Deletion of Wt1 during early gonadogenesis leads to differences of sex development in male and female adult mice

Assessing the role of the WT1 transcription factor (WT1) during early gonad differentiation and its impact on adult sex development has been difficult due to the complete gonadal agenesis and embryonic lethality exhibited by Wt1KO mouse models. Here, we generated Wt1LoxP/GFP;Wt1Cre mice, the first Wt1KO mouse model that reaches adulthood with a dramatically reduced Wt1 expression during early gonadogenesis. Wt1LoxP/GFP;Wt1Cre mice lacked mature gonads and displayed genital tracts containing both male and female genital structures and ambiguous genitalia. We found that WT1 is necessary for the activation of both male and female sex-determining pathways, as embryonic mutant gonads failed to upregulate the expression of the genes specific for each genetic programme. The gonads of Wt1LoxP/GFP;Wt1Cre mice showed a lack of production of Sertoli and pre-granulosa cells and a reduced number of germ cells. NR5A1 and the steroidogenic genes expression was modulated differently in XY and XX Wt1LoxP/GFP;Wt1Cre gonads, explaining the mutant phenotypes. Further studies of the XX Wt1LoxP/GFP;Wt1Cre gonads revealed that deletion of WT1 at an early stage impaired the differentiation of several cell types including somatic cells and the ovarian epithelium. Through the characterisation of this Wt1KO mouse model, we show that the deletion of Wt1 during early gonadogenesis produces dramatic defects in adult sex development.

Assessing the role of the WT1 transcription factor (WT1) during early gonad differentiation and its impact on adult sex development has been difficult due to the complete gonadal agenesis and embryonic lethality exhibited by Wt1KO mouse models. Here, we generated Wt1 LoxP/GFP ;Wt1 Cre mice, the first Wt1KO mouse model that reaches adulthood with a dramatically reduced Wt1 expression during early gonadogenesis. Wt1 LoxP/GFP ;Wt1 Cre mice lacked mature gonads and displayed genital tracts containing both male and female genital structures and ambiguous genitalia. We found that WT1 is necessary for the activation of both male and female sex-determining pathways, as embryonic mutant gonads failed to upregulate the expression of the genes specific for each genetic programme. The gonads of Wt1 LoxP/GFP ;Wt1 Cre mice showed a lack of production of Sertoli and pre-granulosa cells and a reduced number of germ cells. NR5A1 and the steroidogenic genes expression was modulated differently in XY and XX Wt1 LoxP/GFP ;Wt1 Cre gonads, explaining the mutant phenotypes. Further studies of the XX Wt1 LoxP/GFP ;Wt1 Cre gonads revealed that deletion of WT1 at an early stage impaired the differentiation of several cell types including somatic cells and the ovarian epithelium. Through the characterisation of this Wt1KO mouse model, we show that the deletion of Wt1 during early gonadogenesis produces dramatic defects in adult sex development.

Author summary
The WT1 transcription factor (WT1) is a protein expressed during gonad development. WT1 mutations have been reported in several human conditions in which patients present a variable range of genital malformations varying from ambiguous external genitalia to gonadal dysgenesis. Mouse models in which Wt1 has been deleted indicate that WT1 has a critical role in early gonadogenesis. However, assessing the role of this protein in early Introduction Disorders/differences of sex development (DSDs) are congenital conditions in which the development of chromosomal, gonadal or anatomical sex is atypical [1]. During the last years, considerable efforts have been made to understand the genetic factors involved in DSDs. However, many patients with DSDs cannot be provided with an accurate diagnosis as the aetiology of many DSDs remains unknown [2].
Most of the knowledge about the complex process of sex development in mammals comes from the analysis of mouse models in which the expression of key genes involved in sex development has been modified [3]. By embryonic day (E) 10.5 in mice, bipotential gonads arise from the genital ridges, a structure derived from the intermediate mesoderm [3]. Its formation begins with the proliferation of coelomic epithelial cells that give rise to the somatic gonadal precursors [4]. The molecular events that determine gonadal fate in mice occur at around E11.5-12.0, when, following changes in gene expression, the bipotential gonad begins a specific developmental programme that will generate either a testis or an ovary [3]. Sex determination of bipotential gonads and the subsequent testicular or ovarian differentiation are critical steps not only in establishing the sex of an individual, but also in maintaining their reproduction [3].
The Wilms tumor 1 homolog (Wt1) encodes a zinc finger protein whose best-known function is as a transcription factor. WT1 can act as either a transcriptional activator or a repressor depending on its binding partners in a tissue or organ [5]. During gonad formation, WT1 is initially expressed in the intermediate mesoderm and, later on, in the coelomic epithelium of the genital ridge and the somatic cells of the bipotential gonad [6,7]. Following sex determination, WT1 is also expressed in Sertoli cells and granulosa cells [8]. WT1 mutations have been reported in several human conditions including Denys-Drash, Frasier and Meacham syndromes [5]. In all these syndromes, the affected XY children present a variable range of genital malformations varying from ambiguous or female external genitalia to gonadal dysgenesis [5]. Recently, XX patients carrying WT1 mutations have been identified among 46,XX SRY-negative individuals with different features of DSDs, including testicular DSD (TDSD) or ovotesticular DSD and atypical external genitalia [9][10][11].
In the last ten years, new mouse genetic tools have been generated and used in several laboratories interested in different aspects of WT1 biology [12][13][14][15][16][17]. To circumvent the embryonic lethality of the germline Wt1KO mouse model and to uncover new functions of Wt1 in development and adult homeostasis, mice carrying a loxP-flanked allele of the Wt1 gene have been generated and bred with different Cre transgenic animals [12,13,17]. In this study, we generated a new mouse model of the Wt1 gene, the Wt1 LoxP/GFP ;Wt1 Cre mice that display a strong reduction of Wt1 expression during early gonadogenesis. This new Wt1KO mouse model is the first model that reaches adulthood with a dramatically reduced Wt1 expression at an early stage of gonadal development. Interestingly, the analysis of adult Wt1 LoxP/GFP ;Wt1 Cre mice revealed the presence of ambiguous external genitalia that correlated with the absence of mature gonads, as well as the occurrence of genital tracts containing both male and female structures. We found that WT1 is necessary for the activation of both male and female sex-determining pathways, as mutant gonads failed to upregulate the expression of the genes specific for each programme, resulting in dramatic defects in the formation of somatic and germ cells.

Wt1 Cre transgene efficiently recombines in the embryonic gonad
One of the earliest markers that define the somatic cell lineages in the gonads is WT1 [18][19][20]. We decided to study the activity of Wt1 Cre (Tg(Wt1-cre)#Jbeb) in the gonads, a BAC transgenic mouse line that has been extensively used in the cardiovascular field, but whose expression in the embryonic gonads has not been analysed previously ( Fig 1A) [16,21]. This Wt1 Cre line was generated by an insertion of the IRES-EGFP-Cre cassette downstream of the translation stop site of the Wt1 gene. The EGFP protein is undetectable in the transgenic line when assessing with organ section staining or FACS analysis [16]. To monitor Wt1 Cre activity, Wt1 Cre/+ (Wt1 Cre ) mice were crossed with the reporter R26 mTmG/mTmG mice ( Fig 1B). In the Wt1 Cre ;R26 mTmG/+ mouse model, Tomato is expressed ubiquitously. After Cre-mediated loxP recombination, GFP membrane expression is observed. Consistent with WT1 expression, immunostaining analysis of GFP expression in the developing gonads revealed robust Cremediated reporter expression in Wt1 Cre ;R26 mTmG/+ mice (Figs 1C and S1A). At E14.5, most of the lineage-labelled Wt1 Cre cells (GFP-positive) were positive for WT1 ( Fig 1C). Additionally, double immunostaining with antibodies against GFP and the germ cell marker DDX4 demonstrated the absence of recombination in the DDX4-positive germ cells (S1B Fig). Surprisingly, although endogenous WT1 protein is abundantly expressed in the embryonic kidney, almost no GFP-positive cells were detected in this organ, with the exception of a few mesothelial cells surrounding its surface (Figs 1C and S1A). WT1 is expressed in the urogenital ridge that gives rise to the gonads and the adrenal cortex [7]. Wt1 Cre ;R26 mTmG/+ mice constitute a lineage tracing model, which means that all WT1 and WT1-derived cells should be GFP-positive. However, GFP-positive cells were not observed in the adrenal gland of Wt1 Cre ;R26 mTmG/+ mice at E14.5 (S1A Fig). To further verify the GFP immunostaining data with a more quantitative technique, we analysed the percentage of GFP-positive cells in the digested kidneys of Wt1 Cre ; R26 mTmG/+ mice using fluorescence-activated cell sorting (FACS) (S2A and S2B Fig). We also included in our analysis embryonic kidneys from Wt1 GFP/+ mice, a reporter mouse model that has been well characterised in studies on kidney development [22,23]. Our results from the FACS analysis supported the immunostaining data. The percentage of GFP-positive cells in the digested kidneys of Wt1 Cre ;R26 mTmG/+ mice was 0.55+/-0.13% in comparison to the 42.75 +/-4.38% ( � P < 0.05, Student's t-test) found in those of Wt1 GFP/+ mice (S2A and S2B Fig).
Altogether, these results demonstrate that the Wt1 Cre transgene recombines in the embryonic gonad, but its efficiency of recombination is very low in the kidneys and adrenal gland. https://doi.org/10.1371/journal.pgen.1010240.g001 (Fig 2A) [13,14,16]. As a result of this breeding scheme, in Wt1 LoxP/GFP ;Wt1 Cre mice, one copy of the Wt1 allele is flanked by loxP sites, whereas the other copy has a GFP knock-in in exon 1, which disrupts WT1 expression. Immunostaining analysis of WT1 expression in the embryonic gonads at E11.5 revealed a dramatic reduction in WT1 protein expression in the Wt1 LoxP/GFP ; Wt1 Cre mice, suggesting that Wt1 Cre was active from the earlier stages of gonad development ( Fig  2B and 2C). The immunostaining results were confirmed by the quantitative real-time PCR (qRT-PCR) analysis of Wt1 expression at E12.5 ( Fig 2D). However, the Wt1 levels were unaltered in the embryonic kidneys of Wt1 LoxP/GFP ;Wt1 Cre mice, which correlated with the absence of recombination of Wt1 Cre in this organ (Figs 1C, S1 and S2B-S2D). The initial characterisation of the germline Wt1KO mouse model indicated that WT1 has a critical role in early gonadogenesis [20]. However, assessing the role of WT1 in bipotential gonads and its impact on adult sex development has been difficult due to the complete gonadal agenesis exhibited by the conventional Wt1KO mice and the embryonic lethality observed in Wt1KO mouse models [20,24]. The genotyping of the adult mice of our model revealed a sub-Mendelian distribution of the Wt1 LoxP/GFP ; Wt1 Cre mice. The observed lethality of Wt1 LoxP/GFP ;Wt1 Cre mice probably corresponds to the late embryonic stages since the proportion of this genotype was not reduced when embryos were examined on days E14.5 (S1 Table). As Wt1 Cre is expressed in the heart, we suspect that WT1 deletion in the embryonic heart could be the main cause of lethality in some mutants. A very interesting finding was that all the mutant adult Wt1 LoxP/GFP ;Wt1 Cre mice were phenotypically female externally. Based on these results, we suspected a feminisation of XY Wt1 LoxP/GFP ;Wt1 Cre mice, which was confirmed by sex genotyping. The external genitalia in XY Wt1 LoxP/GFP ;Wt1 Cre mice were feminised, making all the XY Wt1 LoxP/GFP ;Wt1 Cre mice indistinguishable externally from the control XX mice with their short anogenital distances (Fig 2E and 2F). Interestingly, a gross post-mortem external examination of the Wt1 LoxP/GFP ;Wt1 Cre mice revealed the presence of structures that resembled a micropenis in both XY and XX mice ( Fig 2G). The analysis of the internal reproductive tract in controls and in XY and XX Wt1 LoxP/GFP ;Wt1 Cre mice revealed that the mutant mice presented two small structures compatible with gonads at the lower pole of both kidneys, with the mutant mice exhibiting reproductive genital tracts containing both male and female structures ( Fig 2H).
Altogether, these data indicate that Wt1 LoxP/GFP ;Wt1 Cre mice represent a new Wt1KO mouse model in which Wt1 is deleted at the bipotential gonad stage.

Wt1 LoxP/GFP ;Wt1 Cre mice represent a new mouse model of DSDs
Next, we performed a detailed histological analysis of the reproductive tract of the adult mutant mice. Interestingly, the analysis of XY and XX Wt1 LoxP/GFP ;Wt1 Cre mice revealed the absence of mature gonads when compared to control mice (Figs 3 and 4). While the testes of control mice exhibited a typical seminiferous tubule morphology with spermatogenic cells, these structures were absent in the gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice (Figs 3A and S4A). To better characterise the mutant gonad phenotype, immunostaining analysis was performed using a panel of different antibodies. The immunostaining of gonads from XY Wt1 LoxP/GFP ;Wt1 Cre mice for laminin validated the histological results. In the testis of control mice, laminin strongly localised at the basement membrane delineating the seminiferous tubule; however, in the gonads of XY Wt1 LoxP/ GFP ;Wt1 Cre mice, the seminiferous tubule was absent and laminin exhibited an irregular distribution pattern (Fig 3B). The immunostaining for DDX4, corroborated the histology data, demonstrating that the defects in the formation of the seminiferous tubule also correlated with an absence of spermatogonia cells in the gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice (Figs 3B and S4A).
The histological examination of adult XX mutant gonads also demonstrated severe ovarian dysgenesis in the XX Wt1 LoxP/GFP ;Wt1 Cre mice (Figs 4A and S4B). Follicles and oocytes were evident in the ovaries of adult control mice, but were absent in the mutant XX gonads (Figs 4A and S4B). Ovaries produce an abundant number of follicles with immature oocytes in the first postnatal week. In line with the data from the histological examination of adult mice, the immunostaining of mutant gonads with DDX4 at postnatal (P5) revealed a dramatic reduction of DDX4 positive oocytes in mutant mice ( Fig 4B). In addition, while each follicle in the ovaries of control mice was clearly delineated by Alpha Smooth Muscle Actin (SMA) immunostaining, the absence of follicles and an atypical SMA staining was clearly evident in the gonads of the mutant mice ( Fig 4B).
The histological analysis also revealed that the remaining genital tract in all the XX and XY Wt1 LoxP/GFP ;Wt1 Cre mice consisted of the simultaneous presence of both male and female genital structures. Both male and female organs, including the oviducts, uterine horns, epididymides, vas deferens, seminal vesicles (SVs) and the prostate, were identified histologically in the XY and XX Wt1 LoxP/GFP ;Wt1 Cre mice (Figs 3A, 4A and S5).
Altogether, these data indicate that the Wt1 LoxP/GFP ;Wt1 Cre mice represent a new mouse model of DSDs that reaches adulthood. Both the XX and XY mutants displayed ambiguous external genitalia, which correlated with defects in the development of mature gonads and the presence of both male and female internal genital tract structures.

Male and female early gonadal development is disrupted in Wt1 LoxP/GFP ; Wt1 Cre mice
Differentiation of the bipotential gonad represents a critical step in the formation of the testes and ovaries [3]. After observing the dramatic DSD features in Wt1 LoxP/GFP ;Wt1 Cre mice, we evaluated whether the phenotype observed in adult mice reflected a defect in gonadal morphogenesis. Compared to control mice, the gross morphological examination of the gonads of Wt1 LoxP/GFP ;Wt1 Cre mice at E14.5 revealed that the XY mutant gonads were morphologically different and had an appearance similar to that of a female gonad. However, no significant differences were observed in the XX mutant gonads (Fig 5A). A disrupted testis cord structure was already evident early on during the morphological analysis of XY Wt1 LoxP/GFP ;Wt1 Cre mice, which was confirmed by laminin immunostaining (S6 Fig). Further examination of the mutant gonads at E14.5 with the anti-WT1 antibody confirmed the reduced protein expression of WT1 observed in the gonads of Wt1 LoxP/GFP ;Wt1 Cre mice at E11.5 (Fig 5B and 5D). Immunostaining for SOX9 confirmed a defect in the testis cord formation and an absence of SOX9-positive Sertoli cells in the gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice ( Fig 5C). Interestingly, FOXL2 immunostaining revealed the absence of FOXL2-positive pre-granulosa cells in the gonads of XX Wt1 LoxP/GFP ;Wt1 Cre mice ( Fig 5C).
Next, we analysed whether the defects in gonad differentiation correlated with changes in the gene expression patterns of each specific programme. The expression of genes that play a relevant role in the initiation of the testis and ovarian developmental programmes was analysed by qRT-PCR at E12.5 (Fig 5E and 5F). Interestingly, the analysis of Sertoli cell markers revealed that Sox9 mRNA expression at E12.5 was almost undetectable in the gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice, which was consistent with the absence of SOX9-positive cells in the mutant gonads at E14.5 (Fig 5C and 5E). The lack of Sox9 expression in the gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice correlated with the downregulation of other Sertoli cell markers such as Sry and the SOX9 downstream target Fgf9 (Fig 5E).   The expression of ovary-specific genes like Wnt4, Rspo1, Fst, Foxl2, Runx1 and Podxl was significantly downregulated in the gonads of XX Wt1 LoxP/GFP ;Wt1 Cre mice compared to control mice, which correlated with the absence of FOXL2-positive cells in the mutant XX mice (Fig 5C and 5F). The initiation of either the testis-or ovary-determining pathway requires both the upregulation of their specific pathway and the repression of the other pathway [25]. Our analysis demonstrated that both the XY and XX Wt1 LoxP/GFP ;Wt1 Cre gonads failed to upregulate their specific sex-determining genetic programme.

Steroidogenic genes are differently regulated in XY and XX Wt1 LoxP/GFP ; Wt1 Cre gonads
The ambiguous external genitalia and the presence of reproductive genital tracts containing both male and female structures observed in the XY and XX Wt1 LoxP/GFP ;Wt1 Cre mice prompted us to analyse the expression of the steroidogenic factor 1 (NR5A1/SF1) in the control and mutant mice. Immunostaining analysis of NR5A1 at E14.5 confirmed previous findings, NR5A1 was abundantly expressed in control testes in comparison with ovaries ( Fig 6A) [26]. In addition, we observed a significant decrease in the number of NR5A1-positive cells in the XY Wt1 LoxP/GFP ;Wt1 Cre gonads. However, we found the opposite findings and we observed the presence of numerous cells that were strongly positive for NR5A1 in XX Wt1 LoxP/GFP ; Wt1 Cre gonad (Fig 6A and 6B).
NR5A1 is a nuclear hormone receptor that controls the expression of the steroidogenic enzymes and cholesterol transporters required for steroidogenesis [27]. Thus, we performed qRT-PCR analysis of several genes involved in different steps of steroid production such as Star, Cyp11a1 and Hsd3b1. Similar to NR5A1, the expression of these genes was downregulated in the gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice (Fig 6A-6C). Interestingly, Star, Cyp11a1 and Hsd3b1 expression was upregulated in the gonads of XX Wt1 LoxP/GFP ;Wt1 Cre mice which correlated also with the increase in NR5A1 observed in mutant XX gonads (Fig 6A-6C).
Müllerian duct regression is induced by the expression of Amh, encoding anti-Müllerian hormone, from the embryonic Sertoli cells [28]. Given the retention of structures derived from the Müllerian ducts in XY Wt1 LoxP / GFP ; Wt1 Cre mice, we also analysed Amh expression in control and mutant gonads at E14.5. As expected, we observed a dramatic reduction in Amh levels in XY Wt1 LoxP / GFP ; Wt1 Cre mice (Fig 6C).
Altogether, these findings demonstrate that Wt1 deletion in the bipotential gonads impairs the typical steroidogenic programme of XY and XX gonads. Our data also indicate that the AMH signalling is disrupted in XY mutant mice.

Deletion of Wt1 during early gonadogenesis leads to a reduction of germ cells in Wt1 LoxP/GFP ;Wt1 Cre mice
To identify the causes of the absence of oocytes and spermatogenic cells in Wt1 LoxP/GFP ;Wt1 Cre mice, we analysed germ cell production in the embryonic gonads. Germ cells in the foetal testes do not enter meiosis and are mitotically arrested, maintaining the expression of pluripotency genes. By contrast, germ cells in the ovaries enter meiosis and express meiotic markers [3]. qRT-PCR analysis of the pluripotency genes Dppa3 and Nanog at E14.5 demonstrated a decreased expression in the gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice compared to control mice ( Fig 7A). Additionally, expression of the meiotic marker Stra8 was downregulated in the gonads of XX Wt1 LoxP/GFP ;Wt1 Cre mice ( Fig 7A). Next, we stained gonads at E14.5 with an antibody against the germ cell marker DDX4. In XY control gonads, DDX4 positive cells were enclosed within the testis cord while in XX gonads, DDX4 positive cells were dispersed within the gonads (Fig 7B). Both XX and XY mutant gonads had reductions in the numbers of DDX4-positive germ cells compared to control mice (Fig 7B and 7C).
Overall, these results indicate that germ cell number is reduced from the embryonic stages in the gonads of XY and XX Wt1 LoxP/GFP ;Wt1 Cre mice.

Abnormal differentiation of postnatal ovaries in Wt1 LoxP/GFP ;Wt1 Cre mice
In contrast to male development, very little is known about the role of WT1 in female gonad development [5]. To gain more insight into the XX mutant gonad phenotype, we decided to first interrogate the nature of the NR5A1-positive cells observed in XX mice at E14.5. Double  (Fig 8A). The abundant number of primary follicles and oocytes observed in the control mice were totally absent in the mutant mice (Fig 8A). To further characterise the impact of Wt1 deletion on ovarian differentiation we stained P5 ovaries with antibodies against E-cadherin (CDH1), DDX4 and RALDH2 (Fig 8B and 8C). Immunostaining of CDH1 and RALDH2 revealed that the abnormal morphology of mutant ovaries correlated with a dramatic downregulation of CDH1 and RALDH2 expression (Fig 8B-8D). Interestingly, the decreased expression levels of RALDH2 observed in the ovarian epithelium (OE) of the mutant gonads correlated with the presence of clusters of somatic cells expressing abundant levels of RALDH2 in the mutant gonads, which probably represent undifferentiated progenitor cells [30]. Altogether, these data suggest that deletion of WT1 in the early progenitor cells of XX gonads dramatically disrupts ovarian development.

Discussion
WT1 mutations have been reported in several human conditions in which patients present a variable range of genital malformations varying from ambiguous external genitalia to gonadal dysgenesis. In this study, we generated and characterised a new Wt1KO mouse model in which Wt1 deletion was accomplished at the early bipotential gonad stage (Wt1 LoxP/GFP ; Wt1 Cre ). The analyses of these mice demonstrated the importance of Wt1 for early gonad differentiation and the impact of this early deletion on the formation of the adult reproductive system. Adult Wt1 LoxP/GFP ;Wt1 Cre mice lacked mature gonads, with both XX and XY mutants displaying genital tracts containing both male and female structures, as well as ambiguous external genitalia.
Despite the impact of WT1 mutations on sex development, many questions about the lossof-function effect of this gene on early bipotential gonad differentiation and its impact on adult sex development remain unanswered. The initial characterisation of the germline Wt1KO mouse model indicated a critical role for WT1 during early gonadogenesis [20]. However, assessing its functions at this early stage and its impact on adult sex development has been difficult due to the complete gonadal agenesis exhibited by the conventional Wt1KO mice and the embryonic lethality observed in Wt1KO mouse models. In an attempt to examine the role of WT1 before sex determination, a ubiquitous tamoxifen-inducible Cre Wt1KO model was recently generated [24]. However, the early lethality of these mutants precluded the analysis of the impact of Wt1 deletion at an early stage on more mature gonad formation and adult sex development. In this study, we used a Wt1 Cre BAC transgenic mouse line that dramatically reduced Wt1 levels in the early bipotential gonads of Wt1 LoxP/GFP ;Wt1 Cre mice. Our mouse model is the first with a Wt1KO in the bipotential gonads in which it is possible to follow the behaviour of the mutant gonads in vivo throughout different developmental phases, circumventing the need for organ cultures, which were required in the case of the tamoxifeninducible Wt1KO mouse model [24]. In this mouse model, as well as in humans with WT1 loss-of-function mutations, the role of WT1 is compromised during the entire lifespan of the individual. In Wt1 LoxP/GFP ;Wt1 Cre mice, Wt1 downregulation was achieved at the same developmental stage in the XX and XY primitive gonads, enabling the analysis of the effects of WT1 deletion on primitive progenitor cells on sex development.
The Wt1 Cre transgenic line used in this study was generated more than ten years ago and has been used by several groups since then [16,21,31,32]. We found that Wt1 Cre efficiently recombines in the embryonic gonads, but, unexpectedly, we observed almost no recombination in the embryonic kidney and adrenal gland. Descendants of WT1-positive cells contribute to the development of the gonads, adrenal gland and kidneys in a temporally distinct manner [6]. Despite the crucial role of WT1 in the formation of several organs, little is known about the molecular mechanisms underlying WT1 regulation. Such diversification in its temporal expression and functions probably reflects the presence of specific regulatory elements of the WT1 gene in the different populations of WT1-positive cells [6,33]. Our data and previous results from other groups suggest that this Wt1 Cre line is able to direct expression in the coelomic epithelium and its derivatives [34].
Many phenotypic features found in the XY Wt1 LoxP/GFP ;Wt1 Cre mice, such as gonadal dysgenesis and ambiguous genitalia, have been identified in XY humans with WT1 mutations, including in those with Denys-Drash syndrome or Frasier syndrome. Male sex differentiation requires androgen activity that must occur within a critical timeframe during early embryo development [3]. A defect in the production of testosterone, due to a reduction in NR5A1 and the expression of several steroidogenic genes, could be one of the main causes of the features of DSDs observed in both our mutant mice and in XY humans with WT1 mutations. The embryonic gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice lacked the expression of early male-specific markers, which correlated with a defect in the production of Sertoli and foetal Leydig cells. Sertoli cells are the first somatic cell lineage to arise in the testes and are believed to play a crucial role in their subsequent differentiation and organisation [3]. Disruption in the formation of early Sertoli cells in Wt1KO gonads could trigger all of the DSD defects observed in the mutant XY mice. Interestingly, despite the gonads of XY Wt1 LoxP/GFP ;Wt1 Cre mice failing to upregulate the testis-specific programme, there was no upregulation of the ovary-determining pathway, with the mutant gonads remaining probably undifferentiated. Furthermore, the XY Wt1 LoxP/ GFP ;Wt1 Cre mice showed reproductive genital tracts containing both male and female structures, which was probably due to an impairment in the regression of the Müllerian ducts. The reduction in Amh expression observed in the XY Wt1 LoxP/GFP ;Wt1 Cre mice could explain the presence of this phenotype.
The gonads of the XY Wt1 LoxP/GFP ;Wt1 Cre mice did not show testis cord formation, which correlated with an impaired production of Sertoli and a reduction in germ cells. Interestingly, we also observed a reduction in the expression of steroidogenic genes, which suggested an impaired production of foetal Leydig cells. Lineage tracing and single-cell transcriptomics of the mouse embryos have indicated that the supporting Sertoli cells and most steroidogenic Leydig cells of the developing testis are derived from common progenitor precursors originating from the coelomic epithelium [18,35]. Based on our results, we postulate that Wt1 deletion in the XY primitive gonad leads to the progenitors of the somatic cells remaining immature.
There are fewer genes known to be associated with XX DSDs compared to XY DSDs, probably because genes and pathways leading to female sex development are poorly understood. Until very recently, patients carrying WT1 mutations and displaying DSDs were XY. However, during the last year, the number of studies reporting XX patients with WT1 mutations has increased [9][10][11]. Our results indicate that Wt1 deletion at the bipotential gonad stage can also impair female sex development. The XX Wt1 LoxP/GFP ;Wt1 Cre mice exhibited several characteristics associated with defects in female sex differentiation that are considered to be DSDs. The XX Wt1 LoxP/GFP ;Wt1 Cre mice in our study showed undifferentiated gonads, reproductive genital tracts containing both male and female structures and ambiguous external genitalia. Intriguingly, some but not all of these features have been recently described in XX humans with WT1 mutations that are considered to be gain-of-function mutations but also loss-offunction mutations [9][10][11].
In recent years, single-cell transcriptomic analyses of the embryonic gonad have revealed new insights into female gonad development [29,36]. These data in combination with lineage tracing analysis have demonstrated the complexity of XX gonad differentiation as well as the presence of different populations of progenitor cells derived from the coelomic epithelium that are positive for WT1 [29]. The increased expression of NR5A1 and some steroidogenic genes observed in the gonads of XX Wt1 LoxP/GFP ;Wt1 Cre mice can contribute to the upregulation of testosterone production as well as the presence of genital tracts containing both male and female structures. Interestingly, the progenitor cells of embryonic gonads express several steroid hormone genes and their increased expression can masculinise female mice [29]. Based on our results, we speculate that the increased expression of the steroidogenic genes and NR5A1 observed in XX Wt1 LoxP/GFP ;Wt1 Cre mice could be due to impaired differentiation of the progenitor cells of the mutant gonads. In line with this observation, we also observed that the NR5A1-positive progenitor cells of early embryonic gonads expressed abundant levels of NR5A1 and PODXL [29].
In XX Wt1 LoxP/GFP ;Wt1 Cre mice, the OE showed abnormal differentiation as the mutant cells were positive for SMA and negative for CDH1. We also observed that RALDH2 was dramatically downregulated in the OE of mutant mice, indicating that the expression of this gene depends on WT1 activation, similar to that observed in the epicardium [37]. In contrast to the cells of the OE, some somatic cells of the mutant ovaries expressed abundant levels of RALDH2, suggesting that the regulation of this gene by WT1 in the gonads could be celldependent. Single-cell RNA sequencing of the mutant gonads is necessary to provide further insights into the stages in which the different populations of progenitor cells are arrested in Wt1 LoxP/GFP ;Wt1 Cre mice.
During embryonic development, WT1-expressing cells contribute to the formation of several tissues. Based on the findings presented here, we propose that a deeper characterisation of the Wt1 LoxP/GFP ;Wt1 Cre mouse model will aid in the understanding of the functions of WT1 in several populations of progenitor cells in different organs and tissues as well as the importance of these cell populations for adult organ formation.

Ethics statement
All animal experiments were carried out in accordance with the regulations of the Animal Experimentation Ethics Committee (CEEA) of the University of Barcelona (ID C121CG2V6), thereby complying with current Spanish and European legislation.

Animal models
The Wt1 GFP/+ , Wt1 Cre (Tg(Wt1-cre)#Jbeb), Wt1 LoxP/LoxP and R26 mTmG/mTmG mice have been described previously [13,14,16,38,39] S2 Table. To generate male Wt1 GFP/+ ;Wt1 Cre mice, Wt1 Cre/Cre males were mated with Wt1 GFP/+ females. Mating between Wt1 GFP/+ ;Wt1 Cre males and Wt1 LoxP/LoxP females generated a new Wt1KO mouse model (Wt1 LoxP/GFP ;Wt1 Cre ) and control animals. To generate Wt1 Cre ;R26 mTmG/+ embryos, Wt1 Cre/+ males were mated with R26 mTmG/mTmG females. Embryos were generated through timed matings, whereby females that had mated during the night were checked for plugs early the following morning. The morning on which a plug was found was considered to be E0.5. Genotyping was carried out using DNA extracted from the tail tip or ear biopsies. The primers used for genotyping and the determination of the presence of the Y chromosome are listed in S3 Table. Tissue preparation for histological analysis Animal adult tissues (around 6 months old) were collected and fixed in 10% neutral-buffered formalin for approximately 24 hours, before being processed and embedded in paraffin. Embedded tissues were cut into 3-to 4-μm sections. Haematoxylin and eosin (H&E) staining was performed at the Histopathology Facility of the IRB, using the standard protocol. Images were acquired using an Olympus CX43 microscope. The histopathological analysis of the adult mice was performed by the Histology Services of the IRB.

Immunofluorescence
Mouse embryos at the indicated stages and dissected postnatal gonads were fixed in 4% or 2% paraformaldehyde (PFA) respectively, embedded in paraffin, and cut into 7-μm sections. Following standard protocols of deparaffinisation, an antigen retrieval procedure was carried out by boiling the samples in a pressure cooker for 15 min in citrate buffer (10 mM tri-sodium citrate 2-hydrate; pH 6.0). The slides were then incubated in blocking serum (2% foetal bovine serum (FBS) in PBS) for 2 h before being incubated overnight with the primary antibody at 4˚C. The slides were then washed with PBS, incubated at room temperature for 2 h with the appropriate secondary antibodies, and stained with DAPI (Thermo, ref. 62249) for 5 min for nuclear staining [40]. For a list of all the antibodies and dilutions used, see S4 Table. Samples were imaged using the Zeiss 880 confocal microscope.
ImageJ software was used for the quantification of Median Fluorescence Intensity (MFI) and the number of DDX4positive (DDX4 + ) and NR5A1-positive (NR5A1 + ) cells. For each image, the number of DDX4 + or NR5A1 + cells versus total cells (DAPI) or the MFI was determined in a minimum of four regions of interest (ROIs). Two images per mouse were analysed. For the quantification of the MFI in the OE of P5 gonads, the OE surface was obtained by the manual drawing of the external contour of this epithelial cell layer. MFI intensity within the OE was quantified and normalized to the OE surface. A minimum of three control and mutant mice were analysed.

FACS analysis
E14.5 embryonic kidneys from Wt1 Cre ;R26 mTmG/+ and Wt1 GFP/+ mice were digested in a collagenase solution (0.01%, dissolved in HBBS; Worthington, LS004196) for 30 min at 37˚C using a shaking block at 1000 rpm (Eppendorf Thermomixer Compact). Collagenase activity was stopped by washing the cells in PBS containing 5% FBS. Cells were pelleted by centrifugation at 300 g for 5 min. Cells were filtered through a 40 μm cell strainer and subjected to FACS. FACS analysis was carried out by gating against negative control littermates. Doublets and dead cells were eliminated, using DAPI staining. The data were analysed using the BD FACS-Diva software (v6.1.3).

Isolation of RNA and real-time PCR
Whole gonads were used to isolate RNA using an RNA Isolation Kit (Invitrogen, 12183025). Purified RNA was used for reverse transcription and cDNA generation with SuperScript III Reverse Transcriptase (Thermo Fisher Scientific, 18080044). Relative expression levels were determined with the SYBR Select Master Mix (Thermo Fisher Scientific, 4472897). Real-time PCR experiments were performed in a LightCycler 480 System (Roche) [40]. The primers used are listed in S5 Table. Quantification and statistical analysis Data are presented as the mean ± standard error of the mean. Statistical significance between two groups was determined by unpaired two-tailed Student's t-test. We applied non-parametric two-way ANOVA followed by Tukey's post-hoc test to evaluate the differences among multiple groups of samples.
Supporting information S1 Fig. Monitoring Wt1 Cre activity in the embryonic urogenital region. (A) Immunofluorescence staining for GFP (green), WT1 (red) and nuclear DAPI staining (blue), using the urogenital region from Wt1 Cre ;R26 mTmG/+ E14.5 mice. Cre activity (GFP-positive cells) is restricted to the gonads, the Müllerian ducts and some mesothelial cells covering the kidneys. AG, adrenal gland; G, gonad; K, kidney. (B) Gonads from Wt1 Cre ;R26 mTmG/+ mice at E14.5 were co-labelled with antibodies against GFP (green) and the marker for germ cells DDX4 (red), as well as stained with the nuclear DAPI dye (blue). Representative immunostaining images from a minimum of three embryos are shown. Scale bars: 150 μm in A; and 50 μm in