Oligogenic analysis across broad phenotypes of 46,XY differences in sex development associated with NR5A1/SF-1 variants: findings from the international SF1next study

Abstract

Background: Oligogenic inheritance has been suggested as a possible mechanism to explain the broad phenotype observed in individuals with differences of sex development (DSD) harbouring NR5A1/SF-1 variants.

Methods: We investigated genetic patterns of possible oligogenicity in a cohort of 30 individuals with NR5A1/SF-1 variants and 46,XY DSD recruited from the international SF1next study, using whole exome sequencing (WES) on family trios whenever available. WES data were analysed using a tailored filtering algorithm designed to identify rare variants in DSD and SF-1-related genes. Identified variants were subsequently tested using the Oligogenic Resource for Variant Analysis (ORVAL) bioinformatics platform for a possible combined pathogenicity with the individual NR5A1/SF-1 variant.

Findings: In 73% (22/30) of the individuals with NR5A1/SF-1 related 46,XY DSD, we identified one to seven additional variants, predominantly in known DSD-related genes, that might contribute to the phenotype. We found identical variants in eight unrelated individuals with DSD in DSD-related genes (e.g., TBCE, FLNB, GLI3 and PDGFRA) and different variants in eight genes frequently associated with DSD (e.g., CDH23, FLNB, GLI2, KAT6B, MYO7A, PKD1, SPRY4 and ZFPM2) in 15 index cases. Our study also identified combinations with NR5A1/SF-1 variants and variants in novel candidate genes.

Interpretation: These findings highlight the complex genetic landscape of DSD associated with NR5A1/SF-1, where in several cases, the use of advanced genetic testing and filtering with specific algorithms and machine learning tools revealed additional genetic hits that may contribute to the phenotype.

Funding: Swiss National Science Foundation and Boveri Foundation Zurich.

Keywords: 46,XY DSD; Differences of sex development (DSD); Oligogenicity; Steroidogenic factor 1 (SF-1/NR5A1).

Fig. 1

Summary of the NR5A1/SF-1 variants of individuals analysed by newly performed WES analysis for this study, derived from international SF1next study cohort. Identified variants in the NR5A1 gene are shown with respect to the gene and protein sequence. The SF-1 protein comprises the DNA-binding domain, which contains two zinc fingers (Zn1 and Zn2), a Fushi-tarazu factor-1 (FTZ-F1) box, the accessory hinge region, and the ligand-binding domain. It harbours two activation functional (AF) domains, activation function 1 (AF-1) and activation function 2 (AF-2). NR5A1, nuclear receptor subfamily 5 group A member 1; UTR, untranslated region.

Fig. 2

Bar plot of the VarCoPP score of 64 rare gene variants found in oligogenic combinations with NR5A1/SF-1 variants in 22 cases with DSD. The pathogenicity score (VarCoPP score) generated by ORVAL’s VarCoPP tool represents the probability (value between 0 and 1) that a variant combination belongs to the disease-causing class. If this score is above 0.4575 (hg38), the model predicts that the combination is disease-causing. For stricter analysis a pathogenicity score ≥0.85 (hg38) (dotted red line) was set as the threshold to include only gene pairs with combinations falling into the 99.9%-confidence zone. For one candidate variant (NR1H2, p.Arg171_Lys172insAsn) no prediction was found in ORVAL. Predicted digenic effect by ORVAL’s digenic effect predictor tool^, for the combination of NR5A1/SF-1 variants of each case with the additional variants are indicated by two colours: True digenic combination (blue), where the simultaneous presence of a pathogenic allele in each gene is necessary for the individual to express the disease phenotype. Monogenic and Modifier combination (violet), where a variant on the major gene induces a disease phenotype, while a mutation in the modifier gene modifies it.

Fig. 3

Stacked bar plot for 22 cases with DSD harbouring NR5A1/SF-1 variants showing the predicted pathogenicity of 65 variants based on Franklin (aqua green) and VarSome (blue) classifications. Clinical significance is given according to ACMG criteria for variants classification: 1 (Benign), 2 (Likely Benign), 3 (Variant of Uncertain Significance, VUS), 4 (Likely Pathogenic), 5 (Pathogenic). Note that none of the variants were predicted pathogenic or likely pathogenic, but need to be included when considering oligogenicity. Symbols indicate model cases that are described in detail in the text.

Fig. 4

Summary of common pathways identified between the NR5A1 gene and 14 other genes, in which additional variants were found in our study participants. The analysis was performed with Reactome and the visualisation with Cytoscape.

Fig. 5

Family trees showing the individual gene variants (left panel) and the predicted oligogenic network (right panel) in five model cases (a–e). The pedigrees depict the inheritance patterns of the identified variants. Note that all variants were observed in a heterozygous state. The networks created for each case by ORVAL inform on predicted gene interactions necessary to reveal a disease phenotype.