Novel De Novo RAI1:c.2736delC Variant Identified in A Child with Features of Smith-Magenis: Insight from Whole Genomes Joint Analysis

1. Introduction

Smith-Magenis syndrome (SMS; OMIM #182290) is a complex neurobehavioral disorder with an estimated prevalence of 1:15,000. This rare syndrome is characterized by neurodevelopmental delay, coarse facial features evolving with age, and maladaptive behaviors that include prolonged temper tantrums and self-injurious behaviors, including hand biting, trichotillomania, and polyembolokoilamania [1]. During infancy, hypotonia, hyporeflexia, and feeding difficulties are commonly seen neurological findings [2,3,4,5]. The maladaptive behavior phenotype is typically recognized after 18 months of age. The classic neurological features include intellectual disability, delayed speech, sensory integration problems, decreased pain sensitivity, and abnormalities in sleep patterns [1]. Additional system involvements, such as musculoskeletal and cardiovascular, are also reported as part of SMS [2,3,4,5,6].

The underlying genomic etiology of SMS is interstitial deletion at 17p11.2 in 90-95% of patients. This genomic location encompasses the retinoic acid-induced 1 (RAI1) gene. RAI1 is the key gene in SMS and plays a regulatory role in healthy skeletal and nervous system development, behavioral maturation, and circadian rhythm. Pathogenic de novo single nucleotide variants (SNVs) in RAI1 are reported in 5-10% of patients with SMS [2,5,6,7]. In this category, missense, nonsense, and splice site variants in RAI1 have been reported [4]. The diagnosis of the SMS syndrome is based on clinical molecular diagnostics tests. The first-tier molecular test is chromosomal microarray analysis (CMA), and if negative, a sequencing-based test is typically performed. Genome studies have reported SMS-like features, including ID, sleep disturbances, and self-injuries behaviors in patients without an SMS molecular diagnosis who harbored deleterious variants in KMTD2, MECP2, KDM5C, IQSEC2, and DEAF1 genes [8,9].

Here, we report a 9-year-old male with global cognitive and psychomotor developmental delay and facial, musculoskeletal, and cardiovascular abnormalities. The patient’s initial chromosomal microarray and karyotype tests were normal. The quad joint WGS analysis of the proband, parents, and brother detected a pathogenic novel de novo variant, which has not been reported previously. This report highlights the importance of sequencing analysis in patients clinically suspected of SMS and with negative conventional diagnostic tests. Additionally, sharing the genotype and detailed phenotype of these relatively rare SMS cases helps understand the genetic and phenotypic spectrum of the disease.

2. Material and Methods

2.1. Participants

The index patient, a 9-year-old male, his parents, and his younger brother were enrolled in the CROseq genome program. The CROseq Genome Program is a collaborative research program between Brigham and Women’s Hospital (BWH) (Boston, USA), and the Department of Pediatrics, University Hospital Centre Zagreb (Zagreb, Croatia), supported by the Mila Za Sve Foundation (Rijeka, Croatia). This program aims to investigate the genomic etiology of disease in patients with complex phenotypes and negative conventional genetic tests. Patient consent, enrollment, and clinical assessment were performed at the University Hospital Center Zagreb.

2.2. Clinical Assessment of Brain Structure and Activity

The brain abnormalities were assessed by MRI scanning (MAGNETOM Trio, A Tim System 3T eco, Zagreb, Croatia). The patient was under general anesthesia during the scan. Age-adjusted repetition time and time-to-echo were applied to create multiplanar T1- and T2-weighted images. The electroencephalogram (Nihon Kohden, Neurofax EEG-1200K; Zagreb, Croatia) examination was performed to evaluate the brain activity. During the record, the sampling rate was 200-10000 Hz. The patient’s forehead is cleaned with alcohol to lower the impedance less than 5 Ω for each electrode. The electrodes were set up in FPz-F9 and FPz-AF7 of the international 10-20 system, which focuses on the brain in the same hemisphere.

2.3. Traditional Genetic Testing

Karyotyping was performed to analyze chromosomal abnormalities according to standard procedures at University Hospital Centre Zagreb. Briefly, the lymphocytes were purified from the peripheral blood sample. Cells were arrested at metaphase and fixed before being harvested for slide preparation. The number and morphology of chromosomes were assessed in at least 20 metaphases. The karyotype result was reported per The International System for Human Cytogenomic Nomenclature (ISCN) 2020.

Chromosome microarray (CMA) was performed according to standard procedures at University Hospital Centre Zagreb. Briefly, genomic DNA was isolated with the FlexiGene DNA Kit (Qiagen) and CMA was performed using Agilent 4x180k aCGH+SNP array (Agilent Technologies, Santa Clara, CA, USA). The genomic DNA of the test and control samples were labeled with fluorescent dyes and hybridized to the array. The array was scanned with a DNA Microarray Scanner (Agilent Technologies, Santa Clara, CA, USA). Data were obtained using the “CytoGenomics 5.1.1.15” software (Agilent Technologies, Santa Clara, CA, USA) to generate the final plots.

2.4. Sample Preparation and Whole Genome Sequencing (WGS)

DNA extraction and WGS were performed at the Medical College of Wisconsin (Milwaukee, USA). Genomic DNA was isolated from peripheral blood samples (2ml) with a purity ratio of 1.75 – 2.0. After robotic DNA library construction, sequencing was conducted on the NovaSeq 6000 platform, with an average depth of 40X.

2.5. Quad Joint Whole Genome Analysis

Genome analyses were performed at the BWH, USA. The quad whole genome analysis was performed on proband, unaffected parents, and the unaffected proband’s brother. Following the quad analysis, a phenotype-based analysis was performed using human phenotype ontology (HPO) terms of the patient’s indications. Variants were prioritized by the gene’s association with the phenotype and variant type. The HPO terms were: Neurodevelopmental delay (HP:0012758), moderate intellectual disability (HP:0002342), delayed speech and language development (HP:0000750), attention deficit hyperactivity disorder (HP:0007018), bicuspid aortic valve (HP:0001647) and stenosis (HP:0001650), failure to thrive (HP:0001508), abnormal facial shape (HP:0001999), unilateral ptosis (HP:0007687), frontal bossing (HP:0002007), relative macrocephaly (HP:0004482), abnormality of mitochondrial metabolism (HP:0003287).

All variants across the genome were included in the investigation. Following the technical assessment, medium, high, and very high-quality variants were further evaluated, and low-quality variants were excluded. The variant frequencies were annotated by gnomAD exome and genome population allele frequencies. The in silico prediction algorithms for missense variants included CADD, REVEL, Polyphen, SIFT, MutationTaster, Mutation Assessor, FATHMM, FITCONS, GENOCANYON, dbscSNV ADA, and dbscSNV RF. The spliceAI prediction score was used in the evaluation of splice variants.

Zygosity analysis was performed by identifying and evaluating compound heterozygous and homozygous variants in the proband. The recessive variants in the proband inherited from unaffected parents and related to the patient’s HPOs were assessed. Separately, de novo analysis was performed to assess variants present in the proband and related to the patient’s HPOs that are absent in the unaffected parents and the brother.

The variant classification was performed according to ACMG-AMP guidelines (25741868). In the attribution of PM2 and BS1 criteria, the gene-specific threshold was applied based on gnomAD (v2.1.1) aggregated allele frequency. The aggregated in silico prediction score was indicative of a deleterious effect if it was higher than 0.7; scores less than 0.15 were assigned for a benign effect. The genes’ tolerance to the loss of function variants was evaluated by using gene constraint metrics of pLI=1 and/or o/e<0.35.

3. Results

3.1. Clinical Description

The proband is a 9-year-old male presented with indications of global cognitive and psychomotor developmental delay, musculoskeletal and cardiovascular abnormalities, and dysmorphic craniofacial features. During his infancy, the patient had generalized hypotonia and feeding difficulties. His brain ultrasound, which was performed at four months of age, was normal. Following intensive physical therapy, he started walking at two years old. The psychomotor assessment revealed that the patient has severe psychomotor retardation. The brain MRI performed at 3.5 years of age showed perinatal hypoxic-ischemic lesions. The electroencephalogram was unremarkable (data not shown).

The detailed clinical evaluation at 9 years of age indicated brachycephaly, broad face, hypertelorism, wide nasal bridge, short and full-tipped nose, midface retrusion, deeply set ears, tented and down-turned upper lip, everted upper lip vermilion, micrognathia, and low hairline (Figure 1). He exhibited global cognitive and psychomotor developmental delay, severe ID, and developmental language disorder. The pervasive developmental disorder was evident with the lack of speech development and poor contact. Behavioral phenotypes included aggression toward others self-injurious episodes, and polyembolokoilamania. He did not acquire sphincter control and exhibited incontinence. He had sleep disturbances, including night awakening, with difficulty falling back asleep. There were skeletal abnormalities including mild scoliosis, thoracic kyphosis, short palm and foot, brachydactyly, and partial syndactyly of both hands’ second and third fingers (Figure 1). He reportedly had a bicuspid aortic valve with mild stenosis and congenital left eyelid ptosis.

3.2. Genomic Findings

CMA and Karyotype Analysis

Karyotype testing was 46, XY. CMA analysis of the proband detected a maternally inherited 2.1 Mb duplication in 22q11.21 (Supplementary Figure S1 and S2). This region contains OMIM genes SLC25A1, CDC45, GP1BB, TBX1, TXNRD2, COMT, TANGO2, RTN4R, SCARF2, PI4KA, SERPIND1, SNAP29, and LZTR1. The mother is unaffected, therefore, this 22q11.21 gain was considered unrelated to the patient’s clinical findings of SMS. The patient was enrolled in the CROseq genome program, and a quad WGS joint analysis was performed.

Quad WGS Joint Analysis

After WGS, a quad-WGS analysis of the proband, unaffected brother, and parents was performed. HPO-based analysis revealed 10,290 variants in 2,567 genes. Compound heterozygous analysis interrogated possible causative variants inherited from the unaffected parents. Quad analysis did not identify deleterious homozygous or compound heterozygous candidate variants in the proband.

De novo analysis interrogated variants related to the HPO that are present in the proband, but absent in the parents and the unaffected sibling. This analysis identified the variant NM_030665.4 (RAI1):c.2736delC (p.Gly913Alafs*37), located in exon 3 of RAI1 (Figure 2). This variant is a 1bp deletion that results in a frameshift and a premature stop codon. Due to its location, this variant likely results in a premature translational stop signal and nonsense-mediated mRNA decay. The gene constraint metrics for RAI1 are pLI = 1, o/e = 0.04. The variant has not been reported in gnomAD, ClinVar, or any published literature. The phasing analysis by quad joint WGS showed that the parents and younger brother did not harbor this variant. No other deleterious variants in genes related to the patient’s HPO list were identified.