Frameshift Variant in ARID2 in a Chilean Individual with Coffin–Siris Syndrome Phenotype

Abstract

Coffin–Siris syndrome (CSS) is one of the several causes of intellectual disability (ID) and, since its first description, has posed diagnostic challenges given its variability and phenotypic overlap with other alterations of chromatin-remodeling-associated syndromes. It is genetically heterogeneous, and causative mutations are detected in less than 70% of cases. The different subtypes of the syndrome described to date are caused by mutations in genes that encode subunits of the SWI/SNF chromatin-remodeling complex, which plays an essential role in the regulation of gene expression during embryogenesis. Whole exome sequencing (WES) has allowed the identification of pathogenic mutations in these genes, including ARID2 . ARID2 is one of the primary components of the SWI/SNF complex and has been associated with ID and phenotypes similar to CSS for the first time in 2015. Fifteen published case reports have identified loss-of-function mutations, suggesting that the underlying pathogenic disease mechanism is haploinsufficiency of ARID2 .

We herein presented the case of an 8-year-old Chilean girl with clinical suspicion of CSS, in whom a novel frameshift variant in ARID2 was identified by WES. She was the first reported case in Latin America to our knowledge and her phenotype displays the main clinical features suggestive of CSS described in other patients with ARID2 variants. However, she did not present behavioral abnormalities, a characteristic frequently reported in the majority of patients with ARID2 variants, and also had some features, such as sparse scalp hair, which is frequently reported as a manifestation of CSS, but is uncommon in this new group of patients.

Introduction

Coffin–Siris syndrome (CSS; Online Mendelian Inheritance in Man [OMIM] #135900) is a rare neurodevelopmental disorder characterized by intellectual disability (ID), sparse scalp hair, coarse facial features, and hypoplasia or aplasia of the nail or the distal phalanx of the fifth fingernails or toenails. It was initially described by Coffin and Siris in the 1970s and since then great effort is made to generate genotype–phenotype correlations trying to delimit the entity. ¹ Most cases occur sporadically and there is both phenotypic variability and genetic heterogeneity, which has led to difficulties in identifying the causes from the initial description of the syndrome to nowadays.

Background

Different forms of CSS are now known to be caused by mutations in genes that encode subunits of the SWI/SNF complex, also known as the BAF complex, which functions as a chromatin-remodeling factor. To date, over 150 patients are reportedly diagnosed with CSS ² and 12 genes are associated with its pathogenesis: ARID1B , ARID1A , ARID2 , SMARCA4 , SMARCB1 , SMARCC2 , SMARCD1 , SMARCE1 , BICRA , DPF2 , SOX11 , and SOX4 . ³ The mutation detection rate underlying the various types of the syndrome is between 54.9 and 71% and most patients harbor de novo autosomal dominant mutations in one of the subunits of the SWI/SNF complex, ^{3 4} with ARID1B being the most frequently mutated gene. ⁵

The differential diagnosis of CSS includes chromatin-remodeling disorders and other syndromes with varied genetic causes such as: Nicolaides–Baraitser syndrome (NCBRS, OMIM #601358), brachymorphism-onychodysplasia-dysphalangism, DOOR—deafness, onychodystrophy, osteodystrophy, and mental retardation—syndrome, hyperphosphatasia-intellectual deficiency syndrome, Borjeson–Forssman–Lehmann syndrome, Wiedemann–Steiner syndrome, Rubinstein-Taybi syndrome, and Cornelia de Lange syndrome.

ARID2 is an essential member of SWI/SNF-B, also called PBAF, and pathogenic variants in this gene associated with ID- and CSS-like phenotypes were first discovered in 2015, with a total of 15 cases published to date. ^{6 7 8 9 10}

The hypothesis that ARID2 may cause a CSS-like phenotype is not conclusive yet and it is unclear whether or not this gene should be considered a true “CSS gene.” The mutations in previously published cases lead to a loss of function of the gene, suggesting that haploinsufficiency of ARID2 is the pathogenic molecular mechanism involved in the development of the phenotype. ⁶

Clinical Manifestation

The main features present in the syndrome are ID, sparse scalp hair, coarse facial features, and hypoplasia or aplasia of the nail or the distal phalanx of the fifth fingernails or toenails. Other more variable features include hypertrichosis, feeding difficulties, growth retardation, craniofacial abnormalities, spinal abnormalities, and congenital heart defects. ^{5 11} The classic phenotype is not always present and there are differences between the phenotypes of patients with different genetic variants. ²

Diagnosis

Neurodevelopmental disorders are a heterogeneous group of conditions with a broad landscape of genetic causes that include aneuploidy, deletions or segmental duplications, numerous monogenic causes, and epigenetic factors. The large phenotypic spectrum of these conditions creates a great challenge in deciding the most appropriate type of genetic test to request in a patient. Nowadays chromosomal microarray and whole exome sequencing (WES) are the main tools to be considered for their diagnostic yield, and recent consensus is placing WES as the first-tier clinical diagnostic test for neurodevelopmental disorders. ¹²

Management

Most neurodevelopmental disorders with dysmorphic features, organ system anomalies, and developmental delay/ID have multiple medical needs that will require a multidisciplinary team of care. There are recommendations on the baseline evaluations that could benefit individuals with CSS and these include: vision and hearing screening, echocardiogram/cardiology evaluation, bone age studies in case of delayed growth, gastroenterology evaluation in case of feeding difficulties or failure to thrive, immunology evaluation in case of history of significant infections, dental exam after teeth eruption, scoliosis evaluation, and developmental assessment/neuropsychiatric consultation. ² Surveillance and treatment of specific manifestations will depend on the initial and ongoing evaluations, as well as the indication of rehabilitation and special education.

Prognosis

Due to the great clinical heterogeneity of the syndrome, it is difficult to determine a life expectancy for the individuals with the syndrome and it has not been reported in the different study cohorts. The general prognosis will depend on the early diagnosis and management of the most severe anomalies that may be present, and on interventions such as physical therapy, speech therapy, and special education so that they can achieve their full potential.

Case Presentation

We presented the case of a Chilean nationality patient with a clinical phenotype of CSS, in whom WES revealed a novel mutation in ARID2 associated with the phenotype.

The patient was an 8-year-old female, the first child of healthy nonconsanguineous parents. She has two younger healthy siblings. She was born at 40 weeks of gestational age by vaginal delivery with a weight of 2,990 g (–0.8 standard deviation [SD]), length of 48.5 cm (–1.26 SD), head circumference of 34 cm (–0.9 SD), and an Apgar score of 9–9. She did not have any congenital anomalies.

At 5 months, she presented a urinary tract infection with ultrasound diagnosis of unilateral grade 4 vesicoureteral reflux. During her first year of life, she was also diagnosed with recurrent obstructive bronchitis and presented an episode of spontaneous pneumothorax.

As for her psychomotor development, she presented a moderate delay in the gross motor area with cephalic support at 5 months, sitting at 8 months, and walking at 20 months of age. She has attended kinesiology therapy since her first year of life. She also presented a mild delay in language development that was noted at the beginning of the preschool age, with difficulties in reading and writing, and she was diagnosed with mild ID at 5 years of age.

She had a delay in eruption of teeth, and throughout her life up until 6 years she presented short stature and low weight, currently with a weight of 22.3 kg (–0.89 SD) and length of 126 cm (–0.33 SD).

As for other medical issues, she presents astigmatism, excessive sweating of hands and feet, and had an adenoidectomy at 7 years of age for snoring.

She was evaluated by a geneticist at 5 years of age for a dysmorphic syndrome and developmental delay. Williams syndrome was ruled out with a normal fluorescence in situ hybridization for chromosome 7. In the neurological sphere, she has not presented seizures; she has a normal brain computed tomography scan and no behavioral manifestations. Her parents and teachers describe her personality as very friendly, talkative, and without other behavior problems. She attends a common school with an integration program.

Her dysmorphic features include sparse scalp hair lighter colored than her relatives, coarse facial features with heavy and thick eyebrows, long eyelashes, downslanting palpebral fissures, broad nasal tip with anteverted nares, wide mouth with full upper lip, small chin, gingival hyperplasia, and macrodontia of central incisors. Bilateral fifth finger mild clinodactyly and small fifth toenails were observed ( Fig. 1 ).

Fig. 1.

( A–D ) Facial characteristics at 8 years of age. Note sparse scalp hair, coarse facial features with heavy and thick eyebrows, long eyelashes, downslanting palpebral fissures, broad nasal tip, wide mouth with full upper lip, small chin, and macrodontia of central incisors. ( E ) Bilateral mild clinodactyly of fifth finger. ( F, G ) Fifth toenail hypoplasia.

Face2Gene application was used as a diagnostic guidance tool through facial recognition and DeepGestalt. The diagnosis of CSS appears within the top 30 suggested syndromes and was the one with the highest agreement in DeepGestalt ( Fig. 2 ). Other syndromes suggested by the application are listed in Table 1 .

Fig. 2.

Heat map of facial recognition analysis by DeepGestalt using the Face2Gene application. On the left, the right side of the proband's face, and on the right the comparative gestalt of Coffin–Siris syndrome (CSS). In red, the facial areas with the greatest overlap with typical CSS characteristics, the area of the eyebrows, mouth, and chin of the patient.

Table 1. Top 10 suggested syndromes by Face2Gene application according to facial gestalt.

1. Coffin–Siris syndrome

2. Mucopolysaccharidosis, type IIIB

3. KBG syndrome

4. Marshall-Smith syndrome

5. Mucopolysaccharidoses

6. Noonan syndrome-like disorder with loose anagen hair

7. Noonan syndrome

8. Turner syndrome

9. Mental retardation, autosomal dominant 5

10. Kleefstra syndrome

WES performed by CENTOGENE AG identified a heterozygous variant c.4527dup (Ala1510 Serfs*14) in the ARID2 gene (NM_152641.3), which causes a change in the reading frame at codon 1510 and the new reading frame ends at a stop codon 13 amino acids downstream ( Supplementary Fig. S1 [available in online version only]). The variant is classified as pathogenic according to the American College of Medical Genetics and Genomics criteria as it is a truncating variant in a gene for which loss of function is a known mechanism of disease according to ClinGen, is absent from control population databases (such as gnomAD exomes and genomes), and is expected to cause nonsense-mediated decay and predicted to be disease causing by the in silico prediction tool Mutation taster ( Supplementary Table S1 , available in the online version only). Truncating variants are the most common among patients with ARID2 -related Coffin–Siris-like phenotype reported to date ( Table 2 ). The variant was confirmed by Sanger sequencing in the proband and parent testing could not be performed.

Table 2. Frequency of clinical features and genotypes of 16 patients (including this report) with ARID2 variants .

Variant type	Truncating variant: 81.25% Deletion: 18.75%
Age (years)	4.5–23 (mean = 9.2)
Gender	Female: 37.5% Male: 62.5%
Prenatal abnormalities	Reported in 8/16 individuals (50%) Most frequently reported: polyhydramnios
Anthropometry at birth	Low birth weight: 25% Low birth height: 43% Microcephaly: 9% a Macrocephaly: 9% a
Intellectual disability	100%
Behavioral issues	81.25%
Language development delay	56%
Dysmorphism	100% Coarse facial features: 62.5%
Ectodermal features	Total: 81.25% Toenail hypoplasia: 50% Blonde hair: 56.25%
Hand, foot, and digital anomalies	43.75%
Other skeletal abnormalities	43.75%
Ophthalmologic abnormalities	Total: 37.5% Refractive errors: 31.25%
Brain malformations	25%
Gastrointestinal abnormalities	31%
Airway/bronchopulmonary abnormalities	25%
Cardiac abnormalities	12.5%
Other neurologic abnormalities	Hypotonia: 12.5% Seizures: 6.25%
Diaphragmatic abnormalities	12.5%
Cleft lip/palate	12.5%
Hernia	12.5%
Endocrinological abnormalities	6.25%
Renal abnormalities	6.25%

Note: Sixteen published cases from: four individuals (Shang et al. 2015 ⁸ ), two individuals (Bramswig et al. 2017 ⁷ ), one individual (Van Paemel et al. 2017 ⁹ ), seven individuals (Gazdagh et al. 2019 ⁶ ), one individual (Khazanchi et al. 2019 ¹⁰ ), and one individual (this case report). More details in the particular anomalies reported in each category available in Supplementary Table S1 (available in the online version only).

^aData not available in all individuals.

Discussion

ARID2 was initially associated in 2015 with ID in a WES study in a cohort of 970 patients with ID. In this first study, four frameshift variants were identified that lead to truncated proteins and therefore were probably deleterious. These patients had developmental delay and presented some characteristics similar to those of CSS and NCBRS. ⁸

In the subsequent years, Gazdagh et al, Bramswig et al, Van Paemel et al, and Khazanchi et al reported another 11 individuals with loss-of-function variants in ARID2 identified by WES or array comparative genomic hybridization, postulating that haploinsufficiency of ARID2 is the pathogenic molecular mechanism involved. ^{6 7 9 10}

The phenotypes of patients with ARID2 loss-of-function variants resemble those of other chromatin-remodeling disorders and some authors propose that CSS, NCBRS, and the newly discovered developmental disorder associated to ARID2 correspond to a spectrum of syndromes with ID related to the SWI/SNF complex rather than singular entities ^{9 13} ; while others postulate that there is sufficient evidence to support the idea that disruptions in ARID2 causes a Coffin–Siris-like phenotype. ¹⁴

The novel variant identified in this case is consistent with the mechanism of haploinsufficiency of ARID2 due to loss of function and the phenotype meets the clinical features suggestive of CSS that are more consistent in the other cases with ARID2 variants. These include ID, hypoplasia of the fifth fingernail, and dysmorphic features ( Table 2 ). Other patient characteristics that overlap with the phenotype of CSS are recurrent respiratory infections, delayed dental eruption, genitourinary abnormalities, and sparse scalp hair, which are not very common in other patients with variants in ARID2. A distinction from most of the other patients reported is that our patient does not present associated behavioral issues that are present in 81% of patients with ARID2 variants.

It is noteworthy that none of the patients with ARID2 loss-of-function variants has hypertrichosis/hirsutism, a common ectodermal feature of CSS. Instead, there is a broad spectrum of skin and hair manifestations where toenail hypoplasia and light-colored scalp hair are the most commonly reported ( Supplementary Table S1 , available in the online version only).

A limitation to this study is that we do not evaluate the patient's extremities or spine for further skeletal anomalies utilizing radiographic images, and mutation segregation studies could not be performed so we cannot assure that the variant found is “de novo.”

Conclusion

The largest number of cases associated with previously reported ARID2 mutations has been identified in large cohorts of patients from ID studies (9 of the total of 15). As the amount of existing data increases with the use of WES and the spectrum of individuals studied is broadened, more precise and definitive correlations between genotype and phenotype can be made. For this purpose, the contribution of new cases from the clinical scenario and of different ethnicities will help to avoid the bias that is generated in the screening studies aimed at the identification of new causal genes, in which patients are selected and usually represent more severe phenotypes within the clinical spectrum.