First report of Coffin-Siris Syndrome with SMARCB1 variant, normal intelligence and mild selective neuropsychological deficits: A case report and literature review

Abstract

Background: The SMARCB1 gene encodes a subunit of the BRG1-Associated Factor (BAF) complex, and mutations in this gene have been linked to Coffin-Siris Syndrome (CSS) type 3. CSS is characterized by a range of developmental disabilities, facial dysmorphic features, and feeding difficulties. There’s been noted genotype-phenotype correlation in CSS, with cases involving SMARCB1 mutations often exhibiting more severe language impairment and intellectual disability. Method: We conducted a review of reported CSS type 3 cases and presented the first instance of CSS associated with a SMARCB1 variant wherein the patient exhibited normal intelligence and only mild selective neuropsychological deficits. The patient underwent evaluation for feeding challenges, growth delay, and dysmorphic features during their second year of life. Subsequently, CSS diagnosis was confirmed due to a de novo heterozygous c.568C > T (p.Arg190Trp) variant in the SMARCB1 gene. Due to learning difficulties, the patient underwent a comprehensive neuropsychological assessment, which was related to the retrospective reconstruction of her medical and developmental history. Results: The patient demonstrated normal intelligence and adaptive functioning, with specific deficits in arithmetic and selective difficulties in verbal learning and long-term memory. Feeding difficulties and language delay observed in early childhood showed significant improvement over time. Discussion: We discuss this case in relation to previously reported CSS type 3 cases, emphasizing neuropsychological aspects. It’s evident that neuropsychological features of CSS can vary among affected individuals, highlighting the importance of personalized support and interventions tailored to specific cognitive and emotional needs by healthcare professionals. Our case suggests avenues for future research to identify specific modifiers of phenotypic expression to explain variability in intellect among patients and pinpoint potential targets for gene therapy.

Keywords:

• Intellectual disability
• neuropsychological deficits
• BAFopathies
• developmental delay
• SMARCB1

Background

Coffin-Siris Syndrome (CSS) is a rare genetic disorder characterized by developmental and intellectual disabilities (ID), as well as various physical anomalies. It was first described in 1970 by Coffin and Siris (1970) as a congenital malformation syndrome characterized by hyposomia, severe ID, and the absence of the terminal phalanx and nail of the fifth finger. In addition to the typical fifth-digit nail/distal phalanx hypoplasia/aplasia, other indicative physical features include a wide mouth with thick, everted upper and lower lips, a broad nasal bridge and nasal tip, thick eyebrows, long eyelashes, hirsutism/hypertrichosis, and sparse scalp hair (OMIM). Other associated characteristics are developmental delay, involving language, motricity and cognition, feeding difficulties and slow growth. Malformations of the cardiac, gastrointestinal, genitourinary, and/or central nervous systems (CNS) also may be present (Schrier Vergano et al., 2013). CSS is an autosomal dominant disease most often caused by ARID1B gene mutation, associated with a typical clinical picture known as CSS type 1 (OMIM 135900). Some patients exhibit less pronounced facial characteristics, with thinner eyebrows and a thin vermillion border of the lips, and are classified as having variant CSS, which are associated to other genes, such as ARID1A, ARID2, DPF2, PHF6, SMARCA2, SMARCA4, SMARCB1, SMARCC2, SMARCE1, SOX4, or SOX11 (Schrier et al., 2012). CSS diagnosis can be confirmed through genetic testing that identifies pathogenic variants in these genes (Schrier Vergano et al., 2013). The BRG1/BRM-associated factor (BAF) complex genes (Santen et al., 2012), including ARID1B, are particularly associated with CSS, with mutations in these genes termed BAFopathies, encompassing CSS, Nicolaides-Baraitser syndrome, and nonsyndromic ID (Aref-Eshghi et al., 2018; Bögershausen & Wollnik, 2018; Machol et al., 2019; Nixon et al., 2019; Wieczorek et al., 2013). Variations in genes encoding SOX transcription factors are also linked to CSS, forming the category of SOXopathies (Al-Jawahiri et al., 2022; Angelozzi et al., 2022; Angelozzi & Lefebvre, 2019; Wieczorek et al., 2013).

Phenotypic variability is common in CSS, with differences observed even between monochorionic twins (LaBrecque et al., 2019), suggesting underlying epigenetic factors at play. Recent literature indicates that studying the epi-signatures of different phenotypes in BAFopathies (including CSS subtypes and Nicolaides-Baraitser syndrome) appears to offer valuable insights into understanding genotype-phenotype correlations (Aref-Eshghi et al., 2018).

Studies on genotype-phenotype correlations in CSS have revealed patterns, with certain gene mutations associated with more or less severe clinical manifestations. In particular, patients with SMARCB1 mutation have a distinct subtype of CSS variant known as CSS type 3, characterized by more severe neurodevelopmental phenotypes, language impairment, structural brain abnormalities, and severe ID (Kosho et al. 2014, Wieczorek et al., 2013, Santen et al., 2013), while those with ARID1A and ARID1B mutations may display a more variable phenotype (Santen et al., 2013). The SMARCB1 gene, identified in Versteege et al. (1998), is located at 22q11.23 and encodes a subunit of the BAF complex. It was initially associated with rhabdoid tumor predisposition syndrome and susceptibility to Schwannomatosis-1, and then associated to CSS type 3 in 2012 by Tsurusaki and colleagues.

We hereafter review in more detail neurologic and psychiatric phenotype of CSS, with a focus on CSS type 3 associated to SMARCB1 mutations. Then, we will comprehensively review documented cases of patients with SMARCB1 mutation described in literature to date, to introduce a new case which challenges current understanding of genotype phenotype correlations in CSS and BAFopathies and whose neuropsychological phenotype is thoroughly discussed in correlation to potential molecular pathogenesis.

Neuropsychiatric features of Coffin Siris Syndrome

Concerning nervous system pathology, CNS malformations are documented in a subset of patients with CSS, and they appear to be more prevalent in those with SMARCB1 mutations. Hindbrain malformations were noted since the first reports of CSS by pathologists who described CSS as a neurocutaneous disorder with hindbrain abnormalities (DeBassio et al., 1985). Among hindbrain malformations, Dandy Walker variant is the most commonly reported, but post-mortem studies have revealed a broader spectrum of hindbrain abnormalities, including cases without Dandy Walker and with abnormalities of the inferior and medial accessory olives, large arcuate nuclei, heterotopic olivary nuclei, and heterotopic nuclei in the white matter of the cerebellum. Furthermore, the presence of complete or partial agenesis of the Corpus Callosum (CC) is frequently observed in CSS, and in a study on subjects with ARID1B mutation CC metrics were found to predict the severity of visuospatial and neuromotor dysfunctions (Demily et al., 2019). Notably, the distinctive expressive language impairment of CSS has not been correlated with specific CNS pathology (Vasko & Schrier Vergano, 2021). In the patients with CSS due to SMARCB1mutation identified and reviewed by Kosho et al. in 2014, CNS malformations were present in all nine patients who underwent neuroimaging. The described findings encompass midline and hindbrain defects, although other abnormalities, such as gyral simplification, have also been reported.

With regard to the neurodevelopmental phenotype, developmental delay is common in CSS, often leading to delayed achievement of motor milestones. Central hypotonia is observed in 75% of patients with CSS (Schrier Vergano et al., 2013). In the largest cohort of BAFopathies to date, which included 208 patients, it was found that the 14 patients with SMARCB1 mutations had the highest rate of impairment in acquiring rolling (57%) and sitting (55%) skills. Independent walking is typically achieved but tends to be delayed, with a mean age of acquisition ranging from 18 to 30 months in CSSs. Patients with ARID2 mutations appear to achieve independent walking at younger ages within this range, while those with SMARCB1 and SMARCE1 mutations achieve independent walking later (Vasko et al., 2021).

Although CSS is typically associated with ID, there have been rare reports of individuals with CSS type 1, who have a normal intelligence quotient (IQ). This has prompted a reevaluation of the assumption that CSS type 1 always involves ID, expanding the understanding of CSS as a spectrum. In the most recent report, it was found that 1 in 70 ARID1B-mutated patients with CSS type 1 had an IQ within the normal range. Currently, ARID1B is considered a gene associated with a range of diseases, from non-syndromic ID to CSS, and its manifestations are not predictable based on the current understanding of gene mutations and molecular pathology. There have been no identified correlations between the neuropsychological phenotype and mutation site or in-frame/out-of-frame mutations of ARID1B (van der Sluijs et al., 2019). No cases with normal IQ have been reported for CSS type 3, which, conversely, includes the most severely impaired patients.

Language impairment in CSS has been a longstanding observation since the initial reports of the syndrome, which appeared to be characterized by an imbalance between expressive and receptive language abilities (Swillen et al., 1995). Expressive language tends to be more severely affected than receptive language, and approximately one-third of all CSS patients are non-verbal (Vasko & Schrier Vergano, 2021). Among individuals with variants in SMARCB1, 78% achieve the milestone of first words, albeit late (Vasko et al., 2021). For non-verbal patients, various augmentative and alternative communication techniques are effectively employed and recommended to alleviate frustration and dysfunctional behaviors (Vasko & Schrier Vergano, 2021). It’s worth noting that the cited study on language by Vasko and Schrier Vergano in 2022 mainly relies on unstandardized parent reports and screening tools such as the Denver Developmental Screening Test.

In a recent cohort, Autism Spectrum Disorder (ASD) has been diagnosed in as many as 45% of CSS patients (Vasko & Schrier Vergano, 2021), although it is not as prominent a feature in SMARCB1-mutated patients. In a previous series of CSS patients predating DSM-5, approximately half exhibited features consistent with ASD (referred to as pervasive developmental disorder features in the text, Swillen et al., 1995). Among the CSS genes, ARID1B has been identified as a high-risk gene for ASD in a comprehensive exome sequencing study of ASD patients (De Rubeis et al., 2014). Furthermore, ARID1B haploinsufficiency has been linked to autism-like behaviors in mouse models (Shibutani et al., 2017).

Psychiatric comorbidities have not been systematically assessed, and the neurobehavioral presentation of CSS has not been extensively studied. Most of the available data pertain to ARID1B mutations, with hyperactivity and aggressive behaviors reported in approximately 10% of cases each, and poor social behavior and rigid behavior reported in around 6% of cases each (Santen et al., 2014). While hyperactivity is also occasionally reported in SMARCB1 cases, it has not been systematically evaluated (Bögershausen & Wollnik, 2018).

A less recent report involving 12 CSS patients with mild to moderate ID, which did not provide information on molecular diagnosis, highlighted the presence of aggressive behaviors in infancy, often preceded by temper tantrums and separation problems during school years. There were also reports of patients with CSS with particularly limited attempts to seek contact with adults, obsessive interests, rituals, and unusual fears, which improved with age (Swillen et al., 1995). A more recent publication has further documented this behavioral presentation (Bender et al., 2011). In one patient with an ARID1B mutation, psychosis has been reported (Mari et al., 2015). Additionally, seizures and tics have been reported but do not exhibit specific characteristics.

Reported cases of Coffin Siris Syndrome due to SMARCB1 mutation

From the initial description in 2012 to the present, literature has reported 49 patients with SMARCB1 mutation-associated CSS. Among the 12 reported patients for whom the degree of ID is specified, 8 have severe ID (Tsurusaki et al., 2012, Santen et al., 2013), 3 have moderate ID (Wieczorek et al., 2013, Gossai et al., 2015), and 1 has mild ID (Cheng et al., 2021). In the single SMARCB1 case from the Mari et al., 2015 cohort, the degree of ID is not specified. In the Tsurusaki et al., 2014 series, all 3 SMARCB1 cases are described as having developmental motor delay. In the Mannino et al., 2018 study, patients were not tested for IQ as they were too young at the time of publication, but 8 SMARCB1 cases exhibited developmental delay, with a mean age at the first word being 12 months and a mean age at walking acquisition being 30 months. One case from the Lee et al., 2021 cohort was not tested for IQ but showed language delay at 1 year. In the recent registry study by Vasko et al., 2021, involving 14 SMARCB1 patients (Vasko et al., 2021), there is no available data on IQ and the age at publication. It is not specified whether other patients underwent structured assessments; however, it is reported that only 3 out of 14 acquired walking, 6 had global developmental delay, 3 had ID, and 2 were diagnosed with ASD. For the 8 cases described by Sekiguchi et al. (2019) from a molecular genetic viewpoint, no clinical information is published. Most of the patients described had at least one brain abnormality. Recently, Schmetz et al. (2024) described two adult cases with SMARCB1 mutations. Both individuals presented with ID, central nervous system malformations, seizures, and were nonverbal.

Four additional patients, described by Diets et al. (2019), presented with the SMARCB1 mutation c.110G > A p.Arg37, which had not been previously reported in CSS type 3 cases. These patients exhibited a severe clinical presentation, including significant speech and motor delays, profound intellectual disability, and hydrocephalus caused by choroid plexus hyperplasia. Six additional patients have been identified in a fetal cohort. Among these prenatal diagnosis cases, intrauterine growth restriction was reported in 50%, and brain anomalies were reported in 33%. One patient exhibited an abnormal cardiac position, and another fetus was diagnosed postnatally with a cardiac septal defect. Thirty-three percent of the patients had hydronephrosis, and both male individuals had cryptorchidism detected postnatally (van der Sluijs et al., 2022).

We aim to present the first case of a CSS variant with a confirmed SMARCB1 mutation exhibiting normal intelligence and only mild neuropsychological deficits. We followed an adolescent patient with dysmorphic features, feeding difficulties, and growth and speech delays in infancy. Considering the rare features observed in this patient, we believe that the detailed description of this case is valuable to provide new insights into the heterogeneity of possible manifestations of a CSS variant associated to SMARCB1 mutation.

Method

We present a case of a patient who was evaluated at our Child and Adolescent Neuropsychiatry Unit, which is part of a tertiary pediatric referral center. The patient was referred to our neuropsychology team by the Medical Geneticist to assess her academic performance, which was slightlty below expectations. She had previously been under the care of our institution’s Medical Genetics Unit, undergoing yearly multidisciplinary assessments for CSS variant due to SMARCB1 mutation, which was diagnosed at 7 years of age. Her genetic evaluation comprised a CSS panel, which was selected based on the patient’s physical features, consistent with CSS, such as a coarse face with a low frontal hairline, broad eyebrows, thick lips, mild macroglossia, and short fifth fingers on both hands. This gene panel includes Next Generation Sequencing (NGS) of ARID1A (NM_006015), ARID1B (NM_020732), SMARCE1 (NM_003079), SMARCA4 (NM_001128849), SMARCB1 (NM_003073), SOX11 (NM_003108), ARID2 (NM_152641), DPF2 (NM_001330308) on genomic DNA of the proband and her parents by using the NimbleGen SeqCap Target Enrichment kit (Roche) according to the manufacture’s protocol on a NextSeq550 platform (Illumina). Proband and her parents underwent genetic testing after a specific informed consent was obtained.

Considering the rare genetic condition and the absence of a recent and comprehensive overall developmental assessment (cognitive, adaptive, linguistic, and emotional-behavioral), it was decided to proceed with a comprehensive neuropsychological and psychiatric evaluation that went beyond academic functioning. The patient’s medical history was reconstructed through chart review.

Psychodiagnostic evaluation

The neuropsychological and emotional-behavioral assessment was conducted entirely in the Italian language, utilizing tools in Italian. Specifically, the assessment included clinical observation, the use of standardized neuropsychological tests, the use of standardized questionnaires, and separate clinical interviews conducted with both the parents and the patient. All the instruments used present normative data calibrated for the Italian population.

Neuropsychological assessment

The neuropsychological assessment encompassed the evaluation of various domains, including cognitive level, adaptive functioning, language abilities, visual-motor integration abilities, long-term memory, and learning abilities. See Table 1.

Table 1. Summary of the neuropsychological assessment.

Domain	Test	Subtest and quotients	Raw score	Weighted score	Z score	Standard score	Qualitative description
Cognitive Level	WISC-IV	Block Design	26	9			Average score
		Similarities	17	9			Average score
		Digit Span	13	8			Average score
		Picture Concepts	20	14			High average score
		Coding	64	15			High average score
		Vocabulary	28	7			Low average score
		Letter-number Sequencing	17	10			Average score
		Matrix Reasoning	18	10			Average score
		Comprehension	16	8			Average score
		Symbol Search	29	13			Average score
		VCI				88	Low average score
		PRI				106	Average score
		WMI				94	Average score
		PSI				123	Above average score
		FSIQ				102	Average score
Adaptive Level	ABAS-2	Communication	71	14			High average score
		Community Use	49	11			Average score
		Functional Academics	66	14			High average score
		Home Living	42	10			Average score
		Health and Safety	52	9			Average score
		Leisure	54	10			Average score
		Self-care	62	7			Low average score
		Self-direction	66	12			Average score
		Social	62	10			Average score
		CC		40		121	Above average score
		PC		20		100	Average score
		SC		37		95	Average score
		GAC		97		105	Average score
Language	BVL 4-12	Naming	60		−0.5		Average score
		Lexical Comprehension	33		−1.5		Below average score
		Grammatical Comprehension	39		0.5		Average score
		Completing Sentences	11		−1		Low average score
Visual-Spatial Abilities	VMI	Integration	19	9	−0.41	94	Average score
		Visual Perception	24	12	0.67	110	High average score
		Motor Coordination	18	6	−1.33	80	Low average score
Long-Term Memory	PROMEA	Spatial Learning (TPo)	36	10			Average score
		Spatial Learning (DRs)	14	10			Average score
		Verbal Learning (TWo)	16	3			Exceptionally low score
		Verbal Learning (DRv)	8	7			Low average score
Reading Abilities	MT-3	Text Reading (Errors)	6.5		−0.84		Average score
		Text Reading (Syll/sec)	2.56		−1.27		Low average score
		Text Comprehension (A) (Accuracy)	4		−1.89		Below average score
		Text Comprehension (B) (Accuracy)	3		−1.73		Below average score
		Text Comprehension (A + B) (Accuracy)	7		−2.15		Exceptionally low score
	DDE-2	Word Reading (Errors)	5		−0.66		Average score
		Word Reading (Time)	126”		−1.15		Low average score
		Non-Word Reading (Errors)	6		−0.25		Average score
		Non-Word Reading (Time)	82”		−0.57		Average score
Writing Abilities	BVSCO-2	Text Spelling (Errors)	5		−0.99		Average score
		Writing Speed(Accuracy)	76		−0.35		Average score
Numerical and Mathematical Abilities	BDE-2	Counting	46	10	−0.03		Average score
		Number Reading	19	5	−1.77		Exceptionally low score
		Number Writing	13	7	−0.94		Average score
		Mental Multiplication	12	6	−1.24		Low average score
		Mental Calculation	8	5	−1.62		Below average score
		Arithmetic Facts	13	6	−1.22		Low average score
		Written Calculation	4	6	−1.14		Low average score
		Triplets	18	11	0.68		Average score
		Insertions	15	6	−1.2		Low average score
		Approximate Calculations	6	5	−1.67		Below average score
		NQ				72	Below average score
		CQ				56	Exceptionally low score
		NSQ				72	Below average score
		TMQ				68	Exceptionally low score

WISC-IV = Wechsler Intelligence Scale for Children – 4th Edition; VCI = Verbal Comprehension Index; PRI = Perceptual Reasoning Index; WMI = Working Memory Index; PSI = Processing Speed Index; FSIQ = Full Scale Intelligence Quotient; ABAS-2 = Adaptive Behavior Assessment System – 2nd Edition; CC = Conceptual Composite; PC = Practical Composite; SC = Social Composite; GAC = General Adaptive Composite; BVL 4-12 = Battery for Language Assessment in children aged 4 to 12 years old; VMI = Developmental Test of Visual-Motor Integration; PROMEA = Assessment of Memory and Learning for Developmental Age; TPo = Total Positions correctly recalled; DRs = Delayed Recall (spatial); TWo = Total Words correctly recalled; DRv = Delayed Recall (verbal); MT-3 = MT Clinical Tests – 3rd Edition; DDE-2 = Developmental Dyslexia and Dysorthography Assessment Battery - 2nd Edition; BVSCO-2 = Writing and Spelling Abilities Assessment Battery - 2nd Edition; BDE-2 = Diagnostic Battery for Developmental Dyscalculia – 2nd Edition; NQ = Numerical Quotient; CQ = Calculation Quotient; NSQ = Number Sense Quotient; TMQ = Total Mathematical Quotient.

Cognitive profile

The Wechsler Intelligence Scale for Children-Fourth Edition (WISC-IV; Wechsler, 2003; Orsini et al., 2012) is an individually administered test battery designed to assess the intellectual abilities of children and adolescents aged 6 to 16 years and 11 months (for more details, see Supplementary Materials).

Adaptive profile

The Adaptive Behavior Assessment System-Second Edition (ABAS-II; Harrison & Oakland, 2003; Ferri et al., 2014) is a parent-report rating scale that evaluates daily living skills in individuals ranging from children and adolescents to adults, spanning from 0 to 89 years of age (for more details, see Supplementary Materials).

Language

The Battery for Language Assessment in Children (BVL 4-12; Marini et al., 2015) assesses language skills in children aged 4 to 12 years. Specifically, the BVL evaluates production, comprehension, and repetition abilities and is divided into three sections (for more details, see Supplementary Materials).

Visual-spatial abilities

The Beery-Buktenica Developmental Test of Visual-Motor Integration (VMI; Beery, 1997; Preda, 2000) is one of the most widely used instruments for assessing visuomotor integration skills in children and adolescents ranging in aged 3 to 18 years (for more details, see Supplementary Materials).

Long-term memory

The Battery for Memory and Learning Assessment in Developmental Age (PROMEA; Vicari, 2007) includes a series of tasks that employ verbal, visual, or spatial stimuli to evaluate both short-term and long-term memory, encompassing both explicit and implicit components (for more details, see Supplementary Materials).

Reading abilities

The MT-3 clinical tests (MT-3)

The MT-3 Clinical Tests (MT-3; Cornoldi & Carretti, 2016) are designed to assess text reading and comprehension skills in students ranging from primary school through the second year of first-grade secondary school (for more details, see Supplementary Materials).

The battery for the evaluation of developmental dyslexia and dysorthograhia (DDE-2)

The Battery for the Evaluation of Developmental Dyslexia and Dysorthography (DDE-2; Sartori et al., 1995) assesses various aspects of reading and writing abilities (for more details, see Supplementary Materials).

Writing abilities

The Writing and Spelling Proficiency Assessment Battery − 2 (BVSCO-2; Tressoldi et al., 2012) assesses various aspects of writing skills and comprises 6 tasks that can be categorized into three groups: graphism, spelling proficiency, and written text production (for more details, see Supplementary Materials).

Numerical and mathematical abilities

The Battery for Developmental Dyscalculia (BDE-2; Biancardi et al., 2016) is designed for the assessment of numerical cognition in children and adolescents, spanning from the third grade of primary school to the third grade of secondary school (for more details, see Supplementary Materials).

Emotional-behavioural assessment

All tests used in the emotional-behavioral assessment were standardized for the Italian population. The emotional-behavioral assessment included conducting a semi-structured clinical interview (to both the child and her parents) and administering parent-report questionnaires that assessed symptoms such as anxiety, inattention, hyperactivity, depressed mood, oppositional behaviors, sleep disturbances, and more. See Table 2.

Table 2. Summary of the emotional-behavioral assessment.

Questionnaire	Subscale	Raw score	T score	Qualitative description
CBCL 6-18	Activities	7	34	Borderline range
	Social	7	44	Normal range
	School	5.5	52	Normal range
	Total Competence	19.5	36	Clinical range
	Anxious/Depressed	3	52	Normal range
	Withdrawn/Depressed	1	52	Normal range
	Somatic Complaints	5	66	Borderline range
	Social Problems	2	52	Normal range
	Thought Problems	0	50	Normal range
	Attention Problems	2	52	Normal range
	Rule-breaking Behavior	3	59	Normal range
	Aggressive Behavior	3	51	Normal range
	Internalizing Problems	9	58	Normal range
	Externalizing Problems	6	53	Normal range
	Total Problems	21	51	Normal range
	Affective Problems	2	56	Normal range
	Anxiety Problems	1	51	Normal range
	Somatic Problems	5	70	Clinical range
	ADHD Problems	2	51	Normal range
	Oppositional-defiant Problems	3	55	Normal range
	Conduct Problems	2	56	Normal range
	Sluggish Cognitive Tempo	0	50	Normal range
	Obsessive-compulsive Problems	2	55	Normal range
	Post-traumatic Stress Problems	2	50	Normal range
CPRS-R	Oppositional	1	41	Normal range
	Cognitive Problems/Inattention	2	44	Normal range
	Hyperactivity	0	40	Normal range
	Anxious-Shy	2	42	Normal range
	Perfectionism	1	40	Normal range
	Social Problems	0	43	Normal range
	Psychosomatic Problems	2	50	Normal range
	ADHD Index	0	40	Normal range
	CGI (Restless-impulsive)	0	40	Normal range
	CGI (Emotional lability)	0	42	Normal range
	CGI (Total)	0	39	Normal range
	DSM-IV Symptom Subscales (Inattentive)	1	42	Normal range
	DSM-IV Symptom Subscales (Hyperactive-impulsive)	1	41	Normal range
	DSM-IV Symptom Subscales (Total)	2	41	Normal range
SDSC	DIMS	6	38	Normal range
	SBD	3	45	Normal range
	DA	3	47	Normal range
	SWTD	6	41	Normal range
	DES	6	46	Normal range
	SHY	2	45	Normal range
	TOT	26	38	Normal range

CBCL 6-18 = Child Behavior CheckList 6-18; ADHD = Attention Deficit/Hyperactivity Disorder; CPRS-R = Conners’ Parent Rating Scale – Revised; CGI = Conners’ Global Index; DSM-IV = Diagnostic and Statistical Manual of mental disorders – 4th Edition; SDSC = Sleep Disturbance Scale for Children; DIMS = Disorders of initiating and maintaining sleep; SBD = Sleep breathing disorders; DA = Disorders of arousal; SWTD = Sleep wake transition disorders; DES = Disorders of excessive somnolence; SHY = Sleep hyperhydrosis; TOT = SDSC total score.

Kiddie-schedule for affective disorders and schizophrenia for school-aged children, present and lifetime version (K-SADS-PL)

The Kiddie-Schedule for Affective Disorders and Schizophrenia for School-aged Children, Present and Lifetime version (K-SADS-PL) is a semi-structured clinical interview to assess current and past psychopathological features and psychiatric disorders in juveniles aged 6 to 18 years according to criteria of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) (Kaufman et al., 1996).

The child behaviour checklist (CBCL)

The Child Behavior Checklist (CBCL, Achenbach & Rescorla, 2001; Frigerio, 2001) is a widely used questionnaire designed to identify behavioral and emotional issues in children and adolescents using a norm-referenced approach (for more details, see Supplementary Materials).

The conners’ parent rating scale-revised (CPRS-R)

The Conners’ Parent Rating Scale-Revised (CPRS-R, Conners et al., 1998; Nobile et al., 2007) is a well-established research and clinical tool used to assess parental reports of behavioral problems associated with ADHD in children and adolescents (for more details, see Supplementary Materials).

The sleep disturbance scale for children (SDSC)

The Sleep Disturbance Scale for Children (SDSC) questionnaire (Bruni et al., 1996) offers a comprehensive assessment of sleep disorders in children and adolescents aged 6 to 15 years (for more details, see Supplementary Materials).

Results

Clinical history

The patient was born to unrelated parents and has a positive family history of Turner syndrome in a 4^th degree relative. She was born at full term via planned caesarean section after an uneventful gestation, without perinatal complications. Her birth weight was appropriate for gestational age. She was initially artificially nursed initially for maternal hypogalactia; then, an intolerance to cow milk’s protein developed, leading to early weaning at 3 months of age. Feeding difficulties and growth delays persisted and were the primary concerns when she was referred to our centerat 14 months of age. At that time, her weight was between the 25th and the 50th percentile, length was between the 3rd and the 10th percentile, and head circumference was between the 25th and 50th percentile.

Phenotypical examination revealed a low frontal hairline, frontal angioma, frontal nevus flammeus, broad sparse eyebrows, thick lips, mild macroglossia, and short 4^th and 5^th metacarpals with a short fifth finger on both hands. Two-dimensional color-Doppler, abdominal ultrasound, ophthalmological, and audiometric evaluations all yielded normal results. Regarding psychomotor development milestones, independent walking was achieved at 14 months, while her first words were articulated at around 30 months, with an expressive language delay. Sphincter control was attained by the age of 3 years.

Regarding language development, delays in the onset of first words and slow vocabulary development were noted. At 40 months of age, mental development was assessed using the Griffith Mental Development Scale Extended Revised (Griffith, 2007), showing adequacy in all areas except for language, where the equivalent age of development was 30.5 months. The overall developmental quotient was 109 (Equivalent Age of 42.5 months). Language assessment was further conducted using the Phono-lexical test (TFL) (Vicari et al., 2008), revealing difficulties in expressive language, particularly in phonological and morphosyntactic aspects.

During a subsequent follow-up at age 5, a Leiter 3 (Cornoldi et al., 2016) was administered, resulting in a non-verbal IQ score of 75. Adaptive abilities, assessed with the ABAS II, were generally within typical expectations. Language was reassessed using the BVL 4-12, with lexical comprehension and morphosyntactic comprehension scores within the normal range. However, phrase repetition was notably deficient (below 2 standard deviations) due to the patient’s difficulty in following test instructions, resulting in the insertion of synonyms, passepartouts, and fillers. Additionally, scores on the Movement Assessment Battery for Children—Second Edition (ABC-2) (Henderson et al., 2007) were obtained, indicating performance at the 9th percentile. Manual dexterity and balance performance fell within the 25^th percentile, while aiming and catching skills were notably low, at the 0.5^th percentile.

Speech therapy, initially coupled with psychomotor therapy, commenced before the age of 4, with sessions held twice a week. Significant progress was observed in speech development, prompting a gradual shift in therapy objectives towards the enhancement of reading and writing skills. Therapy was concluded at 9 years of age upon the achievement of rehabilitation goals.

The patient also experienced delays in acquiring feeding and chewing skills, necessitating specialized evaluation and intensive rehabilitation training. At the age of 4, she underwent the first intensive rehabilitation cycle, enabling her to consume soft adult foods. A second cycle was completed at 5 years of age, focusing on achieving tearing and lateralization skills, as well as refining chewing techniques. While swallowing became effective, anterior chewing predominated, prompting the introduction of posterior and rotational chewing movements during training. Vertical chewing became automated, but rotational chewing required ongoing stimulation and reinforcement. Bottle feeding was discontinued, and parents were educated on promoting proper chewing habits at home.

Upon reassessment at 8 years old, chewing was effective but still exhibited lip compensation, resulting in prolonged meal times. Lingual thrust during swallowing was also observed, which saw improvement with continued at-home training. The patient remained selective in her dietary habits, exhibiting hypersensitivity to new foods. She continuously adhered to an exclusion diet for Food Protein-Induced Enterocolitis Syndrome related to Cow’s Milk.

As contextual socio-environmental factors, it is noted that parents received secondary education, and the socioeconomic status was middle to low. Community resources were limited. The national health system provided speech therapy services in the local area, but did not cover intensive training in her home region.

CNS imaging with Magnetic Resonance performed at 2 years of age showed normal results. No seizures were reported, and EEGs conducted at 3 and 6 years revealed mild nonspecific focal/diffuse anomalies (bursts of sharp waves) without clinical correlation.

Genetic testing, guided by the presence of dysmorphic features, was conducted when she was 7 years old. The CSS panel was chosen since the patient presented with clinical features suggesting CSS, including coarse face with low frontal hairline, broad eyebrows, thick lips, mild macroglossia, and both fifth fingers were short. The diagnosis of CSS was confirmed upon the discovery of a “de novo” heterozygous c.568C > T (p.Arg190Trp) variant in the SMARCB1 gene. The genomic variant was categorized as likely pathogenic (class 4), following the guidelines of the American College of Medical Genetics and Genomics (ACMG) (Richards et al., 2015). Other genetic investigations were performed concomitantly, including standard chromosomal analysis, arrayCGH, and molecular sequencing of the PTPN11 and BRAF genes, and all yielded normal results.

Current neuropsychological assessment

At the age of 10 years and 5 months, she was referred to our Child and Adolescent Neuropsychiatry Unit, specifically to the Learning Disorders service, for an evaluation of her neuropsychological functioning (including learning abilities) and emotional-behavioral profile. At the time of the assessment, she was in the fifth year of Italian primary school (5th grade of education), without requiring special services. Her overall academic performance was reported as good, with selective difficulties in mathematics. No specific cognitive complaints were reported in daily life. Personal autonomies were described in line with chronological age.

No emotional or behavioral problems were reported. Her eating habits were described as selective, while her sleep-wake pattern was noted as regular. Her relationships with peers and personal autonomy were described as normal.

We conducted extensive clinical, neuropsychological, and emotional-behavioral assessments, totaling approximately 9 h, divided into 3 sessions (one per day) (see Tables 1 and 2). In addition to the direct evaluation of the patient, we involved the parents in completing questionnaires related to the emotional-behavioral profile. We also administered the semi-structured K-SADS-PL interview to both the patient and her parents.

During the neuropsychological evaluation, the patient showed a willingness to engage with the clinicians and remained cooperative throughout all assessment phases. Eye contact was present and well modulated for communicative purposes. The speech appeared intelligible, with no phonological deficits present, and discourse organization was age-appropriate. There were no significant fine and/or gross motor coordination issues observed. The prosody of speech appeared consistent with the context.

Her attentional capacities were sufficient for completing the proposed activities, with no indications of motor restlessness. While there was some initial inhibition in interpersonal relations on the first day, the patient appeared more comfortable conversing with the clinicians in the subsequent days. Mild signs of performance anxiety were evident during academic tasks, especially in mathematics, where the patient reported experiencing greater difficulty. However, there were no significant indications of fatigue.

Summary of the Kiddie-schedule for affective disorders and schizophrenia for school-aged children, present and lifetime version (K-SADS-PL)

Following separate interviews with the parents and the patient, no threshold criteria for a DSM-5 psychiatric disorder were identified. Both the patient and the parents demonstrated openness during the interview and willingness to interact with the clinicians.

The patient was described by parents, observers and by herself as well, as having a good attention span, remaining focused in both academic and non-academic activities. There were no reports of hyperactive-impulsive behaviors; she was said to adapt well to the school environment, capable of maintaining attention during lessons and participating in leisure activities calmly. Compliance with rules across various life contexts was considered satisfactory. She was described as having a mildly reserved temperament, but this did not appear to affect her social functioning, despite some selectivity in her social circle, which was, nonetheless, considered adequate. There were no clinically relevant symptoms of anxiety; instead, she occasionally experienced somatic complaints anticipating academic challenges or mild health concerns, likely explainable by her awareness of the genetic condition. Her mood was reported as stable, with no reported depressive symptoms. No motor or vocal tics, nor any obsessive-compulsive symptoms were reported.

Discussion

We present the first case of CSS with a SMARCB1 mutation exhibiting normal IQ and adaptive functioning. The initial clinical presentation was of a female with dysmorphic features, feeding difficulties, and speech delay, who demonstrated remarkable progress as she grew older.

Her neuropsychological assessment at the age of 10 revealed a cognitive profile (i.e. WISC-IV) within the normal range, albeit heterogeneity across its constituent indices. While all indices fell within the normal range, there was a notable peak in information processing speed (i.e. PSI = 123) alongside some difficulties in the verbal area (i.e. VCI = 88). Consistently, this discrepancy could also be observed in some of the other administered tests within the evaluation. Specifically, when examining language abilities (i.e. BVL 4-12), despite an overall normal performance, there was a low average score solely in lexical skills, which in the verbal domain of the cognitive profile can significantly impact performance. These difficulties could be interpreted as residual outcomes of the language delay that characterized the patient’s developmental stages. They were also coherent with selective residual difficulties in verbal learning and memory revealed by further neuropsychological tests. Regarding visuo-spatial abilities and visuo-motor integration (i.e. VMI), the highest performance emerged in the “visual perception” subtest (i.e. standard score = 110), while some difficulties were observed in the “motor coordination” subtest (i.e. standard score = 80). Once again, this result was consistent with the described cognitive profile, highlighting a peak performance in information processing speed. Indeed, the “visual perception” subtest of the VMI could be significantly influenced by the speed of information processing, as it requires quickly discriminating between similar stimuli (similar to the “Symbol Search” subtest of the WISC-IV). Indeed, although the time limit is present in both the motor coordination and visual perception subtests, in the former, the predominant grapho-motor component might interfere with rapid processing skills.

The report of a normal intellect with a cognitive profile within the norm was supported by adaptive functioning (i.e. ABAS-2) in line with chronological age across all its components. Further neuropsychological assessments revealed mild neuropsychological deficits that warrant consideration. Regarding memory skills, a exceptionally low score (i.e. TWo) and a low average score (i.e. DRv) emerged in the verbal subtest, while scores within the norm were observed in the spatial subtest. This profile was consistent with the observed discrepancy in overall cognitive functioning, highlighting that, for the patient, the verbal domain was the most vulnerable area.

Regarding learning skills (i.e. reading, writing, and arithmetic), the patient exhibited widespread difficulties, with more pronounced challenges in text comprehension and numerical/mathematical abilities. Specifically, reading accuracy was within the norm in both text reading and word/non-word reading tasks, while low average scores were observed in reading speed in these tasks (excluding non-word reading). However, text comprehension in both presented passages was significantly impaired. Considering the not severe difficulties in text decoding, it is plausible that the comprehension challenges were largely attributed to linguistic fragility and the limited lexical skills already described in other assessments. Regarding writing, in the spelling task, a low average score emerged for the number of errors, partly influenced by some errors due to dialectal inflections rather than a genuine phoneme-grapheme conversion deficit. Writing speed, however, fell within the norm. As for numerical and mathematical abilities, overall, the patient exhibited widespread impairment (i.e. TMQ = 68), with a more pronounced deficit in acquiring calculation procedures (i.e. CQ = 56) and a milder one in the semantic domains of mathematical skills (i.e. QN = 72 and NSQ = 72).

No accompanying psychiatric comorbidities emerged, and the mild somatic symptoms reported did not significantly impact her family, school, or social functioning. She did not meet the criteria for an anxiety or somatic symptoms disorder as defined by DSM-5.

While normal IQ has previously been reported in a patient with an ARID1B mutation, our findings of a case of CSS with a SMARCB1 mutation and normal intellect challenges the existing knowledge on genotype-phenotype correlations in CSS and in BAFopathies and highlights our still incomplete understanding of their molecular pathology. In interpreting these findings, it must be noted how complex expression regulation and epigenetic mechanisms play a significant role as modifiers of phenotypic expression. Several mechanisms explaining differences in SWI/SNF-related disorders have been documented in the literature. Firstly, the ATP-dependent chromatin remodeling machinery, comprised of proteins encoded by these genes, is dosage sensitive. Secondly, different components within the complex have varying degrees of importance in its overall functioning (Kadoch et al., 2017). Thirdly, mutations at enhancer sites or in transcription factors can have complex effects on overall transcription within the complex (Aref-Eshghi et al., 2018). Furtherly, research into other potential mechanisms explaining gene-phenotype correlations, such as the distinction between in-frame and out-of-frame mutations, has yielded no significant results (Aref-Eshghi et al., 2018). Our report suggests that the neuropsychological phenotype of SWI/SNF-related disorders is various and diversified, even for mutations that have been associated with moderate to severe impairment to date. Further research is needed to investigate and understand the molecular mechanisms underlying this high degree of variation.

Our case is, to the best of our knowledge, the first in which a comprehensive neuropsychological assessment has been conducted in a patient with CSS. This limits direct comparison of visual spatial abilities, long term memory and learning skills of our case with previously reported patients. However, our case occupies a unique position in the current literature, as it presents a genetic etiology consistently associated, until now, with highly impaired neurodevelopmental phenotypes, alongside normal intellect and functioning. This allows us to expand neuropsychological assessment to investigate more subtle deficits in specific functions. In light of this, what we observed as mild and selective impairment in verbal learning and long-term memory may be interpreted as a residual manifestation of the molecular dysfunction of the BAF complex in the CNS. Conversely, the more extensive impairments in neurodevelopment features reported in previously described patients did not occur, purportedly due to the reasons discussed earlier. We propose interpreting the neuropsychological profile of our patient as the mildest manifestation within a distinct neuropsychological spectrum linked to SWI/SNF complex dysfunction. This perspective could be useful in proposing a characteristic “signature” of SWI/SNF complex dysfunction in the context of development and cognition, thereby contributing to a deeper understanding of the role of this molecular machinery in CNS functions. Our interpretation must be weighed against its limitations, inherent in the fact that our is a single report of a rare condition, with a limited number of patients thoroughly described in the literature to date. There might be patients with even milder phenotypes that have not been described yet and who do not present the learning and memory deficits we described. However, it is worth discussing our hypothesis in relation to preclinical data concerning the molecular physiology of the BAF complex. We hypothesize that preclinical studies on BAFcomplex dysfunction can support our suggestion that expressive language delay, verbal memory deficits, and difficulties in verbal learning, with a discrepancy in intelligence scores weighted toward lower verbal comprehension, may represent the mildest and potentially most specific neuropsychological markers of CSS.

The BAF complex exhibits expression in numerous tissues, with recent research highlighting its crucial role in the nervous system from embryonic development onwards. It acts as a regulator of Wnt signaling, influencing neural cell phenotype acquisition by either inhibiting or activating Wnt gene targets through interactions with β-catenin (Vasileiou et al., 2015). SOX11, a gene associated with a mild form of CSS known as type 9, is predicted to operate downstream of the BAF complex, contributing to subsequent stages of neurogenesis by promoting the conversion of postnatal glial cells into neurons (Ninkovic et al., 2013). Its role also extends to neural crest migration, which could potentially explain peripheral clinical manifestations (Chandler and Magnuson, 2016). Focusing on SMARCB1 mutations, murine models have demonstrated its significant role in midline glial cell migration and choroid plexus hyperplasia. Midline defects in preclinical models not only impact the corpus callosum but also affect structures like the hippocampal commissure and thalami (Filatova et al., 2019). Moreover, the role of the BAF complex in neurogenesis persists into adulthood. It has been shown to influence the function of radial glia-like cells implicated in adult hippocampal neurogenesis. Experimental approaches using murine models have established a significant correlation between the expression of a subunit, BAF170, and impaired adult neurogenesis, affecting learning outcomes in tests like the Morris water maze, notably in terms of reversal learning (Tuoc et al., 2017). The complex has definitely been demonstrated to contribute in mature neurons to memory formation, synaptic plasticity, and activity-responsive neurite outgrowth (Vogel-Ciernia et al., 2013). Furthermore, there is evidence of the BAF complex’s direct involvement in histone modifications necessary for long-term potentiation and synaptic plasticity (Vogel-Ciernia et al., 2013). Preclinical research has emphasized its distinctive role in memory formation and learning (Sokpor et al., 2017). This aspect is also discussed in the literature concerning genetic risk factors for ASD and schizophrenia (Son & Crabtree, 2014). In addition, a study on SMARCD1, a gene involved in BAF complex but not correlated to neurodevelopmental disorders to date, has shown on Drosophila models how the gene regulates neurodevelopmental gene expression during a critical time window of juvenile adult brain development when neuronal circuits required for learning and memory are formed (Nixon et al., 2019). We believe that this last observation, particularly in the context of preceding data on BAF complex function in the CNS, helps to correlate the data of SMARCB1 mutation to the clinical neuropsychological picture in our patient. Our patient exhibits a milder degree of neurodevelopmental impairment than previously reported CSS type 3 patients, but perhaps distinctive in defining the mildest end of the spectrum of memory and learning impairment due to SMARCB1 mutations. By evaluating a patient with CSS type 3 and unimpaired general intelligence, we were able to provide a more detailed description of deficits in specific domains. Our case paves a path for future research to identify specific modifiers of phenotypic expression to explain the variability in intellect in patients and individuate potential targets for gene therapy. As also advocated by Nixon and colleagues, understanding the role of SWI/SNF in the postnatal brain could unlock possibilities for postnatal gene therapy, as prenatal intervention in these developments is unlikely (Nixon et al., 2019).

There are however limitations to highlight in the interpretation of our findings. In particular, it is difficult to retrospectively ascertain if oral motor difficulties that also contributed to the feeding disorder could have impact on the language delay described at preschool age. Neurological physical examinations in the first years of life revealed no motor weakness or hypotonia, but deficits in planning and programming of the spatio-temporal parameters of speech movement sequences might have played a role, also considering the difficulties in motor coordination in chewing and milder difficulties in motor cohordination revealed with the ABC-2 retrieved in clinical history. However, both feeding and speech improved impressively at development. Children with childhood apraxia of speech have also more neuropsychological deficits, encompassing verbal and non-verbal memory and learning and persisting until adulthood (Bombonato et al., 2022; Kenney et al., 2006) and their improvement is poorer in adolescence (Lewis et al., 2004).

In conclusion, our findings may enhance the current understanding of the molecular mechanisms related to language, memory, and learning. Importantly, they also have clinical relevance as they could raise awareness about the varied presentations of pathologies related to the BAF complex. This may lead to expanded testing of SWI/SNF complex genes in patients with mild dysmorphism and language impairments.

Authors’ contributions

MA, AB, DM, and MCD designed the study. MA, AB, and MCD drafted the manuscript, with support of EP and SV. DM, SV and MCD supervised the study. All authors contributed to the article and approved the submitted version.

Ethical approval and consent to participate

The Bambino Gesù Children’s Hospital ethics committee provided ethical approval for this study (3231_OPBG_2023). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from the participant’s parents.

Acknowledgments

We would like to thank the participant and her parents for their willingness to participate in the study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical restrictions.

Additional information

Funding

This work was supported by the Italian Ministry of Health with “Current Research” Funds.