Uterine Paramesonephric Cysts in Sprague-Dawley Rats from National Toxicology Program Studies.

Various uterine cysts have been reported in humans, as well as in most domesticated and laboratory animal species. Cysts are commonly associated with cystic degeneration of leiomyomas, adenomyosis, serosal inclusion cysts, and cystic endometrial hyperplasia ( Matalliotakis et al. , 2003 , Mutter et al. , 2007 , Protopapas et al. , 2008 ). In the rat, uterine cysts are predominately of endometrial epithelial origin and are a common age-related change. Development of spontaneous cystic endometrial hyperplasia has occurred in rats exhibiting persistent estrus, hormonal imbalances, and excess estrogen secretion ( Greaves, 2012 , Dixon et al. , 2014 ). Although uncommon, congenital uterine cysts may arise from embryonic (Wolffian or Mullerian) ducts ( Sherrick and Vega, 1962 ). The ovaries, uterus, and upper vagina develop from the paramesonephric (Mullerian) duct ( Goldman and Cooper, 2013 ). The male reproductive tract originates from the mesonephric (Wolffian) duct, which lies parallel to the paramesonephric duct during embryonal development. Traditionally, it has been proposed that it regresses in females in the absence of the production of anti-Mullerian hormone (AMH) by Sertoli cells, and testosterone by interstitial cells of the embryonic testes ( Kobayashi and Behringer, 2003 , Jacob et al. , 2012 , Dixon et al. , 2018 ). However, recently it has been suggested that elimination of the Wolffian duct in female mouse embryos is androgen-independent; and regression is actively promoted by the transcription factor, COUP-TFII (chicken ovalbumin upstream promoter transcription factor II) ( Zhao et al. , 2017 ). In women, congenital cysts of the uterus characteristically have variably thick fibromuscular walls lined by ciliated to non-ciliated cuboidal to columnar epithelium; lack association with the uterine cavity, endometrial glands, and stroma; and are located within the uterine myometrium ( Sherrick and Vega, 1962 ). Distinguishing between benign and neoplastic tissue derived from paramesonephric and mesonephric ducts is difficult, and recent research has focused on identifying immunohistochemical markers specific to each duct. Paired box gene 8 (PAX8) is a transcription factor involved in the organogenesis of paramesonephric derived organs ( Mittag et al. , 2007 ). GATA binding protein 3 (GATA3) transcription factor is involved in the organ differentiation of the urogenital tract originating from the mesonephric duct ( Grote et al. , 2006 ). PAX8 and GATA3 have been identified as immunomarkers to differentiate paramesonephric and mesonephric neoplasms, respectively, in humans ( Tong et al. , 2011 , Howitt et al. , 2015 ). Based on the histological appearance of the uterine wall cysts identified in Sprague Dawley (Hsd: Sprague Dawley SD) rats in our study, we hypothesized that these cysts are remnants of either the paramesonephric or mesonephric ducts. We used routine light microscopy and histochemical and immunohistochemical staining to identify the origin of these congenital cysts.

Uterine subserosal cysts were examined in 20 adult female SD rats (n = 12 rats at 91–93 days of age; n = 7 rats at 266 days of age) from five multi-generation studies conducted by the NTP. Similar appearing subserosal cysts occurred in treated and control rats, and were independent of mating status ( Table 2 ). Grossly identified single cysts were located along the uterine horn or body and were approximately 2–3 mm in diameter, spherical and appeared fluid-filled ( Figure 1A ; Table 3 ). Fifteen of the 20 cysts were observed grossly, while 5 cysts were only identified microscopically. All cysts were unilocular and adjacent to the serosa and the outer myometrial layer of the uterus ( Figure 1B ). The uterus, cervix, and vagina from ten of the rats with adequate cystic tissue available were further examined for this study. Uterine subserosal cysts had well-defined walls of variable thickness with an average measurement of 91 µm. The walls were not associated with the endometrial cavity. The majority of cysts were lined by a single layer of flattened to cuboidal epithelial cells with random, focal cilia, moderate eosinophilic cytoplasm and large oval nuclei containing 1–2 prominent nucleoli. One cyst was lined by more endometrium-like epithelium characterized by pseudostratified tall cuboidal cells with focal cilia and moderate to abundant foamy cytoplasm. Cysts often contained lightly basophilic staining flocculent intraluminal material ( Figures 2A–C ). Staining with Masson’s trichrome showed scant to moderate collagen fibers (blue) and abundant smooth muscle (red) present in the cyst walls ( Figure 2D ). Uterine subserosal cysts were evaluated for immunopositive staining and given a quickscore for each marker. All cyst walls expressed moderate cytoplasmic immunoreactivity for SMA in muscle fibers, and epithelium of all cysts had diffuse, intense cytoplasmic staining for CK-18 ( Figures 2E and F ). A cyst located at the junction of the uterine cervix and uterine horn showed typical histological characteristics of the other observed uterine subserosal cysts ( Figures 3A and 3B ). Evaluation of cysts for hormone receptors showed PR was present in the epithelium and expressed as multifocal weak to moderate nuclear and weak cytoplasmic staining ( Figure 3C , Table 4 ); this staining pattern was observed in of 6 out of 10 cysts. Also, nuclear staining for PR was observed in the smooth muscle of the cyst wall and adjacent endometrial stroma. Additionally, 9 of 10 cysts had weak to intense nuclear staining for ER-α ( Figure 3D , Table 4 ). Each cyst was immunostained for the transcription factors PAX8 and GATA3 to determine paramesonephric or mesonephric origin. We found diffuse, moderate to intense nuclear staining of epithelia for PAX8 in all cysts. The majority of cyst epithelia were negative for GATA3, with the exception of 3 cysts with multifocal, weak nuclear staining in less than 20% of epithelial cells ( Figures 3E–F , Table 4 ). The endometrial epithelial cells of known paramesonephric origin also stained positive for PAX8 and were negative for the mesonephric marker GATA3. Based on IHC quickscores for each marker, characteristically, cyst walls expressed moderate positive staining for SMA; and epithelia stained strongly positively for CK-18 and PAX8, moderately for ER-α, and weakly for PR. Furthermore, with a mean quickscore of 1, GATA3 staining was considered negative for cyst epithelia ( Figure 4 ). The uterus, cervix, and vagina were examined to determine the stage of estrous cycle within the SD rats. Animals presented in various stages of the estrous cycle. Two rats were in proestrus, and the remaining animals were equally divided into estrus and metestrus phases. There was no correlation between estrous cycle stage and expression of the various immunomarkers in the 10 rats evaluated ( Table 4 ).

Subserosal, unilocular cysts present along the uterine horns and uterine body from SD rats used in NTP multi-generation studies were evaluated to determine their origin. In the uterus, cysts are classified as acquired or congenital. There are several causes of acquired cystic lesions. Degeneration within leiomyomas can result in cytic cavities that lack an epithelial lining and are definitively dissimilar to the cysts characterized in this study. Other causes of cyst-like lesions include adenomyosis. which is characterized by the presence of endometrial glands and associated endometrial stroma located within the myometrium and the absence of a smooth muscle wall ( Neri and Eckerling, 1966 , Parrott et al., 2001 ). Adenomyosis is not consistent with the cysts reported herein due to the lack of endometrial stroma directly associated with the cyst epithelium and the presence of a muscular wall underlying the cyst epithelia. The observed cysts are also unlike serosal inclusion cysts that develop from the invagination of the serosa into the adjacent connective tissue and are associated with aged, pluriparous animals ( McEntee, 1990b , Arnold et al. , 1996 , Godfrey and Silkstone, 1998 ). Histologically, uterine serosal inclusion cysts are thin walled, lined by flattened cuboidal mesothelial cells, and contain clear fluid. These cysts may appear similar to those described here; however, the cysts that we characterized did not display the typical mesothelial lining cells or thin walls of serosal inclusion cysts, and were also present in young nulliparous rats. The subserosal uterine cysts in this study are consistent with the criteria for congenital uterine cysts ( Sherrick and Vega, 1962 ). This is supported by the observations that all cysts had smooth muscle walls and were lined by cuboidal epithelium. Cyst epithelia were not associated with the endometrial cavity, glands, or stroma. All cyst walls were located immediately subjacent to or directly associated with the outer myometrium. Therefore, based on our histological evaluations, the subserosal uterine cysts observed in 10 SD rats could be characterized as congenital uterine cysts derived from mesonephric or paramesonephric ducts. Distinguishing between mesonephric and paramesonephric duct derived tissue is difficult due to overlapping histological characteristics and anatomical locations. In humans, mesonephric cysts are described as having a prominent muscular wall lined by flattened epithelium, whereas paramesonephric cysts have less prominent smooth muscle and columnar epithelium ( Sherrick and Vega, 1962 ). In this study, only one of the 20 paramesonephric cysts observed in the rats had columnar epithelium, while the rest generally had cuboidal epithelium. This difference in epithelium in the rats could be a result of compression from intraluminal fluid accumulation or could represent a true species related variation. In women, neoplastic and non-neoplastic tissues derived from the mesonephric duct have not shown immunoreactivity for estrogen or progesterone receptors ( Devouassoux-Shisheboran et al. , 1999 , Silver et al. , 2001 ). This is not true for tissues derived from paramesonephric origin, which have been shown to express PR and ER-α ( Goyal and Yang, 2014 , Wu et al. , 2014 ). Both PR and ER-α are present within paramesonephric derived tissues such as the uterus and oviduct, but have been shown to have variable expression during the estrous cycle in rats ( Wang et al. , 2000 ). Therefore, the lack of immunostaining for steroid hormone receptors does not rule out tissue of paramesonephric duct origin. Although there was variable staining intensity for PR and ER-α, all but one cyst in this study were positive for one or both receptors suggesting that these cysts are most likely of paramesonephric origin. PAX8 is a transcription factor critical in the development of organs derived from the embryonal duct and its expression has been found in paramesonephric-derived epithelium of the uterus and has been used to identify paramesonephric-derived tumors in humans ( Laury et al. , 2011 , Ozcan et al. , 2011 , Heidarpour and Tavanafar, 2014 , Liang et al. , 2016 ). However, it has been shown that mesonephric neoplasms can also be immunopositive for PAX8 suggesting that its expression is present in both paramesonephric and mesonephric derived tissues, and that the use of immunohistochemical staining panels may be most appropriate in identifying neoplastic or nonneoplastic tissue of mesonephric or paramesonepric duct origin ( Roma et al. , 2015 , Goyal and Yang, 2014 ). In a study by Goyal et al. the authors evaluated the expression of PAX8, p16, and ER in differentiating mesonephric proliferations and cervical adenocarcinomas, and concluded that a panel of immunohistochemical stains consisting of PAX8, ER, and p16 is useful in the distinction between mesonephric proliferations and cervical adenocarcinomas ( Goyal and Yang, 2014 ). All cysts in this current study displayed intense immunoreactivity for PAX8 and as mentioned above also expressed PR and ER-α, with the latter receptors reported not to be expressed in mesonephric-derived tissues, thus strengthening support of a paramesonephric duct origin for our identified cysts. The transcription factor GATA3 has an important role in organ development including the urogenital tract originating from the mesonephric duct ( Grote et al. , 2006 ). Developmentally, GATA3 is not expressed in endocervical and endometrial epithelium, but is reported to be expressed in benign and malignant uterine mesonephric lesions in women ( Roma et al. , 2015 ). Others, however, have found negative to focal weak GATA3 expression in endometrial adenocarcinomas ( Miettinen et al. , 2014 , Howitt and Nucci, 2017 ). Both mesosnephric and paramesonephric ducts are derived from the intermediate mesoderm, and coelomic epithelium lines the developing paramesonephric duct ( Mullen and Behringer, 2014 ). Gene expression and signaling events in the intermediate mesoderm determine differentiation of both ducts. Mesenchymal expression of GATA3 regulates mesonephric duct development and migration, while Wnt signaling is required for paramesonephric duct initiation and elongation ( Grote et al. , 2006 , Yucer et al. , 2017 ). Additionally, GATA3 is considered an early marker of mesonephric duct differentiation ( Yucer et al. , 2017 ); therefore, it is plausible that gene expression or signaling alterations affecting the intermediate mesoderm and/or epithelial differentiation could result in GATA3 expression in addition to PAX8 expression in paramesonephric derived epithelium. In our study, the majority of the cysts had no immunoreaction for GATA3 and only 3 cysts showed weak staining in <20% of epithelial cell nuclei, in addition to moderate to intense nuclear staining for PAX8. A recent review of mesonephric proliferations in the reproductive tract of women reported PAX8 and GATA3 expression in mesonephric remnants and cysts, and concluded that mesonephric remnants/hyperplasia is immunohistochemically characterized by strong, diffuse GATA3 and negative ER/PR, as well as CD10 apical expression ( Howitt and Nucci, 2017 ). Based on the findings of the presence of moderate to intense PAX8 and negative to weak focal GATA3 expression in only 3 cysts that also stained positive for PAX8, in addition to the expression of ERα and PR, the cysts described in this paper were considered to be most consistent with a paramesonephric phenotype. The pathogenesis of these uterine cysts is unknown. In females, paramesonephric cysts can arise throughout the reproductive tract, and are thought to be entrapments of the embryological tissue during organ differentiation ( Eilber and Raz, 2003 ). In contrast, it is thought that mesonephric cysts arise within persistent mesonephric ducts and occur in aged animals ( McEntee, 1990b ). In this study, the subserosal uterine cysts did not appear to be an age-related change because all rats were less than one year of age. In male CD-1 mice prenatally exposed to a synthetic estrogen, diethylstilbestrol (DES), cysts occurred in the epididymis associated with estrogen-induced retention of the paramesonephric duct ( Newbold et al ., 1987 ). Similarly, paramesonephric cysts have been reported in the spermatic cord of otters in Sweden, and authors have postulated that environmental endocrine disruptor chemical exposures associated with estrogens or estrogen-like compounds as a potential etiology ( Roos and Agren, 2013 ). Additionally, DES was shown to cause ovarian and vaginal abnormalities, including presumed mesonephric-derived paraovarian cysts, vaginal adenomyosis, and prominent mesonephric duct remnants in the cervicovaginal region in prenatally exposed female CD-1 mice ( Newbold and McLachlan, 1982 , Haney et al. , 1986 ). Furthermore, prominent mesonephric remnants were noted in the uteri of CD-1 mice prenatally exposed to the environmental estrogenic chemical, bisphenol A (BPA) ( Newbold et al. , 2009 ). In our retrospective study, 3 of 10 cysts were in non-treated rats and the remainder of the rats were exposed to various doses of 5 different chemicals. The effects of estrogenic chemicals were not addressed, and therefore we cannot corroborate the role of estrogen in the pathogenesis of these cysts. Congenital uterine cysts arising from mesonephric and paramesonephric ducts are rarely described in the human and veterinary literature, and are often incidental findings ( Sherrick and Vega, 1962 , Gelberg and McEntee, 1986 , McEntee, 1990a ). It is generally thought that these cysts arise primarily from remnants of the mesonephric duct. A recent study of three cases of mesonephric duct remnants in bitches found cysts lined by ectopic endometrial epithelium, glands and stroma which were immunopositive for cytokeratin, ER and PR ( Bartel et al. , 2011 ). In contrast, the epithelial-lined cysts reviewed in our study lacked endometrial glands and stroma. We have confirmed that subserosal uterine cysts with histomorphologic features more consistent with those previously described for mesonephric duct cysts were more phenotypically consistent with cysts derived from the paramesonephric duct based on an immunohistochemical panel of positive staining for PR and / or ER-α, and PAX8 with negative staining for GATA3. It is very difficult to distinguish between cysts of paramesonephric and mesonephric duct origin without immunohistochemical staining, and therefore it is possible that paramesonephric cysts may be misidentified as mesonephric cysts in the rodent uterus. Our findings are important because we have identified and described developmental subserosal uterine cysts in rats that are of paramesonephric duct origin, which has not been previously reported or characterized in the rat pathology literature. Additionally, we present an IHC panel that can be used in rodent tissue to identify cysts of paramesonephric duct origin.

Subserosal uterine cysts were observed grossly and / or microscopically in 20 out of 2,400 SD rats evaluated in 5 multi-generation studies conducted by the NTP. All animals were handled, cared for, and used in accordance with the NRC Guide for the Care and Use of Laboratory Animals ( FDA, 1988 , FDA, 1994 , NRC, 2011 ). These studies were conducted in AALAC-approved and PHS-assured facilities. Uterine tissue from routine necropsies was collected according to the study protocol and fixed in 10% neutral buffered formalin in accordance with U.S. FDA Good Laboratory Practice (GLP) Regulations ( FDA, 1988 , FDA, 1994 ). The female reproductive organs were trimmed according to the Specifications for the Conduct of Studies to Evaluate Reproductive and Developmental Toxicity of Chemical, Biological, and Physical Agents in Laboratory Animals for the National Toxicology Program (May 2011) ( NTP, 2011 ). Briefly, the cervix and vagina were freed en masse from the uterine body to achieve sagittal sections for evaluation. Cross sections of the midpoint of both uterine horns and longitudinal sections of remaining uterine tissue were taken for assessment. Fixed tissues were embedded in paraffin, sectioned at 5 microns, and stained with Hematoxylin & Eosin (H&E). Ten of the 20 animals diagnosed with cysts had adequate cystic tissue available for further evaluation with Masson’s Trichrome and immunohistochemical staining. Uterine, cervical and vaginal tissue sections were examined histologically to determine phase of estrous cycle as outlined in the INHAND female reproductive tract document ( Dixon et al., 2014 ). A total of 9 – 10 uterine cysts were immunohistochemically stained for alpha smooth muscle actin (SMA, smooth muscle marker), cytokeratin 18 (CK-18, epithelial cell marker), estrogen receptor alpha (ER-α, steroid receptor marker), progesterone receptor (PR, steroid receptor marker), PAX8 (paramesonephric tissue differentiation marker), and GATA3 (mesonephric tissue differentiation marker) as outlined in Table 1 . Briefly, formalin-fixed, paraffin-embedded tissue sections were deparaffinized in xylene and rehydrated through graded alcohols. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide before or following heat-induced antigen retrieval using a Decloaker® pressure chamber (Biocare Medical, Concord, CA). Nonspecific sites were blocked using normal horse, donkey, or goat serum, respective to the secondary antibody used, (Jackson Immunoresearch Laboratories, Inc., West Grove, PA) prior to or followed by incubation with an avidin/biotin blocking kit (Vector, Burlingame, CA). All tissues were incubated with the primary antibodies listed in Table 1 for 1 hr at room temperature, followed by incubation with the appropriate secondary antibody for 30 min at room temperature. Negative controls were incubated with normal rabbit IgG (Abcam, Cambridge, MA) or normal mouse IgG1 (BD Biosciences, San Jose, CA) diluted to the same concentration as the primary antibody. Antigen-antibody complexes were labeled using a Streptavidin SS Label (Biogenex Laboratories, San Ramon, CA) or avidin–biotin affinity system (Vectastain Elite ABC kit; Vector Laboratories, Burlingame, CA) and visualized with 3,3’-diaminobenzidine chromogen (DakoCytomation, Carpenteria, CA). Tissue sections were counterstained with hematoxylin, dehydrated through graded ethanol, cleared in xylene, and cover-slipped. Positive controls included rat kidney for PAX8, rat mammary tissue for GATA3, and the uterus, which also served as an internal control for CK18, ER-α, PR, and SMA. All uterine tissues were evaluated by routine light microscopy. Immunohistochemical stains were assessed and a semiquantitative scoring method, which combines the overall percentage of immunopositive cells and average staining intensity, was used to determine a multiplicative quickscore for the wall or epithelium of each cyst at an original magnification of 20X ( Detre et al. , 1995 , Dixon et al. , 2000 ). Briefly, a numerical value was generated for the percent of the cyst showing positive immunostaining (0–4% = 1; 5–19% = 2; 20–39% = 3; 40–59% = 4; 60–79% = 5; 80–100% = 6). This number was multiplied by a numerical value assigned to the intensity of immunostaining in each cyst (negative = 0, weak = 1, moderate = 2, or intense = 3) to obtain a quickscore number ranging from 0–18. The quickscore values were used to assign each cyst to an immunostaining category of: Negative immunostaining = 0–1 (−); weak = 2–6 (+); moderate = 7–12 (++); and intense = 13–18 (+++).