Low‐level germline mosaicism of a novel SMARCA2 missense variant: Expanding the phenotypic spectrum and mode of genetic transmission

Abstract Background Nicolaides–Baraitser syndrome (NCBRS) is a severe neurodevelopmental disorder with multiple abnormalities. To date, all pathogenic variants in SMARCA2 causing NCBRS are de novo and most are missense variants located in the ATPase domain of SMARCA2 protein. Methods In this study, a familial trio whole‐exome sequencing was performed on the proband presenting with intellectual disability, early‐onset epilepsy, and autistic features. A novel missense variant c.553C>G (p.Gln185Glu) in SMARCA2 was identified, which is located in the QLQ domain. The same variant was subsequently also found in the mother's ongoing pregnancy. Samples from accessible tissues such as saliva and sperm other than blood were collected from the parents, and the detection of the target variant was performed by amplicon‐based deep sequencing. Results Low‐level mosaicism of the target variant c.553C>G (p.Gln185Glu) was detected in the father's sperm with allele fraction of 2.8% by amplicon‐based deep sequencing, which was not detected in either parents’ blood or saliva specimens. Heterozygosity of this variant was confirmed in the proband. Conclusion This is the first report of paternal germline mosaicism for a SMARCA2 disease‐causing variant. In addition, the missense variant c.553C>G (p.Gln185Glu) in the QLQ domain causes mainly neurological and developmental phenotypes with unremarkable characteristic facial features and limb abnormalities. Our findings expand the phenotypic spectrum and mode of genetic transmission associated with the SMARCA2 variants.


| INTRODUCTION
Nicolaides-Baraitser syndrome (NCBRS) is an autosomal dominant disease with multiple anomalies characterized by distinctive facial features, abnormalities of trunk and limbs, seizures, severe intellectual disability, and impaired language development (Nicolaides & Baraitser, 1993;Sousa et al., 2009). Seizures often begin in infancy which is difficult to be managed with antiepileptic drugs (Sousa et al., 2009). The signs and symptoms are highly variable in affected individuals. NCBRS is rare and fewer than 100 cases have been reported (Sousa & Hennekam, 2014).
Pathogenic variants in SMARCA2 (OMIM number: 600014) were identified as the underlying genetic causes of NCBRS (Van Houdt et al., 2012). SMARCA2 is located on chromosome 9p24.3, and it encodes a component of SWI/SNF complexes, which play an important role in chromatin remodeling and regulation of downstream gene expression in neural development (Sokpor et al., 2017). To date, de novo pathogenic variants were previously reported in all affected individuals with NCBRS (Bramswig et al., 2015;Ejaz et al., 2016;Hofmeister et al., 2020;Hu et al., 2019;Ma et al., 2020;Mari et al., 2015;Sethi et al., 2019;Sousa & Hennekam, 2014;Tang et al., 2017). Nearly all disease-causing variants are missense changes and clustered in the highly conserved ATPase domain, which is critical for ATP-binding/hydrolysis and DNAbinding in transcriptional regulation (Bogershausen & Wollnik, 2018).
Here, we report a family in which the disease-causing SMARCA2 variant in the proband was found due to lowlevel paternal germline mosaicism. To our knowledge, this is the first case report of germline mosaicism for SMARCA2related disease with a confirmed molecular finding. We also report it as the first disease-causing variant in the QLQ domain associated with solely neurological and developmental phenotypes of NCBRS, which provides new insights into the structure-function relationship of this subunit in the SWI/ SNF complexes.

| Methods
Peripheral blood and saliva samples were collected from both parents, and a semen sample was collected from the father as well. Peripheral blood and amniocytes were collected from the proband and the current pregnancy, respectively. Genomic DNA was extracted using commercial kits according to the manufacturer's instructions (Qiagen, Germany). Amplicon-based deep sequencing was used for mosaicism detection and the estimation of variant allele fraction. Briefly, the genomic region containing the missense SMARCA2 (reference sequence: NG_032162.2, transcript: NM_003070.5) variant was targeted and amplified using specific primers (forward: 5-CCAACAGAGGTCCCTCACCT-3, reverse: 5-GCCTCGGGCCAGCATTTTAT-3) with PCR conditions as following: (1) initiation denaturation at 98℃ for 2 min, (2) 98℃ for 10 s, 70℃ (−1.0℃/cycle) for 30 s, 72℃ for 30 s, 10 cycles, (3) 72℃ for 5 min. Universal sequencing primers were added by the second round of PCR and amplicons were generated with different barcodes for each sample. Multiplexed sequencing was run on the MGISEQ-2000 sequencer (MGI Tech Co., Ltd., China). Sequencing data were analyzed by an in-house developed bioinformatics pipeline. After alignment to the human reference genome (GRCh38) and removal of duplicate reads, the target variant allele fraction was calculated for each sample. DNA from the proband was used as positive control and NA12878 DNA (Coriell Institute) was used as a negative control. Background noise and limit of detection were determined using negative control and a serial dilution (from 0.01% to 10%) of positive control.

| Clinical presentation
The proband is an 8-year-old Chinese boy who presented with intellectual disability, recurrent, and intractable in the QLQ domain causes mainly neurological and developmental phenotypes with unremarkable characteristic facial features and limb abnormalities. Our findings expand the phenotypic spectrum and mode of genetic transmission associated with the SMARCA2 variants.

K E Y W O R D S
amplicon sequencing, germline mosaicism, Nicolaides-Baraitser syndrome, QLQ domain, SMARCA2 gene | 3 of 7 PAN et Al. epilepsy since 2 years after birth. The proband was born to a 26-year-old, G2P0 mother by spontaneous vaginal delivery at term, without pregnancy or birth complications. Developmental delay was noticed when he was about 1 year old, especially on verbal and intellectual development, and the fine motor control and coordination of movement were below average. At 2 years old, the first episode of seizures occurred during sleep, which lasted 1 to 2 min and was associated with jerking of bilateral upper limbs. The seizure was recurrently followed by the frequency increased gradually from ~1/month to ~1/day. At the age of four, seizures started to also occur in the awake period, with the manifestation of fear, startle, and jerking of upper limbs. At the age of seven, the proband manifested intellectual disability, poor verbal expression, and features of hyperactivity disorder, with seizures occurred one to three times per day. The electroencephalogram (EEG) showed abnormal waves at the age of three and seven. History of febrile seizure, brain infection or head trauma was denied, and there is no family history of epilepsy. The administration of antiepileptic drugs such as oxcarbazepine and clobazam showed partial relief on seizure symptoms and frequency. The proband was also evaluated at the autism clinic at age four, with records of limited social communication and repetitive patterns of behavior. No facial dysmorphic features were noticed during the evaluation.

| Genetic testing results
The proband underwent karyotyping and chromosomal microarray analysis (Affymetrix CytoScan 750K array), and no chromosomal abnormalities were detected. A familial trio clinical exome sequencing was performed in which a de novo missense variant in SMARCA2 (NM_003070.5: c.553C>G, p.Gln185Glu) was found in the proband. No other pathogenic or likely pathogenic variants related to the proband's phenotype were identified. Both parents were phenotypically normal and were tested negative for this variant by whole-exome and Sanger sequencing (Figure 1a,b). The c.553C>G (p.Gln185Glu) variant was neither reported in literature or human population databases such as gnomAD. It was predicted as disease-causing by multiple computational tools including PolyPhen-2, SIFT, and MutationTaster, with a REVEL score of 0.831 (Ioannidis et al., 2016). No splicing effects were predicted by in silico tools of splice-site alterations. Based on its mode of inheritance and genotypephenotype correlation, this variant was classified as likely pathogenic according to ACMG guidelines since the criteria of PS2, PM2, PP2, and PP3 were met in this case (Richards et al., 2015). Glutamine residue at codon 185 is highly conserved across different species (Figure 1c), and it is located in the QLQ domain within SMARCA2 protein (Figure 1d). It should be noted that the majority of previously reported SMARCA2 pathogenic variants in the ClinVar database were distributed in the Helicase ATP-binding domain and Helicase C-terminal domain (Figure 1d), and the p.Gln185Glu variant was the first disease-causing variant identified in the QLQ domain, which resulted in the replacement of polar uncharged glutamine by negatively charged glutamic acid, and disturbance of the chemical and structural properties of SMARCA2 protein (Figure 1e).
The proband was then diagnosed with NCBRS based on phenotypic and molecular findings. The proband's mother had an ongoing pregnancy at a gestational age of 9 weeks. Prenatal diagnosis was offered to the family for the fetus after the SMARCA2 variant was identified in the proband. By amniocentesis, it was confirmed that the same variant c.553C>G (p.Gln185Glu) was present in the fetus ( Figure  1a,b). The pregnancy was terminated after genetic counseling and the presence of germline mosaicism was suspected. Both parents were referred for further genetic testing.

| The identification of germline mosaicism
Amplicon-based deep next-generation sequencing (NGS) was performed on samples such as blood and saliva collected from both parents, and sperm collected from the father. DNA isolated from the proband and his fetal sibling were included as positive controls. The NGS coverage of the target locus SMARCA2 c.553C>G (p.Gln185Glu) was more than 5000X in all samples tested. Nearly half of the total reads harbored the variant G allele in the PCR products amplified from the proband's and the fetus' DNA. A low percentage of sequencing reads with the variant G allele were found in the father's sperm DNA, and no such reads were seen in the mother's blood (Figure 2a). The variant allele fraction calculated from sequencing reads indicated heterozygosity of C/G on this position in the proband (52.9%) and the fetus (49.3%). About 2.8% reads of pathogenic allele G were detected in the father's sperm, suggesting the presence of low-level germline mosaicism (Figure 2b). No pathogenic allele was detected in blood or saliva samples from both parents when comparing with the background noise (0.16%) determined by negative control.

| DISCUSSION
In this report, we described the first case where paternal germline mosaicism of a SMARCA2 disease-causing variant was the underlying cause of NCBRS. This disease is considered genetically homogeneous as all patients were found to have de novo pathogenic variants in SMARCA2 (Sousa & Hennekam, 2014). Our finding in this study suggests that germline mosaicism needs to be ruled out in families affected by NCBRS for better planning of future pregnancies. When preimplantation genetic testing is possible, the identification of potential parental germline mosaicism will facilitate the implantation of an unaffected embryo. Prenatal diagnosis for such families is always useful regardless of the conceiving strategy. It is possible that germline mosaicism is present in previously documented NCBRS patients and families, most of which were diagnosed based on exome sequencing results. The limited depth of coverage and sample types (mostly leukocytes) may not be able to reveal the presence of pathogenic variants in germline cells. It may be worth evaluating previously reported NCBRS families in order to obtain the frequency of occurrence of germline mosaicism.
Germline mosaicism is a relatively common underlying mechanism in families who have more than one affected offspring affected by de novo dominant variants born to phenotypically normal parents. It is due to spontaneous pathogenic variants which occur predominately during mitotic or meiotic divisions in oogenesis and spermatogenesis (Spinner & Conlin, 2014). Because of its easier accessibility, studies of mosaicism on sperms are far more than eggs. Sperm mosaicism has been reported in patients with various genetic disorders (Breuss et al., 2020;Pauli et al., 2009;Yang et al., 2017), and the pathogenic allele fractions vary from very low (~0.03%) to moderate level (~39%) (Yang et al., 2017). In our case, the mosaicism appeared to be restricted to the gonadal organ because no pathogenic alleles were detected in other somatic cells such as father's leukocytes or buccal cells. It is also possible that the level of somatic mosaicism was lower than the limit of detection (0.5%). Given the low-level mosaicism (<3%) and two successive offsprings with the same variant in this family, it is of interest to explore if the sperms with the SMARCA2 variant possess increased advantage during competition with sperms containing the normal allele.
The SMARCA2 protein is a subunit of the SWI/SNF complexes which participate in chromatin remodeling in an ATPase-dependent manner. In a review of 61 patients with NCBRS, the vast majority (n = 59) of pathogenic variants are missense and only a few (n = 2) are in-frame deletion (Sousa & Hennekam, 2014). All previously reported pathogenic variants are located through exon 15 to exon 25, which are within the ATPase domain. It is hypothesized that the dominant-negative or gain-of-function effect, but not haploinsufficiency, underlies the role of mutated SMARCA2 in NCBRS (Sousa & Hennekam, 2014;Van Houdt et al., 2012). Herein, we reported the first disease-causing variant in the QLQ domain of SMARCA2 protein. The QLQ domain is featured by the conserved Gln-Leu-Gln residues. For the SMARCA2 protein with Gln185Glu variant, the chemical and structural properties may be severely disrupted by replacing polar uncharged glutamine with negatively charged glutamic acid, but its effects on complexes assembly or ATPase activity are currently unknown. The mutant SMARCA2 protein causes dysregulation of the expression of neural progenitor genes and neural enhancer reprogramming, and genome-wide DNA methylation changes are also relevant to NCBRS pathophysiology (Chater-Diehl et al., 2019;Gao et al., 2019). It should be noted that the SMARCA2 RNA is highly expressed in male testis. Assessment of the SMARCA2 variant c.553C>G p.Gl-n185Glu in neural developmental signaling pathways, DNA methylation signatures, and spermatocytes growth would be of interest in further studies.
In conclusion, this is the first reported germline mosaicism case of SMARCA2-related disease. The inheritance mode, tissue of origin, and level of mosaicism were determined by deep sequencing. It suggests that genetic counseling and recurrence risk estimation could benefit from the germline mosaicism assessment. This is also the first reported diseasecausing variant located in the QLQ domain of the SMARCA2 protein associated with intellectual disability, early-onset epilepsy, and autistic features, suggesting the importance of the high conservation and functional role of this motif.