Speech‐language profiles in the context of cognitive and adaptive functioning in SATB2‐associated syndrome

Abstract SATB2‐associated syndrome (SAS) is a neurodevelopmental disorder caused by heterozygous pathogenic variants in the SATB2 gene, and is typically characterized by intellectual disability and severely impaired communication skills. The goal of this study was to contribute to the understanding of speech and language impairments in SAS, in the context of general developmental skills and cognitive and adaptive functioning. We performed detailed oral motor, speech and language profiling in combination with neuropsychological assessments in 23 individuals with a molecularly confirmed SAS diagnosis: 11 primarily verbal individuals and 12 primarily nonverbal individuals, independent of their ages. All individuals had severe receptive language delays. For all verbal individuals, we were able to define underlying speech conditions. While childhood apraxia of speech was most prevalent, oral motor problems appeared frequent as well and were more present in the nonverbal group than in the verbal group. For seven individuals, age‐appropriate Wechsler indices could be derived, showing that the level of intellectual functioning of these individuals varied from moderate–mild ID to mild ID‐borderline intellectual functioning. Assessments of adaptive functioning with the Vineland Screener showed relatively high scores on the domain “daily functioning” and relatively low scores on the domain “communication” in most individuals. Altogether, this study provides a detailed delineation of oral motor, speech and language skills and neuropsychological functioning in individuals with SAS, and can provide families and caregivers with information to guide diagnosis, management and treatment approaches.

study was to contribute to the understanding of speech and language impairments in SAS, in the context of general developmental skills and cognitive and adaptive functioning. We performed detailed oral motor, speech and language profiling in combination with neuropsychological assessments in 23 individuals with a molecularly confirmed SAS diagnosis: 11 primarily verbal individuals and 12 primarily nonverbal individuals, independent of their ages. All individuals had severe receptive language delays. For all verbal individuals, we were able to define underlying speech conditions. While childhood apraxia of speech was most prevalent, oral motor problems appeared frequent as well and were more present in the nonverbal group than in the verbal group. For seven individuals, age-appropriate Wechsler indices could be derived, showing that the level of intellectual functioning of these individuals varied from moderate-mild ID to mild ID-borderline intellectual functioning. Assessments of adaptive functioning with the Vineland Screener showed relatively high scores on the domain "daily functioning" and relatively low scores on the domain "communication" Lot Snijders Blok and Y. Max Goosen contributed equally in most individuals. Altogether, this study provides a detailed delineation of oral motor, speech and language skills and neuropsychological functioning in individuals with SAS, and can provide families and caregivers with information to guide diagnosis, management and treatment approaches.
clinical management, communication disorder, contextual neuropsychology, neurodevelopmental disorder, SATB2 associated syndrome, speech and language impairments

| INTRODUCTION
The introduction of new DNA sequencing technologies (nextgeneration sequencing) has rapidly improved the identification of genes of which high-penetrance disruptive variants can cause neurodevelopmental disorders. Amongst the most commonly affected genes in neurodevelopmental disorders is SATB2. 1 The neurodevelopmental disorder associated with pathogenic variants in this gene is known as SATB2-associated syndrome (SAS). The exact prevalence of SAS is not known. However, the yield after applying exome sequencing in a large cohort of individuals with undiagnosed developmental disorders showed the frequency of pathogenic variants in SATB2 to be 0.3% (14/4294 probands). 1 SAS presents with marked craniofacial dysmorphisms, intellectual disability (ID), developmental delay, as well as generally restricted or absent speech and severely impaired communicative skills. 2 Individuals with SAS often use communication methods other than (or in addition to) spoken language, such as gestures, sign language and/or augmentative and alternative communication (AAC) devices. In addition to speech problems, other features related to oral motor skills or oral abnormalities are common, including cleft palate, teeth anomalies, drooling and feeding problems. 2 SAS is caused by heterozygous disruptions of the SATB2 gene.
These are mostly variants with a clear loss-of-function effect (frameshift and nonsense variants), but missense variants, variants predicted to affect splicing and copy number variants are reported as well. 3 The SATB2 protein is a transcription factor with important roles in cortical development. 4 One could hypothesize that loss-of-function of SATB2 might disproportionately affect the development of higher cognitive functions, such as attention, memory and executive functioning.
While speech problems are prominent in SAS, there is limited information about mechanisms underlying the oral motor, speech and language impairments observed in affected individuals, other than one recent study on the assessment of speech and language phenotypes in a SAS cohort. 5 That study found that individuals with SAS generally show prominent language impairments, childhood apraxia of speech and various oral motor problems, including hypernasal resonance, pharyngeal phase dysphagia and drooling. 5 The current study aimed to contribute to the understanding of speech and language abnormalities in SAS in the context of general developmental capacities and cognitive and adaptive functioning. The study design included a detailed characterization of oral motor, speech and language profiles combined with neuropsychological testing in 23 individuals with a molecularly confirmed diagnosis of SAS.
2 | METHODS 2.1 | General study design and data collection

| Study design
This study has an observational and cross-sectional study design and was approved by the medical research and ethics committee Arnhem-Nijmegen (CMO Arnhem-Nijmegen; study number NL64562.091.18).
All study procedures were in line with the principles of the Declaration of Helsinki. Recruitment and inclusion for the study took place between April and November 2019. After inclusion and the informed consent procedure, individuals were invited for two testing visits within the Radboud University Medical Center in Nijmegen, the Netherlands: one visit with one of the two speech-language therapists (SLTs), and one visit with a healthcare psychologist. During one of these two visits, a clinical geneticist in training collected details on medical history and growth parameters. In addition to this, parents and/or caregivers were asked to fill in standardized questionnaires about the patient with SAS. Data collection finished in March 2020.

| Individuals
Individuals with SAS from the Netherlands and Belgium were recruited via the Dutch SAS family support group or via the Clinical Genetics department where their SAS diagnosis was established. In order to be eligible to participate in the study, individuals had to meet all the following three criteria: (a) established molecular diagnosis of SATB2-associated syndrome, (b) age of at least 2 years old at time of testing and (c) raised in a Dutch-speaking family with Dutch as first language. There was one exclusion criterion: Individuals with SAS who also had another molecular diagnosis that likely contributed to their developmental phenotype were excluded from participation, for example, individuals with larger copy number variants not only affecting SATB2 but also encompassing additional neurodevelopmental disorder-associated genes. In total, 23 individuals were included for participation in the study.

| General data collection
Data on developmental and medical history were collected via medical file notes and a standardized medical history during one of the visits.
Growth parameters were measured during the visit or, if this was not possible, derived from recent measurements in another context. All official molecular test reports with the SATB2 diagnosis were collected, and variant details were converted into standardized nomenclature using hg19 as a reference genome and NM_001172509.1 (SATB2 isoform 1) as the standard transcript. All data were deidentified and stored in a secure and study-specific Castor 9 yielding a vocabulary quotient. The Schlichting tests for language comprehension and language production 10 were used to measure receptive and expressive language skills. These norm-based standard scores or Q scores have a mean score of 100 (SD 15), with a score of 85-115 representing average range performance.
When the administration of the Schlichting tests was not possible due limited language and/or understanding, the Dutch Nonspeech Test (NNST) 11 was used. This test comprises a receptive scale and an expressive scale. Scores on both scales were expressed in percentile scores, with a mean score of 50.
Subtests of the Dutch version of the clinical evaluation of language fundamentals (CELF) 12 were used instead of the Schlichting tests when individuals had a sufficient level of language. The subtests "concepts and following directions," "expressive vocabulary," "recalling sentences," and "formulating sentences" were administered.
The Q scores and percentile scores of all the language assessments were interpreted as mild (1-1.5 SD below mean), moderate (1.5-2 SD below mean) and severe (>2 SD below mean).

| Feeding and oral motor evaluation
A specifically designed questionnaire for problems with swallowing related to different consistencies of food was used in all individuals. It also included questions regarding drooling and dental problems. This semi-structured questionnaire is used in earlier studies where it has demonstrated its usefulness and importance to differentiate dysphagia characteristics. 16,17 Problems with only chewing (refers to problems in the oral phase) and chewing and choking (refers to problems in the oropharyngeal phase) were scored with a five-point scale and recoded into two categories (À) no problems or (+) problems to a certain extent (2 = less than once a day, 3 = once every day, 4 = several times a day, 5 = food is not offered).
Structural or functional impairments of the oral region were assessed with the self-composed oral-facial motor assessment for children (OMAC). This assessment tool examines oral motor function (e.g., face, lips, tongue, velum, jaw), oral-facial structural integrity (e.g., symmetry, lip seal), strength (e.g., eye closure, lip closure, tongue, jaw) and the saliva swallow (e.g., slurping, swallowing on demand) by observation. Problems with the performance or imitation of the items were scored and recoded in the category (À) no problems and (+) problems to a certain extent. was not possible due to limited language and/or understanding, the WPPSI-III-NL was administered. Raw scores were converted into developmental age equivalents ranging from "below 2;7" to "above 7;10." Although test administration was performed according to standard procedures, slight alterations were made to compensate for language problems of the individuals. For instance, individuals were allowed to respond using Dutch Sign Language and/or using AAC when verbal responses were required and extra verbal cues and explanation were given to engage individuals further when non-compliant (i.e., "testing the limits").

| Adaptive functioning
Adaptive behavior has been described as the combination of conceptual, social and practical skills acquired to function adequately in daily life. 21 The level of adaptive functioning was measured using the Vineland Screener 0-6 years, 22 filled out by parents. This questionnaire is a Dutch screener version of the gold standard Vineland adaptive behavior scales 23 and consists of 72 questions, providing a total score and four domain scores: communication, social functioning, daily functioning and motor skills. Raw scores were converted to developmental age scores (in months), reflecting the level of adaptive functioning. 22 To enable inter-individual comparison of Vineland profiles, individual Vineland scores (age equivalents) per domain were normalized by dividing each score by the total Vineland score (age equivalent) of the same individual. A normalized score of 1.0 indicates that the age equivalent of the domain score is similar to the age equivalent of the total score of this individual.

| Behavioral problems
The presence of behavioral problems was measured by parent-based reports, using age-specific versions of the Achenbach system of empirically based assessment 24 : the Dutch versions of the child behavior checklist (CBCL/1,5-5 25 and CBCL/6-18 26 ) and the proxy version of the adult behavior checklist (ABCL/18-59). 27 These parent-based questionnaires consist of 100, 113 and 134 items, respectively and provide a total score for observed behavioral problems, scales for internalizing (i.e., anxiety, depression and withdrawal) and externalizing (i.e., aggressive behavior, conflict with others/social mores) problems and several syndrome subscales. In this study, only the syndrome scales were included that were present in all three versions: somatic, anxious, withdrawn, attention and aggression problems. Raw scores were converted to standardized T-scores. For the total score and internalizing and externalizing scales, a score of 64 and higher is considered to be in the clinical range (i.e., consideration of professional help is warranted), for the syndrome scale the cut-off for a score in the clinical range is a T score of 70. [25][26][27] 3 | RESULTS

| Individuals and characteristics
In total, 32 individuals were examined for eligibility to participate in the study. Nine were not included, because the parents/caregivers decided not to participate after being informed about study details (n = 6), because the child was not raised with Dutch as first language (n = 2) or because the SATB2 disruption was part of a large microdeletion with many other genes possibly affecting neurodevelopment (n = 1). A total of 23 individuals started participation in the study, all of whom completed it; 70% of these individuals were male. The age of individuals at inclusion varied from 2;10 to 40;8 years old (median age 11;7). Growth parameters and other baseline characteristics are included in Table 1.
Details on the SATB2 variants in the individuals are included in Table S1. In short, the majority of individuals (21/23; 91%) had a heterozygous single nucleotide variant (SNV) affecting SATB2; two individuals (9%) had a de novo 2q33.1 microdeletion (Table 1). Almost all variants (21/23; 91%) were confirmed to be de novo, hence not present in blood-derived DNA of either of the two parents of the individuals. Two individuals were siblings and carried the same de novo variant, suggesting germline mosaicism in one of the parents. Constitutive mosaicism was not detectable by Sanger sequencing of parental blood samples. In one individual, the SATB2 variant was found to be a mosaic variant and present in 32 of 143 exome sequencing reads ($22%). The age at which the molecular diagnosis of SATB2 was established in each individual varied between 0;5 years and 38;11 years, with a median of 10;10 years (Table 1).

| Language
Receptive language abilities were measured in 21 individuals. Two individuals were not assessed because test procedures were not Mean height corrected for age (SD)  Individual Vineland scores (age equivalents) per domain were normalized by dividing each score by the total Vineland score (age equivalent) of the same individual: a score of 1.0 indicates that the age equivalent of the domain score is similar to the age equivalent of the total score of this individual. (E) Age equivalents of Vineland total scores (months) obtained in nonverbal and verbal individuals versus chronological age (months). (F) Age equivalents of test scores of three different receptive language (sub)tests, compared with Vineland total score age equivalents. The gray triangle indicates a ceiling effect for the Vineland test score, as the maximum score of 68 months was obtained for this test developmentally appropriate or individuals were not able to be tested.
All the tested individuals showed severe receptive language deficits when compared with age-related peers, except for one (individual 5) with a mild deficit. Expressive language could be measured in nine verbal individuals. Eight of them had a severe expressive language deficit and one had a moderate to severe deficit (individual 10) when compared with age-related peers.

| Feeding and oral motor evaluation
Feeding and swallowing problems were common in the total group of individuals with 87% affected (n = 20), while in the remaining three individuals no feeding problems were mentioned. In the nonverbal group, all individuals had feeding problems. In the verbal group, 80% exhibited feeding problems. For the feeding problems in the nonverbal group, 25% involved swallowing problems in the oral phase (e.g., chewing problems and overstuffing) and 75% involved the oropharyngeal phase (e.g., choking, aspiration). This is in contrast to the swallowing problems in the verbal group where 87.5% suffered from oral phase problems and only 12.5% showed oropharyngeal phase problems.

| Neuropsychological functioning
Observation of behavior during testing procedures showed clear differences with regard to task understanding and concentration, both of which likely mediated task compliance that was further hampered in case of increased restlessness. We classified the results of formal neuropsychological testing using Wechsler scales in three groups ( Figure 1B): a group in which age appropriate administration of In the second group, consisting of eight individuals (35%), nonage matched Wechsler administration of several subtasks of the WPPSI-II-NL were derived. These eight individuals had a chronological age of 9;3 to 40;8 and age equivalents calculated based on Wechsler subtask scores ranged from <2.7 to <7.1 years. Of these eight individuals, five were classified as verbal and three as nonverbal ( Figure 1B; Table 2).
The last group consists of three individuals (13%), who were noneligible for testing in either form, because of a lack of understanding and cooperation. These three individuals were all classified as nonverbal. relatively low compared with the total score ( Figure 1D). When distinguishing between verbal and nonverbal individuals, identical patterns were seen across subdomains ( Figure 1D). The overall levels of adaptive functioning seem to be higher in the verbal group compared with the nonverbal group ( Figure 1C).

T A B L E 2 Results per individual
When comparing the results of receptive language tests (converted to age equivalents) with the age equivalents matching the Vineland adaptive functioning total score, the results of these language tests seem to align with the estimated level of adaptive functioning ( Figure 1F). One individual (5%, nonverbal) scored within the clinical range for attention problems and one (5%, verbal) for aggression problems (Table 2).

| Genotype-phenotype comparison
In terms of genotype-phenotype relations, we looked more specifi- Generally, it is difficult to assess the IQ levels in individuals with ID by using conventional methods that are based on the normal popu- give rise to distinctive profiles of adaptive functioning, which might also be partly age-dependent. [30][31][32] The relative weakness of communication in the adaptive functioning profile observed in the Vineland scores within our study is in line with the findings from direct speech and language assessments, as well as the literature on SAS so far. 33 Variations in adaptive functioning domains with relatively strong daily-living skills based on Vineland questionnaires are commonly reported in other neurodevelopmental disorders. 32,34,35 As already shown for other genetic syndromes, the assumption that cognitive functioning is strictly related to all adaptive functioning domains does also not apply to SAS. 36 Classifying an accurate level of ID based on the required equal weighting of intellectual functioning (i.e., IQ) and level of adaptive functioning 37 is therefore a challenge and more indepth analysis of intellectual functioning and adaptive functioning is required.
In Chow et al. 42 shows that receptive language skills in particular have a strong predictive property when it comes to challenging behavior, and that improving (receptive) language skills can have a mitigating effect on the development of behavioral problems. Although our results do not directly support this link between problems in language and behavior, it is possible that the relatively low levels of frustration observed in individuals in our study and a related lack of initiation contribute to the severe speech phenotype. Findings like these warrant a broad and strong dimensional approach to clinical assessment using gold-standard instruments and a careful consideration of contextual factors to correctly interpret a particular behavior as part of the SAS phenotype profile. 39,40 Research has also shown that it is necessary to interpret challenging behavior in ID in relation to contextual variables, in order to establish an effective intervention plan. 43 Different types of heterozygous SATB2 disruptions were found in the individuals included in our study. While there is some evidence that missense variants of SATB2 might be associated with milder phenotypes, 3 functional characterization of effects of variants in this gene has so far been limited. It is therefore unclear whether missense variants have different effects from the loss-of-function that is assumed for most other variants. 3,44 As SATB2 encodes a transcription factor that can have pleiotropic effects on multiple different pathways and developmental processes in the brain, it is important to realize that many different factors (e.g., stochastic developmental factors) might ultimately contribute to the phenotypic presentation, even between individuals with identical pathogenic variants.
In addition to individuals with single nucleotide variants affecting Our study has some limitations that should be taken into account.
First, because of the low prevalence of SATB2 variants in the population, it is not possible to study a large cohort of affected individuals with the same native language in the same age range. The consequent differences in chronological ages in our cohort, as well as the varying levels of cognitive functioning, made systematic testing using comparable tests more difficult and in some cases impossible, leading to suboptimal data collection. In line with this, to show true capacities of all individuals in this study, modifications to the standardized study assessments had to be made, which might have influenced the results In summary, with this study we provide a delineation of speech, language and oral motor skills in individuals with SAS, combined with emerging data on neuropsychological functioning. While overlapping and highly recurrent features were seen for both the speech and language domain and the adaptive functioning profile, there was also a high variability observed, mainly in severity of features. This study can provide families, speech therapists, psychologists and other caregivers with the necessary information to guide diagnostic and treatment approaches in order to obtain the best functional outcomes in individuals with SATB2 associated syndrome.

ACKNOWLEDGMENTS
We thank all individuals and their families and caregivers for their con-