Sex-Reversal in XX (PIS-/-) Goats Occurs Early in Gonad Development and Affects Supporting Cell Lineage
Detailed knowledge about the genetics of the polled mutation in goats was used to produce intersex fetuses at crucial steps of sexual differentiation (36 dpc, 40 dpc, 56 dpc, 70 dpc). Reports in the scientific literature previously have stressed the extremely masculinized phenotype of adult XX sex-reversed polled goats compared with XX sex reversal in other mammalian species such as dogs and pigs (Soller et al., 1969; Meyers-Wallen and Patterson, 1988; Pailhoux et al., 1997). Accordingly, we found a very early onset of sex reversal in goat XX (PIS-/-) gonads. Histologic observations showed that cord formation was only delayed by 4–5 days compared with normal XY testis differentiation (40 dpc vs. 35 dpc). Thereafter, the structure of the XX (PIS-/-) testes was very similar to XY testes until birth. These histologic data were confirmed by the transcription analysis of several genes involved in the sex determination cascade.
The first effect of the PIS mutation was a reduction in aromatase expression as early as 36 dpc. Aromatase (CYP19), which converts androgens into estrogens in follicle precursor cells in the ovarian context, has been shown to play a very important role in many vertebrate species. In turtles and birds, aromatase inhibitor administration leads to gonadal sex reversal (Abinawanto et al., 1996; Belaid et al., 2001; Vaillant et al., 2001). In bovids, in contrast to mice, CYP19 expression was detected early in the fetal ovary, at the same stage as AMH in the fetal testis (Payen et al., 1996; Quirke et al., 2001). CYP19 presents with a limited expression window in ruminants, starting from the sex-determination period and culminating before the onset of meiosis. Its down-expression in goat XX sex-reversed gonads is correlated with the first histologic sign of gonadal de-feminization: the reduction of the ovarian cortex. Mice lacking aromatase activity (ARKO) have been produced by targeted disruption of Cyp19 (Toda et al., 2001) and showed a depletion of ovarian follicles without any sign of sex reversal. The interspecific discrepancy described here between rodents and ruminants could result from the absence of aromatase expression in mice before birth (Greco and Payne, 1994).
Parallel to this study, we have recently shown that expression of FOXL2 was dramatically inhibited as early as 36 dpc in XX (PIS-/-) sex-reversed gonads and was probably the cause of the sex-reversal phenotype (Pailhoux et al., 2001). FOXL2 normally expressed in follicular cells (Crisponi et al., 2001; and our unpublished observations) could directly or indirectly regulate the expression of aromatase in these cells.
In our study, the up-regulation of “male” genes in the XX (PIS-/-) gonads was observed between 36 and 40 dpc instead of 32–35 dpc during normal testicular differentiation. SOX9, which is a primary testis-differentiating gene, is capable of triggering male differentiation (Foster et al., 1994; Wagner et al., 1994; Morais da Silva et al., 1996; Vidal et al., 2001), directly activating AMH gene transcription (De Santa Barbara et al., 1998; Arango et al., 1999). In contrast to SRY, SOX9 is highly conserved among amniotes and expressed exclusively in Sertoli cells. To date, no intermediate has been found between SRY and SOX9. The up-expression of SOX9 very early in XX (PIS-/-) fetuses is consistent with the fact that the supporting cell lineage is first affected by PIS mutation and that gene(s) disrupted by this mutation rank(s) high in the cascade of sex-determination. After the increase of SOX9 expression, AMH is also strongly up-regulated in XX (PIS-/-) intersex gonads as early as 40 dpc. AMH has been shown to inhibit CYP19 expression in rat and sheep fetal ovaries (Vigier et al., 1989; Rouiller-Fabre et al., 1998), but here the down-regulation of aromatase expression precedes the AMH appearance. Despite this second rank, AMH could participate in the complete cutoff of CYP19 expression. Absence of aromatase activity results in the secretion of testosterone instead of estrogens. AMH and testosterone then work together to constitute the male internal and external phenotype.
In contrast to supporting cell markers, the expression of steroidogenic-cell specific genes such as WNT4 and 3β-HSD was not clearly affected at 36, 40, and 44 dpc. A significant difference in their expression was only observed at 56 dpc. Accordingly, the steroidogenic cells are affected by the PIS mutation after Sertoli cell differentiation that occurred at 40 dpc.
Among other cells contributing to the interstitium, vascular endothelial cells and peritubular myoid cells migrate into the testis from the mesonephros (Buehr et al., 1993; Martineau et al., 1997). Capel and coworkers have shown that Sry is essential to this mesonephric cell migration (Capel et al., 1999; Schmahl et al., 2000) that is necessary for testicular cord formation and is absent in the ovary-determining pathway. From sections of XX (PIS-/-) testes followed by Tuchmann du Plessis coloration, peritubular myoid cells could be observed surrounding the cords producing the basal lamina in association with Sertoli cells. We show, therefore, that mesonephric cells migrated into the gonads without the presence of SRY. This result seems to indicate that FGF9, the growth factor responsible for this migratory process (Colvin et al., 2001), is up-regulated in the testis by a gene acting downstream of SRY. SOX9, which induces a complete masculinization of the gonads in XX transgenic mice (Vidal et al., 2001), seems to be the best theoretical candidate for FGF9 up-regulation.
In the PIS phenotype, sex reversal occurred very early in the supporting cells of the developing gonad and is completely different from freemartinism, where masculinization results from AMH exchanges much later in fetal life (Vigier et al., 1984). In this later case, AMH destroyed meiotic germ cells leading to follicular/ Sertoli cells transdifferentiation.
How to Explain the Variability of Phenotypes Visible as Early as 56 dpc?
In the polled goat model, the same mutational event induces a variable masculinization of XX (PIS-/-) sex-reversed gonads detectable as early as 56 dpc. Contrary to the dog XX sex-reversal model (Meyers-Wallen et al., 1994), this variability in goats was not due to a delay in the timing of the mutated gene expression, because all five sex-reversed fetuses studied at 40 dpc presented a similar degree of sex-reversal. The starting point was the same but, with time, the magnitude of the defects varied.
Variable phenotypes have been encountered in many cases of sex-reversal, like Sry-associated sex reversal in mice (Eicher et al., 1982), SOX9-associated sex reversal in humans (Sinclair, 1998), and recently, Fgf9-associated sex reversal in mice (Colvin et al., 2001). In this case, Fgf9 disruption in XY fetuses leads to phenotypes ranging from testicular hypoplasia to complete sex reversal. The sex-determining pathway is governed by several genes having a very subtle gene dosage such as WT1, SOX9, DAX1 (Veitia et al., 2001), and all slight modifications of this dosage could affect gonadal differentiation and be at the origin of the variability of masculinization observed here.
Why Did Germ Cells Die in XX (PIS-/-) Sex-Reversed Gonads?
An invariable feature of XX (PIS-/-) sex-reversed gonads is the fact that testicular parts are always devoid of germ cells in the adult. One aim of this study was to determine the critical developmental stage of germ cell loss. Surprisingly, germ cell loss seems to be an ongoing process occurring from the moment when female germ cells enter meiosis (56 dpc) until birth. Similar results have been described in both XXSxr and XXY mouse testes, where germ cell number significantly decreased in the early stages of testis differentiation and was absent after birth (McLaren, 1981, 1983; Hunt et al., 1998). It seems that germ cell loss could principally be due to an altered X-chromosome dosage. Indeed in normal females, the XX PGCs undergo X-inactivation at the onset of migration from the gut endoderm and re-activate the silenced X-chromosome when they enter the urogenital ridge. In XX sex-reversed testis, germ cells do not enter meiosis, they undergo X-chromosome reactivation at the same time as XX germ cells in ovary (McLaren and Monk, 1981). It has been postulated that the presence of two active X-chromosomes was incompatible with both prenatal and postnatal male germ cell development, suggesting also that the regulation of X-chromosome activity is independent of ovarian morphogenesis.
Another factor that could explain the loss of germ cells in these XX (PIS-/-) gonads is the presence of AMH. Indeed, it has been demonstrated that fetal rat ovaries cultured in the presence of AMH show massive loss of germ cells around the time of onset of meiotic prophase (Vigier et al., 1987). Similar results have also been observed in mouse by transgenesis experiments (Behringer et al., 1990; Lyet et al., 1995). The inhibitory effect of AMH on meiotic germ cells could be responsible for the loss of germ cells located in the ovarian part of ovotestes. Indeed, we showed that these germ cells enter meiosis as in normal female gonads (around 56 dpc) but concomitantly, XX Sertoli-like cells present in the testicular part of the same gonad produce AMH. Meiotic germ cells, therefore, could be destroyed by AMH (synthesized by Sertoli cells), causing prefollicle cells to disappear or to transdifferentiate. This mechanism could explain why the ovarian parts observed during fetal life were generally missing at adulthood.
In conclusion, female-to-male sex reversal in polled goats results from a unique mutational event that affects early gonadal differentiation. In XX (PIS-/-) gonads, ovarian differentiation does not occur, either because it is not initiated, or because it is arrested by a missing element of the cascade. The second hypothesis seems to be more likely, because the aromatase gene is faintly expressed in XX (PIS-/-) gonads at 40 dpc, attesting the beginning of ovarian determination. After the female pathway blockage, the testis-determining pathway is started and SOX9 seems to be the conductor. This finding suggests that the missing elements in XX (PIS-/-) gonads enabled male genes to be expressed and that their normal counterpart inhibited male differentiation during ovary development. These findings are in complete agreement with the Z hypothesis (McElreavey et al., 1993), except that Z in this case is not a gene but a regulatory element acting on at least two genes (Pailhoux et al., 2001). Investigations are currently in progress to understand the role of the three different actors of the PIS locus: the PIS regulatory element, PISRT1, and FOXL2.