Ontogeny of estrogen receptors in human male and female fetal reproductive tracts.

Since crown-rump measurements were not possible, the convention for age determination is heel-toe length. Examination of wholemount images of human fetal reproductive tract organs staged by this method reveal a clear increase in size and morphological complexity based upon heel-toe age determinations with very few exceptions ( Robboy et al, 2017 ; Shen et al, 2018 ). However, the accuracy of designated ages of all studies using this convention is likely plus/minus 1 week at best. It is informative to review a quote from the famous Carnegie collection of human embryos that is relevant to the staging of human embryos. “An embryo is assigned a Carnegie stage (numbered from 1 to 23) based on its external features. This staging system is not dependent on the chronological age or the size of the embryo. The stages are in a sense arbitrary levels of maturity based on multiple physical features. Embryos that might have different ages or sizes can be assigned the same Carnegie stage based on their external appearance because of the natural variation which occurs between individuals ( Smith, 2016 ).” With this cautionary note, we describe the ontogeny of estrogen receptors in the human fetal male and female reproductive tracts. Most of the human female reproductive tract is derived from the Müllerian ducts, which at ~8 weeks of gestation fuse in the midline to form the uterovaginal canal ( Mutter and Robboy, 2014 ), the precursor of uterine corpus, uterine cervix and a contributor to human vaginal development ( Cunha et al, 2018a ; Kurita and Terakawa, 2020 ; Robboy et al, 2017 ). The cranial unfused portions of the Müllerian ducts form the uterine tubes. The caudal end of the uterovaginal canal makes contact with the female urethra near the introitus whose epithelium is derived from endodermal urogenital sinus (UGE). UGE also contributes to development of human vaginal epithelium ( Cunha et al, 2018a ; Robboy et al, 2017 ). Indeed, Bulmer asserted that upgrowth of UGE “forms the whole of (vaginal) epithelial lining” ( Bulmer, 1957 ). Bulmer’s theory is supported by an immunohistochemical study with antibodies that recognize Müllerian epithelium (PAX2) versus UGE (FOXA1) ( Cunha et al, 2018a ; Robboy et al, 2017 ). This background of the developmental biology of the human female reproductive tract is critical in understanding the ontogeny of estrogen receptors over the time frame of 8 to 21 weeks of gestation. Estrogen receptor alpha (ESR1) was expressed in epithelium of the cranial portion of the Müllerian duct (precursor of the uterine tube) at 8 weeks of gestation ( Figs. 1A & 10A ). Epithelium and mesenchyme of the newly formed uterovaginal canal, derived from midline fusion of the Müllerian ducts, was ESR1-negative in 8- to 10-week specimens and is surrounded with an ESR1-negative mesenchyme ( Figs. 1B ). ESR1 was first detected at 12 weeks of gestation in the mesenchyme of the cranial end of the uterovaginal canal (uterine corpus precursor) ( Figs. 1C – D ), while the epithelium of the uterovaginal canal remained ESR1-negative ( Fig. 1D ). Examination of the mesenchyme of the cranial end of the uterovaginal canal reveals two distinct zones: (a) the zone adjacent to the epithelium was ESR1-positive, the presumed precursor of endometrial stroma, while (b) the peripheral ESR1-negative zone adjacent to the serosal surface is undoubtedly the myometrial precursor ( Figs. 1C – D ). At 12 weeks, the lumen of the caudal aspect of the uterovaginal canal becomes occluded to form the solid vaginal plate ( Robboy et al, 2017 ), which makes contact with UGE-derived (FOXA1-positive) endodermal introital epithelium ( Figs. 1C & E ). The epithelium of the solid vaginal plate and associated mesenchyme were ESR1-negative at 12 weeks of gestation ( Figs. 1C & E ). At 14.5 weeks of gestation, expression of ESR1 was increased in mesenchyme associated with the cranial aspect of the uterovaginal canal (uterine corpus precursor) ( Figs. 2A – C ). Intense ESR1-reactivity was observed in the prospective endometrial stroma ( Fig. 2B , double-headed green arrow) adjacent to the Müllerian epithelium in the region of the uterine corpus and uterine cervix, while the prospective myometrium (outer zone) was ESR1-negative ( Figs. 2A – B , double-headed red arrow). At the apex of the uterovaginal canal (uterine fundus), the Müllerian epithelium was a mixture of ESR1-positive and ESR1-negative cells ( Fig. 2B , inset). This region is at/near the uterotubal junction. The Müllerian epithelium situated more caudally within the uterovaginal canal (at or near the uterine corpus/uterine cervical boundary) was ESR1-negative ( Fig. 2C ), even though the uterine corpus/uterine cervical boundary cannot be ascertained with certainty at this age ( Robboy et al, 2017 ). The epithelium and associated mesenchyme of the caudal solid vaginal plate (near the introitus) were ESR1-negative ( Figs. 2A & D ), while the more cranial portion of the solid vaginal plate contained ESR1-positive epithelial cells in association with ESR1-negative mesenchymal cells ( Figs. 2A & E ). At 16 weeks of gestation the epithelium of the uterotubal junction was strongly ESR1-positive, while the surrounding mesenchyme was mostly ESR1-negative ( Figs. 3A – B ). This observation contrasts with the observation at 14.5 weeks ( Fig. 2B ) in which the epithelium of the uterotubal junction was a mixture of ESR1-positive and ESR1-negative cells as reported previously ( Cunha et al, 2017a ). The epithelium of the uterine corpus was for the most part ESR1-negative (not illustrated) as is the epithelium of the uterine cervix ( Fig. 3C ). The junction of the canalized uterovaginal canal and the solid vaginal plate showed an abrupt shift in ERS1 expression with the epithelium of the canalized uterovaginal canal being ESR1-negative and the epithelium of the solid vaginal plate being ESR1-positive ( Figs. 3A , D – E ). The mesenchyme associated with the uterine corpus and uterine cervix (prospective endometrial stroma) was ESR1-positive ( Figs. 3A – C ) as was the case for the 12 and 14.5 week specimens ( Figs 1C – D & 2A – B ). The mesenchyme associated with the solid vaginal plate was a mixture of ESR1-positive and ESR1-negative cells ( Figs. 3D – E ), while the prospective myometrium was ESR1-negative ( Fig. 3A ) as was the case for the 12 and 14.5 week specimens ( Figs 1C – D & 2A – B ). The pattern of ESR1 at 18 weeks of gestation was similar to that seen at 16 weeks. Epithelia of the uterine corpus and uterine cervix were ESR1-negative ( Figs. 3F – H ), while the solid vaginal plate exhibited zones of ESR1 reactivity (caudally) and ESR1-negativity cranially ( Figs. 3F & I ). At 18 weeks of gestation the uterine corpus and uterine cervix are distinct. Gland rudiments have appeared in the cervix ( Figs. 3F & H ) but not in the uterine corpus ( Fig. 3F ). The vaginal fornices are obvious at 18 weeks (asterisks in Fig. 3F ). In the uterine corpus, ESR1-reactive endometrial stroma is in intimate association with the ESR1-negative uterine epithelium ( Figs. 3F – G ). The mesenchyme associated with cervical epithelium was predominantly ESR1-negative ( Fig. 3H ) with rare ESR1-positive cells. At 21 weeks of gestation, transverse sections of the human fetal uterus reveal an irregular lumen lined by ESR1-negative luminal epithelial cells ( Fig. 4B ). Uterine glands penetrating peripherally into the endometrial stroma were predominantly ESR1-negative (but contained rare ESR1-positive cells), and the endometrial stroma was a mixture of ESR1-negative and ESR1-positive cells ( Fig. 4B and inset). In humans, endogenous estrogens are elevated at 21 weeks of gestation ( Oakey, 1970 ) and have elicited thickening and differentiation of the vaginal epithelium ( Figs. 5 & 6 ). The multi-layered human vaginal epithelium ( Fig. 5 ) exhibited strong ESR1 immunostaining in its basal layer and slightly weaker ESR1 immunostaining in supra-basal layers ( Fig. 6F , see figure 5 for the approximate region labeled “F”). In the uterine corpus and uterine cervix the endometrial stroma ( Fig. 5 , double-head green arrow) was ESR1-positive, while the myometrium ( Fig. 5 , double-head red arrow) was predominantly ESR1 negative. The area in figure 5 labeled “6B&D” represents an interface between a thick vaginal-like epithelium and a thin pseudostratified columnar epithelium, which is seen at high magnification in figures 6B and D . We interpret this as the squamo-columnar junction between the vagina and cervix. Note that the stroma across this junction is ESR1-positive as is the thick vaginal epithelium ( Figs. 6B & D ). The pseudostratified columnar epithelium at this interface is ESR1 negative ( Figs. 6B & D ). See Table 1 for summary. At 8 weeks of gestation transverse sections depict the caudal aspect of the uterovaginal canal flanked by the Wolffian ducts ( Fig. 7 ). All three epithelial structures were ESR2-positive and were surrounded by ESR2-negative mesenchyme ( Fig. 7A ). At 9 weeks of gestation transverse sections of the uterovaginal canal were obtained from cranial ( Fig. 7B ) and caudal ( Fig. 7C ) levels. In both areas ESR2 immunostaining was seen in the epithelium of the uterovaginal canal and Wolffian ducts ( Figs. 7B – C ). The mesenchyme associated with these epithelial structures was strongly ESR2-positive in the caudal region ( Fig. 7C ), but in the cranial regions was weakly positive/negative ( Figs. 7B ). At 10 weeks of gestation ( Fig. 7D ) the epithelium of the uterovaginal canal and Wolffian ducts remained ESR2-positive, while the mesenchyme was a mixture of ESR2-positive and ESR2-negative cells. Note that the Wolffian ducts appear to be initiating regression in the 9- ( Fig. 7B ) and 10 -week ( Fig. 7D ) specimens. At 12.5 weeks of gestation, the epithelium of the uterovaginal canal was strongly ESR2-positive throughout as was its associated mesenchyme ( Fig. 8A & B ), which in the uterine and cervical regions could be interpreted as endometrial stroma. Mesenchyme peripheral to the prospective endometrial stroma, that is presumptive myometrium, was ESR2-negative ( Fig. 8A ). Epithelium of the solid vaginal plate at 12.5 weeks and its associated mesenchyme was ESR2-positive ( Fig. 8A & C ). At 16 weeks of gestation ( Figs. 8D – H ) the boundary between the uterine corpus and uterine cervix remained indistinct ( Fig. 8D ), even though positions of these organs can be roughly inferred. Epithelium of the uterine corpus was pseudostratified columnar and weakly ESR2-positive and was surrounded with prospective endometrial stroma containing ESR2-positive and ESR2-negative cells ( Figs. 8D – E ). In the cervical region, the epithelium transitions to a multilayered stratified epithelium, which was ESR2-positive ( Fig. 8D , F & H ) and was surrounded by mesenchyme that contained ESR2-positive and -negative cells. More caudally the solid vaginal plate exhibited areas of ESR2-positivity and ESR2-negativity ( Figs. 8D , F – H ). At 21 weeks uterine luminal and glandular epithelial cells were ESR2-negative as was the associated endometrial stroma, while the prospective myometrium was ESR2-positive ( Fig. 4A ). The caudal segment of the fetal female reproductive tract is represented as the solid epithelial vaginal plate, which at 17 weeks showed evidence of proliferation and differentiation into a mature vaginal epithelium ( Figs. 9A – B ), presumably under the influence of endogenous estrogens ( Oakey, 1970 ). Figures 9A – B reveal this differentiation process; both the differentiating vaginal epithelium and surrounding stroma are ESR2 positive, at least in this most caudal position at 17 weeks of gestation. At 18 weeks, the process of vaginal epithelial differentiation had advanced ( Figs. 9C – D ), and the vagina had developed a lumen filled with sloughed squamous cells ( Fig. 9C ). The differentiating vaginal epithelium was strongly ESR2-positive, especially in the basal and suprabasal layers, and the surrounding mesenchyme was a mixture of ESR2-positive and negative cells ( Fig. 9D ). At 21 weeks of gestation the vaginal epithelium, especially that near the vaginal fornices, was a mixture of ESR2-positive and ESR2-negative cells ( Fig. 6A & C , arrowheads). In the vagina proper vaginal epithelial and stromal cells were mostly ESR2-negative ( Fig. 6E ). At the squamocolumnar junction between vaginal and cervical epithelia ( Fig. 6A & C ), the psuedostratified columnar cervical epithelium was ESR2-negative ( Fig. 6A & C ). See Tables 2 and 3 for summary. At 8 to 9 weeks of gestation the epithelium of the human uterine tube (unfused Mullerian duct) was ESR1-positive ( Figs. 1A & 10A – B ) and remained ESR1-positive at all stages examined (8 to 20 weeks) ( Fig. 10A – H ). The surrounding mesenchyme was weakly ESR1-positive/negative at 8 to 9 weeks, but at later stages the mesenchyme was a mixture of ESR1-positive and ESR1-negative cells. It is possible that the position within the uterine tube (uterine end versus ovarian end) may influence the intensity of ESR1 expression. See Table 4 for summary. The pattern of ESR2 expression in the developing human uterine tube was almost identical to that seen for ESR1. At 8 to 9 weeks the epithelium of the human uterine tube was ESR1-positive ( Figs. 11A – B ), while the surrounding mesenchyme was weakly positive/negative. From 10 to 20 weeks both the uterine tubal epithelium and mesenchyme were ESR2-positive ( Figs. 11C – G ). See Table 4 for summary. The verumontanum ( Fig. 12A , red lines and arrows) is a distinctive landmark for the developing human prostate at 8 to 9 weeks, which is before prostatic buds have appeared ( Cunha et al, 2018b ). The Wolffian and fused Müllerian ducts (prostatic utricle) join the epithelium of the verumontanum and can be seen in figure 6 of Cunha et al (2018). At 9.5 weeks of gestation the epithelium of the prospective prostatic urethra, the verumontanum, the Wolffian and fused Müllerian ducts (prostatic utricle) as well as their associated mesenchyma were ESR1-negative ( Fig. 12A ). Thereafter (11, 15 and 21 weeks of gestation), when prostatic buds were present, ESR1 remained undetectable in epithelium of the prostatic urethra ( Fig. 12C ), in epithelium of solid prostatic ducts ( Figs. 12B – D ), in canalized prostatic ducts (not illustrated) and in the mesenchyme associated with these epithelial structures ( Fig. 12 ). See Table 5 for summary. Estrogen receptor beta (ESR2) was observed at 9.5 weeks of gestation in epithelial cells in the mid-dorsum of the prospective prostatic urethra (verumontanum) and in mid-dorsal mesenchyme between the verumontanum and the Wolffian ducts and prostatic utricle (fused Müllerian ducts) ( Fig. 13A ). Most of the remaining epithelium of the prospective prostatic urethra was ESR2-negative with scattered weakly ESR2-positive cells ( Fig. 13A ). At 11 ( Fig. 13B ) and 15 weeks ( Fig. 13C ) solid and canalized prostatic ducts were ESR2-positive, while the surrounding mesenchyme was ESR2-negative ( Figs. 13B – C ). At 22 weeks solid (and canalized, not illustrated) prostatic ducts remined ESR2-positive, while the mesenchyme contained a mixture of ESR2-positive and ESR2-negative cells ( Fig. 13D ). See Table 5 for summary. ESR1 staining was performed on epididymal specimens at 8, 9, 12, 14, 16, 17, 18, 19 and 20 weeks of gestation. For all ages the epididymes were section longitudinally. However, this plane of section is not optimal for identification of the Wolffian duct within the urogenital ridge at 8 and 9 weeks of gestation. The longitudinal sections of the 8–9 week urogenital ridge revealed an abundance of ducts, most of which are mesonephric ducts and perhaps a portion of the Wolffian duct destined to form the coiled epididymis. ESR1 staining was not apparent in both the 8 and the 9 week specimens ( Fig. 14A – B ). In contrast, at 12 weeks of gestation weak ESR1 staining was seen in ducts of the head and tail of the epididymis ( Fig. 14 C – D ). For specimens at 14 weeks and older, a consistent staining pattern was seen in that ducts within the head of the epididymis were of small diameter, closely associated with each other and consistently ESR1-positive, while surrounded by an ESR1-negative mesenchyme ( Figs. 14E – F , H – I , K – L ). In contrast, ducts within the tail of the epididymis were of larger diameter, less densely associated with each other, consistently ESR1-negative, and surrounded by an ESR1-negative mesenchyme ( Figs. 14E , G , H , J , K , M ). Due to variable section orientation, the body of the epididymis was generally not discernable, but was identified in the 17 week specimen ( Fig. 14H ). The body of the epididymis appeared to be a transition zone of ESR1 expression and thus exhibited a mixture of small ESR1-positive ducts and larger ESR1-negative ducts (Fig. H). See Tables 6 and 7 for summary. ESR2 staining was performed on specimens at 8, 9, 12, 14, 16, 17, 18, 19 and 20 weeks of gestation. Longitudinal sections of the urogenital ridge at 8 and 9 weeks of gestation revealed an abundance of ducts, most of which are mesonephric ducts. The Wolffian duct is difficult to identify in such longitudinal sections. ESR2 staining was not apparent in the many ducts seen in both the 8- and the 9-week specimens ( Fig. 15A – B ). At 12 weeks of gestation ducts of the head and tail of the epididymis are apparent and are ESR2-positive ( Figs. 15C – E ). The pattern of ESR2 staining varied somewhat with age. Ducts in the head and tail of the epididymis were strongly ESR2-positive at 14 and 16 weeks ( Figs. 15F – H ). At 17 to 19 weeks of gestation ducts of the head of the epididymis were weakly ESR2-positive, while of the tail of the epididymis were distinctly ESR2-positive ( Figs. 15I – N ). From 12 to 19 weeks the surrounding mesenchyme contained a predominance of ESR2-positive cells ( Figs. 15C – N ). See Tables 6 and 7 for summary. On the whole, expression of ESR1 was restricted to rare subsets of cells in the human fetal penis. The main structures within the developing human penis are the corporal body, the glans, the urethra and the developing preputial lamina ( Figs. 16A & C ). At 11.5 weeks of gestation a small focus of ESR1-positive epithelial cells was seen at the junction of the ventral preputial lamina with the ventral epidermis ( Figs. 16A – B , arrowheads). At this stage (11.5 weeks) none of the other tissues/cells exhibited ESR1 immuno-reactivity, including the corporal body that remained ESR1-negative at all ages examined. At 12.5 weeks the basal epidermal cells near the urethral meatus exhibited ESR1 immuno-reactivity ( Figs. 16C – D , arrowheads). In addition, mesenchymal cells associated with the proximal aspect of the preputial lamina were ESR1-positive ( Fig. 16E , arrowheads). At 14 to 15 weeks, basal epidermal cells near the ventral attachment of the preputial lamina with the epidermis were ESR1-positive ( Figs. 17A & C ), and thus exhibited a pattern similar to that seen at 11.5 and 12.5 weeks of gestation (see figure 16 ). A few epithelial cells of the dorsal preputial lamina were also ESR1-positive ( Fig. 17B , arrowheads). At 15 weeks discrete foci of ESR1-positive epithelial cells ( Figs. 17D – F ) were observed in the canalizing urethral plate (double-headed green arrow in 12D) during formation of the urethral meatus ( Figs. 17D – F , red arrow in E=urethral meatus) and in basal epithelial cells of the preputial lamina ( Fig. 17F ). At 16 weeks ( Fig. 18 ) ESR1-positive basal epidermal cells were seen near the peno-scrotal junction ( Fig. 18A – B , arrowheads). As reported at 14 weeks, patches of epithelial cells in the dorsal preputial lamina were also ESR1-positive (not illustrated). See Table 8 for summary. Estrogen receptor beta (ESR2) was widely expressed in the developing human penis. At 11.5 weeks of gestation ESR2 was strongly expressed in the corporal body ( Fig. 19A – B ), in the epithelium of the urethral groove and urethra ( Figs. 19A – B , green arrowheads), in penile epidermis ( Fig. 19B , black arrowheads), in mesenchyme of the glans ( Fig. 19B ), and in the epithelium of the preputial lamina ( Fig. 19B , red arrowhead). ESR2-positive cells/tissues seen at 11.5 weeks of gestation remained ESR2-positive at 12.5 weeks ( Figs. 19C – E ). The corporal body ( Fig. 19C ) remained ESR2-positive at all ages examined. At 12.5 weeks of gestation the urethral meatus (green arrow in Fig. 19E ) is evident, and its epithelium and associated mesenchyme were ESR2-positive ( Fig. 19E ). The epithelium and associated mesenchyme of the preputial lamina were also ESR2-positive at 12.5 weeks of gestation ( Fig. 19D ). At 14 weeks of gestation ( Figs. 19F – H ), epithelial cells of the preputial lamina were ESR2-positive ( Figs. 19G – H ), while the corporal body exhibited weak ESR2-reactivity ( Fig. 19F ). At 16 weeks ( Fig. 20 ) the urethral epithelium and surrounding mesenchyme was ESR2-positive ( Fig. 20A – B ). Epidermal cells and associated mesenchymal cells were ESR2-positive at the penoscrotal junction ( Fig. 20A & C ). See Table 8 for summary. As is the case for the developing human penis, ESR1 expression in the fetal human clitoris is meager. At 9.5 weeks of gestation, a patch of ESR1-positive mesenchymal cells was observed ventral to the corporal body ( Figs. 21A – B ). At 12.5 weeks of gestation ( Figs. 21C – D ) ESR1 was expressed in basal epidermal cells near the attachment of the ventral preputial lamina ( Figs. 21C – D , black arrowheads). At 16 weeks of gestation ESR1 was detected in a subset of mesenchymal cells associated with the female urethra ( Figs. 21E – F , arrowheads). See Table 9 for summary. ESR2 was broadly expressed in the developing human clitoris. At 9.5 weeks of gestation para-sagittal sections revealed localization of ESR2 to the edge of the vestibular groove ( Fig. 22A ). The epithelial cells of the vestibular groove as well as a subset of clitoral epidermal cells were ESR2-positive ( Figs. 22A – B , green arrowheads in C ). At 9.5 weeks of gestation the corporal body was ESR2-negative ( Fig. 22A ). At 12.5 weeks of gestation ESR2 was detected in the corporal body ( Fig. 22D – E ), in the preputial lamina and in the epidermis ( Figs. 22D – E ). At 16 weeks of gestation ESR2 was also detected in the corporal body, the preputial lamina and in the epidermis ( Figs. 22F – G arrowheads). Across all of these ages the mesenchyme of the glans exhibited sparse expression of ESR2 ( Fig. 22 ). See Table 9 for summary.

Human fetal reproductive tract organs (8–22 weeks of gestation) were collected devoid of patient identifiers after elective termination of pregnancy (Committee on Human Research at UCSF, IRB# 12–08813). Gestational age of fetal specimens was estimated using heel-toe length as described previously ( Drey et al, 2005 ; Robboy et al, 2017 ; Shen et al, 2018 ). Specimens were fixed in 10% buffered formalin, embedded in paraffin and serially sectioned at 7μm. Every 20 th section was stained with hematoxylin and eosin to assess histology. Intervening paraffin sections were immunostained with antibodies to ERS1 (Abcam, Ab16660, 1/100) and ESR2 (Leica, NCL-ER-beta, 1/50) as described previously ( Rodriguez et al, 2012 ). Immunostaining was detected using horseradish peroxidase-based Vectastain kits (Vector Laboratories, Burlingame, CA). For negative controls the primary antibodies were omitted. Some of the data reported in this paper has been reported previously but is augmented by additional unpublished data to present a comprehensive temporal/spatial ontogeny of ESR1 and ESR2 in developing male and female human reproductive tracts. This study is based upon analysis of 121 human fetal specimens 8 to 21 weeks of gestation compiled from several previous publications and augmented with additional specimen acquired recently. A set of 13 specimens were dedicated exclusively for histological examination over the age range of 8 to 22 weeks; another set of 108 specimens were immunostained for ESR1 and ESR2 as indicated in Tables 1 – 9 .

All studies dealing with the ontogeny of estrogen receptors within human male and female reproductive tract organs, past and present, have inherent limitations on sample size as well as on the reason for pregnancy termination, as well as genetic issues or health issues with the pregnancy. We presume that for the vast number of specimens analyzed that genetic issues or health issues were not the reason for termination of pregnancy, and thus our data is representative of normal ontogeny of human fetal estrogen receptors. Moreover, it should be noted that our study is based upon 121 specimens, the largest collection analyzed to date. An understanding of the ontogeny of estrogen receptors within human male and female reproductive tract organs is essential for understanding the role of estrogens in normal development and the various malformations induced by exogenous estrogens in human male and female reproductive tracts. There are two estrogen receptors: (a) the classic estrogen receptor alpha, now designated ESR1 and (b) estrogen receptor beta now designated ESR2. These two receptors are expressed individually in unique temporal and spatial patterns in developing human male and female reproductive tract organs, which give clues as to which receptors may be involved in estrogen-induced malformations and possible roles in normal development. This report provides the most detailed account of ESR1 and ESR2 expression in developing human male and female reproductive tracts to date, extending our earlier studies ( Baskin et al, 2020 ; Cunha et al, 2017a ; Cunha et al, 2018b ) and filling in gaps in the literature. Glatstein and Yeh were the first to report estrogen receptor alpha in the human fetal female reproductive tract and described ESR1 expression in uterine mesenchyme in 17 to 22 week fetuses ( Glatstein and Yeh, 1995 ), an observation confirmed subsequently ( Cunha et al, 2017a ). The present study extends these findings by describing ESR1 and ESR2 expression in epithelia and mesenchyma of uterovaginal canal, uterine tube, uterine corpus, uterine cervix and vagina of human female fetuses from 8–21 weeks ( Cunha et al, 2017a ). Glatstein and Yeh (1995) also reported an absence of estrogen receptor alpha immunostaining in uterine epithelium, a finding consistent with our previous report ( Cunha et al, 2017a ) and confirmed in the present study. This observation contrasts with ESR1 expression in adult human uterine epithelial cells ( Kurita et al, 2005 ; Lessey et al, 1988 ), but is consistent with observations in mice that ESR1 is first expressed in uterine mesenchyme and only later in uterine epithelium ( Bigsby and Cunha, 1986 ; Bigsby et al, 1990 ; Cooke et al, 1997 ; Yamashita et al, 1990 ). Thus, the temporal absence of ESR1 in developing uterine epithelium is a feature common to both embryonic/neonatal mice and fetal humans. The current study provides additional observations on the ontogeny of estrogen receptors in the human fetal female reproductive tract, most importantly observations on the ontogeny of ESR2. The ontogeny of estrogen receptors in the developing human female reproductive tract is a particularly important topic and provides the mechanistic underpinning for the structural anomalies reported in DES daughters ( Jefferies et al, 1984 ; Kaufman et al, 1980 ; Kaufman et al, 1977 ; O’Brien et al, 1979 ), who were exposed in utero to this potent estrogen. Mice treated with exogenous estrogens in the perinatal period provide a remarkably comparable model of estrogen-induced teratogenesis, and such mouse studies demonstrate that estrogen-induced anomalies are mediated via ESR1 in so far as malformations seen in perinatally estrogen-treated wild-type mice are absent in ESR1-KO mice ( Couse et al, 2001b ; Couse and Korach, 2004 ). Endocrinologic studies support this conclusion ( Nakamura et al, 2008 ). However, ESR2 has been shown to be important in persistent inhibition of postnatal uterine epithelial cell proliferation, as uterine and vaginal cell proliferation is elevated in adult ESR2-knockout versus wild-type mice ( Nakajima et al, 2015 ). The presence of uterine ESR1 and ESR2 are presumed to be the basis of prenatal human uterine growth which is most intense between the 16th and 24th weeks of gestation ( Mukerje, 1980 ; Pietryga and Wozniak, 1992 ). Initial expression of epithelial ESR1 varied in an organotypic manner. ESR1 was initially detected in epithelia of uterine tube at 8 weeks. Epithelia of the uterine corpus and cervix was consistently ESR1-negative for the most part from 12 to 20 weeks with few exceptions, but at 20 to 21 weeks uterine glandular epithelium was a mixture of ESR1-positive and ESR1-negative cells. ESR1 was detected in epithelium at the uterotubal junction at 16 and 20 weeks. It is noteworthy that expression of ESR1 in the Mullerian duct and its derivatives (uterovaginal canal, uterovaginal junction, uterine fundus, uterine corpus and uterine cervix) varies on a cranial-caudal basis. The epithelium of the cranial portion of the Mullerian duct (uterine tubal rudiment) expressed ESR1 as early as 8 weeks of gestation, while the epithelium of the more caudally situated uterovaginal canal was ESR1-negative as was epithelium of the uterine corpus and uterine cervix at all ages examined (12 to 21 weeks). Only the most cranial epithelia of the uterine fundus and uterovaginal junction expressed ESR1. Thus, cranial-caudal position appears to influence the expression of ESR1. The solid vaginal plate exhibited ESR1 at 16 and 18 weeks, and human fetal vaginal epithelium exhibited histologic maturation at 17 to 21 weeks coincident with expression of ESR2 and ESR1. Additional timed vaginal specimens would be useful to correlate estrogen-induced vaginal epithelial maturation with the timing of expression of ESR1 and ESR2. It is puzzling that mature vaginal epithelium was ESR2-positive at 18 weeks ( fig. 9D ), but mostly ESR2-negative at 21 weeks ( Fig. 6E ). Initial mesenchymal expression of ESR1 varied in an organotypic manner. ESR1 was definitely detected in mesenchyme of the uterine tube at 10 to 11 weeks, while initial expression of ESR1 in the uterine/cervical mesenchyme began at 12 weeks, and vaginal mesenchyme first expressed ESR1 at 16 weeks. For all female reproductive tract organs ESR1 was continuously maintained after initial expression. ESR2 was detected in epithelia of the Müllerian ducts and uterovaginal canal as early as 8 weeks of gestation and continued to be expressed in epithelia of developing human female reproductive tract organs (uterus, cervix and vagina) thereafter. Mesenchymal expression of ESR2 began as early as 9 weeks in association with the newly formed uterovaginal canal, was apparent in the developing uterine tube at 10 weeks and was maintained thereafter at variable levels within the organotypic derivatives of the uterovaginal canal. On the whole, the ontogenic expression of ESR1 and ESR2 were remarkably similar in the developing uterine tube. The ontogenic organotypic profiles of ESR1 and ESR2 coincide with the period (7 to 15 weeks of gestation) in which DES elicited structural anomalies of reproductive tracts of women whose mothers received DES during pregnancy ( Jefferies et al, 1984 ; O’Brien et al, 1979 ). A consistent finding was strong ESR1 and ESR2 staining in the prospective endometrial stroma and an absence of staining in the prospective myometrium. Our identification of prospective myometrium was based solely upon it position near the serosa, but is verified by smooth muscle actin staining reported earlier ( Cunha et al, 2018a ; Fritsch et al, 2013 ). Xenografts of human fetal female reproductive tracts (5 to 17 weeks post fertilization) grown in DES-treated athymic mouse hosts exhibited suppression of normal reproductive tract development and exhibited defects in the uterine tube, endometrial stroma, myometrium cervix and vagina including vaginal adenosis ( Cunha et al, 2017b ; Robboy et al, 1982 ; Taguchi et al, 1983 ). Thus, the ontogeny of estrogen receptors (ESR1 and ESR2) in the developing human female reproductive tract coincides remarkably well with the “window of sensitivity” to the adverse effects seen in DES daughters. Moreover, the spatial organotypic distribution (epithelium and mesenchyme) of both ESR1 and ESR2 are consistent with the structural anomalies within the human female reproductive tract that were elicited by prenatal exposure to DES ( Jefferies et al, 1984 ; Kaufman et al, 1980 ; O’Brien et al, 1979 ; Sandberg, 1976 ). Structural malformations of the reproductive tract of DES daughters are manifestations of architectural alterations in epithelium and stroma, with perhaps stromal alterations being of paramount importance. Major malformations such as T-shaped uterotubal junctions, abnormal uterine luminal contours, and incompetent cervix ( Jefferies et al, 1984 ; Kaufman et al, 1986 ) are predominately defined by abnormalities in the shape of the fibromuscular wall (stroma). It appears that the epithelium simply covers (lines) the abnormally shaped stromal contours. Our mechanistic hypothesis is that DES acting via uterine/cervical mesenchymal ESR1 and ESR2 has perturbed the morphogenetic and (quite likely) the differentiation programs inherent in the conversion of Müllerian duct mesenchyme into the properly shaped fibromuscular stroma of Müllerian duct organ derivatives, the uterine tube, uterine corpus and cervix. Significantly, ESR1 and ESR2 was expressed in Müllerian mesenchyme of the uterine tubal rudiment beginning at 9 to 10 weeks ( Fig. 7 ), and at 12 weeks in mesenchyme of the uterine corpus and persisting to at least 21 weeks in the developing human female reproductive tract. Thus, the timing of expression of both ESR1 and ESR2 corresponds with the “window of sensitivity” (7 to 15 weeks ) for induction of structural anomalies in DES daughters ( Jefferies et al, 1984 ). Neonatal rats and mice treated with DES from birth exhibited alterations in the spatial organization of uterine mesenchyme within 3 to 5 days of initiation of DES treatment, which culminated in adulthood as disruption of myometrial development, increased vimentin and decreased smooth muscle alpha actin throughout the fibromuscular wall of the uterus ( Brody and Cunha, 1989 ). A comparable study of neonatally DES-treated mice described a similar disruption in myometrial development with uterine glands penetrating into the disrupted myometrium, a condition called adenomyosis ( Huseby and Thurlow, 1982 ). Adenomyosis in mice has been attribution to hyperprolactinemia, a consequence in mice of neonatal estrogen treatment ( Mori and Nagasawa, 1983 , 1988 ). Malformation of the uterine tube has an additional level of complexity, which may be related to the T-shaped uterotubal junctions seen in DES daughters. The human uterine tube is a complex structure having 3 anatomically defined portions from medial to lateral: (a) the isthmus, (b) the ampulla, and (c) the infundibulum ( Clemente, 1985 ). Mesenchymal expression of ESR1 in the uterine tube begins as early as 10 weeks but may vary by medial-lateral location as mesenchyme at the uterotubal junction was weakly ESR1-positive at 16 weeks and strongly ESR1-positive at 20 weeks ( Figs. 3 ). In contrast, the epithelium of the developing uterine tube expressed ESR1 from 8 weeks onward to 20 to 21 weeks, with ESR2 exhibiting an identical epithelial expression profile. Whether subtle differences in ESR1 and ESR2 expression varies across isthmus, ampulla, and infundibulum of the uterine tube, perhaps in a temporal fashion in each of these areas, needs to be explored more fully. Also one must be cognizant of the idea that malformation of the uterotubal junction may be due to actions mediated via estrogen receptors in the epithelium, the mesenchyme or both in combination via endocrine and/or paracrine mechanisms. It is notable that ESR1 was detected in epithelium of the uterine fundus (near the uterotubal junctions) at 14 weeks and in the uterotubal junction per se at 16 weeks, but not within epithelium of the uterine corpus at all stages examined except for rare ESR1-posistive epithelial cells in uterine glands at 21 weeks of gestation. Thus, mesenchymal and/or epithelial ESR1 present in the developing uterine tube, uterus and cervix during the “window of sensitivity” (7 to 15 weeks ) may play a role in the structural tubal anomalies seen in DES daughters. Experimental animal studies have shown that developmental exposure of the rodent prostate to exogenous estrogens can have long-lasting pathologic effects, including increased susceptibility to carcinogenesis with aging ( Ho et al, 2006 ; Nelles et al, 2011 ; Prins et al, 2007 ; Risbridger et al, 2005 ). Such experimental animal studies raise the question of the ontogeny of ESR1 and ESR2 in the human fetal prostate. Previous reports demonstrate ESR2 in epithelial and stromal cells of human fetal prostate from 7 weeks through mid-gestation ( Adams et al, 2002 ; Shapiro et al, 2005a ). These studies demonstrate that ESR2 is the predominant estrogen receptor in the human fetal prostate. Indeed, by immunohistochemical methods ESR2 was detected in epithelium of the prostatic urethra (verumontanum) and in solid and canalized prostatic ducts from 9.5 to 22 weeks of gestation and was weakly expressed in prostatic mesenchyme at 22 weeks of gestation. This pattern of expression of ESR2 is maintained into neonatal and prepubertal periods ( Adams et al, 2002 ), and in adulthood ESR2 is the predominant estrogen receptor being expressed in basal epithelial cells and stromal cells. ESR1 was not detected until 15 weeks of gestation with sparse staining of the prostatic utricle ( Shapiro et al, 2005a ). At 19 weeks Shapiro reported ESR1 in fetal prostatic stromal cells, in urogenital sinus epithelium, ejaculatory ducts near the prostatic utricle and occasionally in prostatic ducts ( Shapiro et al, 2005a ). In definitive prostatic structures (ducts and associated stroma) ESR1 immunostaining was sparse. In the present study ESR1 was not detected in the human fetal prostate from 8 to 22 weeks, a finding in agreement with Adams et al (2002) . Based upon these observations, the role of ESR1 in human prostatic development remains to be determined with the caveat that expression of ESR1 apparently has not been studied in the third trimester when human prostatic epithelium exhibits squamous metaplasia attributed to high level of endogenous estrogen. Human prostatic epithelium is known to be responsive to endogenous and exogenous estrogens. During the third trimester, human fetal prostatic epithelium undergoes squamous metaplasia in response to high endogenous estrogen levels ( Oakey, 1970 ; Zondek and Zondek, 1979 ). Estrogenic induction of prostatic squamous metaplasia can be elicited experimentally in xenografts of human fetal prostate grown in DES-treated hosts ( Sugimura et al, 1988 ; Yonemura et al, 1995 ). Whether such squamous metaplasia is mediated by ESR1 or ESR2 in the human fetal prostate remains to be determined, even though ESR2 is the predominant estrogen receptor in the human fetal prostate as discussed above. However, in mice estrogen-induced prostatic squamous metaplasia is dependent upon signaling via ESR1 and not ESR2 ( Risbridger et al, 2001 ), a conclusion consistent with estrogenic imprinting studies in mice ( Prins et al, 2001 ). Prins et al has suggested that estrogen imprinting of the mouse prostate is mediated via stromal ESR1 ( Prins et al, 2001 ). This conclusion is supported by tissue recombinant studies carried out in mice by Risbridger et al, who further asserted that full and uniform estrogen-induced prostatic epithelial squamous metaplasia requires ESR1 in both the epithelium and stroma, that is, mouse prostatic squamous metaplasia only occurs when both epithelium and mesenchyme are wild-type ( Risbridger et al, 2001 ). There are many reports of persistent adverse effects in prostates of mice and rats treated perinatally with exogenous estrogen including predisposition to cancer ( Arai, 1970 ; Chung and MacFadden, 1980 ; Prins et al, 2007 ; Prins et al, 2006 ; Prins et al, 2008 ; Rajfer and Coffey, 1978 ; Santti et al, 1994 ; Timms et al, 2005 ; vom Saal et al, 1997 ). Whether such reports employing laboratory animals have relevance to human pathogenesis remains to be determined especially in regard to environmental estrogenic exposure. However, it is sobering to note that expression of ESR1 and ESR2 in rats and mice differs substantially from that reported herein in the human fetal prostate ( Shapiro et al, 2005a ). In laboratory rodents ESR1 is prominently expressed in prostatic mesenchyme ( Cooke et al, 1991 ; Prins et al, 2001 ; Timms et al, 1999 ). Notably, exposure to estrogens at physiologic levels activates ESR1 in mouse prostatic mesenchyme and increases methylation of the ESR1 promoter ( Bhandari et al, 2019 ). Prins reported that ESR1was observed in developing mouse prostatic stromal cells, but not in epithelial cells based upon immunohistochemistry and concluded that estrogenic imprinting of the developing mouse prostate is mediated via stromal ESR1 ( Prins et al, 2001 ). In contrast, Cooke et al, employing autoradiography with 3H-estradiol, reported estrogen binding to developing prostatic epithelial cells ( Cooke et al, 1991 ). Unfortunately the autoradiographic technique does not discriminate between ESR1 and ESR2. Further investigation will be required to resolve differences between human and animal studies and the relevance of animal observations to human prostatic pathogenesis. A considerable literature documents the adverse effects of exogenous estrogens on penile development in animals (see reviews by Cunha et al, 2015 ; Goyal et al, 2007 ). These animal studies, in combination with human epidemiologic studies, have led to the concept that human hypospadias is due in part to estrogenic endocrine disruptors ( Baskin et al, 2001 ; Carmichael et al, 2012 ; Klip et al, 2002 ; Palmer et al, 2005 ; Pons et al, 2005 ; Steinhardt, 2004 ) and emphasizes the need to explore the temporal/spatial distribution of ESR1 and ESR2 in the developing human penis. Both ESR1 and ESR2 are expressed in the developing penis, with ESR1 expression being consistently less prominent than that of ESR2. Kalloo et al failed to detect ESR1 in human fetal corpus cavernosum at 18 to 22 weeks of gestation ( Kalloo et al, 1993 ), an observation confirmed in this study. Crescioli et al subsequently reported ESR1 in blood vessels associated with the 12 week human fetal corpus cavernosum, in smooth muscle cells, and in stromal cells associated with skin and the urethral epithelium; these estrogen receptor-positive cells were somewhat rare and sparse ( Crescioli et al, 2003 ). ESR2 (but not ESR1) was detected in the corporal body. In addition, ESR1 and ESR2 are expressed in adult human penile urethral epithelium ( Dietrich et al, 2004 ). We concur with Crescioli et al that ESR1 is sparsely expressed in a very limited and sparse subset of cells within the developing human fetal penis based upon our analysis of 11.5 to 16 week specimens ( Baskin et al, 2020 ), while ESR2 is more extensively expressed in the developing human penis as shown in figures 19 – 20 . We noted rare, (a) scattered ESR1 expression in epithelium of the preputial lamina and in a subset of mesenchymal cells associated with the preputial lamina, in basal epidermal cells associated with the urethral meatus, and in the canalizing urethral plate within the glans ( Baskin et al, 2020 ). In addition, para-sagittal sections slightly lateral to the midline revealed ESR1-positive epidermal cells at the penoscrotal junction. This strategic position of ESR1 positive epidermal cells may have two interesting implications. (a) ESR1 at the penoscrotal junction may be related to penile webbing ( Bonitz and Hanna, 2016 ). (b) Alternatively, ESR1 expression at the penoscrotal junction may be indicative of estrogen receptor expression in the penile raphe where the right and left urethral folds have fused to form the urethra. From examination of scanning electron micrographs ( Baskin et al, 2020 ; Liu et al, 2018b ), it is evident that the penile raphe is not always located in the precise mid-ventral position and in some cases curves away from the midline. This constellation of ESR1-positive cells is located in a strategic position consistent with the anomalous morphology that constitutes human hypospadias. ESR2 is broadly expressed in (a) the developing corporal body, (b) epithelium of the urethral groove, (c) preputial lamina, and (d) epidermis. Thus, possibly in concert with ESR1, ESR2 is a mechanistic candidate for estrogen mediated hypospadias. Given the interest and knowledge in the role of estrogen in reproduction and fertility in males and females, a comprehensive mapping of the developmental expression of ESR1 and ESR2 in different penile and clitoral tissues and cells is required for understanding congenital malformations of the external genitalia ( Mowa et al, 2006 ). Penile hypospadias is one of the most common congenital malformations with an incidence as high as 1:200–1:300 male births ( Baskin, 2000 ). The development of this common congenital malformation has been attributed to genetic susceptibility and exposure to estrogenic endocrine disruptors ( Baskin et al, 2001 ; Carmichael et al, 2012 ; Steinhardt, 2004 ). Human hypospadias consists of the following elements: (a) a urethral defect, (b) a preputial defect (usually a ventral deficit in the prepuce), (c) chordee (abnormal curvature of the penis) and (d) local absence or hypoplasia of the corpus spongiosum (Baskin et al., 1998). In hypospadias the urethral meatus is located at various points proximal to its normal distal position at the tip of the glans. This malformation is frequently associated with failure of fusion of the urethral folds, a process that converts the urethral groove into the tubular penile urethra. However, normal human penile urethral development occurs via two entirely different morphogenetic mechanisms. Within the penile shaft urethral fold fusion converts the open urethral groove to the tubular penile urethra. Within the glans, the urethra forms by direct canalization of the solid urethral plate ( Baskin et al, 2018 ; Li et al, 2015 ; Liu et al, 2018a ). Formation of the urethra within the glans is associated with formation of the ventral aspect of the prepuce ( Cunha et al, 2020b ; Liu et al, 2018b ), which is deficient in hypospadias. Expression of ESR1 and ESR2 in the human fetal preputial lamina and associated mesenchyme provides a mechanistic scenario for perturbation of normal development of the prepuce in response to exogenous estrogenic endocrine disruptors as well as suggesting a role of ESR1 and ESR2 signaling in normal preputial development. In this regard, perinatal treatment of mice with DES elicits malformation of the external prepuce and the urethral meatus ( Mahawong et al, 2014a , b ). The role of estrogen signaling in normal development of the human clitoris is unclear at this time. Moreover, from all of the reports on malformations in DES daughters, there is no mention of clitoral malformations including from colposcopic studies in which such malformations could have been detected if present. Studies on the development of the mouse clitoris are not informative in regard to human clitoral development as the development and adult morphology of the mouse clitoris is completely non-comparable to that of the human ( Cunha et al, 2020a ). The epididymis develops from that part of the Wolffian duct lying adjacent to the testis. During development, under the influence of fetal testicular androgens ( Jost, 1970 ), this straight duct elongates and folds upon itself to form a compact organ of tortuous tubules, containing over 6 meters of a continuous duct ( Hinton et al, 2011 ; Rosse and Gaddum-Rosse, 1985 ). The length of the human fetal urogenital ridge and the straight Wolffian duct at 8 weeks of gestation is ~1.43mm based upon measurement from serial sections. Thus, from 8.5 weeks of gestation to adulthood this segment of the Wolffian duct increases in length by a factor of ~4300 times. Considerable growth in length of the epididymal segment of the Wolffian duct occurs during fetal life as indicated by the dozens of ductal profiles seen in longitudinal sections, a process dependent upon androgen action ( Joseph et al, 2009 ; Murashima et al, 2015 ) and mediated by androgen receptors present in the human fetal epididymis ( Shapiro et al, 2005b ). Epididymal growth during fetal stages varies considerably between the head versus the tail of the epididymis. For example, at 14 weeks of gestation sections of the head of the epididymis contains 40 to 50 ductal profiles, while the tail of the epididymis contains only ~6 ductal profiles. Thus ductal growth within the human fetal epididymis is vastly higher in the head versus the tail. Limited studies have documented the expression of estrogen receptors in the developing and adult epididymis of human and animal species. Using steroid autoradiography estrogen receptors were detected in epithelium and mesenchyme of the neonatal mouse epididymis ( Cooke et al, 1991 ). With the advent of antibodies to ESR2, Takeyama et al reported ESR2 in epithelium and mesenchyme of the 20 week human fetal epididymis, but ESR1 was not detected ( Takeyama et al, 2001 ). Shapiro et al reported ESR1 in human epididymal epithelium (presumably of the head of the epididymis), but not mesenchyme at 16 and 22 weeks of gestation. However, ESR1 staining was not observed in the tail of the epididymis, an observation confirmed in this study. In contrast, ESR2 was detected in human epididymal epithelium and mesenchyme at these ages ( Shapiro et al, 2005b ). The present study confirms and extends previous studies on ESR1 and ESR2 through analysis of fetal epididymis from 8 to 19 weeks. There are several reports of ESR1 and ESR2 in the adult epididymis including in human, and one consistent finding is high ESR1 and/or ESR2 expression in the head of the epididymis and low or absence of ESR1 and/or ESR2 expression in the tail of the epididymis ( Atanassova et al, 2001 ; Carpino et al, 2004 ; Hess et al, 2011 ; Joseph et al, 2011 ). This pattern of high estrogen receptor expression in the epididymal head and low or an absence of estrogen receptors is seen in the present study for ESR1 as early as 12 weeks of gestation, while ESR2 was observed at comparable levels in the epithelium of both the head and tail of the epididymis. The function of ESR1 and ESR2 in development of the human fetal epididymis remains an open question, which is beyond the scope of the present paper. However, epididymes form in ESR1- and ESR2-null mice ( Couse et al, 2001a ; Couse et al, 1997 ; Eddy et al, 1996 ; Joseph et al, 2010 ), and thus development of the epididymis (in mice) is not dependent upon signaling through either of these estrogen receptors. Exposure of the developing epididymis to exogenous estrogen elicits a range of irreversible deleterious effects. Anatomic/histologic effects of exogenous estrogens during development results postnatally in formation of epididymal cysts ( Dunn and Green, 1963 ; Gill et al, 1981 ), reduction in basal epithelial cells secondary to reduced testosterone action ( Atanassova et al, 2001 ), underdevelopment of the epididymal duct epithelium ( McKinnell et al, 2001 ), multilayering of basal epididymal epithelial cells ( Atanassova et al, 2005 ), and epididymal inflammation ( Miyaso et al, 2014 ; Nanjappa et al, 2019 ; Palmer et al, 2009 ). Exogenous estrogens during development also elicit permanent effects on the molecular pathways within the epididymis. Molecular effects of developmental estrogen treatment include postnatal reduction in the number of H(+)-ATPase positive epididymal epithelial cells required for luminal acidification of the epididymis ( Fisher et al, 2002 ), postnatal suppression of androgen receptor mRNA ( Yamamoto et al, 2009 ), and postnatal downregulation of aquaporin 9 which plays a role in fluid and solute transport ( Pastor-Soler et al, 2010 ; Wellejus et al, 2008 ). The temporal ontogeny of ESR1 and ESR2 in the human fetal epididymis coincides with the period of susceptibility of the human reproductive tract of exogenous estrogen and provides the molecular mechanism for long-term deleterious effects. The ontogeny of ESR1 and ESR2 in human male and female reproductive tract organs has been sparsely covered in the literature both in terms of the various fetal organ rudiments and the age range of observations. The current paper provides a comprehensive coverage of this broad topic and discusses the implications of ESR1 and ESR2 relative to normal development and malformations elicited by exogenous estrogens.

Sensitivity to the hormonal milieu (especially androgens and estrogens) is a critical signature of developing human (and animal) internal and external genitalia. Both androgens and estrogens elicit their effects via specific hormone binding proteins (androgen and estrogen receptors) whose expression begins during embryonic/fetal periods. This report focuses on the ontogeny of estrogen receptors, estrogen receptor alpha (ESR1) and estrogen receptor beta (ESR2) in human fetal male and female reproductive tracts, a topic for which there are major gaps in our knowledge. Thus, the goal of this paper is a comprehensive temporal/spatial mapping of human ESR1 and ESR2 expression in organs of human male and female internal and external genitalia over the time frame that encompasses the early ambisexual stage (8 weeks of gestation) to 22 weeks when morphogenesis and differentiation are advanced. The role of estrogens in normal development of human and animal male and female internal and external genitalia is unclear. Studies of ESR1-KO and ESR2-KO mice demonstrate that male and female internal and external genitalia form normally ( Krege et al, 1998 ; Lubahn et al, 1993 ), that is, all male and female reproductive tract organs form in these mutant mice. Nonetheless, effects of ESR1-KO are seen postnatally as reduced prostatic ductal branching ( Chen et al, 2009 ) and impaired postnatal growth of uteri and mammary glands ( Bocchinfuso and Korach, 1997 ; Couse and Korach, 1999 ). In humans, male and female fetuses are exposed to high levels of endogenous maternal estrogens beginning at the end of the second trimester and throughout the third trimester ( Oakey, 1970 ). Endogenous estrogens elicit squamous metaplasia of human fetal prostatic epithelium ( Zondek and Zondek, 1979 ), lactation and formation of breast nodules in newborn males and females ( Madlon-Kay, 1986 ), and proliferation and differentiation of vaginal epithelium ( Cunha et al, 2018a ). These effects in both male and female fetuses are transient and disappear after birth. However, exogenous estrogens can have long-term irreversible deleterious effects both in humans and laboratory animals. The pioneering studies of Howard Bern, John-Gunnar Forsberg, John McLachlan and Retha Newbold, and Gail Prins attest to the plethora of teratogenic and carcinogenic effects of exogenous estrogens on male and female genitalia in laboratory animals ( Bern and Talamantes, 1981 ; Forsberg and Kalland, 1981 ; McLachlan and Newbold, 1987 ; McLachlan et al, 2001 ; Newbold, 2008 ; Prins, 2008 ). The human counterpart to these animal studies is the well-known human diethylstilbestrol episode ( Herbst et al, 1975 ; Herbst et al, 1971 ). Diethylstilbestrol (DES), a synthetic estrogen, was administered to pregnant women from the mid-1940s to 1971 and induced a variety of malformations of the human female reproductive tract as well as clear cell adenocarcinoma of the vagina ( Herbst and Anderson, 1990 ). Such prenatal treatment with DES resulted in a spectrum of malformations of the uterine tubes, uterine corpus, cervix and vagina, which included T-shaped uterotubal junction, malformed incompetent cervix, abnormally shaped endometrial cavity, and vaginal adenosis ( Hoover et al, 2011 ; Jefferies et al, 1984 ; Rennell, 1979 ; Robboy et al, 1984 ; Robboy et al, 1977 ; Stillman, 1982 ; Titus-Ernstoff et al, 2010 ). Normal morphogenesis of male and female mouse external genitalia is also perturbed by developmental exposure to exogenous estrogens ( Mahawong et al, 2014a , b ; Sinclair et al, 2016 ), an observation that agrees with an increased incidence of hypospadias in sons of women treated prenatally with DES ( Klip et al, 2002 ; Palmer et al, 2005 ; Pons et al, 2005 ). Accordingly, human hypospadias has been attributed in part to exposure to estrogenic endocrine disruptors ( Baskin et al, 2001 ; Carmichael et al, 2012 ; Steinhardt, 2004 ; Yiee and Baskin, 2010 ). Timing of developmental exposure has been shown to be an important factor in estrogen-induced pathogenesis. The onset of sensitivity to endogenous or exogenous estrogens is clearly related to the ontogeny of expression of ESR1 and/or ESR2, which varies in different organs, tissues and cells. From animal studies, it is apparent that estrogen-induced pathogenesis is governed by a “window of sensitivity”, before which and after which estrogenic induction of pathogenesis is not possible or greatly diminished. For instance, prenatal treatment of mice with DES elicits a 10% incidence of vaginal adenosis, while neonatal DES treatment elicits a 75% incidence of this lesion ( Newbold and McLachlan, 1982 ). Adenosis can be elicited in mice when DES is administered from birth to 4 days postnatal. However, when DES treatment is initiated on day 6, vaginal adenosis was not detected ( Forsberg and Kalland, 1981 ). Kimura and Nandi described an abnormal form of vaginal epithelial differentiation, namely ovary-independent persistent vaginal cornification. A 100% incidence of this abnormality was observed when estradiol was administered from birth to day 5, but was not observed when initiation of estrogen treatment was delayed by 2 days ( Kimura et al, 1967 ). In humans, the “window of sensitivity” for DES-induced structural anomalies within the female reproductive tract is 7 through 15 weeks of gestation ( Jefferies et al, 1984 ), keeping in mind that estimates of gestational age are notoriously inaccurate, especially when based upon last menstrual period. After the “window of sensitivity” for irreversible adverse effects, organs/tissues respond appropriately to estrogenic stimulation, but effects are manifest only during the continued presence of the estrogenic ligand. Thus, effects disappear rapidly after removal of the estrogenic stimulus and thus has no long-term effects as is the case with adult exposure to exogenous estrogens. Given the close correspondence between the teratogenic and carcinogenic effects elicited by exogenous estrogens in prenatal/neonatal animals and prenatal humans, a comprehensive ontogeny of ESR1 and ESR2 in human male and female internal and external genitalia provides a mechanistic understanding of the effects of endogenous estrogen during normal development and the adverse effects of exogenous estrogens on development of male and female human reproductive tract organs. As indicated in several of the companion papers in this special issue, estrogenic environmental endocrine disruptors have been implicated for possible adverse effects in humans. Thus, it is timely to review the extant literature on the ontogeny of estrogen receptors in developing human male and female internal and external genitalia and to present new data filling in gaps in our knowledge.