Missense variant contribution to USP9X-female syndrome

Abstract

USP9X is an X-chromosome gene that escapes X-inactivation. Loss or compromised function of USP9X leads to neurodevelopmental disorders in males and females. While males are impacted primarily by hemizygous partial loss-of-function missense variants, in females de novo heterozygous complete loss-of-function mutations predominate, and give rise to the clinically recognisable USP9X-female syndrome. Here we provide evidence of the contribution of USP9X missense and small in-frame deletion variants in USP9X-female syndrome also. We scrutinise the pathogenicity of eleven such variants, ten of which were novel. Combined application of variant prediction algorithms, protein structure modelling, and assessment under clinically relevant guidelines universally support their pathogenicity. The core phenotype of this cohort overlapped with previous descriptions of USP9X-female syndrome, but exposed heightened variability. Aggregate phenotypic information of 35 currently known females with predicted pathogenic variation in USP9X reaffirms the clinically recognisable USP9X-female syndrome, and highlights major differences when compared to USP9X-male associated neurodevelopmental disorders.

Fig. 1. Conservation and protein location of likely pathogenic USP9X-female variants.

a Cross species protein alignment of USP9X showing conservation of altered amino acid residues. b Location of female likely pathogenic variants on the USP9X protein structure. USP9X variation tolerance landscape is provided (see Materials and Methods). Locations of male likely pathogenic variants and predicted deleterious missense somatic cancer variants (extracted from COSMIC database with CADD score ≥30) are shown for comparison.

Fig. 2. USP9X-female missense variants display in silico signatures of pathogenicity.

a Aggregate comparison of common variants extracted from gnomAD (allele frequency >1:100000), against females missense variants, likely pathogenic male missense variants, and variants found in cancer (extracted from COSMIC database) using a suite of in silico prediction tools. Box-whisker plots are defined as follows: centre line, median; box limits, upper and lower quartiles; whiskers, min and max values. *significantly different from common variants p < 0.05 by two-tailed equal variance Student’s t-test. b Comparison of CADD and PROVEAN scores reveal clustering of variants all female missense variants in the upper-right quadrant consistent with pathogenicity (CADD > 25, PROVEAN > 0.565). Scores of common variants are significantly correlated (Pearson’s correlation given). Colour scheme as in a. Inset identifies each variant in the ‘pathogenic quadrant’. Graphs show percent of each type of variant, and the overall composition of variant types within the pathogenic quadrant.

Fig. 3. Structural modelling of USP9X-female variants located in the catalytic domain.

a Homology model of USP9X (grey) with catalytic site (magenta), likely pathogenic female variants (red) and location of predicted deleterious cancer variants (blue; extracted from COSMIC database with CADD score >30). Interaction with ubiquitin is shown. Likely pathogenic variants are positioned in regions of well‐ordered secondary structure or flexible regions involved in zinc-binding. b Insets indicate local structural effects of indicated likely pathogenic female USP9X variants. All native amino acid side chains are represented as grey sticks. Variant amino side chains are indicated by red sticks. Side chains of the amino acids forming the core catalytic site are indicated by magenta sticks. Location of cancer variants is highlighted in blue. Zinc ion is represented with a yellow sphere. Hydrophobic van der vaals radii are indicated by dots and charge–charge interactions are shown by dotted lines.

Fig. 4. Phenotypic features of females with USP9X missense and single amino acid deletion variants.

a Constellation and frequency of major clinical features. b Magnetic resonance imaging (MRI) of brains of affected individuals. Note prominent extra-axial spaces in all, hypoplastic corpus callosum in Females 25 and 29, optic nerve atrophy and in Female 25, Dandy Walker malformation and Blake’s pouch cyst in Female 26, and ventriculomegaly of the 3rd and 4th ventricles in Female 29. c Images of affected females showing facial dysmorphisms, with common features including deep-set eyes, telecanthus, blepharophimosis, a broad nasal tip with wide alae and short collumnella, and low set and dysplastic ears. Written consent was obtained for the publication of photographs.

Fig. 5. Comparison of phenotypic features of females with different USP9X variant types and with males.

a Location of bona fide loss-of-function variants in individuals with USP9X-female syndrome. Three novel variants described in this study are annotated. b Images showing facial dysmorphisms of females affected by novel nonsense and frameshift USP9X variants. Note resemblance to individuals in Fig. 4c. c Comparison of clinical features of females with missense and single amino acid deletions (n = 12) variants with females harbouring null alleles including all previously published cases and three novel cases identified in this study (n = 23). Sankey plot highlights overlap across all features except short stature. d Comparison of clinical features of a combined female cohort (n = 35) with phenotypes of all published males with likely pathogenic variants (n = 16). Sankey plot highlights overlap across neurological features but not in other major female associated congenial phenotypes. Thickness of each stream is proportional to the percentage of cases with each feature, which is also provided numerically at the terminal nodes.