A tale of two tracts: history, current advances, and future directions of research on sexual differentiation of reproductive tracts

Abstract

Alfred Jost's work in the 1940s laid the foundation of the current paradigm of sexual differentiation of reproductive tracts, which contends that testicular hormones drive the male patterning of reproductive tract system whereas the female phenotype arises by default. Once established, the sex-specific reproductive tracts undergo morphogenesis, giving rise to anatomically and functionally distinct tubular organs along the rostral–caudal axis. Impairment of sexual differentiation of reproductive tracts by genetic alteration and environmental exposure are the main causes of disorders of sex development, and infertility at adulthood. This review covers past and present work on sexual differentiation and morphogenesis of reproductive tracts, associated human disorders, and emerging technologies that have made impacts or could radically expand our knowledge in this field.

This review covers past and present work on sexual differentiation and morphogenesis of reproductive tracts, associated human disorders, and emerging technologies that have made impacts or could radically expand our knowledge in this field.

Introduction

Sexually dimorphic establishment of reproductive tracts is a critical step in developing reproductive tract organs anatomically and functionally distinct between male and female. Before sexual differentiation, an embryo possesses both female and male reproductive tract progenitors, which are also known as Müllerian and Wolffian ducts, respectively (Figure 1A). To establish sexual dimorphism of reproductive tract system, the embryo eliminates one of the two primitive ducts and maintains the other (Figure 1B). As a result, an embryo retains only one reproductive tract corresponding to its sex: Müllerian duct for the female and Wolffian duct for the male [1]. Müllerian ducts in the female eventually differentiate into the adult female reproductive tracts, which include oviduct, uterus, cervix, and the upper part of vagina (Figure 1C). These specialized components of the female reproductive tract create an environment that allows mating, fertilization, embryo transport, embryo implantation and development, and delivery of fetus. On the other hand, Wolffian ducts in the male develop into the adult male reproductive tract, which consists of epididymis, vas deferens, and seminal vesicle (Figure 1C). These male-specific organs facilitate the maturation, transport, storage, and delivery of the sperm. This review covers past and present work on sexual differentiation and morphogenesis of reproductive tracts, and related human developmental disorders. It also discusses unanswered questions and future opportunities brought forth by emerging technologies.

Figure 1.

Establishment of the sex-specific reproductive tract system. (A) Before sexual differentiation, the embryo possesses both primitive female and male reproductive tracts, which are known as Müllerian (represented in pink) and Wolffian ducts (in blue), respectively. (B) During sexual differentiation, the embryo eliminates one of the two primitive ducts and maintains the other. As a result, the XX embryo retains the Müllerian ducts and the XY embryo retains the Wolffian ducts. (C) Müllerian ducts in the XX embryo eventually differentiate into the adult female reproductive tracts, which include oviduct, uterus, cervix, and the upper part of vagina. On the other hand, Wolffian ducts in the XY embryo develop into the adult male reproductive tract, which consists of epididymis, vas deferens, and seminal vesicle.

Jost's paradigm: hormonal regulation of sexual differentiation

Jost's remarkable microsurgical experiments on rabbit embryos laid the foundation for our current knowledge of mammalian sexual differentiation [2, 3]. By removing the gonads of rabbit fetuses before the onset of sexual differentiation, Jost discovered that embryos without gonads, regardless of their genetic sex (XX or XY), acquired a female phenotype: regression of Wolffian ducts and maintenance of Müllerian ducts. He then grafted a fetal testis adjacent to the ovary of an XX embryo and found that the female pattern was reversed to the male pattern with the maintenance of Wolffian ducts and regression of Müllerian ducts. From these observations, Jost concluded that the female phenotypic development is a default process and masculinization of reproductive tract system is imposed by fetal testicular hormones.

At the time of Jost's experiment, testosterone was the only known hormone secreted by the testis for its ability to restore maleness in castrated animals and humans [4, 5]. By 1935, three research teams led by Adolf Butenandt, Kàroly Gyula, and Leopold Ruzicka, successfully synthesized the testicular hormone, ultimately named testosterone [6]. When Jost grafted a testosterone implant, instead of a fetal testis, to the XX embryo, Wolffian ducts were maintained but Müllerian ducts regression did not occur. This observation not only ascertained the function of testosterone on maintaining the Wolffian duct, but also implied that another testicular factor(s) must be responsible for inhibiting the survival of Müllerian ducts. Jost proposed the name “Müllerian inhibiting substance” or MIS for such factor, which is now known as anti-Müllerian hormone (AMH).

How do androgens promote the maintenance of Wolffian ducts?

In addition to Jost's experiments, the involvement of androgens on Wolffian duct maintenance was supported by various models of testicular feminization due to androgen insensitivity, including human [7], cattle [8], rat [9], and mouse [10]. Affected individuals were genetically XY that produced normal testicular androgens; however, they failed to be virilized. Genetic inheritance studies in mice revealed that testicular feminization was caused by an X-linked gene, which was named Tfm (testicular feminization mutation) [10]. After the cDNA of androgen receptor was cloned and mapped to the X chromosome in human and rat [11, 12], it became clear that the Tfm mouse was caused by a frameshift mutation in the androgen receptor gene, resulting in premature protein translation [13].

Androgen receptor (AR, or NR3C4) is a member of nuclear receptors with the conserved ligand binding domain and cysteine-rich DNA-binding domain [14]. Androgen receptors first appear in the Wolffian duct mesenchyme and later extend to both Wolffian duct epithelium and mesenchyme in mice and rat [15, 16]. However, in human embryos, androgen receptors seemed to be only expressed in Wolffian duct epithelium [17]. The transcriptional regulation of androgen receptor depends on androgen ligand activation. In the absence of androgens, androgen receptor is localized in the cytoplasm; upon binding to androgens, androgen receptor is translocated to the nucleus to regulate gene transcription [14]. The biosynthesis of androgens occurs in fetal Leydig cells [18]. The action of androgen receptor on Wolffian ducts must be imposed within a specific fetal programming window, which is E15.5–17.5 in rat and predicted to be approximately 8–12 weeks of gestation in human [18]. Inhibition of androgen receptor actions during this masculinization programming window disrupted stabilization and differentiation of Wolffian ducts [19–21].

The maintenance of Wolffian ducts is a product of androgen receptor action in the mesenchyme (Figure 2). Global knockout or loss-of-function mutation of androgen receptor resulted in failure to maintain Wolffian ducts in the XY mice [15, 22, 23]. However, ablation of androgen receptors only in the Wolffian duct epithelium had no impact on Wolffian duct maintenance, elongation, and coiling, supporting that action of androgen receptor occurs primarily in the Wolffian duct mesenchyme [15]. Mesenchymal cells of Wolffian duct also play instructing roles in Wolffian duct differentiation. When the epithelium from the upper Wolffian duct (future epididymis) was combined with the lower Wolffian duct mesenchyme (future seminal vesicle), the epithelium lost its epididymal identity and became seminal vesicle-like structures [24]. These results indicate that the fate and differentiation of Wolffian ducts are dictated by signals from the Wolffian duct mesenchyme.

Figure 2.

Mechanisms underlying the sex-specific establishment of the Wolffian duct. In the XX embryo, COUP-TFII in the Wolffian duct (or WD) mesenchyme suppresses Fgf expression and in the absence of FGF signaling and the Wolffian duct epithelium degenerates. On the other hand, in the XY embryo, testis-derived androgens activate androgen receptor (AR) and initiate the survival program via mesenchymal–epithelial crosstalk. The activated AR induces expression of growth factor EGF (epidermal growth factors) or other unknown factors in the mesenchyme. EGF binds its receptor EGFR in Wolffian duct epithelium, which presumably actives p-ERK and ETV4/5-mediated survival pathway in the epithelium. Additional AR action in the mesenchyme somehow antagonizes COUP-TFII’s inhibition of Fgf. As a result, FGF production is enhanced, which may promote Wolffian duct survival via the FGFR2/p-ERK pathway. Although EGFR and FGFR1 are expressed in the mesenchyme, it remains unclear about specific effects of EGF and FGF on mesenchymal cells during AR promoted Wolffian duct maintenance.

The mesenchymal androgen receptors control differentiation and survival of the Wolffian duct epithelium through induction of paracrine factors such as epidermal growth factor (EGF) (Figure 2). Androgen antagonist treatment to the XY mouse embryo decreased Egf expression [25], whereas androgen treatment to the XX embryos increased Egf and EGFR expression in the Wolffian duct [25, 26]. Additionally, in the absence of androgen, EGF treatment exerted androgenic effects on Wolffian duct maintenance and differentiation in organ culture and in vivo conditions [27]. Blocking EGF and EGFR action by their specific antibodies prevented androgen-induced effects on Wolffian ducts in organ culture [26, 27]. These results suggested that EGF/EGFR could be a downstream effector of androgen actions (Figure 2). Compound mutations in Egfr led to male infertility in mice but it is unknown whether Wolffian duct derivative tissues were abnormal in this mutant mouse [28]. In addition, Egf knockout mice had no abnormal phenotypes [29], suggesting that the EGF/EGFR pathway may be compensated by other growth factor signaling or alternatively dispensable for Wolffian duct development in vivo.

Fibroblast growth factor (FGF) is another candidate responsible for mediating mesenchymal AR actions (Figure 2). FGF7 and FGF10 are the major FGF ligands expressed in Wolffian duct mesenchyme [30–33]. Additions of FGF7, FGF10, or both were able to maintain Wolffian duct in the absence of testes in organ culture [32]. Among four receptors for FGFs in mammal, Fgfr2 is the major FGF receptor in Wolffian duct and its ablation led to caudal Wolffian duct regression [33]. While these observations implicate the involvement of FGF signaling in Wolffian duct maintenance (Figure 2), the connection between FGFs and androgen signaling remains to be determined.

Besides FGFs, growth hormone (GH), insulin-like growth factor (IGF1), and transformation growth factor beta 2 (TGFβ2) were implicated in Wolffian duct maintenance based on organ culture or ex vivo experiments. Both GH and its downstream mediator IGF1 were able to promote the maintenance of Wolffian ducts in the organ culture condition [34]. Igf1 knockout mice displayed underdevelopment of Wolffian duct derivative tissues including epididymis, seminal vesicle, and prostate. However, production of androgen, the driving hormone for Wolffian duct tissue development, was also reduced in Igf1 knockout males [35]. Therefore, it is unclear whether these hypoplastic Wolffian duct phenotypes in Igf1 knockout mice are the results of local IGF1 deficiency, androgen deficiency, or a combination of both. In the case of Tgfβ2, only one out of five examined Tgfβ2 knockout male mice had testis hypoplasia and vas deferens dysgenesis [36]. Further investigations are needed to affirm the physiological importance of these growth factors in Wolffian duct maintenance.

Although the maintenance of Wolffian duct during sexual differentiation is regulated through the action of mesenchyme, its survival prior to sexual differentiation depends on epithelial transcription factors, including homeobox gene Emx2 and paired box gene Pax2. In the absence of Emx2, although the Wolffian duct was initially formed, its structure degenerated in both sexes before the onset of sexual differentiation [37]. Likewise, loss of Pax2 caused degeneration of Wolffian duct and Müllerian ducts before the initiation of sexual differentiation despite normal-looking gonads, indicating an autonomous function of PAX2 in maintaining Wolffian ducts (as well as Müllerian ducts) [38]. These results implicate that maintenance of Wolffian duct epithelium is initially controlled by its own transcriptional programs until the onset of sexual differentiation, when androgen receptor signaling in the mesenchyme kicks in.

How does anti-Müllerian hormone induce the regression of Müllerian ducts?

Anti-Müllerian hormone belongs to the transforming growth factor beta (TGFβ) gene family with a C-terminal domain highly conserved with human TGFβ [39]. AMH is secreted specifically by Sertoli cells in both fetal and adult testes and granulosa cells of postnatal ovaries [40]. Purified AMH from incubation media of bovine fetal testicular tissue caused the elimination of Müllerian ducts in fetal XX rat mesonephros in culture [41, 42]. AMH cDNA and AMH gene in bovine and human were later cloned [39, 43]. Transfection of human AMH gene in Chinese hamster ovary cells led to AMH protein production, which induced Müllerian duct regression in organ culture assay [39]. Transgenic XX mouse embryos that express ectopic human AMH have no Müllerian duct derivative tissues [44], whereas Amh deficient XY mice retained female reproductive tract organs [45]. In humans, mutations in AMH are associated with persistent Müllerian duct syndrome in XY individuals [46]. The collective work unequivocally demonstrates the governing role of AMH in inducing Müllerian duct regression.

Being a member of the TGFβ gene family, anti-Müllerian hormone signals via heterotetrameric complexes of type I and type II kinase receptors in which the type II receptor phosphorylates and activates the type I receptor upon binding of the ligand [47]. The signaling is then transduced to the nucleus by members of the receptor-activated (R)-Smad family, which form trimeric complexes with the common mediator SMAD4 to regulate specific gene expression (Figure 3) [47]. The cDNA of type II receptor for AMH (Amhr2) was cloned from Sertoli cell, fetal ovaries, and fetal rat reproductive tracts [48–50]. The expression of the type II receptor gene is localized to the Müllerian duct mesenchyme during embryogenesis, and Sertoli cells and granulosa cells in fetal and adult testes and ovaries [48–51]. Expression of Amhr2 is directly controlled by the transcription factor Wilm's tumor or WT1 in the mesenchyme [52] (Figure 3). The functional significance of AMHR2 in transducing AMH signal was validated by the genetic mouse model: Amhr2 knockout XY mice exhibited Müllerian duct maintenance phenotype indistinguishable from Amh knockout and Amh/Amhr2 double knockout male mice [53]. These observations support that the pathway activated by AMH consists of only the ligand AMH and one type II receptor AMHR2. This notion was corroborated by clinical findings that XY patients with AMH or AMHR2 gene mutations caused the similar phenotypes of persistent Müllerian ducts [54].

Figure 3.

Mechanisms underlying sex-specific establishment of the Müllerian duct. In both sexes, Müllerian duct (or MD) mesenchyme expresses AMHR2, which is regulated by Müllerian duct epithelium-derived WNT7A and mesenchymal transcriptional factor WT1. It is unknown which specific FZ receptor (s) in the mesenchyme mediates WNT7A signaling to control Amhr2 expression. In the XX embryo, due to a lack of AMH, AMHR2-mediated signaling in Müllerian duct regression is not activated and as a result, Müllerian duct epithelium survives. It is unclear whether any mesenchyme-derived survival factor(s) are present to promote Müllerian duct maintenance. On the other hand, in the XY embryo, testis-derived AMH engages with AMHR2 and ALK2/3. Upon binding of AMH, AMHRII phosphorylates and activates ALK2/3. The signaling is then transduced to the nucleus by members of the receptor-activated (R)-Smad family (SMAD1/5/8), which form trimeric complexes with the common mediator SMAD4 to regulate specific gene expression, including WNT ligands Wnt4 and Wnt5a, WNT inhibitors such as Wif1, the transcriptional factor Osx, and the matrix metalloproteinase Mmp2. By unknown mechanisms, these molecules induce apoptosis or epithelial–mesenchymal transition (EMT), two major processes in Müllerian duct regression. Additionally, in the mesenchyme, secreted WNT ligands induce CTNNB1 (β-catenin)-dependent signaling in the nucleus, where the stabilized β-catenin engages and/or regulates TCFs and LEF1. These β-catenin-activated transcriptional factors regulate Osx and other unknown factors that participate in the regression of Müllerian duct epithelium.

The type I receptors for AMH and downstream R-Smads play redundant roles in mediating AMH-induced Müllerian duct regression. Two putative type I receptors for AMH in reproductive tract differentiation were identified as Alk3/Bmpr1a and Alk2/Acvr1, which are specifically expressed in Müllerian duct mesenchyme (Figure 3) [55–57]. Alk2 and Alk3 ablation in Amhr2+ mesenchymal lineages led to Müllerian duct retention in 0 and 50% examined male mice, respectively. However, the 100% penetrant phenotype was observed in Alk2 and Alk3 double mutant males [57]. The action AMH on Müllerian duct regression was abolished only when all the three R-Smad effectors (SMAD1, SMAD5, or SMAD8) were ablated in the Müllerian duct mesenchyme [57]. Ablation of the R-Smad common mediator Smad4 in MD mesenchyme led to only partial Müllerian duct retention on one side or both sides in XY mice [58].

Besides R-Smad effectors, another putative downstream transcriptional effector of AMH signaling in inducing Müllerian duct regression is the zinc figure transcriptional factor Osterix (Osx/Sp7) (Figure 3) [59]. Osx/Sp7 is specifically expressed in Müllerian duct mesenchyme in XY embryos. Amhr2 ablation in XY mouse embryos abolished Osx/Sp7 expression while ectopic expression of human AMH in XX mouse embryos induced its expression in Müllerian duct mesenchyme, implicating Osx as the direct downstream effectors of AMHR2 signaling. Ablation of Osx in Müllerian duct mesenchyme only caused delayed Müllerian duct regression in XY embryos, suggesting that Osx mediates only a part of AMH downstream signaling in inducing Müllerian duct regression.

It remains unclear how the AMH/AMHR2 signaling in the mesenchyme promoted the demise of Müllerian duct epithelium. Matrix metalloproteinases (MMPs) are calcium-dependent zinc-containing endopeptidases, which were hypothesized to promote apoptosis in Müllerian epithelial cells either by cleaving substrates on the epithelial cells or by modifying survival or death-inducing factors derived from the mesenchyme [60]. Among MMPs, Mmp2 was of great interest. Mmp2 expression was upregulated in male Müllerian duct mesenchyme and abolished in the absence of AMH, indicating Mmp2 was a potential downstream target of AMH signaling. However, Mmp2 expression was not regulated by AMH downstream transcriptional factor Osx based on the observation that Osx ablation in Müllerian duct mesenchyme did not change Mmp2 expression [59]. In ex vivo organ culture conditions, mature MMP2 led to increased incidence of Müllerian duct regression in the XX mouse embryos while Mmp2 knockdown partially blocked Müllerian duct regression in XY mouse embryos [60]. However, Mmp2 knockout mice appeared overall normal [61], suggesting that other Mmp members might compensate for the loss of Mmp2.

Crosstalk between AMH/AMHR2 and WNT signaling in regulating the fate of Müllerian ducts

The regression and maintenance of Müllerian duct during sexual differentiation is also dependent on WNT signaling, which interact with the AMH/AMHR2 pathway. WNT signal transduction cascade includes a WNT ligand (at least 19 Wnt genes were found in most mammalian including human and mouse), a heterodimeric receptor complex consisting of a seven-transmembrane receptor Frizzled (Fz), and a LRP5/6 coreceptor protein [62]. The WNT ligand Wnt7a is specifically expressed in Müllerian duct epithelium and required for Amhr2 expression in the mesenchyme (Figure 3). Wnt7a knockout XY mice had persistent Müllerian ducts due to the loss of Amhr2 in Müllerian duct mesenchyme [63].

In addition to Wnt7a, multiple WNT ligands and WNT ligand inhibitors have been suggested to involve in Müllerian duct fate decisions (Figure 3). Wnt4 is a downstream factor of AMH signaling and specifically expressed in Müllerian duct mesenchyme of XY embryos but not of XX embryos [64]. Wnt5a is another ligand expressed in the mesenchyme of developing murine Müllerian duct [65]. However, loss of Wnt4 [64] or Wnt5a [66] in Müllerian duct mesenchyme had normal Müllerian duct regression in XY embryos. Wnt4 and Wnt5a double ablation in Müllerian duct mesenchyme caused endometrial gland development and causes partial Müllerian agenesis in female mice but it was not reported whether Müllerian duct was persistent in the XY mice [67].

Secreted WNT inhibitors could participate in regulating the fate of the Müllerian duct (Figure 3). A WNT inhibitory factor Wif1 was identified as an AMH-regulated gene in Müllerian duct mesenchyme [68]. Knockdown Wif1 in organ culture condition led to partial or complete Müllerian duct maintenance in XY [68]. The secreted frizzled-related proteins (SFRPs) are a family of soluble proteins that are structurally related to frizzled (Fz) proteins. Because of their homology with the WNT-binding domain on the Fz receptors, SFRPs are able to inhibit WNT signaling [69]. Sfrp1 was expressed in the mesenchyme of both Müllerian duct and Wolffian duct [70]; Sfrp2 and Sfrp5 exhibited sexually dimorphic expression in Müllerian duct prior to the onset of regression [71]. These expression patterns suggest their potential roles in sexual differentiation of Müllerian ducts. However, knockout Sfrp1 and Sfrp2 caused a slight delay in Müllerian duct regression [70] and ENU-induced null mutations in Sfrp2 and Sfrp5 did not cause abnormalities in reproductive tract development at all [71]. These observations indicate that these WNT inhibitors could play redundant roles in regulating WNT signaling in Müllerian duct regression (Figure 3).

While essential roles of WNT ligands in regulating the fate of Müllerian duct are well established, receptor(s) for these WNT ligands have not been identified (Figure 3). There are 10 mammalian Fz receptors for WNT binding. So far Fz1 was the only receptor detected in Müllerian duct tissues by both RT-PCR and in situ hybridization. Fz1 was specifically expressed in Müllerian duct epithelium and its mesenchyme [72]. However, Fz1 knockout males were fertile, which indicates that they may have normal Müllerian duct regression [73]. Fz9 expression pattern in mesonephros was not reported during sexual differentiation [74] and Fz9 knockout had no Müllerian duct-related phenotypes [75]. Fz10 is specifically expressed in Müllerian duct epithelium [76], and its loss does not lead to Müllerian duct phenotype in mice (MGI: 3604450). Fz receptors may play redundant roles in mediating WNT signaling in the Müllerian ducts. Other than the Fz receptors, the glycosylphosphatidylinositol-anchored protein RECK was suggested to be a selective WNT7a receptor, and its binding with G protein-coupled receptor GPR124 modulated WNT7A-induced signaling in rat brain [77]. Reck is expressed in mesonephric mesenchyme during sexual differentiation of reproductive tract in mouse based on GUDMAP online database [78, 79]. These observations suggested a potential role of Reck in mediating WNT7a signaling in Müllerian duct differentiation. Since Reck knockout mice died at E10.5 [77], conditional ablation of Reck in Müllerian duct mesenchyme using Amhr2-Cre could shed light onto its potential role in Müllerian duct regression.

Upon activating membrane receptors, WNT ligands induce three major intracellular downstream pathways: the canonical pathway induced by β-catenin stabilization and two noncanonical β-catenin independent pathways, the WNT/Ca²⁺ and planar cell polarity (PCP) pathways [80]. Involvement of noncanonical β-catenin independent pathways in Müllerian duct regression has not been reported. On the other hand, the function of the β-catenin-dependent pathway in Müllerian duct fate decisions has been well documented (Figure 3). In response to the canonical WNT ligands, ligand-activated Fz receptors phosphorylate LRP, which breaks down the destruction complex and stabilizes β-catenin. The stabilized β-catenin accumulates and enters the nucleus where it engages DNA-bound TCF/LEF family transcription factors to drive target gene expression (Figure 3) [62]. Ablation of β-catenin in Müllerian duct mesenchyme led to Müllerian duct maintenance in XY mouse embryos [64]. In the absence of β-catenin, expression of the WNT target molecule, LEF1, was abolished in Müllerian duct mesenchyme [64]. Therefore, in the normal condition, β-catenin could regulate and engage LEF1 for target gene expression to induce Müllerian duct regression (Figure 3) [64]. However, Lef1 null XY mice had normal Müllerian duct regression [81], suggesting that action of β-catenin may be mediated by other TCF/LEF family transcriptional factors (Figure 3). Counterintuitively, when β-catenin was constitutively activated in Müllerian duct mesenchyme, focal maintenance of Müllerian duct in XY mouse embryos was observed [82]. These observations indicate that a properly tuned activity of β-catenin is critical for Müllerian duct regression.

The WNT/β-catenin and AMH/AMHR2 pathways interact and converge on β-catenin intracellularly in regulating the Müllerian duct fate. Nuclear staining of β-catenin was promoted by AMH signaling in Müllerian duct mesenchyme of XY embryos [83]. In the absence of AMH downstream effectors Smads, β-catenin expression was locally reduced (Figure 3) [58]. These observations indicate that AMH/AMHR2/SMAD modulates β-catenin expression and/or stabilization [64]. On the other hand, specific ablation of β-catenin in the Müllerian duct mesenchyme reduced expression of the AMH/AMHR2 downstream effector Osx [59]. However, β-catenin ablation did not affect the expression of Amhr2 and the AMH downstream signaling Wnt4 [64]. Another AMH downstream signaling Mmp2 [60] is a well-known target of Wnt/β-catenin signaling in cancer cells [84] but whether Wnt/β-catenin signaling regulates Mmp2 in Müllerian duct regression is unknown. Therefore, β-catenin as the core player of the canonical WNT pathway modulates partial AMH/AMHR2 signaling pathway for Müllerian duct regression (Figure 3).

Revisiting Jost's paradigm: hormone-independent sexual dimorphism of urogenital traits

While Jost's paradigm holds true in both animal studies and clinical cases, emerging evidence prompts some reconsideration of this dogma. The Jost paradigm contends that the female pattern of reproductive system occurs by default as the presence of Müllerian duct and absence of Wolffian duct being the product of a lack of fetal testes and their hormone-product androgens and AMH. However, the discovery of an unexpected role of a mesenchymal transcription factor COUP-TFII (chicken ovalbumin upstream promoter transcription factor II or NR2F2) added new twists and insights into this process. When Coup-tfII was inactivated in the Wolffian duct mesenchyme in the XX embryos, Wolffian ducts were maintained despite the absence of androgen and its action [32]. As a consequence of Coup-tfII inactivation, Fgf 7 and Fgf10 expression was increased along with the activation of its receptor (FGFR2) and downstream pERK-mediated cell survival signaling in the Wolffian duct epithelium, leading to the maintenance of Wolffian duct. These findings imply that in the androgen-free female embryo, the Coup-tfII/Fgf/pERK pathway inhibits Wolffian duct survival (Figure 2) and COUP-TFII serves as a “Wolffian inhibiting factor.” Being a member of the orphan nuclear receptor family, COUP-TFII is not known to be activated by endogenous ligands. It was reported that COUP-TFII can be activated by retinoic acid (RA) in multiple cell lines [85]. During sexual differentiation, RA is produced at high levels in the mesonephros where the Wolffian ducts develop, suggesting that RA could potentially activate COUP-TFII [86].

Hormone-independent establishment of reproductive traits, similar to the maintenance of Wolffian ducts without androgen action in the Coup-tfII knockout XX mouse embryos, is observed in other occasions. For example, female spotted hyenas and elephants develop male-type external genitalia prior to gonadal differentiation, suggesting that the masculinization of external genitalia in these two species is not regulated by fetal testes [87]. In wallabies, sexual dimorphisms are already present before differentiation of the gonads, with the scrotum developing in males and the pouch and mammary glands in females. The sexually dimorphic development of these tissues is believed to be controlled by yet-to-identified gene(s) on the X chromosome [88]. These cases of hormone-independent sexual dimorphisms argue against the simplicity and generality of Jost's paradigm, which holds true in majority of cases but need to be refined in specific contexts.

Molecular pathways that direct rostral-caudal patterning of the reproductive tract

Once sexual dimorphism of reproductive tract system is established, the stabilized Wolffian duct and Müllerian duct differentiate to morphologically and functionally distinct tubular organs in XY and XX embryos. From rostral to caudal regions, Müllerian ducts eventually differentiate into oviduct, uterus, cervix, and the upper part of vagina in XX (Figures 1 and 4); on the other hand, Wolffian ducts in XY develop into epididymis (consisting of four major regions: initial segment, caput, corpus, and cauda), vas deferens, and seminal vesicle (Figures 1 and 4). Despite the fact that maintenance and regression of two reproductive tracts is governed by testicular hormones, rostral-caudal patterning of two reproductive tracts is less dependent on gonadal hormones and their signaling pathways, and becomes predominantly regulated by epithelial–mesenchymal interactions.

Figure 4.

Putative morphogenetic factors and their expression in rostral-caudal patterning of the reproductive tract. During rostral-caudal patterning, the male reproductive tract differentiates into epididymis (consisting of four major regions: initial segment, caput, corpus, and cauda), vas deferens, and seminal vesicle (A); the female reproductive tract gives rise to oviduct, uterus, cervix, and the upper vagina (B). Mesenchymal–epithelial interaction is the core mechanism in establishing rostral-caudal patterning of the two reproductive tracts. The model depicts expression localization of mesenchymal factors (marked in purple rectangles) and epithelial factors (in green rectangles) along the rostral-caudal axis. Inverted triangle indicates gradient expression from rostral to caudal region. Hox, homeobox; Inhba, inhibin beta A; Wnt, wingless/integrated; Bmp, bone morphogenetic protein; Ar, androgen receptor; Pkd1, polycystic kidney disease 1; components in RTK signaling, receptor tyrosine kinase) [Ros1 (Ros proto-oncogene 1, also known as cRos, an orphan receptor tyrosine kinase), Shp1 (Src homology region 2 domain-containing phosphatase-1), ERK (also known as MAPK, mitogen-activated protein kinase), Dusp6 (dual specificity phosphatase 6) and Pten (Phosphatase and tensin homolog)], p63, tumor protein p63; RA, retinoic acid; PCP proteins, planar cell polarity; Vangl2, Vang-like protein 2; and Celsr1, cadherin EGF LAG seven-pass G-type receptor 1.

Epithelial–mesenchymal interaction is the fundamental mechanism in patterning tubular structures [89]. During organogenesis, regional specification of the epithelium depends upon the identity of the adjacent mesenchyme. The critical role of the mesenchyme in the epithelial differentiation of male and female reproductive tracts has been established by tissues recombination studies [24, 90, 91], where mesenchyme of a specific origin and epithelium of a different origin were recombined and cultured to determine whether the fate of the epithelium can be altered. For example, when cultured with uterine mesenchyme, vaginal epithelium differentiates into uterine epithelium. In the reciprocal tissue recombinants, vaginal mesenchyme transformed the uterine epithelium into a stratified squamous vaginal epithelium [90]. Similar recombination experiments using Wolffian duct tissues have been conducted wherein the upper Wolffian duct (future epididymis) was combined with the lower Wolffian duct mesenchyme (future seminal vesicle). The epithelium of the upper Wolffian duct lost its epididymal identity and became seminal vesicle-like structures in response to induction by lower Wolffian duct mesenchyme [24]. Therefore, during the reproductive tract patterning, mesenchyme is probably differentiated first and then the preceding specified mesenchyme induces epithelial region-specific phenotype. Despite the critical roles of the mesenchyme in inducing epithelial differentiation, the interaction between the mesenchyme and epithelium is not a one-way street. Epithelium-derived signaling molecules are important for proper mesenchymal differentiation as well (see the details in next two sections). Although epithelial–mesenchymal interaction is the common mechanism underlying patterning of two reproductive tracts, the detailed pathways in patterning the two tubular organs are not exactly the same.

Patterning of the male reproductive tract

Mesenchymal differentiation by Homeobox genes along rostro-caudal axis is at the core of Wolffian duct patterning (Figure 4A). Hoxa9 and Hoxd9 are expressed in the epididymis and vas deferens (Figure 4A) [92]. Approximately half of the Hoxa9/Hoxd9 double knockout mice were fertile in these mice [93], suggesting that Wolffian duct development was probably normal and loss of Hoxa9 and Hoxd9 might be compensated by other Hox genes. The Hoxa10 expression is mainly in the mesenchymal compartments of caudal epididymis throughout the distal male reproductive tract during morphogenesis (Figure 4A) [94]. Hoxa10 mutant males had anterior transformation of the cauda epididymis and the proximal vas deferens, diminished stromal clefting of the seminal vesicles, and decreased size and branching of the coagulating gland [94, 95]. Hoxa11 is expressed in the mesenchyme of fetal Wolffian duct and adult vas deferens (Figure 4A). Either individual Hoxa11 knockout or Hoxa11 and Hoxd11 (a paralog to Hoxa11) double knockout caused malformations of vas deferens that resembled an epididymis [96, 97]. Hoxa13 and Hoxd13 are expressed in the distal part of Wolffian duct tissues (Figure 4A) [98]. Disruption of Hoxd13 alone led to hypoplastic seminal vesicles, which become more severe in Hoxa13^+/– and Hoxd13^–/– compound mutant mice [99]. In addition to Hox9–13, comparing gene transcriptional profiles of different Wolffian duct parts within the morphogenesis window has identified additional region-specific Hox transcripts, such as Hoxc4 and Hoxc9 in epididymis [100]. Specific functions of these newly identified Hox genes in Wolffian duct differentiation remains to be studied. The importance of Hox genes in determining the rostral-cauda patterning of the Wolffian duct is well recognized; nevertheless, it is not known how these Hox transcription factors induce morphology and phenotype variation in different parts of the Wolffian duct.

The specified mesenchyme provides signaling inputs via paracrine signaling to instruct epithelial differentiation in the male reproductive tract. Inhibin beta A (Inhba), a component of inhibins and activins (members of TGFβ superfamily ligand), is highly expressed in the mesenchyme of the anterior Wolffian duct and diminishes posteriorly before the onset of Wolffian duct morphogenesis (Figure 4A). Inhba mutant mice failed to develop epididymal coiling due to a dramatic decrease in epithelial proliferation [101], indicating its critical function in inducing the highly convoluted epididymal structure. In addition to inhibin beta A, BMPs and WNTs are other putative mesenchymal factors that control epididymal patterning: the epididymal epithelium failed to coil in the absence of Pkd1, a large structural membrane spanning glycoprotein involved BMP/Tgfβ signaling transduction [102], or β-catenin [103](Figure 4A).

Although epithelial differentiation is regulated by paracrine signaling from the mesenchyme, signaling pathways innated to the epithelium are also necessary to achieve its proper and complete differentiation. Androgen receptor in the Wolffian duct epithelium controls differentiation of basal cells in the epididymal epithelium by regulating p63, a driver for basal cell differentiation (Figure 4A) [15]. Although androgen receptor is expressed in other parts of Wolffian duct derivative tissues, androgen receptor regulation of basal cell differentiation is specific to the epididymis. The specific function of androgen receptor in the epididymis could be achieved by the pioneering factor AP-2α (activating enhancer binding protein 2α), which guides AR to specific genomic loci for epididymal gene regulation [104]. However, AP-2α knockout mice do not seem to have any reproductive defects (MGI: 104671), indicating that AP-2α can be compensated by other pioneering factor(s) or is dispensable for androgen receptor action in inducing basal cell differentiation in the epididymis.

Additionally, receptor tyrosine kinase (RTK) signaling in the epithelium also regulates establishment and/or maintenance of epithelial identity in the epididymis, espeically the initial segment. Multiple receptor-associated components in RTK pathways are highly expressed in the epithelium of initial segment of the mouse epididymis (Figure 4A), including an orphan receptor tyrosine kinase Ros1 (also known as c-Ros) [105], its negative regulator Shp1 [106], and fibroblast growth factor receptor substrate 2 (Frs2) [107]. Both Ros1 mutation and a naturally occuring mutation in Shp1 caused the complete loss of the regional characteristics of the initial segment. The proximal epididymal epithelium lacking Frs2 gave rise to abnormal shape of promixal epididymis with a range of variety in a subgroup of mice (∼10% examined animals) [107]. In addition, downstream pathways of RTK, such as MAPK/ERK and PIP3/AKT pathways, exhibited higher activity in the initial segment, including p-ERK1/2 and Src (components in ERK pathways), Dusp6 (a negative regulator of MAPK/ERK), and Pten (a negative regulator of PIP/AKT signaling) [108]. p-ERK1/2 and Dusp6 were critical for the proliferation of the initial segment and the caput & corpus regions, respectively [108]. Pten is required to maintain the epithelial fate in the initial segement because postnatal ablation of Pten (from postnatal P17 onward) induced dedifferentiation of the initial segment [109]. Even though multiple RTK pathways are well characeterized in regulating differentiaton or/and maintenance of the initial segment, specific factor(s) that engage and activate tyrosine kinase receptors has yet to be discovered.

Patterning the female reproductive tract

Similar to that in the male, region-specific Hox genes determine mesenchymal differentiation during the female reproductive tract patterning as well. Comparable to Hox gene expression along the anterior-posterior axis of the male reproductive tract, expression of Hoxa9, Hoxa10, Hoxa11, and Hoxa13 in the female reproductive tract is restricted to the oviduct, the uterus, the posterior uterus and cervix, and the cervix and upper vagina, respectively (Figure 4B) [110]. Genetic polymorphisms in HOXA9 have been linked to Müllerian duct abnormalities in humans [111] even though Hoxa9 knockout mice did not have any defects in urogenital organs (MGI: 96180). Hoxa10 and Hoxa11 are expressed in the mesenchyme of Müllerian ducts region that eventually becomes uterus. Both Hoxa10 and Hoxa11 null mutation led to partial morphological shift of the uterus towards the appearance of the rostral oviducts [95, 96]. Absence of Hoxa13 led to agenesis of cauda Müllerian ducts [99, 112]. Hoxa13^+/− and Hoxd13^−/- (a paralog to Hoxa13) compound mutations led to homeotic transformation of cervix to uterus [99].

Besides Hox genes, RA signaling regulates mesenchymal differentiation of Müllerian duct. Retinoic acid signaling is transduced by the two families of RA nuclear receptors (RARs and RXRs) during development. High expression of RA producing enzymes and transactivation activity was detected in the proximal Müllerian ducts that develop into the uterus (Figure 4B) [113]. In organ culture of undifferentiated Müllerian ducts, RA treatment induced uterus-specific Hox genes and uterine stromal differentiation, whereas inhibition of RA receptor signaling induced vaginal stromal differentiation [113]. These in vitro culture studies indicate that RA signaling determines the border between uterine and vaginal mesenchyme differentiation [113]. Ablation of RA receptors in mice would help us further elucidate significant roles of RA signaling in Müllerian duct mesenchymal differentiation in vivo. However, RAR mutations and ablation of RAR with RXR in mice led to differential Müllerian duct agenesis [114, 115], preventing the investigation of RA signaling in Müllerian duct patterning in vivo.

Mesenchymal regulation of epithelial differentiation is mediated in part by WNT signaling. Mesenchyme-derived Wnt5a is required for patterning of multiple regions in the female reproductive tract (Figure 4B). Wnt5a null mutation caused short and coiled uterus with no glands and the loss of defined cervical/vaginal structures [66]. Another Wnt ligand of note is Wnt4, which is expressed in uterine mesenchyme at birth prior to Müllerian duct morphogenesis (Figure 4B) [65]. Postnatal ablation of uterine Wnt4 led to transformation of the columnar epithelium to a stratified epithelial layer, a characteristic of the cervix and vagina [116], suggesting that mesenchyme-derived Wnt4 regulates uterine epithelial fate.

While it is well known that mesenchyme-derived WNT signaling plays significant roles in regulating epithelial differentiation, how the WNT signaling exerts their actions in the epithelium has yet to be fully understood. WNT ligands signal through three major well-characterized intracellular pathways: a canonical β-catenin-dependent pathway and two non-canonical β-catenin independent pathways, the WNT/Ca²⁺ and PCP pathways [80]. It is noteworthy that PCP proteins are expressed in Müllerian duct epithelium during morphogenesis and PCP gene mutations led to impaired Müllerian duct patterning (Figure 4B) [117, 118]. The WNT/PCP pathway regulates cell motility and polarity through cytoskeletal changes [119]. VANGL2 is an archetypal WNT/PCP protein, and a point mutation in Vangl2 led to dislocalized VANGL2 and in turn altered epithelial polarity. The female reproductive tract with mutant Vangl2 had septate vaginas, uncoiled oviduct, hyperplastic myometrium, and hypoplastic stromal mesenchyme in the uterus [117]. Besides, loss of a VANGL2-associated PCP gene Celsr1 (cadherin EGF LAG seven-pass G-type receptor) resulted in discontinuities of oviduct and uterus, aberrant epithelial folding, shape and arrangement in the oviduct [118]. Undoubtedly, the noncanonical WNT/PCP pathway is critical for appropriate establishment of epithelial polarity in Müllerian duct patterning. Phenotypes of PCP mutant mice are dissimilar to Wnt4 or Wnt5a null mice, suggesting that mesenchymal WNT signaling via WNT4 or WNT5a is probably not transduced through the PCP intracellular pathway in the epithelium.

Besides WNT, FGF7/10, BMP4, and Activin A from the mesenchyme also regulate epithelial patterning, particularly epithelial stratification in the cervix and vagina (Figure 4B). These mesenchymal factors signal through their respective receptors and intracellular pathways FGF/MAPK, BMP/Smad4, and Activin A/RUNX1, respectively, to cooperatively induce cervical and vaginal stratification [120]. Disruption in either of these pathways converts vaginal to uterine epithelial cell fate by inducing P63 expression [120]. P63 is the transcriptional factor that induces the epithelial stratification program [121]. Cervical and vaginal epithelia lacking p63 differentiated into columnar, uterine-like epithelium [122],

Although epithelial differentiation is regulated by mesenchymal signals, interaction between these compartments in the female reproductive tract is a two-way street. The epithelium of Müllerian ducts expresses Wnt7a, which becomes more restricted to oviductal and uterine parts before morphogenesis (Figure 4B) [65]. Wnt7a null mutation in mice caused posterior transformation of oviduct to uterus and uterus to vagina [123], implicating that epithelium secreted Wnt7a is critical for mesenchymal differentiation of the oviduct and uterus. Absence of β-catenin in the mesenchyme of the female reproductive tract do not recapitulate Wnt7a phenotype [72, 124], indicating that the action of WNT7A on mesenchymal differentiation is mediated by an unknown β-catenin independent mechanism.

Disorders of sexual differentiation of reproductive tracts in humans

Sexually dimorphic establishment of reproductive tracts can be perturbed by genetic alteration and environmental exposure, leading to human disorders that feature absence and persistence of reproductive tracts. Phenotypes of these human disorders are consistent with Jost's paradigm.

Androgen insensitivity syndrome (AIS) is a disorder of sexual differentiation due to inactivating mutations in androgen receptor gene that confer a spectrum of androgen resistance in the XY patients. AIS is clinically subclassified into three categories: (1) complete form (CAIS) that exhibits normal female external genitalia and a complete absence of Wolffian duct-derived organs, (2) mild form (MAIS), which present normal male external genitalia but may show gynecomastia and impaired spermatogenesis, and (3) partial form (PAIS) that presents phenotypes between CAIS and MAIS (please see reviews [125–127]). Fetal exposure to endocrine disrupting chemicals in humans is associated with phenotypes similar to androgen insensitivity syndromes [128, 129].

Persistent Müllerian duct syndrome (PMDS) is defined as the presence of ectopic female reproductive tract organs, including uterine and fallopian duct tissue, in XY patients that are normally virilized [46]. In these XY patients, undescended testes and abnormalities in male excretory ducts are frequent, and infertility is the most common complication of PMDS. Mutations in AMH/AMHRII were found in most PMDS patients [46]. In patients where no mutation in AMH/AMHRII were found, mutations in distal promoter or intron of AMH/AMHRII, or other alterations in genetic pathways implicated in Müllerian duct development, were thought to be responsible for these idiopathic PMDS. Prenatal exposure to a synthetic estrogen diethylstilbestrol increased the incidence of presence of Müllerian duct remnants in human [130], suggesting that estrogenic environmental chemicals are a major risk factor that contribute to persistent Müllerian duct syndrome. PMDS was also associated with prostate cancers in two clinical cases [131, 132]. In mice, residual Müllerian duct mesenchyme contributes to prostate tissues, and Lkb1 deletion in Müllerian duct mesenchyme led to the formation of benign prostatic hyperplasia [133]. This finding in mice raises the possibility that prostate tumors in reported PMDS patients may be of Müllerian duct mesenchyme origin.

Mayer-Rokitansky-Küster-Hauser (MRKH) syndrome is a disorder that affects 1 in 4500 females, and is characterized by the absence of Müllerian duct structures [134]. It can occur independently or coupled with various abnormalities. Whole-exome sequencing indicated that loss-of-function variants in OR4M2 (olfactory receptor, family 4, subfamily M, member 2) and PDE11A (phosphodiesterase 11A), nonsynonymous variants in LRP10 (LDL receptor related protein 10) and DOCK4 (Dedicator of cytokinesis 4), as well as deletions at 16p11.2, 15q11.2, 19q13.31, 1p36.21, and 1q44 were associated with MRKH syndromes [135, 136]. Exposure of rat fetuses to phthalate can cause uterine agenesis, a characteristic phenotype in MRKH syndrome [137]. However, in humans, exact contributions of nongenetic factors to the pathogenesis of MRKH are unclear [138].

Gartner's ducts are the vestigial remnants of Wolffian ducts in the vaginal regions of XX patients [139]. The presence of Gartner ducts is usually clinically asymptomatic, but they can grow large enough to form symptomatic cysts. The Gartner's duct cyst can mimic pelvic organ prolapse because both syndromes may exhibit a cystic mass posterior/anterior to vagina [140, 141]. Malignant transformation of Gartner cysts is very rare but a case was reported in the literature, which was treated successfully with surgical excision and radiation therapy [142]. In utero exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin at the time of sexual differentiation led to persistent Wolffian duct remnants in the rat vagina [143]. However, the association of dioxin exposure with Wolffian duct maintenance has not been reported in humans.

Remaining questions and future opportunities presented by emerging technologies

Visualizing the tridimensional cellular and morphological processes during sexual differentiation and morphogenesis of reproductive tracts

Generating a precise cellular and molecular cartography of reproductive tract system is critical for the understanding of the mechanisms of its organogenesis. Tridimensional imaging techniques will help uncover novel cellular and morphological events that cannot be visualized in histological and 2D microscopic analyses. For example, combination of whole mount immunostaining, solvent-based clearing, and light-sheet imaging has successfully built a 3D cellular map of the human development during the first trimester of gestation [144]. The 3D imaging of human reproductive tract system during sexual differentiation windows shed some interesting insights [144]. In contrast to Wolffian ducts that were vascularized in both sexes, the 3D imaging technique revealed sexually dimorphic vascularization surrounding the Müllerian ducts, which was unsheathed by vascular network in females but avascular in males [144]. This observation suggested that lack of vascularization may facilitate Müllerian duct regression in human males. Confocal imaging in combination with 3D analysis was also successfully applied to identify and quantify morphological changes in the luminal structure of murine uterus during embryo implantation, discovering a stereotypical reorientation of glandular ducts towards the site of implantation [145]. In addition to confocal imaging, optical coherence tomography is another powerful imaging technique, which penetrates more deeply into the sample to capture micrometer-resolution 2D or 3D images within live biological tissues. Given that this imaging method has successfully created high-resolution 3D in vivo imaging of mouse oviduct and visualize cellular events in oviducts [146–148], it can be applied to obtain live images of developing reproductive tract organs at fetal and postnatal stages. Hopefully, utilization of these 3D imaging approaches will create a molecular, cellular, and morphological atlas of reproductive tract organs, which will be of paramount importance to understanding its organogenesis in normal and pathological conditions.

Deciphering the molecular mechanisms underlying human reproductive tract differentiation

Although anatomical, histological, and microscopic analyses of human fetal tissues and the genetics of developmental disorders in reproductive tracts have helped us understand how human reproductive tracts form and develop [81, 149–151], knowledge about molecular mechanisms underlying this organogenesis in human is still limited. One major challenge is that we lack an in vitro system, which does not rely on materials from human fetuses, in order to model in vivo human organogenesis. Tridimensional organoid culture is a technique that can overcome this challenge. Organoids are progenitor cell-derived or stem cell-derived 3D structures that, on much smaller scales, re-create important aspects of the 3D anatomy and multicellular repertoire of their physiological counterparts and that can recapitulate basic tissue-level functions [152]. Endometrial and oviductal organoids deriving from adult mouse and human adult tissues have been successfully established [153–156]. In both species, continuous growth and differentiation of uterine and oviductal organoids depend on WNT and Notch signaling, respectively, suggesting conserved roles of these two pathways in female reproductive tract differentiation in mouse and human [153–156]. In addition to progenitor cells in adult tissues, organoid culture can derive from human-induced pluripotent stem cells (iPSCs). Based on results from studying in vivo mouse or human organ development, researchers have successfully directed differentiation of human iPSCs through intermediate mesoderm, coelomic epithelium into fallopian tube organoid [157], or uterine endometrial stromal fibroblasts [158]. By introducing specific signaling cues or alterations in the process of directed differentiation, we can investigate functional significances of these signaling factors underlying fetal human reproductive tract differentiation that otherwise cannot be studied due to tissue inaccessibility.

Determining the roles of epigenetic/chromatin-based mechanisms in sexual differentiation and morphogenesis of reproductive tracts

Signaling pathways and transcriptional regulation in reproductive tract development are well characterized (see the reviews [81, 151, 159]). It is now widely accepted that chromatin accessibility and its spatial 3D architecture play a crucial role in transcriptional regulation and subsequent cellular differentiation. Specialized packaging (or “closing”) of the chromatins sequesters inactive genomic regions while its “opening” exposes genomic active regions (promoter, enhancers and other regulatory elements) accessible to transcription machinery [160]. The spatial organization of chromosomes arrangement facilitates communication between genes and their regulatory elements, which can be located elsewhere along the chromosome [161]. Experimental and computational methods in chromatin biology, such as ATAC-seq for assessing chromatin accessibility, and chromosome conformation capture (3C) and 3C-derived methods for examining spatial chromatin arrangements, have provided a rich trove of information about promoter, enhancer, other regulatory elements, and their interactions in gene expression [161, 162]. Genome-wide chromatin structure changes has been suggested to involve in Müllerian duct differentiation in chicken, where chromatins of Müllerian ducts at different stages show different affinity with an antibody against chick newborn oviducts [163]. This observation indicates that chromatin changes may determine the developmental programming of accurate genomic expression and hormone responsiveness in the sexual differentiation tissues. The established techniques in chromatin biology will be powerful tools to test this notion. In addition to probe chromatin-based mechanisms in normal condition, another application of techniques for chromatin analysis is to understand chromatin organization alteration in pathological conditions. As discussed in the previous section, chromosome deletions are known to be associated with certain human disorders of reproductive tract development. While loss of genes within deleted regions can result in pathogenesis, it remains unclear whether the loss of noncoding chromosomal regions can disorganize chromatin arrangement and thus cause misregulation of expression of critical genes localized outside the deletion regions.

Categorizing the heterogeneous populations of cells in reproductive tract organs

Understanding cellular heterogeneity, especially in the mesenchyme and epithelium, will help us understand how mesenchymal–epithelial interactions along the rostral-caudal axis differ and drive regionalization. A powerful tool for dissecting heterogeneity in tissues is single cell RNA-seq (scRNA-seq), which permits transcriptomic analysis of individual cells. Analyzing postnatal mouse uterus by scRNA-seq uncovered a novel subpopulation of early progenitor in the epithelium [164] and chemokine (C-X-C motif) receptor 4; chemokine (C-X-C motif) ligand 12 signaling axis in mesenchymal–epithelial interaction for uterine gland formation [165]. Additionally, scRNA-seq analysis can provide a panorama of the entire cell types/subtypes in addition to epithelium and mesenchyme, which is essential to fully categorize the building blocks of reproductive tracts. Cell heterogeneities of human female reproductive tract have been revealed in a scRNA-seq analysis of human uterus, where 11 transcriptionally distinct clusters of epithelia, 6 clusters in stromal cells, 5 in endothelial cells, 2 in smooth muscle cells, 2 myofibroblasts, and 6 in immune cells have been found (https://doi.org/10.1101/267849). Single-cell transcriptomic analysis of fetal reproductive tract tissues will enable us to understand cellular heterogeneities and provide a rich resource for investigating signaling regulation network, especially in mesenchymal–epithelial interactions.

Establishing the roadmap of cell lineage specification during reproductive tract differentiation

A cell lineage is the developmental history of a differentiated cell as traced back to its progenitor from which it arises. Resolving lineage relationships between cells in the developing reproductive tracts, and between an early progenitor and its descendants, will illuminate mechanisms underlying normal development. A popular method for tracing lineage in genetically accessible model organisms such as mouse is the fate-mapping technique, which leverages the expression of recombinase enzymes in a cell-specific or tissue-specific manner to activate the expression of a reporter [166]. As a result, the cell/tissue and their progeny will be labeled with the reporter for tracking their fate as they differentiate. This approach has been used to investigate lineage differentiation in female reproductive tract formation in two studies. One study found that labeled secretory cells in the oviduct had the capacity to self-renew and gave rise to another oviductal cell type, ciliated cells, implicating critical roles of secretory cells in epithelial homoeostasis [167]. The other study discovered that Müllerian duct epithelial cells were partly derived from Wnt4+ lineage [168], suggesting that multiple lineages contribute to Müllerian duct cells. Noticeably, simultaneous lineage tracing of large-scale number of cells has been developed to depict lineage trees in zebrafish [169] and mouse [170]. The main principle is to utilize barcodes that consist of multiple CRISPR/CAS9 target sites to label cells. The barcodes will progressively and stably accumulate unique mutations over multiple cellular divisions and targeted sequencing analyses of the patterns of shared mutations will elucidate lineage relationships of analyzed cells. If inheritable cellular barcodes are transcribed, scRNA-seq can be used to simultaneously read the lineage barcodes and provide transcriptome information of single cells. Combination of computational analysis of lineage barcodes and single-cell RNA-seq has enable organization of the taxonomy of cell types into lineage tree in zebrafish [171–173]. If eventually applicable to mouse organism, this systematic and simultaneous assay of lineage relationship and cell types will provide unparalleled information on reproductive tract development.

Conclusions

Like a prologue in a play, Jost's classic experiments in the 1940s trailblaze the new frontier the field of sexual differentiation. Since then, significant insights into the establishment of sexually dimorphic reproductive tract system were gained from ex vivo experiments, genetic animal models, and human clinical cases. With the technological advancements and multidisciplinary collaborations, the field is poised to connect epigenetic mechanisms, molecular pathways, cellular events, and tridimensional morphological changes underlying the process of reproductive tract morphogenesis. The yielded knowledge will enable us to not only decipher the fundamental process of dimorphic establishment of reproductive tracts, but also provide insights into how defects and diseases originate from impaired fetal development.

Notes

Edited by Dr. Romana Nowak, PhD, University of Illinois Urbana-Champaign

Author Biographical

Humphrey H-C Yao, Senior Principal Investigator, and Fei Zhao, NIH Intramural Postdoctoral Fellow, in the Reproductive Developmental Biology Group at NIEHS/NIH. Dr. Humphrey H-C Yao received his doctoral degree at the University of Illinois in Urbana-Champaign in 1999 under the guidance of Janice Bahr, and then completed his postdoctoral training in Blanche Capel's lab at Duke University Medical Center in 2002. He became Assistant Professor in the Department of Comparative Biosciences at University of Illinois in Urbana-Champaign in 2003 and received Associate Professor tenure in 2009. Dr. Yao was recruited to National Institute of Environmental Health Sciences (NIEHS/NIH) in 2010 and was promoted to Senior Investigator in 2018. The main thrust of Dr.Yao's group is to define the normal process of how gonads and reproductive tracts form during embryogenesis and investigate the impacts of environmental exposure to harmful chemicals on these process. Dr. Yao is the corresponding author of multiple major publications on sexual differentiation of gonads and reproductive tracts, including Archambeault and Yao, PNAS, 2010 and Zhao et al., Science, 2017.

Dr. Fei Zhao is the postdoctoral fellow in Dr. Yao's lab and the first author of their recent publication in Science. Dr. Zhao received his doctoral degree at the University of Georgia in 2014 under the guidance of Dr. Xiaoqin Ye, and joined Dr. Yao's lab in 2014. Dr. Zhao received multiple awards during his postdoctoral training, including K99/R00 award in 2018.