Retinal dystrophy in an individual carrying a de novo missense variant of SMARCA4

Abstract Background Coffin–Siris syndrome (CSS) is characterized by intellectual disability, dysmorphic facial features, growth deficiency, microcephaly, and abnormalities of the fifth fingers/toes. CSS is caused by mutations in several genes of the BRG1‐associated factor pathway including SMARCA4. Methods Whole‐exome sequencing was performed on a 14‐year‐old female individual who presented with mild intellectual disability and dysmorphic features, tooth abnormalities, and short stature. She had brachydactyly but no aplasia or hypoplasia of the distal phalanx or nail of the fifth digit. She was also found to have retinal dystrophy that has not been previously reported in CSS. Results The individual presented herein was found to harbor a previously unreported de novo variant in SMARCA4. Conclusion This case expands the phenotypic spectrum of CSS manifestations.

variants are mostly missense and are localized within the three central domains of the protein Helicase/SANT-associated, DEAD-like helicase and Helicase C-terminal domains (Bramswig et al., 2015;Errichiello et al., 2017;Kosho, Okamoto et al., 2014;Tsurusaki, Okamoto, et al., 2014;Tzeng, du Souich, Cheung, & Boerkoel, 2014). Here, we report an individual harboring a de novo missense SMARCA4 variant falling outside the central domains who presented with mild intellectual disability, mild and not distinctive dysmorphic features, short stature, tooth agenesis, and retinal dystrophy. This case expands the spectrum of phenotypic abnormalities of CSS.

| CASE REPORT
The child was born to non-consanguineous parents by caesarean section after 39 weeks of gestation with a birth weight of 2.700 g. She was able to walk independently at 13 months; by 18 months she could say a few words, and she was toilettrained at 3 years of age. She had learning difficulties and at the time of the last evaluation, when she was attending third year of high school, she could read and write but she had difficulties with simple calculations. A WISC-III testing performed when she was 10 years old revealed an IQ of 59. She has good social interactions and had no history of seizures.
At the age of 14 years and 5 months, her weight was 58 kg (75 th centile), height 148 cm (2 nd centile) and occipitofrontal circumference 55 cm (55 th centile). She showed mild facial dysmorphic features with bulbous nasal tip and micrognathia, and brachydactyly without hypoplasia of the distal phalanx or hypoplastic nails of hands or feet ( Figure 1a). She had tooth abnormalities and agenesis confirmed by X-rays (Figure 1b), and she never underwent dental extraction procedures. She had normal sweating.
By the age of 9 years, she was noted to have night-blindness and at the age of 14 years, she was found to have a best-corrected visual acuity of 20/40 in both eyes (right eye [RE], cylinder −1 alpha [axis] 100° left eye [LE], sphere + 1.25 = cylinder −2.25 alpha 90°), abnormal color vision, and normal ocular motility. Lenses were clear. Fundus examination revealed a pink optic disc, widespread dystrophy of retinal pigment epithelium (RPE) with pigment deposits resembling bone spicules in mid-periphery. Retinal vasculature was normal. Fundus autofluorescence showed a ring of macular hyper-autofluorescence ( Figure 1c). Optical coherence tomography (OCT) performed with the spectral domain OCT (Cirrus HD-OCT; Carl Zeiss, Dublin, CA) showed RPE dystrophy with vitreo-retinal interface syndrome ( Figure 1d). Electroretinography revealed scotopic and photopic traces below noise level. Auditory brain steam response and EEG did not detect any abnormalities. Brain MRI performed when she was 10-years old only showed an arachnoid cyst at the ponto-cerebellum junction. The echocardiogram revealed a mild pulmonary insufficiency while abdomen ultrasound was normal.  Chromosome microarray analysis did not detect pathogenic chromosomal rearrangements and Next Generation Sequencing (NGS) and MLPA of a panel of genes responsible for ectodermal dysplasia was negative. NGS of 137 genes responsible for inherited retinal dystrophies also did not detect pathogenic variants. Following informed consent, the proband was enrolled in the Telethon Undiagnosed Diseases Program and genomic DNA from the proband and both her parents underwent whole-exome sequencing (WES). Genomic DNA was enriched using SureSelect Clinical Research Exome (Agilent, Technologies, Santa Clara, CA) and sequenced with the NextSeq500 sequencing system (Illumina, San Diego, CA). A custom pipeline based on Burrows-Wheeler Alignment tool (BWA) Genome Analysis Toolkit (GATK), and ANNOVAR (Wang, Li, & Hakonarson, 2010) were used to call, annotate, filter, and prioritize variants. WES identified in the proband a de novo heterozygous variant c.4297G >A in exon 31 of SMARCA4 resulting in the p.(Glu1433Lys) amino acid change (NM_001128849, chr19: g.11,152,013G > A). The variant was confirmed by Sanger sequencing in the proband and was found to be absent in both parents. The affected amino acid residue: (a) is located between the Helicase C-Terminal domain and the Bromodomain, (b) it affects an evolutionarily highly conserved residue, (c) is not reported neither in ExAc nor in GnomAD, (d) is predicted to be pathogenic by Polyphen, and (e) has a CADD score of 28.8. Taken together and based on current guidelines (Richards et al., 2015), this SMARCA4 variant is classified as likely pathogenic.

| DISCUSSION
Components of BAF complex modulate gene expression and cell differentiation via nucleosome remodeling. Therefore, loss of functional BAF complex can affect gene expression resulting in pleiotropic phenotypic manifestations . The individual herein reported presented with a mild CSS phenotype and showed retinal dystrophy that has not been previously reported in CSS. She had brachydactyly but no aplasia or hypoplasia of the distal phalanx or nail of the fifth digit. Moreover, her dysmorphic features were mild and non-specific. Minor dental anomalies such as small, conic, pointed, widely spaced teeth along with delayed dentition have been reported in CSS individuals (Hoyer et al., 2012;Wieczorek et al., 2013) including subjects carrying SMARCA4 variants (Errichiello et al., 2017). Therefore, tooth abnormalities we observed in our case are part of the spectrum of dental anomalies of CSS. Noteworthy, variants in genes responsible for oligodontia and ectodermal dysplasia (e.g., PAX9, EDA, MSX1, AXIN2, EDARADD, NEMO, and KRT17) (Ye & Attaie, 2016) were not detected by either the targeted sequencing or WES.
Eye abnormalities have been reported in CSS individuals including subjects with SMARCA4 variants, and they include severe myopia, strabismus, microphthalmia, and spherophakia (Errichiello et al., 2017;. However, retinal pigmentary defects were not previously reported neither in individuals harboring SMARCA4 variants nor other genes of the BAF complex. Both targeted sequencing of genes responsible for inherited retinopathies and WES ruled out inherited retinal disorders in our case. Interestingly, the retinal defect detected in the individual herein described is paralleled by a disorganized retinal structure, abnormal retinal lamination, and disrupted RPE pigmentation observed in smarca4 null zebrafish model (Zhang et al., 2014;Zhang, Bonilla, Chong, & Leung, 2013), suggesting that retinal dystrophy is indeed due to SMARCA4 defects.
SMARCA4 missense variants in CSS individuals are largely clustered within the three central domains of the protein (Helicase/SANT-associated, DEAD-like helicase and helicase c-terminal) (Tsurusaki, Okamoto, et al., 2014), whereas the variant detected in our case did not affect any of the central domains and is among the most C-terminal missense variants reported so far. The location of the variant might explain the atypical phenotype of our case presenting with retinopathy but also without the typical dysmorphic features and digital abnormalities of CSS. Nevertheless, further individuals harboring SMARCA4 variants are needed to confirm whether retinal abnormalities are dependent on SMARCA4 defects. Moreover, further studies are needed to evaluate whether retinal changes are specific for SMARCA4 defects or might be due also to the defects of other genes of the BAF complex causing CSS.