosis (DIE). Quantitation showed no differences for K18, K19, and MUC1 between endometrium with and without endometriosis. Also, K18 was not different between endometrium and endometriotic lesions. In contrast, K19 and MUC1 were modestly but significantly decreased in the endometriotic lesions compared to endometrium. However, the maintained expression of epithelial markers in all investigated tissues, regardless of the pathological condition, clearly indicates no loss of the epithelial phenotype. This is further supported by the reduced presence of epithelial vimentin in endometriotic lesions which is in contrast to an increase in stromal vimentin in ectopic endometrium, especially in ovarian endometriosis. The ZEB1 increase in endometriotic lesions, especially in DIE, on the other hand suggests a role of partial EMT in the development of endometriotic lesions, possibly connected with the gain of invasive capabilities or stemness. Taken together, although we found some hints for at least a partial EMT, we did not observe a severe loss of the epithelial cell phenotype. Thus, we propose that EMT is not a main factor in the pathogenesis of endometriosis.