Craniosynostosis, inner ear, and renal anomalies in a child with complete loss of SPRY1 (sprouty homolog 1) function

Introduction SPRY1 encodes protein sprouty homolog 1 (Spry-1), a negative regulator of receptor tyrosine kinase signalling. Null mutant mice display kidney/urinary tract abnormalities and altered size of the skull; complete loss-of-function of Spry-1 in humans has not been reported. Methods Analysis of whole-genome sequencing data from individuals with craniosynostosis enrolled in the 100,000 Genomes Project identified a likely pathogenic variant within SPRY1. Reverse-transcriptase PCR and western blot analysis were used to investigate the effect of the variant on SPRY1 mRNA and protein, in lymphoblastoid cell lines from the patient and both parents. Results A nonsense variant in SPRY1, encoding p.(Leu27*), was confirmed to be heterozygous in the unaffected parents and homozygous in the child. The child’s phenotype, which included sagittal craniosynostosis, subcutaneous cystic lesions overlying the lambdoid sutures, hearing loss associated with bilateral cochlear and vestibular dysplasia and a unilateral renal cyst, overlapped the features reported in Spry1−/− null mice. Functional studies supported escape from nonsense-mediated decay, but western blot analysis demonstrated complete absence of full-length protein in the affected child and a marked reduction in both parents. Conclusion This is the first report of complete loss of Spry-1 function in humans, associated with abnormalities of the cranial sutures, inner ear, and kidneys.


INTRODUCTION
Studies in Drosophila first identified sprouty as an important negative regulator of fibroblast growth factor receptor (FGFR) and other receptor tyrosine kinase (RTK) signalling pathways. 1 There are four mammalian sprouty homologues (encoded by SPRY1-SPRY4), that in humans range in size from 288 to 319 amino acids (aa), and share three regions of homology: a canonical Cbl-tyrosine kinase-binding domain (aa 51-57), a serine-rich domain (aa 112-131) and a cysteine-rich domain (aa 181-306; numbering refers to Spry-1, see online supplemental figure 1A). 2 Sprouty proteins lack enzymatic activity but decrease activation of the RAS-MAPK pathway, potentially by binding and sequestering the adapter protein GRB2 following sprouty phosphorylation by RTKs at key tyrosine residues, thus preventing membrane localisation of SOS, a RAS guanine exchange factor. 3 Studies in mice have shown that homozygous deletion of the entire coding sequence of Spry1 (orthologous to human SPRY1) affects multiple developmental processes; heterozygous mice (Spry1 +/− ) are phenotypically normal. In the original report, 4 95% of homozygous neonates (mixed genetic background) displayed unilateral or bilateral ureter and kidney malformations, and 71% died within 5 months. These abnormalities were principally attributable to loss of negative feedback regulation of the RET RTK in the Wolffian duct, 4 and could be phenocopied by homozygous substitution of the critical tyrosine residue (Tyr53Ala) within the Cbl-tyrosine kinase-binding domain. 4 Notably, crossing onto a different mouse strain background (FVB) was associated with reduced penetrance of urogenital malformations in Spry1 −/− animals. 5

WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ Loss of sprouty homolog 1 (Spry-1) function in the mouse causes defects in the kidneys and urinary tract, altered skull size, and other phenotypes, but the consequence of loss of the human orthologue has not been described. Subsequently, detailed studies of Spry1 −/− mice have revealed additional defects in multiple organs and cell types, including rostral cortex, 6 mammary gland, 7 mesenchymal stem cells 8 and muscle satellite cells. 9 These phenotypes are exacerbated by adding loss-of-function (LoF) alleles of other Spry genes (notably Spry2), indicating partial functional redundancy between Spry proteins. 10 Depending on strain background, both increased and decreased skull size were documented in Spry1 −/− mice, but craniosynostosis was not reported. 11 Bona fide disease-causing variants of human SPRY1 have not been reported and therefore the consequence of complete LoF of the human orthologue was unknown. Here, we describe a homozygous SPRY1 nonsense variant, identified in a patient with syndromic craniosynostosis. We show that this results in complete loss of Spry-1 and is also associated with severe defects of the inner ear and mild kidney abnormalities.

Genetic investigation and variant filtering of whole-genome sequencing data
Following normal routine genetic investigations, the family was enrolled into the Genomics England 100,000 Genomes Project and Illumina whole-genome sequencing (WGS) was performed on the proband and both parents. 12 No clinically reportable variants were identified. The participant variant call format files were examined in the Genomics England Research Environment, as described. 13 Following annotation with Ensembl Variant Effect Predictor, the variants were filtered using a gnomAD (v.2.1.1) allele frequency of ≤0.01, 14 Combined Annotation Dependent Depletion (CADD) score ≥20 and no intersection with low complexity regions. Given the reported consanguinity, variants within homozygous regions identified using Illumina's ROHcaller were prioritised.

Functional analysis of cDNA and protein
Fresh blood was obtained from the proband and both parents to prepare genomic DNA and lymphoblastoid cell lines. Dideoxysequencing of genomic DNA was used to confirm the SPRY1 variant. RNA was obtained from lymphoblastoid cells, following which SPRY1 cDNA analysis (including deep sequencing) was performed by reverse transcriptase PCR (RT-PCR); protein was extracted for quantification of Spry-1 by western blotting. The antibody used to detect Spry-1 was rabbit mAb D9V6P (#13013, Cell Signaling Technology), detected using donkey anti-rabbit-horseradish peroxidase (HRP) (ab97085, abcam); anti-GAPDH-HRP (14C10, Cell Signaling Technology) was used as a loading control. Complete methods, including all primer sequences (online supplemental table 1) and PCR amplification conditions for genomic and cDNA analysis are provided in the online supplemental material.

Targeted resequencing of SPRY1
A search for additional variants in SPRY1 was undertaken by targeted resequencing of 617 individuals with previously undiagnosed craniosynostosis (online supplemental methods and online supplemental table 2).

Case report
The index patient (II-1) is the first-born male child of consanguineous parents (half-first cousins; online supplemental figure 2A) of Indian descent. Apart from familial hypercholesterolaemia in the mother (I-2), both parents are healthy and phenotypically normal.
Antenatal scans had shown increased nuchal translucency at 12 weeks and dolichocephaly from 20 weeks. Array comparative genomic hybridisation (aCGH) on an amniocentesis sample was normal. He was born at 39 weeks' gestation and noted at birth to have scaphocephaly, turricephaly, posterior ridging of the sagittal suture and bitemporal narrowing, with tense, symmetrically positioned cystic lesions (3×1 cm) over the parietal bones bilaterally. At 12 weeks of age, his length and weight were around the 50th centile; his head circumference could not be accurately measured owing to the cystic lesions. Distinctive facial features (figure 1A, 1B) included hypertelorism, a broad base to his nose, large ears with earlobe creases, and naevus flammeus over his philtrum and central forehead. There was a capillary malformation on his upper back. In his extremities, there was soft skin on his hands (figure 1C) with deep palmar and plantar creases, and a proximally placed overriding second toe on the right. CT scan of his head (aged 10 weeks) showed sagittal synostosis and bilateral subcutaneous cystic lesions over the lambdoid sutures. Aged 4 months, he had spring-assisted cranioplasty (removed at 8 months) and scalp lesions removed (shown histologically to be dermoid cysts).
He failed his newborn hearing screen and was subsequently found to have bilateral profound sensorineural hearing loss. MRI scan of his inner ear (aged 8 months) demonstrated bilateral inner ear dysplasia. On the right, there was cochlear hypoplasia with anterior off-set markedly hypoplastic apical turn (cochlear hypoplasia 4), similar to the unwound cochlea found in branchiootorenal syndrome. The vestibule and semi-circular canals were relatively normal. On the left there was an incomplete partition Developmental defects type 1 malformation of the cochlea with abnormal vestibule and semi-circular canals ( figure 1D,E). The VIII nerves were normal bilaterally. He underwent bilateral cochlear implantation in a two-stage procedure (aged 15 and 18 months). MRI of the brain showed a small corpus callosum, which was dysmorphic posteriorly ( figure 1F).
A renal ultrasound and renogram showed a simple left renal parapelvic cyst, while the right kidney appeared normal. He has normal renal function and is normotensive with no microalbuminuria; testes and external genitalia are normal. Echocardiogram at 3 months of age showed a small atrial septal defect (which had resolved spontaneously by 4 years) and mild supravalvular pulmonary stenosis of no clinical significance.
His speech and language development were delayed, consistent with his hearing loss. However, his motor and cognitive development is normal. At 4 years of age, he uses three-word sentences and follows complex instructions. Ophthalmology review was normal, with no concerns regarding his vision.

Molecular analysis
Initial genetic investigations of the proband included normal postnatal aCGH and craniosynostosis gene panel testing. Analysis of WGS data from the 100,000 Genomes Project showed that his genome harboured five regions of homozygosity ≥2 Mb (online supplemental figure 2B); after filtering three homozygous variants remained, including one stop-gain (in SPRY1) and two predicted missense variants (in CCDC80 and RASSF6). Further evaluation of the CCDC80 and RASSF6 homozygous variants, as well as two heterozygous de novo missense variants (in ATXN10 and POMP), did not support pathogenicity of these variants (online supplemental table 3), leaving the SPRY1 variant as the sole candidate to be prioritised for further analysis. Dideoxy-sequencing confirmed the homozygous nonsense variant (c.80T>A, encoding p.(Leu27*) in SPRY1), which was heterozygous in both parents ( figure 2A).
The premature termination codon, p.Leu27*, arises in the last and only coding exon of SPRY1 (online supplemental figure 1A) and is therefore predicted to escape nonsense-mediated decay (NMD). 15 To confirm this, we undertook RT-PCR analysis of RNA extracted from lymphoblastoid cells and, in support, we identified RNA products in all three family members ( figure 2B). The presence of the variant in the cDNA in approximately equal quantity to the normal allele was confirmed by deep sequencing of RT-PCR products from the parental samples, consistent with normal production of mutant RNA transcript and escape from NMD in lymphoblastoid cells (online supplemental table 4).
To assess the effect of the variant on Spry-1 protein synthesis, we undertook western blot analysis and identified a band of ~35 kDa in all healthy controls and both parents, corresponding to the expected size of Spry-1 (figure 2C). Of note, the antibody used was raised to an epitope comprising residues surrounding aa 70 of human Spry-1, located C-terminal of p.Leu27 (online supplemental figure 1A). Accordingly, the predicted truncated protein appeared entirely absent in the proband; band intensity was considerably weaker in both heterozygous parents compared with controls, consistent with loss of one copy of full length Spry-1. Results of replicate analyses and additional controls are shown in online supplemental figure 3.

DISCUSSION
To our knowledge, this is the first description of a patient harbouring a homozygous nonsense variant in SPRY1. This variant is predicted to affect all known SPRY1 splice forms and to lead to a severely truncated Spry-1 protein that has lost all conserved features. The complete loss of full-length functional Spry-1 protein in the proband was corroborated by western blot analysis ( figure 2C). The parents, who were phenotypically normal, were heterozygous for the variant and exhibited diminished amounts of full-length Spry-1 protein.
Major points of interest arising from this work are the delineation of human SPRY1 LoF phenotypes in homozygous and heterozygous states, and comparison with the mouse null mutant. The patient shares many phenotypes previously reported in mouse Spry1 -/mutants (see 'Introduction' section), but with apparent differences in relative severity. Renal and ureteric anomalies, which are a major feature of the murine Spry1 -/phenotype and often lead to reduced viability, were manifested in the patient only as a unilateral simple parapelvic cyst, which did not compromise renal function. By comparison, the patient exhibited major developmental abnormalities of the inner ear, affecting both the cochlea and vestibular systems; whereas inner ears of Spry1 -/mice are normal, 16 but manifest similar anatomical features when crossed with Spry2 LoF alleles. 17 Given current evidence that the renal and inner ear abnormalities are predominantly driven by dysregulated signalling through different growth factor/RTK axes (GDNF/RET and FGF/FGFR, respectively), it is possible that these signalling pathways differ in their relative sensitivity to Spry-1 inhibition in the mouse and human. A failure to repress FGF/FGFR signalling is consistent with the presence of sagittal synostosis in the patient, since dysregulated FGF/FGFR signalling is a well-documented cause of this phenotype. 18 The patient manifested several additional phenotypes including dysmorphic facial features, bilateral dermoid cysts over the scalp and minor brain anomalies; the identification of additional cases is required to determine whether these fit into a consistent pattern.
We are aware of two previous reports of putatively pathogenic heterozygous SPRY1 variants-a frameshift (p.(Gln6fs)) 19 in a patient with sagittal synostosis, and a stop-gain (p.(Glu79*)) 20 in a patient with Noonan syndrome (OMIM #609942). Two arguments suggest that these variants alone are unlikely to account for the associated phenotypes reported. First, the gnomAD dataset (v.2.1.1) identifies 13 individuals out of ~140 000 sequenced, including normal controls, heterozygous for LoF variants in SPRY1. This relatively high frequency conflicts with such alleles having highly penetrant pathogenic effects and accords with the low probability of LoF intolerance score (pLI=0). Second, in the current study, both parents were heterozygous for the LoF variant in SPRY1 but displayed no relevant clinical phenotype. Importantly, no homozygous LoF variants are recorded in the gnomAD v.2.1.1 or v.3.1.1 datasets.
Based on similarities of the patient's phenotype with the mouse Spry1 -/mutant and absence of other credible pathogenic variants in the patient, we propose that complete loss of Spry-1 likely accounts for the abnormal phenotypes described. Further corroboration of genotype-phenotype correlations will require the identification of additional individuals harbouring homozygous variants in SPRY1; however, our attempts to identify such individuals through GeneMatcher, or by sequencing of 617 patients with unsolved craniosynostosis, were unsuccessful.
In conclusion, we present the first evidence of a SPRY1 knockout in humans. Although these observations are based on a single case study, there are substantial overlaps with the reported phenotypes in Spry1 -/mutants in mice. Together with functional evidence showing loss of full-length protein within the proband, and lack of any other identified genetic pathology, we propose that the SPRY1 variant is likely causative.

Developmental defects
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Open access This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/ licenses/by/4.0/.