Social, neurodevelopmental, endocrine, and head size differences associated with atypical deletions in Williams–Beuren syndrome

Abstract

Williams–Beuren syndrome (WBS) is a multisystem disorder caused by a hemizygous deletion on 7q11.23 encompassing 26–28 genes. An estimated 2–5% of patients have “atypical” deletions, which extend in the centromeric and/or telomeric direction from the WBS critical region. To elucidate clinical differentiators among these deletion types, we evaluated 10 individuals with atypical deletions in our cohort and 17 individuals with similarly classified deletions previously described in the literature. Larger deletions in either direction often led to more severe developmental delays, while deletions containing MAGI2 were associated with infantile spasms and seizures in patients. In addition, head size was notably smaller in those with centromeric deletions including AUTS2. Because children with atypical deletions were noted to be less socially engaged, we additionally sought to determine how atypical deletions relate to social phenotypes. Using the Social Responsiveness Scale-2, raters scored individuals with atypical deletions as having different social characteristics to those with typical WBS deletions (p = .001), with higher (more impaired) scores for social motivation (p = .005) in the atypical deletion group. In recognizing these distinctions, physicians can better identify patients, including those who may already carry a clinical or FISH WBS diagnosis, who may benefit from additional molecular evaluation, screening, and therapy. In addition to the clinical findings, we note mild endocrine findings distinct from those typically seen in WBS in several patients with telomeric deletions that included POR. Further study in additional telomeric deletion cases will be needed to confirm this observation.

1 |. INTRODUCTION

Williams–Beuren syndrome (WBS, “Williams syndrome”; OMIM #194050) is an autosomal dominant, multisystem genetic disorder caused by a hemizygous deletion of 26–28 genes on chromosome 7q11.23. The severity of phenotypes varies between patients, but salient manifestations include characteristic facies, cardiovascular abnormalities such as hypertension and arterial stenosis (Pober, 2010), endocrine irregularities involving hypercalcemia (Sindhar et al., 2016), hypothyroidism (Cambiaso et al., 2007), early, but not precocious puberty (Partsch et al., 2002), abnormal glucose tolerance (Cherniske et al., 2004), and a predictable pattern of neurodevelopmental and cognitive strengths and weaknesses (Cherniske et al., 2004; Davies, Udwin, & Howlin, 1998; Klein-Tasman, Li-Barber, & Magargee, 2011; Kopp, Parrish, Lugo, Dougherty, & Kozel, 2018). The recurrent deletion is produced through nonallelic homologous recombination at low-copy repeat sites at each end of the Williams–Beuren syndrome critical region (WBSCR), with 24.9% of “transmitting” parents harboring an inversion in this area that predisposes to the gene loss (Bayés et al., 2003; Hobart et al., 2010). While the typical deletion size of 1.5–1.8 Mb occurs in the majority of patients, an estimated 2–5% of patients possess an “atypical” deletion, which can be longer or shorter than this characteristic deletion (Bayés et al., 2003; Pober, 2010). In the case of a longer deletion, additional gene loss may extend beyond the WBSCR toward the centromere, the telomere, or bidirectionally.

The cognitive and behavioral profile associated with WBS is well-established, with a mean intelligence quotient (IQ) of 50–60 and noted cognitive impairment of visuospatial abilities. This is balanced by relative strengths in language skills and facial processing abilities (Bellugi, Adolphs, Cassady, & Chiles, 1999; Martens, Wilson, & Reutens, 2008). Autistic features, most notably repetitive behaviors and impaired social cognition (van der Fluit, Gaffrey, & Klein-Tasman, 2012), and mild to moderate intellectual disabilities (Martens et al., 2008; Meyer-Lindenberg, Mervis, & Berman, 2006) are common in typical WBS. Patients with WBS are described as having a “peaks and valleys” social profile demonstrating strong social motivation combined with weaknesses in social awareness and social cognition, and a predilection toward restricted, repetitive behaviors (Klein-Tasman et al., 2011; Kopp et al., 2018). Other neurological manifestations of WBS include Chiari malformation type I (with an unknown prevalence, but estimates run as high as 10%) (Pober, 2010; Pober & Filiano, 1995). More importantly, seizures, including infantile spasms, or types of focal or generalized epilepsy are an uncommon presentation in patients with typical WBS deletions (Nicita et al., 2016).

Attempts at elucidating genotype–phenotype correlations in patients with atypical deletions are complicated by the variability of deletion length among individuals. A limited number of patients with atypical deletions have been described in the literature; characterization of these patients shows increased rates of autism spectrum disorder (ASD) (Edelmann et al., 2007; Fusco et al., 2014), infantile spasms (Marshall et al., 2008; Röthlisberger, Hoigne, Huber, Brunschwiler, & Capone Mori, 2010), epilepsy (Fusco et al., 2014; Ramocki et al., 2010), and more severe intellectual disabilities than typically seen in WBS (Ramocki et al., 2010). Seizures/infantile spasms have most commonly been associated with atypical distal deletions affecting the MAGI2 gene (Fusco et al., 2014; Marshall et al., 2008); however, MAGI2 may not be the sole contributor as other studies demonstrated that deletions sparing MAGI2 also result in infantile spasms (Marshall et al., 2008; Röthlisberger et al., 2010).

By adding 10 new patients with large WBS deletions to the literature, we aimed to ascertain phenotypic hallmarks of patients with atypical deletions and to expand upon or confirm findings in the existing literature. Furthermore, we intended to identify and quantify social differences that would further differentiate between patients with atypical and typical deletions in the clinical setting. In elucidating these differentiators, our goal is to assist physicians in recognizing patients with atypical WBS based on their clinical presentation. Many individuals with WBS were—and still are—diagnosed with fluorescent in situ hybridization (FISH) testing, a test type that does not yield information about deletion size. Because the deletion size is unknown in these cases, counseling often reflects only the features of typical WBS and lacks discussion that may be relevant to those with an atypical deletion size. These individuals would benefit from additional, high-resolution molecular testing, as increased knowledge about deletion size will better inform physicians and families about further deletion specific health maintenance needs.

2 |. MATERIALS AND METHODS

2.1 |. Human subjects

The human studies were conducted in accordance with the established protocols maintained by the Institutional Review Board at Washington University School of Medicine in St. Louis, MO and the National Heart, Lung, and Blood Institute, National Institutes of Health (NIH) in Bethesda, MD.

Individuals with WBS and their caregivers were recruited through our clinics (St. Louis Children’s Hospital, St. Louis, MO or NIH Clinical Center, Bethesda, MD) or the Williams Syndrome Association Family Conventions (July of 2014 and 2016). Consent was obtained from parents or legally authorized representatives of 106 individuals with a confirmed diagnosis of WBS. When possible, assent was obtained from cognitively capable individuals. Ninety-three individuals (44 males, 49 females) between the ages of 3 months and 65.5 years (M = 17.14; SD = 13.28) were ultimately studied after 14 individuals were removed due to absent data or secondary genomic abnormalities as outlined below.

Upon consent, parents or legally authorized representatives completed a medical history questionnaire through Research Electronic Data Capture (REDCap), an electronic data capture tool (Harris et al., 2009). Annual follow-up with consented individuals allowed for retrieval of up-to-date information and tracking of pertinent changes relating to the consented patients’ health. Guardians signed medical release forms for the collection of Supporting Information meant to complement the questionnaire.

Upon study enrollment, participants provided a sample of DNA via blood or saliva collection. WBS diagnosis was confirmed through chromosomal microarray analysis (CMA) (either previously performed clinically or research-based, see below for details) or quantitative PCR copy number (CN) assays (TaqMan probes (Thermo Fisher Scientific, Waltham, MA): AUTS2 (Hs04984177_cn), CALN1 (Hs04946916_cn), FZD9 (Hs03649975_cn), CLIP2 (Hs00899301_cn), POM121C (Hs07529820_cn), and HIP1 (Hs00052426_cn)). CN analysis was performed according to the manufacturer’s instructions and output data were analyzed using the CopyCaller Software (Thermo Fisher Scientific, Waltham, MA). Clinical arrays varied among individuals and are not all defined here (see Supporting Information Table for array types used in patients with atypical deletions). Research arrays used either CytoScan HD arrays analyzed via Chromosome Analysis Suite (Applied Biosystems, Foster City, CA), or Infinium OmniExpressExome-8 v1.6 arrays analyzed via GenomeStudio 2.0 (Illumina, San Diego, CA).

For the majority of subjects, including patients without a clinical or research array assessment, the deletion size was also confirmed via exome sequencing data (details on exome sequencing of these samples can be found in Kopp et al., 2018). To determine deletion size in WBS exomes on 7q11.23, we first extracted reads from BAM files using the CopywriteR package (Kuilman et al., 2015), calculated the log2 ratio of the reads from our WBS patients, and compared these with the average reads taken from a similarly sequenced bank of 36 non-WBS exomes. We then smoothened the log2 ratios for each individual and clustered these smoothened data in a heat map. Visualization of the heat map allowed detection of atypical deletions. In general, deletion size was well correlated with available CMA and CN analysis results (data not shown); in situations where exome data were of insufficient quality to determine the deletion size (n = 3, all of whom had previous CN analysis), a new sample was sent for CMA to confirm deletion.

The deletion size was considered atypical if it extended beyond the low-copy-number repeats on the centromeric (chr7:72,349,928) or telomeric (chr7:75,115,576) ends (hg19) or was smaller than the minimal 26 WBSCR genes (Pober, 2010). Individuals with atypical deletions that included Online Mendelian Inheritance in Man (OMIM) genes beyond the typical, 26–28 gene encompassing WBSCR, were grouped based on the direction of the increase (centromeric, telomeric, or bidirectional). From our data cohort, we initially identified 106 individuals with a completed Social Responsiveness Scale, Second Edition (SRS-2) form (see below), REDCap clinical information, and genetic testing (clinical or research CMA or CN analysis). Thirteen patients with shorter than typical deletions or with a known pathogenic secondary deletion or duplication were excluded from analysis.

2.2 |. Social Responsiveness Scale, Second Edition

The SRS-2 is a 65-item screening questionnaire that uses a Likert scale (1 = “not true” through 4 = “almost always true”) to quantify both the presence and severity of social impairment, with the total score serving as a screening tool for ASD (Constantino, 2012). Subscale categories include social awareness, social cognition, social communication, social motivation, and restricted, repetitive behaviors. The instrument is stratified into three age-specific categories (preschool, school, and adult) and can be used for individuals as young as 2.5 years old and up to 89 years of age. Total raw scores are transformed to T-scores, which are scaled to adjust for gender- and age-related differences. T-score values allow for stratification of severity, with higher scores indicating a greater degree of social impairment. T-scores less than or equal to 59 are considered to be within normal limits. T-scores of 60–65 are classified as mildly impaired, 66–75 as moderately impaired, and 76 and above as severely impaired. The instrument has been utilized previously in individuals with WBS, revealing a characteristic “peaks and valleys” pattern of strengths and challenges (Klein-Tasman et al., 2011; Kopp et al., 2018).

2.3 |. Literature review

To further define potential genotype–phenotype relationships with genes of interest outside of the WBSCR, we conducted a thorough review of the established literature. Criteria for inclusion of reported patient data in our study cohort included: (a) presence of an established WBS diagnosis with a noted atypical deletion pertinent to our areas of interest, and (b) absence of a reported secondary deletion or duplication. Atypical deletions of interest were defined as deletions larger than the WBSCR that covered the same or a smaller genomic space (Chr7:64,150,565–81,226,064; hg19) as the newly described patients, allowing for optimal genotype–phenotype correlations.

2.4 |. Statistical analysis and results presentation

All statistical evaluation was performed with SPSS Statistics software (IBM, Armonk, NY). With two-sample t-tests, we compared the means of the atypical- and typical-deletion groups. Nonparametric testing was used when indicated. Multivariate analysis of variance (MANOVA) was used to test any difference in multigroup means between SRS-2 subscores in typical versus atypical patients. Graphs and tables were prepared with Prism7 (GraphPad, San Diego, CA).

3 |. RESULTS

3.1 |. Description of subjects with atypical deletion size

Our final cohort consisted of 93 individuals (83 with a typical WBS deletion and 10 with a larger than typical—henceforth referred to as an “atypical”—WBS deletion) with no known pathogenic secondary deletions or duplications. Of the 10 identified atypical deletions, 4 extended in the centromeric direction, 5 in the telomeric direction, and 1 bidirectionally (Supporting Information Table). Patients D and E were from the same family and were both categorized in the centromeric-deletion group. Their deletions were noted to spare the GTF2i gene within the WBSCR on the telomeric side and extend centromerically to the region just outside of AUTS2, sparing it as well.

We then reviewed data from the published literature and found 17 individuals who met our criteria for atypical WBS deletions (Fusco et al., 2014; Komoike et al., 2010; Marshall et al., 2008; Röthlisberger et al., 2010). This cohort included 2 centromeric deletions, 10 telomeric deletions, and 5 bidirectional deletions. When available, the coordinates of these deletions were converted to hg19 and plotted alongside the coordinates of our patients (Figure 1). The OMIM-designated genes in the region are depicted at the top of the figure and are color coded based on their gnomAD probability of loss-of-function intolerance (pLI) score to depict the genes most likely to impact deletion phenotype. Eighteen additional individuals from the literature review were excluded due to secondary deletions or 7q11.23 deletions extending beyond the areas of interest.

FIGURE 1.

Large atypical deletions detected in individuals with WBS. Collection of Williams–Beuren syndrome (WBS) deletion sizes from our cohort (darker colors; n = 10) and from individuals identified through literature review (lighter colors; n = 17), with OMIM genes in the area represented at the top of the figure. The genes are color coded by gnomAD probability of loss-of-function intolerance (pLI) score: pLI ≤0.1: light gray, pLI 0.11–0.89: dark gray, pLI ≥0.9 black. Dotted sidebar: deletion area of uncertainty based on older testing methods, White sidebar: low-copy-number repeats in the WBSCR (not well characterized by chromosomal microarray analysis) shown on the typical deletion

3.2 |. Clinical descriptions of new subjects with atypical WBS deletions

3.2.1 |. Patient A (bidirectional, 17.1 Mb)

Female child with a significant history of infantile spasms and epilepsy. Symptoms were noted at 3 months of age, with clusters of flexor spasms lasting around 30 min. Hypsarrhythmia was noted on electroencephalography, which also showed multifocal epileptiform spike discharges consistent with epileptic encephalopathy. MRI demonstrated increased folding in the frontal lobes with small gray matter heterotopias along the anterior frontal horns bilaterally. Multiple medication regimens were attempted for seizure control until she was stabilized on clobazam. She had microcephaly on a typical growth chart and her head circumference was characterized as ~−3 SD on the WBS-standardized occipitofrontal circumference (OFC) curve with concern for secondary craniosynostosis. The patient displayed severe acne at around 2 years old. Given the POR deletion, endocrine evaluation was undertaken, which demonstrated a 17-OH progesterone level of 354 ng/dL (upper limit of normal [ULN]: 100 ng/dL for prepubertal females). Repeat evaluation showed a somewhat lower level, which was felt to represent the heterozygous POR state. Her morning cortisol level showed appropriate elevation. There was no evidence of genital malformations, though she was eventually diagnosed with precocious puberty and was started on Supprelin at age 6 years. Clinical gene sequencing revealed no concerning variants in the remaining POR allele. An evaluation at 5 years of age showed global developmental delay with developmental skills equivalent to a 6-month level using the Mullen Scales of Early Language, a standardized developmental test for children birth through 68 months (Mullen, 1995). Her communication skills were minimal, both verbally and using picture boards. Neither eye contact nor social interaction was noted by the evaluators. Additionally, she was unable to transfer items between hands and was nonambulatory. She continually demonstrated repetitive movements and self-injurious acts. She had sleep problems comprised of sleep apnea with frequent nightly awakenings. She had a history of pulmonary valve stenosis with mild supravalvular aortic stenosis (SVAS), neither of which required intervention. She passed away unexpectedly at 8 years of age due to a gastric perforation and sepsis following a brief illness with associated dehydration.

3.2.2 |. Patient B (centromeric, 8.0 Mb)

Male teenager with a history of ASD and intellectual disability. The patient was nonverbal and communicated by mechanically pointing to objects, and by using a communication board and limited sign language. He experienced chronic sleep problems including delayed onset and frequent nighttime waking. He had notable microcephaly on a typical growth chart and his head circumference was at ~−3 SD on the WBS OFC growth chart with bitemporal narrowing; his facies were atypical for WBS. There were no abnormal cardiovascular findings. There was no history of infantile spasms or seizures and no diagnosis of psychiatric conditions other than the ASD, although the latter was difficult to assess given his nonverbal state.

3.2.3 |. Patient C (centromeric, 8.3 Mb)

Male infant born at 36 2/7 weeks with multiple congenital anomalies including hypoplastic left heart syndrome (HLHS) and aortic arch with discrete coarctation, ectopically placed right kidney with duplicated collecting system, microphallus, and rocker bottom feet. There was prenatal concern for intrauterine growth retardation. At birth, he showed microcephaly on a typical growth chart and his head circumference was at −2 SD on the WBS OFC chart. The patient underwent stage I of the hybrid procedure for HLHS. His postoperative period resulted in a deteriorating clinical picture leading to extended ICU admission, until he passed away at age 2 months. He had cardiac failure secondary to HLHS with profound pulmonary and hepatic congestion noted as the immediate cause of death. No infantile spasms or other seizure-like activity were noted. Developmental outcomes were unknown given the circumstances.

3.2.4 |. Patient D (centromeric, 3.6 Mb)

Male child with a history of severe SVAS with repair in early infancy, and strabismus, also corrected surgically. He was diagnosed with attention deficit hyperactivity disorder (ADHD), generalized anxiety disorder, oppositional defiant disorder, and anger outbursts. Some self-injurious behavior was noted. Evaluation at age 6 established his IQ to be in the borderline range (Full Scale Intelligence Quotient [FSIQ]: 77); per Wechsler Preschool and Primary Scale of Intelligence Scale, Fourth Edition (Wechsler, 2012). (Standard scores have a mean of 100 and an SD of 15), with low average scores in verbal comprehension, perceptual reasoning, working memory, and processing. Of note, similar to his family member (Patient E), his facial features were not typical for WBS. His head circumference was at −2 SD on a typical growth chart and at ~−1 SD on the WBS OFC chart.

3.2.5 |. Patient E (centromeric, 3.6 Mb)

Female diagnosed in adulthood with WBS, who had an overall mild clinical course including no significant cardiovascular disease. Her facies were not typical for WBS. Her medical history was notable for cholelithiasis, acid reflux, menorrhagia, and difficulties with stress, anxiety, specific phobias, and sleep disturbances. Wechsler Adult Intelligence Scale, Fourth Edition (Wechsler, 2008) (mean = 100, SD of 15), established her IQ as borderline (FSIQ: 75), with similar borderline scores in verbal comprehension and processing speed, and low average performance in perceptual reasoning and working memory. She received special education instruction throughout her schooling. There was no history of infantile spasms, seizures, or ASD diagnosis. She was microcephalic on a typical growth chart, but her head circumference was at ~−1 SD on the WBS OFC chart. Early developmental milestones are unknown.

3.2.6 |. Patient F (telomeric, 4.0 Mb)

Male child with a history of mild branch pulmonary artery stenosis and pulmonary hypertension diagnosed in infancy. He had a cervical spine stenosis at C1–C3, and systemic hypertension managed with medication. His medical history included buried penis, penoscrotal webbing, and phimosis (with repair at 9 months old). The patient was noted to have generalized hypotonia with no evidence of muscle weakness or neuropathy-like clinical presentation other than a history of abnormal urodynamic studies. Developmental milestones were consistent with WBS established timelines at 1.5 years of age. There was no known history of infantile spasms, seizures, or psychiatric conditions. His head circumference was in the normal range on a typical growth chart and was at +1 SD on the WBS OFC chart.

3.2.7 |. Patient G (telomeric, 3.6 Mb)

Middle-aged female with cognitive impairment and late diagnosis of WBS (in adulthood). She had hypertension, which was treated with multiple antihypertensive medications. She had no known vascular stenosis, or genitourinary or renal anomalies. From a psychiatric perspective, she had ADHD, anxiety, and a specific phobia. Notably, she had had limb and facial (bilateral frontalis muscle) weakness likely present since birth—she was originally diagnosed as having cerebral palsy. However, her complaints of weakness increased in adulthood and evaluation of this finding revealed diminished lung capacity. She received a tracheostomy in her 40s and required ventilation at night. There was no history of infantile spasms or seizures, and her developmental milestones were unknown. Caregivers, however, reported a stroke or white matter change on an MRI performed in her 40s. The family did not recall a specific event that might have been associated with a stroke. Her head circumference was in the normal range on a typical growth chart and at +1 SD on the WBS OFC chart.

3.2.8 |. Patient H (telomeric, 3.4 Mb)

Male child with a history of mild right pulmonary artery stenosis, delayed gastric emptying with oral aversions, and episodic hypercalcemia. He was noted to have microphallus with hypospadias, which was corrected at 18 months of age. His 17-OH progesterone level was mildly elevated to 165 ng/dL (ULN: 110 ng/dL for prepubertal males) on one occasion, but was lower upon re-evaluation. His morning cortisol level was not suppressed. He had chronic mild elevation of serum transaminase levels, but no liver failure. A liver biopsy was nondiagnostic, but showed portal vein dilatation with ductular reaction and pericholangitis. MRI of the liver was normal. There was a significant developmental delay, predominantly in language and social interaction domains, with no spoken words by his last evaluation at age 5. The patient uses one sign in communication. Minimal interaction and engagement with people were noted, including reduced eye contact. His medical history was notably absent for microcephaly, infantile spasms, seizures, or psychiatric disorders (anxiety, phobias, and depression), but this was difficult to assess given his limitations with communication. His head circumference was in the normal range on a typical growth chart and at +2 SD on the WBS OFC chart.

3.2.9 |. Patient I (telomeric, 4.7 Mb)

Female child with a history of mild, stable SVAS, and chronic constipation requiring bowel cleanout and daily laxative regimen when younger. An extensive neuropsychological workup at age 9 years old, which showed her IQ and other measures within the Wechsler Intelligence Scale for Children, Fifth Edition (Wechsler, 2014) (mean = 100, SD = 15) to be extremely low (FSIQ: 48). She was noted to have delayed language development that met diagnostic criteria for expressive and receptive language disorder, which improved with therapeutic intervention. She had a history of generalized anxiety disorder, specific phobia, ADHD, and impulse control difficulties, with aggression toward peers in some instances. She had early adrenarche; however, hormone testing was not consistent with puberty and her bone age was reduced. Because of her POR deletion, the 17-OH progesterone level was evaluated and was found to be elevated at 218 ng/dL. Her morning cortisol level was not suppressed. Earlier in her life, there was a concern for seizures, but the workup was negative, with the episodes being more consistent with clonus. Her head circumference was in the normal range on a typical growth chart and at +1.5 SD on the WBS OFC chart.

3.2.10 |. Patient J (telomeric, 2.4 Mb)

Female child with a mild history of hypercalcemia and hypercalciuria, exotropia and astigmatism, and feeding intolerance with chronic constipation. There were no abnormal endocrine laboratory values or early puberty. Her developmental delay was not greater than that of an individual with a typical WBS deletion. She was diagnosed with ADHD that was well controlled with medication. She had no history of infantile spasms, seizures, or other known psychiatric conditions including specific phobias or sleep abnormalities. Her head circumference was in the normal range on a typical growth chart and was at about the 50th percentile on the WBS OFC chart.

3.2.11 |. Genotype Phenotype Correlation with previously described phenotypes

Medical histories for the newly described patients were evaluated for findings established in the medical literature (Table 1). Epilepsy has been linked to numerous genes distal to the WBSCR, including HIP1 (Fusco et al., 2014; Ramocki et al., 2010), MDH2 (Ait-El-Mkadem et al., 2017), YWHAG (Fusco et al., 2014; Komoike et al., 2010; Ramocki et al., 2010), and MAGI2 (Marshall et al., 2008; Röthlisberger et al., 2010), as well as to one gene from the centromeric side, KCTD7 (Kousi et al., 2012; Staropoli et al., 2012; Van Bogaert et al., 2007). After examining our cohort, only Patient A had a history of infantile spasms and seizures. Her seizures were refractory to multiple medications. Her deletion covered all of the potentially associated epilepsy-related genes. However, she was the only individual in our atypical-deletion cohort who had a deletion involving MAGI2. Four other patients with at least a partial YWHAG deletion (Patients F–I) and five others with at least a partial HIP1 deletion (Patients F–J) had no history of seizures, including infantile spasms. Similarly, Patient B, whose deletion included KCTD7, was also seizure free. These results were similar to those described by Marshall et al. (2008), and were consistent with reports showing that seizures, for most of these genes, have been associated with biallelic mutations (Kousi et al., 2012; Staropoli et al., 2012; Van Bogaert et al., 2007). There was no statistically significant difference in the presence of seizures between our atypical- and typical-deletion groups (p = .47), when taken as a whole. However, there was a notable significant difference (p < .001) when Patient A was added to the patients with MAGI2 deletions from the literature (n = 7) and this group was then compared with our remaining cohort (n = 92) of MAGI2 intact individuals plus the other patients without a MAGI2 deletion shown in the table (n = 11).

TABLE 1.

Phenotypic variables of interest related to associated deleted genes in Williams–Beuren syndrome (WBS) patients with large atypical deletions

		Diagnosis
	Patient	Age	Method	Microcephaly	Seizures	Development	Cardiovascular	Psychiatric history	Genital and endocrine	Other
Bidirectional	Marshall 11		Clinical		++	++ severe global DD
	Marshall 16		Clinical		++	++ severe ID
	Marshall 14		Clinical		++	++ severe psychomotor delay
	Marshall 13		Clinical		++	++ severe psychomotor delay, nonverbal	++ WPW
	Röthlis-Berger	4 mo	FISH	+ −1 SD (WBS)	++	++ severe motor DD	+	−	+ hypothyroidism
	A	1 mo	FISH	++ −3 SD (WBS) 2° craniosynostosis	++	++ global DD	+	++ self-injurious behaviors	++ ↑ 17-OH-Prog; precocious puberty	Gastric perforation; deceased
Centromeric	Marshall 1		Clinical						+ hypercalcemia
	Komoike 1	1 mo	FISH	+ −1 SD (WBS)	++	++ severe ID, Motor DD	+	−
	B	18 mo	FISH	++ −3 SD (WBS)	−	++ ID, nonverbal at 16 years	−	+ ASD	−
	C	1 d	CMA	++ −2 SD (WBS)	−	−	++ HLHS, coarctation	−	++ microphallus	Deceased
	D	7 mo	FISH	+ −1 SD (WBS)	−	+	+	++ ODD, self-injurious behav, ADHD, anxiety	−
	E	23 yr	FISH	+ −1 SD (WBS)	−	Early milestones unknown	−	+	−
Telomeric	Fusco 179		FISH/MLPA	+ −1 SD (WBS)	−	+	+	++, self-injuries ADHD, PDD	−
	Fusco 160		FISH/MPLA	++ −2.2 SD (WBS)	++	++ “severe” NOS	−	+ anxiety, obsessions	−
	Komoike 2	<18 mo	FISH		++		+	−
	Marshall 5		Clinical			++ severe ID
	Marshall 6		Clinical			++ severe ID
	Marshall 7		Clinical			++ severe ID
	Marshall 8	3 wk	FISH	−		++ severe ID	+			PNH
	Marshall 10	15 mo	FISH	−	++ IS, hypsarrhythmia, atypical	++ severe DD (unspecified)	+			1° craniosynostosis
	Marshall 12		Clinical		++ IS, seizure
						++ severe DD (unspecified)
	Marshall 15		Clinical		++ IS, focal seizures	++ severe ID
	F	6 mo	CMA	−	−	+	+	−	++ Penoscrotal web, phimosis	Cervical spine stenosis
	G	32 yr	FISH	−	−	Early milestones unknown	−	+ anxiety, ADHD, phobia	−	Weakness, poss. stroke, tracheostomy
	H	15 mo	CMA	−	−	++ nonverbal at 5 years	+	−	++ ↑17-OH prog microphallus, hypospadias	↑ serum transaminases
	I	4 mo	CMA	−	−	++ speech delay	+	+ anxiety, ADHD	++ ↑ 17-OH prog
	J	2 yr	FISH	−	−	+	+	+ ADHD	−

Note: Patients with WBS from our cohort (darker colors; n = 10) and from patients identified through literature review (lighter colors; n = 17). “+” indicates a positive clinical history, including features seen in WBS patients with a typical deletion. “++” indicates a more severe or unusual finding than seen in WBS patients with a typical deletion; specific, individual findings are noted within respective cells. “–” indicates there was no finding in this category. Gray cells represent an unknown history (not reported by original authors).

Abbreviations: ↑, increased; 17-OH prog, 17-OH progesterone; ADHD, attention deficit hyperactivity disorder; ASD, autism spectrum disorder; behav, behavior; d, day; DD, developmental delay; ID, intellectual disability; IS, infantile spasm; MLPA, multiplex ligation-dependent probe amplification; mo, month; NOS, not otherwise specified; ODD, oppositional defiant disorder; PDD, pervasive developmental disorder; PNH, periventricular nodular heterotopia; Poss, possible; wk, week; WPW, Wolf Parkinson White; yr, year.

3.3. |. Phenotypic differences in patients with typical versus atypical WBS deletions

In addition to evaluation of previously published phenotypes, we also looked for increased frequency or severity of typical WBS traits in individuals with an atypical deletion from our cohort. Based on data collected through parental report questionnaires, there were no appreciable differences noted between the typical- and atypical-deletion groups involving: (a) the timing of developmental milestones including rolling over (p = .14), sitting (p = .11), walking (p = .649), running (p = .53), transferring objects between hands (p = .68), and first word (p = .76); (b) the frequency and presence of cardiovascular abnormalities including SVAS (p = .83), supravalvular pulmonic stenosis (p = .87), and systemic hypertension (p = .23); (c) the presence of endocrine abnormalities including hypothyroidism (p = .37), hyperglycemia (p = .98), obesity (p = .86), hypercalcemia (p = .80), and hypercalciuria (p = .43); and (d) the presence of various psychiatric conditions including ADHD (p = .18), depression (p = .98), and anxiety (p = .11). The presence of phobias, however, was significantly different between the two groups (p = .04), with patients with typical deletions having a higher prevalence. After correction for multiple testing, the result did lose statistical significance.

Because not all data were available in the literature, statistical analyses could not be performed on the larger combined data set. However, examining our new cases with atypical deletions together with the patients from the literature review revealed some features uncharacteristic of WBS. While mild to moderate developmental delays are expected in WBS (Mervis & Velleman, 2011), more severe instances, such as complete lack of verbal communication in teenage years and/or developmental regression, were witnessed in several patients (Patients A, B, H, Marshall 10, 11 and 12, Röthlisberger, Komoike 1, and Fusco 160). Additional unusual findings included HLHS along with rocker bottom feet in Patient C, gastric perforation in Patient A, Wolff–Parkinson–White syndrome in Marshall Patient 13, periventricular nodular heterotopia in Marshall patient 8 (also described in [Ferland et al., 2006]), and craniosynostosis in Marshall patient 10 (also described in [Morimoto et al., 2003]).

An additional area of attention was the psychiatric history. From our cohort, two participants (Patients A and D) had a history of self-injurious behavior. Patient A was perceived by family to have a very high pain tolerance, which they believed contributed to her self-injury, while Patient D carried an additional diagnosis of oppositional defiant disorder with anger outbursts. Fusco et al. (2014) reported that Patient 179 in their cohort had a history of significant aggressive behaviors, self-injurious acts, ADHD, temper tantrums, and pervasive developmental disorder. Deletion breakpoints were dissimilar in these three patients. Consequently, no one specific gene could be implicated in the phenotype.

Early death was seen in two individuals in the total (n = 27) atypical deletion cohort (p = .15 as compared to the typical deletion cohort in which one death was reported). Significant congenital heart disease contributed to the death of Patient C, and a fatal acute gastric perforation occurred in Patient A. These two individuals had a deletion of 8 and 17 Mb, respectively.

We also reviewed the patients’ diagnosis data and found that children with atypical deletions tended to have a similar diagnostic pattern as those with typical deletions, with no difference in age at diagnosis (typical 3.2 ± 6.4 years [mean ± SD], atypical 4.3 ± 11.5 years, p = .13 Mann–Whitney). However, two of the patients with atypical deletions were diagnosed quite late (in adulthood), while the remainder where data were available were diagnosed under 18 months. An assessment could not be made on choice of molecular method for diagnosis (a potential surrogate for clinician suspicion of atypical deletion) because many participants, both with typical and atypical deletions, were initially diagnosed prior to the clinical availability of CMA. Forty percent of the new cohort had CMA as their diagnostic test.

3.4 |. Social features of patients with atypical WBS deletions

As previously described, individuals with typical WBS deletions reveal a “peaks and valleys” performance on the SRS-2 with: more profound impairment in social cognition and restricted, repetitive behaviors; less profound impairment in social awareness and social communication; and overall preservation of social motivation (Klein-Tasman et al., 2011). Our typical-deletion cohort (n = 83), which is also described in Kopp et al. (2018), demonstrated strikingly similar patterns (average T scores for each group shown in Figure 2). The percentage of patients with moderate-to-severe impairment in total SRS-2 T score was 53.01%. The percentage of patients with moderate-to-severe impairment in SRS-2 T subscores were as follows: 37.3% in awareness, 61.5% in cognition, 41.0% in communication, 8.4% in motivation, and 54.2% in restricted, repetitive behaviors.

FIGURE 2.

Social Responsiveness Scale T-scores for individuals with typical or atypical Williams–Beuren syndrome (WBS) deletions. (a) SRS-2 subscale T scores (social awareness, social cognition, social communication, social motivation, and restricted, repetitive behaviors) are shown for patients with typical WBS deletions (n = 83, gray) and for patients with atypical WBS deletions (n = 8, black). For each patient, the subscores are connected by a line revealing the previously reported peaks and valleys pattern for typical deletion patients but not in those with atypical deletions (p = .001). (b) SRS-2 subscale T score means and SDs for each subgroup. The dotted line demonstrates the threshold of clinical significance as scores less than or equal to 59 are considered within normal limits. **Atypical deletions were associated with higher T-scores (more abnormal) in the subscale “social motivation” (p = .005)

The atypical-deletion group in our cohort, however, exhibited a different pattern. Although the total T score was similar in both groups (data not shown), patients with atypical deletions (n = 8; the other two patients were too young at the time of evaluation) did not show the usual WBS peaks and valleys pattern on the SRS-2 T scores (p = .001, Figure 2a). Subsequent subgroup testing showed that individuals with had a higher mean T-score (62.00; SD = 12.75) for the subscale “social motivation”—i.e., they were more affected—than those with typical deletions (M = 52.25; SD = 8.58; p = .005; Figure 2b). Differences in other treatment subscores were not observed.

3.5 |. Subgroup comparisons

Because of small patient numbers within subgroups (centromeric vs. telomeric vs. bidirectional), a statistical subanalysis could not be performed. However, when viewed together, it appeared that each of the larger deletion types had a negative impact on social motivation, with the largest (bidirectional) deletion having the most abnormal result (Supporting Information Figure A). The centromeric-deletion group in our cohort, however, required additional investigation. This group consisted of three individuals (Patient C was too young to complete the SRS-2), one with a deletion spanning the entire WBSCR but extending beyond AUTS2 (Patient B) and two individuals with a deletion that spared GTF2i within the WBSCR and AUTS2 (Patients D and E). Deletions and/or point mutations in GTF2i and AUTS2 have been associated with impaired social functioning (Beunders et al., 2016; Borralleras, Sahun, Perez-Jurado, & Campuzano, 2015; Crespi & Hurd, 2014; Jolley et al., 2013). When the centromeric-deletion group was divided, the patient with the deletion containing GTF2i and AUTS2 (Patient B) had a social motivation score similar to the patient with the bidirectional deletion, while the patients with a deletion sparing GTF2i and AUTS2 (Patients D and E) had lower T scores (less impairment) in all areas than the typical-deletion group (Supporting Information Figure B).

4 |. DISCUSSION

WBS is a well-described genetic disorder affecting all major organ systems with a predictable collection of features. Currently, many patients are diagnosed using advanced molecular methods. However, for a variety of social and economic reasons, some patients have a clinical diagnosis or FISH as the only testing modality. The ability to identify patients who may deviate from typical features of WBS will allow clinicians to determine when additional testing is indicated, providing a more accurate diagnosis and improved health maintenance. In this study, we focus on patients with atypical deletion size and specific features that should suggest evaluation for larger centromeric and/or telomeric deletions, but countless other phenotype combinations exist in which a patient may manifest additional clinical features not commonly seen WBS. In those cases, it should not be assumed that the patient merely exhibits a rare presentation of a rare disease. Rather, the existence of a second diagnosis should be considered and additional molecular diagnostics sought, either through the use of CMA, focused gene panel testing, or exome. Studies in 22q11.2 deletion syndrome, for example, have shown a second diagnosis rate of nearly 1% in their patients (Cohen et al., 2018).

Previous reports suggested that individuals with larger than typical deletions have increased risks of seizures, developmental delay and ASD (Edelmann et al., 2007; Fusco et al., 2014; Komoike et al., 2010;Marshall et al., 2008; Röthlisberger et al., 2010). Our study confirmed these findings with the addition of 10 newly reported individuals with large atypical deletions. Marshall et al. (2008) proposed an association of MAGI2 deletion with infantile spasms or seizures given that 15 of their 16 patients with a WBSCR deletion expanding in the telomeric direction to include MAGI2, had such a phenotype. When MAGI2 was not included in the deletion, seizures were not typically present (Marshall et al., 2008; Röthlisberger et al., 2010). Additional studies expanded on this work by offering further exploration into the biological mechanisms of such a deletion (Komoike et al., 2010), as well as theorizing about other candidate genes in addition to MAGI2 that may contribute to the seizure phenotype. Our findings paralleled those in the existing literature, with a history of seizures only reported in our one patient with a bidirectional deletion that included MAGI2. When combined with the cases from the literature in whom MAGI2 was affected, the frequency of seizures in this group was greater than in our cohort of patients with a typical deletion.

Previous literature described “severe developmental delay” in patients with larger deletions, but additional skill-specific data were not readily reported (Marshall et al., 2008). As in these studies, our data also showed more severely affected neurodevelopmental presentations in patients with atypical deletions. Several of our patients remained nonverbal well into their school-age years, while others demonstrated expressive and receptive language disorders. In our survey, no statistical difference was noted for timing of early developmental milestones between those with typical and WBS deletions. These results were likely limited by sample size and precision of recall, especially for older children. Formal prospective developmental testing in a cohort of patients with atypical WBS deletions may yield clearer results.

To quantify differences in social functioning in our cohort, we used the SRS-2. While subjects with typical WBS deletions presented with a relative strength in social motivation, they had increased challenges in social cognition and displayed restricted, repetitive behaviors, as was also shown in previous studies (Klein-Tasman et al., 2011; Kopp et al., 2018). The patients with atypical deletions, however, had higher (more abnormal) social motivation scores than those with typical deletions. This suggested that deletion of other high/moderate probability of loss of function intolerance (pLI) score genes, in particular of AUTS2 on the centromeric side and/or of HIP1, YWHAG, and MAGI2 on the telomeric side, changes the behavioral profile of WBS. In addition, lower SRS-2 scores (patient is less affected) across multiple subscales for the two individuals with atypical deletions that spared both AUTS2 and GTF2i, added credence to the role of GTF2i in cognitive and social functioning (Barak et al., 2019; Dai et al., 2009; Sakurai et al., 2011). Although AUTS2 and GTF2i appear to play a role in social function, the fact that larger deletions in either direction lead to impairments in social domains suggests that these findings may be nonspecific and may simply reflect a larger burden of gene loss that contributes to decreased IQ (Huguet et al., 2018). In addition to providing more details on social functioning and developmental outcomes in individuals with atypical WBS deletions, we identified new features in these cohorts that could potentially be assigned to the additionally deleted genes. The manifestation of more severe psychiatric presentations, such as self-injurious behaviors, anger outbursts with aggressive behaviors, and oppositional defiant disorder, could be associated with a YWHAG deletion given its previously reported association with neurodevelopmental abnormalities (Coe et al., 2019; Fusco et al., 2014; Kim et al., 2019; Komoike et al., 2010; Ramocki et al., 2010). Deletion of YWHAG, however, does not fully explain this phenotype as Patient D from our cohort demonstrated these behaviors with preservation of the gene.

Additionally, we presented three patients with POR hemideletions who had higher than expected 17-OH progesterone levels. While these levels were not high enough to warrant a diagnosis of Antley-Bixler syndrome—which is typically caused by a biallelic mutation in POR (low pLI) rather than mere hemizygosity—some hormonal differences may be present in hemizygous individuals and monitoring them for evidence of sex hormone disturbance may be warranted. Additionally, stress hormone evaluation using a morning cortisol or ACTH stimulation test could be considered.

In our cohort, microcephaly was overrepresented in patients with centromeric deletions and was more severe in those whose deletions include AUTS2. While all of our patients with telomeric deletions were normocephalic, Fusco et al. (2014), however, reported microcephaly in two of the 10 patients with telomeric deletions from the literature.

One patient (C) with a centromeric deletion presented with HLHS, a feature not typically ascribed to WBS. None of the genes known to be deleted have been reported to cause HLHS. It is possible that the HLHS was caused by genetic changes outside of the WBSCR.

Additional features such as early death, age of diagnosis, and diagnostic test type were investigated but differences between groups could not be detected. These findings were likely obscured due to sample size, with year of diagnosis likely contributing as well.

While our contribution expands the knowledge base, especially pertaining to the association of loss of social motivation with an atypical deletion and the relationship of MAGI2 deletion with infantile spasms and seizures, there are numerous avenues for continued research expansion. An important confounding factor potentially contributing to the social impairment we observed—and possibly to psychiatric manifestations—is IQ (Hus, Bishop, Gotham, Huerta, & Lord, 2013). Additionally, prospective studies are needed to further delineate the relationship between deletion size and cognitive-behavioral profile outcomes. Our findings, however, improve the ability of clinicians and healthcare providers (especially those still dependent upon older diagnostic tools) to identify atypical deletions in WBS patients. Early recognition of atypical presentations of WBS will further guide decision-making toward advanced diagnostic methods and will facilitate access to interventional services designed to improve health maintenance. This in itself offers a novel, meaningful approach in the management of this unique subset of WBS patients.