From Enrico Sertoli to freemartinism: the many phases of the master testis-determining cell^†

Abstract

Sertoli cells, first identified in the adult testis by Enrico Sertoli in the mid-nineteenth century, are known for their role in fostering male germ cell differentiation and production of mature sperm. It was not until the late twentieth century with the discovery of the testis-determining gene SRY that Sertoli cells’ new function as the master regulator of testis formation and maleness was unveiled. Fetal Sertoli cells facilitate the establishment of seminiferous cords, induce appearance of androgen-producing Leydig cells, and cause regression of the female reproductive tracts. Originally thought be a terminally differentiated cell type, adult Sertoli cells, at least in the mouse, retain their plasticity and ability to transdifferentiate into the ovarian counterpart, granulosa cells. In this review, we capture the many phases of Sertoli cell differentiation from their fate specification in fetal life to fate maintenance in adulthood. We also introduce the discovery of a new phase of fetal Sertoli cell differentiation via autocrine/paracrine factors with the freemartin characteristics. There remains much to learn about this intriguing cell type that lay the foundation for the maleness.

When Enrico Sertoli stared at the “cellule ramificate,” or branched cells, in a slice of human testicle under the microscope in 1865 [1], he was unaware that his legacy will live through the age of microscopy to the era of single-cell sequencing in the twenty-first century. Sertoli cells, the testis-specific cell type that was named after him, are not only the “mother” cells that nurse male germ cells but also the leading actors that usher in the pathway toward maleness. In the adult testis, Sertoli cells attach to the basement membrane of seminiferous tubules and extend their cell body toward the lumen while enveloping all stages of male germ cells. A critical component of the hypothalamic–pituitary–gonadal axis, adult Sertoli cells acquire their unique features and ability to provide the niche for spermatogenesis through their response to pituitary hormones and local factors [2]. Adult Sertoli cells facilitate an orderly progression of spermatid morphogenesis and sperm production, and at the same time, create a barrier that allows the meiotic and haploid male germ cells to escape from the surveillance of the immune system. Adult Sertoli cells were considered a terminally differentiated cell type; however, they were later found to retain the plasticity to decommit to the male identity and transdifferentiate into granulosa cells, their female counterparts in the ovary [3]. In the early 1990s, the identification of the testis-determining gene SRY (Sex-determining region of the Y chromosome) further revealed the formative role of fetal Sertoli cells in testis morphogenesis [4–6]. Sertoli cells, the only cell type in the fetal testis that produce SRY, orchestrate formation of seminiferous tubules, establishment of testis-specific vasculature, appearance of fetal Leydig cells, and regression of the female reproductive tract [4–8].

Fate specification and establishment of Sertoli cell lineage in fetal testis

The appearance or fate specification of Sertoli cells is initiated by SRY in most mammals with some exceptions like the Amami spiny rat and the vole Ellobius [9, 10]. SRY is exclusively expressed in the bipotential supporting cell progenitors that give rise to either Sertoli cells or its female counterpart granulosa cells. In the mouse XY embryos, Sry expression first appears in the center of the gonadal primordium around embryonic day (E) 10.5 [6, 11]. Initial Sry expression starts cell-autonomously without the involvement of external factors. The histone demethylase JMJD1A mediates H3K9me2 demethylation of the mouse Sry locus, which leads to Sry expression [12]. Sry expression in the mouse XY gonads is transient, spreading from the center to the poles of the gonadal primordium [11]. SRY, along with other transcription factors such as NR5A1 (or Steroidogenic factor 1), directly induces the expression of transcription factor Sox9 [13], a conserved testis-inducing factor in many non-mammalian and mammalian species [13–15]. SRY induces Sox9 expression by binding to the enhancer element TESCO (testis-specific enhancer of Sox9 core), Enh13, and other regions upstream of the transcription start site of Sox9 [13, 16]. SOX9 further enhances its own expression through the action of NR5A1 and SOX8 [13, 17, 18]. In addition to the action mediated through transcription activities, the SRY/SOX9 pathway also induces production of secreted factors from pre-Sertoli cells, including fibroblast growth factor 9 (FGF9) and prostaglandin D2 [19–22] (Figure 1A). FGF9 and its receptor FGFR2, both present in pre-Sertoli cells, maintain SOX9 expression and induce proliferation of pre-Sertoli cell population [23]. Pre-Sertoli cells also express prostaglandin D synthase (Ptds), an enzyme responsible for prostaglandin D2 (PGD2) production which enhances Sox9 transcription levels independently of FGF9 [20, 22, 24].

In addition to inducing Sertoli cell differentiation, SRY/SOX8/9 also antagonize the molecular pathways responsible for the differentiation of granulosa cells. These “pro-ovary” pathways include the WNT pathway (ligand WNT4, secreted WNT activator R-spondin1 or RSPO1, and its intracellular effector beta-catenin) and the transcription regulators FOXL2 and RUNX1 [21, 25–28]. These pro-ovary factors direct the bipotential supporting cell progenitors toward the granulosa cell program by repressing the testis program [26, 29]. Conversely, ectopic activation of the components of these pro-ovary pathways in the pre-Sertoli cells disrupts SOX9 expression and specification of Sertoli cells, leading to testis-to-ovary sex reversal and ovotestis formation [26, 27]. The antagonist action of SRY/SOX9 is essential to suppress expression of pro-ovary genes and pathways to avoid their feminizing effects. CBX2, a member of the Polycomb Repressive Complex 1 (PRC1), stabilizes the testis pathway downstream of SRY/SOX9 by repressing the expression of the components of the WNT pathway [30]. With the activation of the Sertoli cell program and inhibition of the granulosa cell program, the bipotential progenitors acquire their identity as Sertoli cells.

Fate maintenance of Sertoli cells in adult testis

As the supporting cell progenitors commit to either the testis or ovary fate, they were assumed to acquire terminally differentiated sex-specific gene expression and phenotypes. However, several findings in mice suggest otherwise. In adult XX mice with either disabled production [31] or disabled responsiveness to estrogens [32], granulosa cells in the adult ovary lost their identity and transformed into Sertoli-like cells with expression of Sox9. Loss of Foxl2, a conserved transcription factor for granulosa cell differentiation in vertebrates, and in particular mammals [33, 34], recapitulated similar granulosa-to-Sertoli cell transdifferentiation in adult XX mice [35]. On the other hand, when transcription factor Dmrt1 was deleted in Sertoli cells, adult XY Sertoli cells gained Foxl2 expression, lost their polarity, and gradually transdifferentiated into granulosa-like cells [3, 36]. Dmrt1 expression in adult Sertoli cells requires the combined action of SOX8/9. In adult Sertoli cells, in the absence of Sox8/9, Dmrt1 expression was downregulated, Sertoli-to-granulosa cell transdifferentiation occurred, and disintegration of seminiferous tubules was observed [37]. These observations implicate that adult Sertoli cells require SOX8/SOX9/DMRT1 to maintain their identity in the testis (Figure 1C) whereas such action is antagonized by the estrogen/FOXL2 pathway in the adult ovary. The bivalent status of key genes for Sertoli cells and granulosa cells during fetal life appears to be retained in adulthood, and may contribute to the lineage plasticity of differentiated Sertoli and granulosa cells.

Uncovering a new phase of Sertoli cell fate maintenance

The discovery of SRY and its role in testis determination in the 1990s supports that the specification of Sertoli cell lineage in most mammals occurs cell-autonomously via the transcription action of SRY in the bipotential progenitor cells. However, these findings seem contradictory to the intersex freemartin cases in cattle, sheep, pigs, and goats, which had been documented since ancient Greece [38–43]. In freemartin cases where the pregnant female carries one XX and one XY dizygotic twin, the XX twin, despite a lack of SRY, is often masculinized with retention of the Wolffian ducts, regression of the Mullerian ducts, and appearance of seminiferous cord structures in the ovaries. Frank Lillie and Tandler and Keller in the early 1900s independently observed that the freemartin twins share placental circulation through anastomosis, prompting them to postulate that hormones from the XY twin, through the shared placenta, contribute to the masculinization of the XX twin [38, 39]. At that time, androgen was the only known testis-derived hormone responsible for the masculinization of Wolffian ducts and external genitalia; however, androgens are not responsible for the appearance of testis structures in the freemartin ovary [44, 45]. The second testis-produced hormone was not discovered until the 1970s by Alfred Jost. Using the rabbit embryo as the model, Jost found a non-androgen, testis-produced factor(s) that induced Mullerian duct regression regardless of the genetic sex of the embryos [46]. This factor was later named as Mullerian inhibiting substance (MIS) or anti-Mullerian hormone (AMH, the currently used acronym) by Josso and colleagues [47, 48]. In contrast to androgens, AMH induced partial freemartin phenotypes, such as loss of female germ cells and appearance of seminiferous tubule structures, in the fetal rat and ovine ovary in vitro [49–51]. Although the naturally occurring freemartin cases were not reported in mouse embryos, mouse fetal ovaries developed freemartin phenotypes when they were transplanted under the kidney capsule of adult male recipients [52–56], indicating that in mice, adult testes produce the freemartin factors.

The focus on the freemartin factors has been on their ability to masculinize female reproductive organs; however, the action of freemartin factors on their source, the fetal testis, was completely unknown. One would expect that if the freemartin factors had such potent sex-reversing impacts on the fetal ovary, they must play a role in formation of the fetal testes. AMH, the first hormone, and putative freemartin factor produced by fetal Sertoli cells, was dispensable for normal testis morphogenesis at least in mice and humans [7]. Humans lacking AMH or its receptor, AMHR2, exhibit a disorder of sex development almost identical to the knockout mouse models, in which the Mullerian ducts in the affected XY individuals fail to regress, eventually forming a uterus and fallopian tubes (reviewed in [57]). Sertoli cells of mouse fetal testes also produce another hormone activin B [8]. Similar to the absence of Amh, XY mouse embryos lacking Inhbb, the gene encodes the beta subunit of activin B, developed normal testes with minor defects in vasculature [8]. We predicted that Sertoli cell–derived AMH and activin B, both members of TGF-beta superfamily proteins, together ensure normal fetal testis morphogenesis and when both factors are absent, testis morphogenesis would be compromised. Disappointingly, the initial testis formation in the Amh/Inhbb double-knockout XY embryos was normal with proper gene expression and establishment of testis cords and other cell types [58]. However, 4 days after the initiation of testis morphogenesis, the double-knockout testes gradually transformed into ovotestes where SOX9+ Sertoli cells in the poles transdifferentiate into their female counterpart FOXL2+ granulosa cells. The transdifferentiated Sertoli cells also lost their polarity, leading to disintegration of the testis cords on the poles. The testis cords in the center of the ovotestis remained intact, flanked by clusters of FOXL2+ granulosa cells and meiotic germ cells. Fetal Leydig cells were abundantly dispersed in the interstitium of the testicular domain, producing sufficient androgens to maintain the Wolffian ducts and masculinized the external genitalia. As a result, the double-knockout XY embryos were intersex with the presence of ovotestis, Mullerian duct derivatives, and Wolffian duct derivatives. Such intersex phenotypes lasted to adulthood with the coexistence of mature male and female reproductive tracts and ovotestes with seminiferous tubules in the centers and follicles in the poles. Due to the presence of the female reproductive tracts, the ovotestes were cryptorchid, resulting in defects of spermatogenesis and few mature sperm in the epididymis. Follicles in the ovarian domain developed to the preantral stages. These follicles responded to the hormone stimulation for superovulation but never ovulated naturally, probably due to a male-like hypothalamic–pituitary axis.

The transformation of fetal testes into ovotestes in the absence of AMH and activin B sheds new light on the normal process of sex determination. First, the transdifferentiation of fetal Sertoli cells into granulosa cells few days after initial testis formation indicates that once their lineage is specified by SRY and its downstream components, the actions of AMH and activin B are required to maintain their identity (Figure 1B). AMH and activin B serve as autocrine and paracrine factors that act on their source Sertoli cells, which express the receptors for both factors [58–62] to suppress the molecular programs for granulosa cells. Second, a lack of complete testis-to-ovary sex reversal of the double-knockout testes implies that other yet-to-be-identified factors compensate partially for the loss of AMH and activin B. Third, AMH and activin B have the freemartin capability based on the findings that without these two factors, the ability of adult Amh/Inhbb double-knockout males to masculinize the fetal ovaries was significantly reduced [58]. The fact that the freemartin effect was not completely abolished in the adult double-knockout XY mice further supports the presence of other freemartin factors such as androgens and other TGF-beta family proteins [56, 63–66].

Perspectives

Enrico Sertoli probably had no idea that the cell type named after him turned out to be the master regulator of testis morphogenesis and male reproductive functions. Sertoli cells, the first somatic cell type that appears during testis formation, determine the male phenotypic characteristics. Without Sertoli cells, seminiferous tubules do not form, spermatogenesis never occurs, androgen-producing Leydig cells and their hormonal effects fail to appear, and the female reproductive tracts remain. In contrast to many non-mammalian species where sex determination is controlled by hormones, Sertoli cell lineage in most mammals arises cell-autonomously through the transcription action of SRY and its downstream effector SOX9. However, the freemartin phenomena suggest that mammalian Sertoli cells maintain a hormone-sensitive nature. We argue that AMH, which is the testis-determining, Sertoli cell–inducing factor in teleost fish, like tilapia and pejerrey [67, 68] and possibly alligator [69], takes on a different role in mammalian sex determination. At least in mice, AMH and other TGF-beta family members maintain fetal Sertoli cell identity and retain the freemartin ability to induce Sertoli cell appearance in the fetal ovary. Whether such dual functions of AMH and activin are present in other species remains to be investigated. Freemartinism is clearly not an aberrant phenomenon; it reminds us not only of how sex determination is evolutionarily conserved and divergent but also of how much can still be learned from the master cell of testis determination discovered more than 150 years ago.

Author contribution

H.H-C.Y. and K.R. wrote the manuscript and agree to the order of authors.

Conflict of Interest: The authors have declared that no conflict of interest exists.

Footnotes

References