NR5A1 gene variants repress the ovarian‐specific WNT signaling pathway in 46,XX disorders of sex development patients

Abstract Several recent reports have described a missense variant in the gene NR5A1 (c.274C>T; p.Arg92Trp) in a significant number of 46,XX ovotesticular or testicular disorders of sex development (DSDs) cases. The affected residue falls within the DNA‐binding domain of the NR5A1 protein, however the exact mechanism by which it causes testicular development in 46,XX individuals remains unclear. We have screened a cohort of 26 patients with 46,XX (ovo)testicular DSD and identified three unrelated individuals with this NR5A1 variant (p.Arg92Trp), as well as one patient with a novel NR5A1 variant (c.779C>T; p.Ala260Val). We examined the functional effect of these changes, finding that while protein levels and localization were unaffected, variant NR5A1 proteins repress the WNT signaling pathway and have less ability to upregulate the anti‐testis gene NR0B1. These findings highlight how NR5A1 variants impact ovarian differentiation across multiple pathways, resulting in a switch from ovarian to testis development in genetic females.

Despite these known causes, SRY-negative 46,XX (ovo)testicular DSDs have a diagnostic rate much lower than that of other DSDs (Eggers et al., 2016) highlighting a need to identify novel genes or genetic variants underlying these DSDs. Recently, five reports have described a recurrent heterozygous variant in the Nuclear Receptor Subfamily 5 Group A Member 1 (NR5A1) gene (c.274C>T; p.Arg92Trp) in 12 cases of 46,XX (ovo)testicular DSD (Baetens et al., 2017;Bashamboo et al., 2016;Domenice et al., 2016;Igarashi et al., 2017;Takasawa et al., 2017), although the underlying mechanism for how this NR5A1 variant causes sex reversal in these cases is still unclear. NR5A1, or steroidogenic factor-1 (SF1, Ad4BP; MIM# 184757), is a nuclear receptor transcription factor with a key regulatory role in development of the hypothalamus-pituitarygonadal and hypothalamus-pituitary-adrenal axes (Morohashi et al., 1996;Parker et al., 1997;Val, Lefrancois-Martinez, Veyssiere, & Martinez, 2003), and thus in the development of the male and female gonads (reviewed in Ferraz-de-Souza, Lin, & Achermann, 2011;Lin et al., 2008). Human variations in NR5A1 have been widely characterized in 46,XY DSDs (gonadal dysgenesis, hypospadias, undervirilization, and male infertility; Robevska et al., 2018;Ropke et al., 2013) and 46,XX primary ovarian insufficiency (Voican et al., 2013) and commonly show variable phenotypic expressivity and incomplete penetrance. In a recent study, we applied massively parallel sequencing of a targeted gene panel to a DSD cohort, which included 26 patients with 46,XX (ovo)testicular DSD (Eggers et al., 2016). From this study, we identified three new cases of 46,XX (ovo)testicular DSD with the NR5A1 c.274C>T; p.Arg92Trp variant. Furthermore, we report an additional NR5A1 variant, c.779C>T; p.Ala260Val, in a case of 46,XX ovotesticular DSD. To our knowledge, this is the first instance of an NR5A1 variant that does not affect codon 92 occurring in an individual with 46,XX (ovo)testicular DSD. We have used protein modeling, protein localization, and luciferase assays to propose mechanistic insights for NR5A1-mediated XX sex reversal in these cases.

Ethical approval and patient recruitment
Patient recruitment, consent, and DNA extraction were carried out as described previously (Eggers et al., 2016). Ethical approval for this study was obtained from the Human Ethics Committee of the Faculty of Medicine at the Royal Children's Hospital, Melbourne, Victoria, Australia (HREC22073).

Massively parallel sequencing
Total genomic DNA was sequenced on a targeted panel (HaloPlex, Agilent, Santa Clara, CA) that includes 64 diagnostic DSD genes (described in Eggers et al., 2016). This included six genes that have been implicated in 46,XX (ovo)testicular DSD (FGF9, RSPO1, SOX3, SOX9, SOX10, and WNT4). NR5A1 variant numbering is based on Gen-Bank reference DNA sequence NM_004959.4, with the A of the ATG initiation codon designated +1. NR5A1-predicted protein annotations are based on NP_004950. Analysis for genomic modifiers was performed by filtering variants using a list of 116 NR5A1-related genes alongside our previously reported filtering criteria. The initial gene list (N = 576) was compiled from data in STRING (https://string-db.org/) and NR5A1 overexpression/knockdown assays (Doghman, Figueiredo, Volante, Papotti, & Lalli, 2013), 116 of these genes were covered by our targeted gene panel (N = 1,024).

Luciferase assays
Luciferase reporter assays were performed in two cell lines (HEK 293-T or COS-7) in 96-well plates using Lipofectamine 2000 and the Dual-Luciferase Reporter 1000 Assay System Kit (Promega; Fitchburg, WI), with co-transfection of Renilla (pRL-TK) as a marker of transfection efficiency. To assay the NR5A1-mediated transactivation of mTESCO, COS-7 cells were co-transfected with a reporter construct, pGL4-mTESCO (75 ng/well), pRL-TK (10 ng/well), SOX9 (50 ng/well), NR5A1 (WT, p.Arg92Trp, p.Ala260Val and p.Arg84His which was used as a control; 50 ng/well), and FOXL2 (50 ng/well); cells were harvested 24 hours post transfection. The mTESCO reporter was used to assay NR0B1-mediated repression of SOX9, as previously described . Interaction of NR5A1 with NR0B1 was assayed in HEK 293-T cells using a pGL4-hDAX1 reporter as previously described (Mizusaki et al., 2003). To assay the ability of NR5A1 proteins to regulate the canonical WNT pathway, we used the TOPFlash-TCF reporter as previously described (Bashamboo et al., 2016). Luciferase activity was measured on an Infinite M200 Pro plate reader (Tecan; Männedorf, Zürich, Switzerland); data represent the mean with standard error of three independent experiments, each performed in triplicate.

Clinical phenotype
All of the individuals presented have a 46,XX karyotype (SRY-negative), with varying degrees of virilization (Table 1). Patient 1 was born to nonconsanguineous parents and recruited as an adult male with azoospermia and bilateral testes; he had undergone an orchidopexy in childhood. Patient 2 presented at birth with ambiguous genitalia, a phallus measuring 2 cm, perineal urethral opening, and bilaterally palpable gonads. Histological evaluation showed them to be bilateral ovotestes. There were no Müllerian structures detected, and testosterone production declined over time. The child underwent orchidopexy and was raised as a female. Patient 3 presented with virilization and was raised as a male; he had a micropenis (stretched penile length of 3 cm at 5 years), hypospadias, and small underdeveloped scrotum with no evidence of Müllerian structures and a positive testosterone response to stimulation. This individual also has a seizure disorder managed by neurologists. Patient 4 was initially raised as a male, then as female from 4 years of age. The patient presented with ambiguous genitalia, a small phallus and vagina (4 cm deep), with separate urethral and vaginal openings. The right gonad was an ovotestis with sporadic germ cells in the tubules, calcifications, and primordial follicles in ovarian part. The left gonad was ovarian tissue with both primordial and developing follicles. Patients 1-3 were found to have normal adrenal function based on hormonal profiling at the time of sample collection, these data were not available for Patient 4.

Identification of novel and known variants in NR5A1 associated with 46,XX (ovo)testicular DSD
46,XX (ovo)testicular DSD cases were screened using two assays.
Patient DNA was run on our custom multiplex ligation-dependent probe amplification (MLPA) assay (previously described in Ohnesorg et al., 2017), which screens 10 genes implicated in DSD for copy number variants, including four probes in the enhancer region of SOX9. The MLPA showed that Patients 1-4 are negative for common copy number variants in known DSD genes. We then performed massively parallel sequencing on a targeted DSD gene panel (previously described in Eggers et al., 2016) to screen the patients for single nucleotide variants and small insertions/deletions in 64 diagnostic DSD genes. A pre-viously reported variant in exon 4 of NR5A1 (NM_004959.4:c.274C>T, p.Arg92Trp; Baetens et al., 2017;Bashamboo et al., 2016;Igarashi et al., 2017) was identified in three cases of 46,XX (ovo)testicular DSD (Table 2; Patients 1-3). The change was found to be maternally inherited in Patient 3. The mother was subfertile, having 4-6 menses per year and had difficulty falling pregnant. Interestingly, the maternal aunt and uncle of Patient 3 also experienced fertility issues. A further variant (NM_004959.4:c.779C>T, p.Ala260Val) was found in exon 4 of NR5A1 in a single case of 46,XX ovotesticular DSD (Patient 4). The NR5A1 p.Ala260Val variant was extremely rare in the population (gno-mAD frequency: 4.13e-6) and predicted to be damaging in two of the four in silico algorithms used. It has been previously implicated in 46,XY DSD (Chan et al., 2015). These two NR5A1 variants were found in heterozygous form and flagged for functional validation.

NR5A1 variants have unaffected protein expression and subcellular localization
To assess whether the NR5A1 variants affected protein localization or expression, we used immunofluorescence staining ( Figure 2b).
The wild-type and p.Arg92Trp variant NR5A1 protein showed strong nuclear expression with nucleolar exclusions (Figure 2b: ii and iv), consistent with previous reports (Baetens et al., 2017;Kohler et al., 2008).
We showed that the p.   Figure 3a). This loss of activation was also evident in the presence of the SOX9-inhibiting factor Forkhead Box L2 (FOXL2), indicating that the variant NR5A1 proteins do not affect the FOXL2-mediated repression of SOX9 (Figure 3a).

NR5A1 variants alter ovarian pathway activation via reduced WNT signaling activity
As the above assay does not explain why the male pathway is activated and ovarian pathway repressed in these 46,XX (ovo)testicular patients, we turned to female pathways. Previous functional assays on the NR5A1 p.Arg92Trp variant show disruption to both ovarian and testicular pathways (Bashamboo et al., 2016;Igarashi et al., 2017).
NR0B1 is a key gene involved in repression of the testis pathway; its dysregulation may explain these XX phenotypes. In ovarian development, NR5A1 transactivation of SOX9 is repressed by NR0B1 Ludbrook et al., 2012) proposed that mutant NR5A1 is less responsive to NR0B1. We assessed this by co-transfecting HEK 293-T cells with NR0B1, SOX9, and NR5A1 constructs as well as the mTESCO Sox9 enhancer (Figure 3b). In the presence of wild-type NR5A1 and SOX9, addition of NR0B1 resulted in dosage-dependent repression of Sox9 mTESCO. Similarly, using both mutant forms of NR5A1, Sox9 mTESCO was also repressed in a NR0B1-dosage-dependent manner (Figure 3b), although the initial activation of the reporter was lower for both variants as seen in Figure 3a. Thus, these variants are still responsive to NR0B1-mediated repression of SOX9. Bashamboo et al. (2016) also investigated NR0B1 dysregulation.
It has been suggested that mutant NR5A1 causes dysregulation of NR0B1 and thus loss of SOX9 repression in XX gonads (Bashamboo et al., 2016). We assessed the effect of mutant NR5A1 on NR0B1mediated repression of SOX9. To assay any direct effect of these variants on the NR0B1 promoter, we used a reporter construct containing the upstream NR0B1 promoter region (994 bp). HEK 293-T cells were co-transfected with this NR0B1 reporter and constructs for wild-type and variant NR5A1 (Figure 3c). Wild-type NR5A1 upregulated NR0B1 promoter activity and both NR5A1 variants showed similar upregulation of the NR0B1 promoter, consistent with the above assay showing that the variants do not change NR5A1-mediated NR0B1 repression of SOX9. To test the NR5A1/ -catenin complex in NR0B1 upregulation, we introduced -catenin. -catenin and wild-type NR5A1 showed a fivefold increase in NR0B1-reporter activity compared the empty vector controls, but the patient variants significantly repressed this activity (Figure 3d). To further test whether the NR5A1 variants are repressing the WNT/ -catenin pathway, we looked at the effect of wild-type and variant NR5A1 on canonical WNT activity using the TOPFlash reporter system (Figure 3e). The TOPFlash reporter shows a 20-fold induction upon the transfection of the wild-type -catenin construct. Introduction of the wild-type NR5A1 repressed WNT signaling activity induced by -catenin, and we found that both variant forms of NR5A1 repressed this further (P < 0.0001). As both variants present in the patient are heterozygous, we also transfected each variant with an equal amount of wild-type construct. Increased repression of WNT signaling was still observed when each variant NR5A1 was transfected with or without the wild-type form, indicating some dominant negative effect of the variant allele on the WT. Taken together, these results indicate that variants in NR5A1 show no reduced NR0B1-dependent repression of SOX9 expression or changes in NR0B1 promoter activation, yet these variants have increased repression of the ovarianspecific WNT signaling pathway. This may underlie testis development in an XX background.

Additional genomic variants may contribute to oligogenic inheritance
Given the wide phenotypic variation observed in individuals with NR5A1 variants, recent reports hypothesize that oligogenic inheritance is likely to be at play (Camats, Fernandez-Cancio, Audi, Schaller, & Fluck, 2018;Robevska et al., 2018). We filtered our massively parallel sequencing data with a list of 116 NR5A1-related genes to identify variants that may act additively with the NR5A1 variants in these four patients. Sixteen NR5A1-associated variants were identified and summarized in Supporting Information

DISCUSSION
During gonadal differentiation, NR5A1 is involved in both activation and repression of the testis pathway. Variants in the NR5A1 gene can cause a wide variety of DSDs including 46,XY gonadal dysgenesis, 46,XX premature ovarian insufficiency, and recently a single variant (p.Arg92Trp) has been implicated in 46,XX (ovo)testicular DSD (Baetens et al., 2017;Bashamboo et al., 2016;Domenice et al., 2016;Igarashi et al., 2017;Takasawa et al., 2017  mutant is unresponsive to NR0B1, resulting in a loss of SOX9 repression in 46,XX (ovo)testicular DSD cases. By contrast, we showed that both forms of mutant NR5A1 are still responsive to NR0B1 and still allow repression of SOX9 activity, suggesting that this anti-testis interaction is maintained in the 46,XX DSD patients. These contradictory results may be in part due to different cell types being used, performing this assay in the native (human fetal ovary) cells would be necessary to establish whether NR0B1 de-repression is truly playing a role in these phenotypes. Bashamboo et al. (2016) proposed an alternate mechanism whereby the NR5A1 p.Arg92Trp variant has less ability to upregulate the anti-testis NR0B1, perhaps via a loss of synergy with -catenin. Ultimately, this would mean that testis factors, such as SOX9, are no longer suppressed. We found that both NR5A1 variants showed similar activation of the NR0B1 promoter. However, we was assessed in COS-7 cells with an NR5A1 antibody (green). Cells were transfected with an equal amount of NR5A1 expression vector (wild-type or variant). Nuclear counterstaining was performed with DAPI (blue). Wild-type NR5A1 showed strong nuclear staining with nucleolar exclusions (b: i and ii). The variant NR5A1 protein expression and localization was unaffected (b: iii-vi) did find that compared to wild-type, the two variants repressed the -catenin-mediated activation of this promoter construct, indicating a repression of WNT signaling might be at play. In ovarian differentiation, NR5A1 is also necessary for induction of female pathway genes Wnt4 and Rspo1 (Combes et al., 2010). Bashamboo et al. (2016)  . This is also observed when the female pathway SOX9 repressor, FOXL2, is also transfected. (b) Co-transfection of HEK 293-T cells with NR5A1, SOX9, and increasing concentrations of NR0B1 showed that mutant NR5A1 does not affect NR0B1-mediated repression of SOX9. SOX9 activity was measured using the TESCO reporter. (c) Co-transfection of COS-7 cells with wild-type or mutant NR5A1 shows no change in activity of the NR0B1 promoter for both NR5A1 mutants (d) Co-transfection of COS-7 cells with wild-type or mutant NR5A1 and -catenin results in repression of the NR0B1 promoter for both NR5A1 mutants. (e) TOPFlash activation is reduced when HEK 293-T cells are transfected with -catenin and mutant NR5A1 compared to wild-type NR5A1. Data represent the mean with the standard error of three independent experiments performed in triplicate. An unpaired t-test was applied to obtain P-values, ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05; ns = P > 0.05 frequency in large population databases (e.g., gnomAD) and individuals with the NR5A1 p.Arg92Trp variant have diverse ethnic backgrounds, including African, Hispanic, European, and Asian, suggesting that the variable phenotypic expressivity observed is not a product of founder effect or genetic background. Additional genomic variants may act as modifiers of expression and thus explain some of the variable expressivity observed. A recent report on 46,XY DSDs with NR5A1 variants explored oligogenic inheritance by filtering exome sequencing variants with a list of NR5A1-associated genes (Camats et al., 2018). Using a similar, albeit targeted, approach, we identified an additional 2-6 variants per patient that may modify NR5A1 expression or act additively with NR5A1 to generate the wide phenotypic variation observed (Supporting Information Table S2 variants have been reported in a case of 46,XX testicular DSD before . Given that NR5A1 and POR are both involved in steroidogenesis, these two variants may have an additive effect on steroidogenic function; this may apply to other variants we report in known steroidogenic genes including STAR, PTCH1, and HSD3B1. Functional assessment would be required to confirm their pathogenicity and interaction with NR5A1 in these phenotypes. In future, exome sequencing on these and additional individuals (affected and unaffected) with NR5A1 variants would enable a genome-wide and unbiased approach to investigating oligogenic inheritance.
In our cohort of patients with SRY-negative 46,XX (ovo)testicular DSD, NR5A1 variants contribute to 15% (4/26) of cases. This is comparable to SOX9 enhancer duplications, which underlie 19% (5/26) of cases in this cohort (Croft et al., 2018;Ohnesorg et al., 2017). Consequently, NR5A1 gene variants should be considered an important cause underlying cases of SRY-negative 46,XX (ovo)testicular DSDs. Screening for NR5A1 gene variants should be included in routine genetic testing for these patients.