Testis formation in XX individuals resulting from novel pathogenic variants in Wilms’ tumor 1 (WT1) gene

Significance Sex development involves a precise spatiotemporal expression and interactions of numerous genetic factors, including the WT1 (Wilms tumor 1) gene. Complete and partial loss-of-function WT1 variants are associated with 46,XY disorders/differences of sex development (DSD). Some 46,XX individuals develop testis in absence of the testis-determining gene SRY. We describe a genotype/phenotype association where variants impacting the C-terminal zinc finger (ZF4) of WT1 cause testis development in 46,XX individuals. XX mice carrying a pathogenic variant of ZF4 display masculinization of the fetal gonads. Testis formation may be due to inappropriate interaction between the mutated WT1 and an essential ovarian determinant β-CATENIN. These data show that variants affecting a specific domain of a developmental transcription factor can switch organ fate.

was observed at birth and she was referred with diagnosis of congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency because of increased serum 17Hydroxyprogesterone levels. However, molecular study of CYP21A2 gene was negative. Genital examination showed 3 cm length phallus, single perineal orifice, scrotal folds fused in midline and non-palpable gonads. External masculinization score (EMS): 7/12. Karyotype was 46,XX and FISH excluded an SRY translocation. A human chorionic gonadotropin (hCG) stimulation test was performed at age of 3.5 years, resulting in a serum testosterone level of 2.37 ng/ml and anti-Müllerian hormone (AMH) of (34 pmol/l), both in female reference range. Müllerian structures were detected on ultrasound. Laparoscopy and bilateral gonadal biopsy showed the presence of testicular tissue. Histology confirmed the presence of testicular parenchyma with signs of dysgenesis and scarce germ cells. The patient was lost in follow-up (Table 1).
Patient 3 is the first child of nonconsanguineous healthy parents, born full-term with adequate weight birth. Familial history is unremarkable. At birth genital ambiguity was noticed. Cytogenetic studies showed a 46,XX SRY-negative karyotype. The baby was assigned the female sex. Genital examination at 3 months of age showed 2.2 cm length phallus with chordee, complete labio-scrotal fusion, single and small peno-scrotal orifice and nonpalpable gonads. EMS: 4/12. Testosterone after hCG stimulation test was 3.79 ng/ml; AMH was between the male and female reference values (164.2 pmol/l) and GnRH stimulation test showed a predominant LH response (29.4 MIU/ml).
Müllerian structures were detected on ultrasound. Laparoscopy revealed intraabdominal gonads and two hemi Müller. Histology of bilateral gonadal biopsy showed dysgenetic testicular parenchyma with scarce germ cells. The baby was reassigned the male sex at 3 months of age. Hypospadias repair and orchidopexy were performed between 1-3 years of age (Table 1; Figure 1B).
Patient 4 is the first child of healthy nonconsanguineous parents. Infertility was reported in some relatives of the baby´s father. At birth ambiguous genitalia was observed. On genital exam the patient presented a 3 cm length phallus with well-developed corporal tissue, pigmented labioscrotal folds fused in the midline, a single phallo scrotal orifice and nonpalpable gonads. EMS: 7/12. Karyotype was 46,XX and SRY-negative. At 12 days of life, testosterone after hCG stimulation test was 4.91 ng/ml; AMH was low for the male reference range (82.1 pmol/l) and GnRH stimulation test showed a predominant LH response (20.2 MIU/ml). Pelvic ultrasound revealed an apparently normal uterus and both gonads with follicles. Laparoscopy and bilateral biopsy were performed. Histological analysis showed that both gonads were ovotestes with gonadoblastoma and dysgerminoma and gonadectomy was performed. The patient was assigned the female sex (Table 1; Figure 1C). Patient 5a was born at 38 weeks of gestation after an uneventful pregnancy. Baby was a first child of healthy nonconsanguineous Hungarian parents. Atypical genitalia was noticed at birth, with 3 cm length and 1.5 cm wide phallus like clitoris, single penoscrotal orifice, complete labioscrotal fusion and non-palpable gonads (Prader IV, EMS 7/10). Congenital adrenal hyperplasia was excluded and the cytogenetic testing revealed a 46,XX, SRY-negative karyotype. The gender of rearing was female. At three months of age explorative laparotomy identified a uterus and two macroscopically undifferentiated gonads. At the age of six month, she underwent feminization genioplasty, her androgen levels were still significantly elevated with testosterone 152 ng/dl (2-15 ng/dl) and androstenedione 81 ng/dl (5-40 ng/dl). At the age of four years the basal testosterone level was low (<10 ng/dl) and in the hCG stimulation test it reached 42 ng/dl. At age of 12 years she underwent a bilateral gonadectomy and histological examination revealed ovotestis in both gonads without germ cells or sign of malignancy. A diagnosis of 46,XX ovotesticular DSD was indicated (Table 1; Figure 1D). Patient 5b is the brother of Patient 5a, was born at 38 weeks of gestation. His birth weight was 3,200 g and birth length was 57 cm. At birth he presented with male external genitalia, but testes were not palpable in the scrotum nor in the inguinal canal. After birth he underwent surgery for a diaphragmatic hernia. His karyotype was 46,XY. At age 1.2 years his basal FSH was 0.7 IU/L and testosterone <0.1 nmol/L, but after the hCG stimulation the testosterone level increased properly (11.1 nmol/L). At 1.8 years of age laparoscopy identified a rudimentary testis on the right side. Orchidectomy was performed on the left side and histological examination of this gonad did not find any testicular tissue, but pieces of funiculus spermaticus and epididymis tissue. At age of 9 years his hormone levels were in prepubertal range (FSH: 0.6 IU/L, LH <0.11 IU/L, T < 0.43 nmol/L, E2 < 92 pmol/L). The combination of diaphragmatic hernia with 46,XY DSD indicated a diagnosis of Meacham syndrome (Table 1; Figure 1E). Figures of the Superposition of ZF4 of the KTS-structure on the partial ZF4 structure of the KTS+ structure was performed with Lsqkab (40,41). In an effort to minimize overinterpretation of the structural alignment, no refinement was performed. Predictive structures were analysed for the p.Lys491Gln, p.Arg495Gly and p.Arg495Gln proteins.

Plasmid construction.
Vector containing full-length WT1(+KTS) isoform in a pSG5 vector was kindly provided by Dr Maëlle Pannetier and Dr Eric Pailhoux. The pCMX-NR5A1 vector with human NR5A1, the pCMV6-FOXL2, the pIRES-hrGFPII-Gata4, pCS2+Fog2, pCDNA-SOX9-Flag and Tesco reporters have been previously described (42)(43)(44). To assay the ability of WT1 proteins to modulate the canonical WNT pathway, we used the TOPFlash-TCF reporter plasmid, as previously described (16). Vector containing a 1-kb fragment of the caprine FOXL2 promoter from -842 to +168 was cloned upstream of the luciferase gene in a pGL3basic vector (Promega) and have been previously described (45) and kindly provided by Dr Maëlle Pannetier and Dr Eric Pailhoux.

Site-Directed Mutagenesis.
WT1 expression vectors containing the p.Arg495Gly mutation was generated by site-
Results are shown as the mean ± SEM of at least three independent experiments, each performed at least in quadriplicate.

Transfections in KGN-1 cells line.
Using GeneJuice® Transfection Reagent, 500ng or 1ug of plasmids were transfected in KGN cells (WT1-wild-type, WT1p.Arg495Gly and empty vectors). After two-days of transfection, cells were collected in TRIzol™ Reagent (Invitrogen™) and stored at -80°C prior to analysis. RNA extraction was performed according to the manufacturer recommendations. Both quality and quantity were assessed through the Agilent 2100 Bioanalyzer System (Agilent Technologies).

Collection of human fetal gonads and Immunofluorescence.
Human fetal ovaries were isolated from material available following elective termination of pregnancy during the first trimester at the Department of Gynaecology at Copenhagen University Hospital (Rigshospitalet) and Hvidovre Hospital, Denmark.
Fetal age determined by scanning crownrump length and by evaluation of foot length (46). Sex of foetal samples was determined by PCR for SRY as previously reported (47).

Immunofluorescence on human fetal gonads.
Tissue was fixed in formalin immediately after dissection and dehydrated, paraffin embedded and sectioned (4 µm) using standard procedures. Immunofluorescence was performed as previously described (48). Primary antibodies used was WT1 (Abcam,

Statistical analysis.
Statistical analyses were carried out using GraphPad Prism 6 software (GraphPad Software). Quantitative data were subjected to a one-way ANOVA (or student ttest) followed by Bonferroni comparison.

Design, preparation and microinjection of sgRNAs mRNA.
Single guide RNA (sgRNAs) was designed (www.crispr.mit.edu) to target in proximity to the desired SNP, located in exon 10 of the mouse Wt1 gene (MN_144783, chr 2), by CRISPR/Cas9 genome editing. sgRNA was prepared as described elsewhere (49).
Briefly, the complementary pair of oligos (Table S1) was phosphorylated, annealed and cloned into pX459-V2 plasmid (Addgene #62988). A T7-sgRNA PCR product was amplified with a T7 promoter sequence introduced on the forward primer in conjunction with a universal reverse primer sgRNA-uni.R (Table S1). This product was used as the template for in vitro transcription (IVT) using the MEGAshortscript T7 IVT kit (Life Technologies). The sgRNA was purified using the MEGAclear kit (Life Technologies) and eluted in RNase-free water. Cas9 mRNA was obtained commercially (Tebu-Bio).
Single-strand DNA oligo (ssOligo) (120 bp) was designed so that it contains the desired mutation of a G instead of a C, converting amino acid Arginine at position 495 into a Glycine. The oligo also contained 3 silent mutations that alter the DNA sequence of exon 10 but will not change the amino acid sequence of the WT1 protein ( Figure S5).
The silent mutations were inserted in order to prevent from the CAS9 protein from cutting the homology dependent repair (HDR) oligo that was inserted instead of the original Wt1 sequence. The ssOligo was synthesized by IDT. A mix of ssOligo sgRNA and CAS9 mRNA (100 ng/μl ssOligo, 50 ng/μl sgRNA, 100 ng/μl CAS9 mRNA) were injected into the cytoplasm of F1 (C57BL/6J x CBA) one-cell stage embryos as described (49). Injected zygotes were surgically transferred on the same day into oviducts of pseudopregnant CD1 recipient females.

Genotyping the genetically altered mice.
Founders with the targeted insertion of the HDR allele were identified amongst live born mice using miSeq next generation sequencing of a PCR fragment surrounding the targeted region (all genotyping primers are listed in Table S1). Both founders and F1 mice were genotyped using miSeq. The following generations of mice were genotyped using either Sanger sequencing of a PCR fragment that contains the desired SNP/wildtype allele or using real time PCR with specific probes designed for the SNP/wild-type allele (Transnetyx, Cordova, TN). This was performed using genomic DNA extracted from tail tissue of embryos or ear punch tissue of adult animals. Chromosomal sex was determined by PCR genotyping with primers Sex-F and Sex-R (50) or using real time PCR with specific probe targeting the Y chromosome (Transnetyx, Cordova, TN). Two founders that contained one HDR-targeted allele and a second wild-type allele were used to transmit the HDR allele and a stable line was generated ( Figure S5) Carrying the Arg495Gly WT1 mutation in two copies is embryonic lethal and embryos die just before birth and no adult homozygous mouse was even generated. Hence, homozygous embryos were generated by breeding two heterozygous mice.

Next-generation sequencing to screen CRISPR deleted mice.
Illumina MiSeq platform was used to sequence founder mice, because they are often mosaic, with several different alleles. The F1 mice were sequenced in the same way.
The primers used for Sanger sequencing of the WT1 SNP PCR were also used for the miSeq PCR, however the adapter sequence was added to the 5' and 3' ends of the forward and reverse primers, respectively ( using the NCBI blastn.

Timed mating and tissue preparation.
Embryos and animals carrying Wt1 SNPs were produced by crossing heterozygotes.

Quantitative Real-Time Polymerase chain reaction (qRT-PCR).
qRT-PCR reactions were performed in duplicate using SYBR Green PCR master mix (Invitrogen) and 150 nM each of forward and reverse primers, and analyzed on the Applied Biosystems 7500 Real-Time PCR System (Thermo Fischer Scientific).
Primers are listed in Table S2.   The KTS+ isoform binds the minor groove, while the KTS-isoform binds the major groove. It has been suggested that ZF4 plays a more substantial role in stabilization of DNA-binding in the KTS-isoform than it does in the KTS+ isoform (39-41). The in silico models suggest that mutations of either Arg495 or Lys491, or other mutations that affect ZF4 in KTS+ would impact nucleic acid binding if ZF4 assumes the position observed in the KTS-structure. In any scenario, the fact that two patients presented with complete absence of ZF4 as a result of frameshift mutations indicates that the function of the ZF4 is abolished.       Thickness of both cortex and medulla is reduced in Wt1 Arg495Gly/Arg495Gly kidneys compared to wild-type. E16.5 XY and XX Wild-Type Kidney: These sections exhibit a moderately mature kidney containing well-developed renal corpuscles (Malpighian corpuscle) with a defined capsule of Bowman and investing a glomerular tuft. Glomerular tufts are surrounded by an adequate urinary/glomerular space and the urinary and vascular poles are also discernible. The macula densa is frequently observed. Although incompletely developed and immature, proximal and distal convoluted tubules are morphologically distinguishable. The outer cortex and juxtamedullary cortex are discernible. The corticomedullary junction is not sharply demarcated and the pelvis is discernible. Mesenchymal cells are abundant. E16.5 XY and XX heterozygous (Wt1 Arg495Gly/+ ) kidney: No histological differences are observed in comparison with the wild-type kidney. E16.5 XY and XX homozygous (Wt1 Arg495Gly/Arg495Gly ) kidney: Age-matched kidneys are significantly smaller. The cortex is architecturally homogeneous with no clear boundary between the outer cortex and juxtamedullary cortex. There is a significant lack of developed renal corpuscles and glomerular spaces are not observed. Furthermore, the number of renal corpuscles is subjectively lower than in the wild-type and Wt1 Arg495Gly/+ kidneys. Similarly, tubules are more immature in comparison with the wild-type and Wt1 Arg495Gly/+ kidneys. Morphological differences indicating differentiation towards proximal or distal convoluted tubules are not observed. Mesenchymal cells are abundant. Microscopic examination reveals significant morphological differences with wild-type and Wt1 Arg495Gly/+ kidneys. Although histologically normal, and considering that, as per history, they are age-matched organs, E16.5 Wt1 Arg495Gly/Arg495Gly kidneys do not contain developed glomerular corpuscles and tubules are underdeveloped. The low number and small size of both renal corpuscles and tubules account for the relative smaller size of the examined kidneys in comparison with Wt1 Arg495Gly/+ and wild-type. Otherwise there is no evidence of degeneration or inflammation in the examined sections.