Comprehensive study of 28 individuals with SIN3A-related disorder underscoring the associated mild cognitive and distinctive facial phenotype

Witteveen-Kolk syndrome (OMIM 613406) is a recently defined neurodevelopmental syndrome caused by heterozygous loss-of-function variants in SIN3A. We define the clinical and neurodevelopmental phenotypes related to SIN3A-haploinsufficiency in 28 unreported patients. Patients with SIN3A variants adversely affecting protein function have mild intellectual disability, growth and feeding difficulties. Involvement of a multidisciplinary team including a geneticist, paediatrician and neurologist should be considered in managing these patients. Patients described here were identified through a combination of clinical evaluation and gene matching strategies (GeneMatcher and Decipher). All patients consented to participate in this study. Mean age of this cohort was 8.2 years (17 males, 11 females). Out of 16 patients ≥ 8 years old assessed, eight (50%) had mild intellectual disability (ID), four had moderate ID (22%), and one had severe ID (6%). Four (25%) did not have any cognitive impairment. Other neurological symptoms such as seizures (4/28) and hypotonia (12/28) were common. Behaviour problems were reported in a minority. In patients ≥2 years, three were diagnosed with Autism Spectrum Disorder (ASD) and four with Attention Deficit Hyperactivity Disorder (ADHD). We report 27 novel variants and one previously reported variant. 24 were truncating variants; three were missense variants and one large in-frame gain including exons 10–12.


Introduction
Witteveen-Kolk syndrome (OMIM #613406) was first described in 2016 with characteristic distinctive facial features, microcephaly, short stature, mild intellectual disability (ID) with delayed cognitive and motor development and subtle anomalies on MRI-brain imaging [1]. Although sparsely reported, frameshift as well as missense variants in the Switch-insensitive 3 transcription regulator family member A (SIN3A) (OMIM *607776) have been described in larger neurodevelopmental disorder cohorts, with an overall mild clinical picture [2]. Narumi-Kishimoto et al. presented a further patient with SIN3A frameshift variant and facial features of Witteveen-Kolk with relatively mild ID and normal growth [3].
The SIN3A gene is located in the chromosome 15 band q24 and is within the shortest region of overlap of various reported 15q24 microdeletions, therefore, is thought to be the critical gene for the atypical 15q24 microdeletion syndrome [4]. SIN3A encodes a transcriptional regulatory protein, which is associated with scaffolding in the core histone deacetylase complex [5]. In our earlier study, we showed that SIN3A is involved in cortical neurogenesis, supporting the hypothesis that variants in the gene that adversely affect its function lead to a broad range of developmental and neurological problems. We identified an additional 28 patients with SIN3A variants in order to comprehensively define the phenotype with a focus on both developmental and behavioural aspects, as well as investigating genotype-phenotype correlations.

Methods
We collected the molecular and clinical features on 28 unpublished individuals with SIN3A variants by a collaboration facilitated by Deciphering Developmental Disorders (DDD study) [6], GeneMatcher [7] and DECIPHER (DECIPHER v9.24: https://decipher.sanger.ac.uk/) [8], in which, several clinical groups independently identified individuals with developmental delay/intellectual disability (DD/ID) and related phenotypes with rare variants in SIN3A during routine diagnostic genetic testing. An application to the DDD study for a Complementary Analysis Project was made, allowing access to anonymised details of patients with SIN3A variants identified through this study (https://www.ddduk.org/). Clinicians of selected patients were then contacted to invite patients and their families to be recruited. Clinical analysis of these patients was performed during regular consultations focusing on medical history, physical examination and observational analysis of behavioural features along with reported history by the family. In all patients, exome sequencing and variant filtering were performed, according to the routine protocol and diagnostic procedures at each institute.
Identified patients with a class 4 or 5 variants (likely pathogenic or pathogenic variants) in SIN3A, according to the American College of Medical Genetics (ACMG) criteria, were approached to participate in this study. Informed consent for publication was obtained from all patients and/ or their guardians. The responsible clinician reviewed medical records of each participant in order to comprehensively document the phenotype.
The clinical significance of the variants identified was interpreted according to the ACMG [9] guidelines and further review publications [10]. Excluded from the study were: patients with an additional proven genetic diagnosis where the SIN3A variant was not thought to be contributory or the sole pathogenic finding, those with a chromosomal anomaly explaining or likely to be explaining the phenotype, and those in whom SIN3A variants were of uncertain clinical significance with no convincing clinical correlation.

Molecular genetics
In this study, we reported 28 patients with variants that adversely affect protein function and which are classified as pathogenic or likely pathogenic variants in SIN3A. The findings of this study expand the variant spectrum previously reported in SIN3A ,3 . 27/28 patients in this study have novel variants. Patient 15 with c.3310C>T, p. (Arg1104*) variant is a family member of a patient we have previously reported in Witteveen et al. [1]. Predominantly, variants found in our patient cohort are truncating and predicted to result in a protein loss-of-function (25/28). All 24 truncating variants have been classified as variants that adversely affect function (pathogenic) using the ACMG and ACGS (The Association for Clinical Genomics Science) variant classification guidelines using criterions: PVS1, PM2, and PS2 where appropriate [9,10]. Of the four remaining variants, three were missense (patients 4, 18 and 26; Table 1) and one was a large in-frame gain which included the whole of exons 10, 11 and 12 (patient 24; Table 1). These four variants were all de novo in our patients with a specific and consistent phenotype to SIN3A and were absent from control population data sets in gno-mAD (https://gnomad.broadinstitute.org/). We have also used criterion PM1 at moderate level for two of the three reported missense variants, c.377C>T, p.(Ala126Val) and c.463A>G, p.(Lys155Glu), as they were located within a protein functional domain. These results show that missense causative variants are not clustered in a hot spot within SIN3A ( Fig. 1 provides a schematic SIN3A structure and variant locations which, as demonstrated, are distributed throughout the gene). Details of the variants, ACMG criterions and classification are listed in Table 1.

Assessment of pathogenicity of missense variants in SIN3A
In order to assess if missense variants in our cohort were predicted to adversely affect the critical functional domains in SIN3A, we used a 3D model based on the solution  structure of mouse SIN3A PAH1 bound to the Sin3 interaction domain (SID) of SAP25 (SIN3A Associated Protein 25) [11]. The pair of amphipathic helices (PAH) domains are predicted as important for the recruitment by and interaction with a diverse number of transcription factors and therefore, critical for protein function [12,13]. Both altered residues by c.377C>T, p.(Ala126Val) and c.463A>G, p.(Lys155Glu) were predicted to be part of the SAP25 SID-binding surface. We predicted that SIN3A Ala126 with its small hydrophobic side chain formed a pocket for the larger hydrophobic side chain of SAP25 Leu142 and was invariant across species [11]. SIN3A Lys155 is also predicted to form a 2.2A hydrogen bond with the polar Gln143 of SAP25 (orange line in Fig. 2) and was also conserved in the PAH1 domain across species.
Along with the lack of normal variation in this region (using missense constraint data in SIN3A from Decipher: https:// decipher.sanger.ac.uk/gene/SIN3A#overview/protein-info), we applied criterion PM1 at moderate level for the classification of both c.377C>T, p.(Ala126Val) and c.463A>G, p.(Lys155-Glu). We also suggest that PM1 can be applied at a moderate level for missense variants within residues p.119-189 of the SIN3A protein providing that the change caused by the variant was not present in other species and also meets the PM2 criteria (see Fig. 2).
Finally, the c.3317T>C, p.(Met1106Thr) is within the Cterminal domain of SIN3A, this region is not well characterised and the in silico pathogenicity prediction programs suggests a benign effect (REVEL score of 0.105 shows a benign effect). However, this variant was confirmed to be de novo in patient 18, novel in control populations in gnomAD and with a consistent phenotype to that of the rest of our cohort. No other variants were found in this patient on exome sequencing. Therefore, we classified this variant in addition as likely pathogenic, but without the use of PM1. Table 2 shows a summary of the characteristics of 28 patients (17 males, 11 females) with variants in SIN3A who were included in this retrospective study. The mean age of participants was 8.2 years (range 0.6-67 years). Table 3 provides an overview of salient features described in this cohort in comparison to published literature and Table 4 has detailed phenotypic information on this cohort of patients.

Development
Of 28 patients, 15 (56%) had global developmental delay. Out of 16 patients ≥8 years old, (one patient was not formally assessed) seven (44%) had mild intellectual disability (ID), four had moderate ID (25%), and one had severe ID (6%). Four (25%) did not appear to have any cognitive impairment. In 21 patients ≥5 years old, some form of motor developmental delay was reported in 13 patients (62%) and in 16 patients (76%) there was some form of language developmental delay. Intelligence Quotient (IQ) was only formally measured in 6/28 patients and the score ranged from 60 to 100. This suggests that patients within this cohort have low normal intelligence.
Autism spectrum disorder (ASD) is rarely diagnosed before the age of 24 months [14]. In one patient aged 14 months, ASD was thought possible but not formally assessed due to the young age. In three of 23 patients ≥2 years (13%), ASD was diagnosed. Attention Deficit Hyperactivity Disorder (ADHD) is most commonly diagnosed in children between 6 and 12 years, though this can be diagnosed in a younger age group. ADHD was diagnosed in four patients ≥2 years (17%).
Growth and feeding difficulties 13/28 (46%) patients in our cohort had head circumference at least two standard deviations (SD) below the mean. Weight was less than two SD below the mean in 8/28 patients (29%) and height was less than two SD below the mean in 5/28 patients (18%).15/28 (54%) of patients had feeding difficulties with at least two patients documented to require nasogastric tube feeding mainly in the neonatal period.

Craniofacial features
Fourteen (50%) patients were reported to have craniofacial dysmorphism. In ten patients, the facial gestalt was    Fig. 3). The facial gestalt appears to be similar and potentially recognisable but only in the context of reverse phenotyping with published images, confirming previous findings. Three patients had a palatal defect with one also having a bifid uvula.

Other clinical manifestations
In terms of neuroimaging, given the milder developmental delay phenotype, many patients in the cohort have not had MRI-brain imaging (19/28, 68%). Nine out of a total of nine patients that had brain imaging performed had reports available and in 7/9 (78%) abnormalities were seen. Most common was ventriculomegaly in 2/9, and a hypoplastic/ dysplastic corpus callosum in 2/9. Cerebellar atrophy was reported in two patients and there was one patient had a Chiari 1 malformation. 6/28 (21%) patients were noted to have seizures currently or in their past medical history. Hypotonia was a common manifestation as well, present in 12/28 (43%) patients.
Of note, 2/28 patients were reported to have a malignancy, including sigmoid adenocarcinoma in the 67-year old patient which may be an age-related cancer and 1 younger patient with Burkitt lymphoma (Patient 16 at age  eight). Patient 26 was noted to have T-cell lymphopenia at age 11 associated with immunoglobulin deficiency and significant bronchiectasis.

Discussion
Heterozygous loss-of-function variants in SIN3A were recently described to result in a novel neurodevelopmental syndrome comprising intellectual disability and varying degrees of developmental delay. This syndrome defined as Witteveen-Kolk syndrome was further characterised by subtle brain abnormalities, including corpus callosum dysgenesis and ventriculomegaly, distinctive facial features (a broad, tall forehead; small mouth, thin upper lip with pointed chin and down-slanting palpebral fissures), hyperlaxity and short stature. Furthermore, we showed previously that in vivo functional knockdown of SIN3A leads to reduced cortical neurogenesis, altered neuronal identity and aberrant cortico-cortical projections in the developing mouse brain. Therefore, it is likely that the aberrant cortical development underlies impaired neurodevelopment and leads to cognitive and behavioural problems in patients, varying from paediatric to adult-onset. Here, we summarise the neurodevelopmental and facial phenotype in an additional 28 patients with SIN3A-related disorder. Table 2 provides an overview of the patients reported here and all previously reported patients with SIN3A structural variants alone. Supplementary table provides a comprehensive review of all the clinical information available on this cohort. Below is extracted information from these tables.

Genotype-phenotype correlation in SIN3A
Most of the patients in this cohort (25/28; 89%) have truncating or frameshift variants in SIN3A except three patients with missense variants (Patients 4, 18 and 26) which have been classed as a class 4 (likely pathogenic) variant based on a combination of evidence as previously described (see variant table for further information). It is likely that as previously described haploinsufficiency is the predominant likely mechanism in SIN3A-related disorder but it may be that the phenotype may differ depending on the nature of the SIN3A variant. From the evidence gathered so far, there is no apparent correlation between severity of phenotype and genotype.

Neurodevelopment
As noted in our earlier study [1], the overall intellectual disability seems to be mild, with 11 of the 16 patients (69%) over the age of eight years having either no ID or mild ID. The low prevalence of severe intellectual disability different from the neurodevelopmental phenotype of moderate or severe intellectual disability associated with 15q24 microdeletion syndrome, suggests that additional genes contribute to the cognitive phenotype in the 15q24 deletion syndrome [15].
The median intelligence is relatively high with IQ of 74. Interestingly, all patients in whom intelligence was tested reported a higher verbal than performal intelligence score. This is an important finding that clinicians should be aware of, since disharmonic intelligence profiles easily lead to overestimating of the self-management capabilities of patients. A further study is planned to undertake detailed psychometric assessments and IQ measurements in this cohort.

Behavioural phenotype
Overall, a third of the cohort had a psychiatric or behavioural condition reported, including ADHD, aggressive behaviour, OCD, depression, psychosis, anxiety and schizoaffective disorder. In three patients, ASD was concurrent with a psychiatric diagnosis. This is a significant finding, since such neuropathology has a significant impact on the quality of life of patients especially in those adults with milder neurodevelopmental phenotypes. Knowing that patients with variants in the SIN3A gene are at risk for such concerns, early intervention is important to ensure optimal treatment and outcomes.
Of those patients with psychiatric disease, only two had brain imaging done. Interestingly, both those MRIs showed ventriculomegaly, while one also showed delayed myelination. The younger patients in the cohort did not have any neuroimaging done given the milder clinical presentation. However, it is likely that once the diagnosis of a SIN3Arelated disorder is made, imaging of the brain should be offered in the context of neurological symptoms rather than routine work-up.

Craniofacial dysmorphism
Our previous study presented evidence for a characteristic facial appearance associated with SIN3A. Some of the patients in this cohort have clear similarities. As with many of the mild and variable neurodevelopmental phenotypes, it remains to be seen whether the facial gestalt is easily identifiable in clinical practice. However, there appears to be a common, emerging facial phenotype with a tall, broad forehead, down-slanting palpebral fissures, triangular face with a pointed chin and a thin upper lip, based on the patient's photographs (both included and unpublished but shared with the authors due to parental consent for publication of photos being declined).
Interestingly, one patient was first suspected of progeria, because of the typical shape of his neurocranium (Patient 25: J). Following genetic testing, he was diagnosed with a variant in SIN3A, demonstrating that this may be part of the spectrum of the syndrome. Patient 12: G also appears to have a progeric face. Sparse hair and reduced subcutaneous tissue was reported in 3/28 (10%) of patients in this study and note the progeric appearance in at least two of the patients in this cohort. However, no other ectodermal features were identified. This leads to the possibility of differential diagnoses including progeroid group of conditions; however, SIN3A-related disorder does not appear to present predominantly with a progeroid phenotype from the large cohort described here.
Other clinical manifestations 14/28 (50%) patients in our study population had either epilepsy, hypotonia, or both. Two out of the three patients with epilepsy who had a brain MRI (of a total of five patients with epilepsy) had abnormalities found. None of the patients with epilepsy from this cohort reported any psychiatric disease.
Interestingly, other commonly reported symptoms in patients with intellectual disability were not reported in our patient population. Constipation for instance, was only reported in one patient, while hearing loss and refraction abnormalities were also not as prevalent as in other intellectual disability cohorts or the 15q24 microdeletion syndrome [15,16].
Further information also needs to be collected to ascertain whether malignancy is a significant association or merely an observation with a large cohort of patients with Witteveen-Kolk Syndrome. In addition, as described above, additional features appear to be emerging from the larger cohort of patients published here and further follow-up is required to see if this is a consistent part of the phenotype.

Conclusion
Patients with disease causing variants in SIN3A usually have mild global developmental delay/ID, with some even having tested IQs in the normal range with variable penetrance. There are similar facial features in around half of patients for which a targeted molecular evaluation would be feasible. However, it is likely that diagnostic evaluation and identification of SIN3A variants in suspected individuals will be performed using large ID panels or WES/WGS. There is evidence to suggest these patients are at risk for psychiatric-and neurological conditions and therefore, a multidisciplinary team approach should be considered in caring for these patients. Data collected so far seems to suggest additional features such as hypotonia, seizures along with the previously well described neurodevelopmental association with this disorder.
There is no apparent genotype-phenotype correlation and/ or missense variant hot spot within SIN3A and the missense variants appear to be distributed throughout the gene based on observation of this cohort. As expected, majority of patients in published literature and this cohort appear to have truncating variants reinforcing haploinsufficiency as likely mechanism of pathogenicity, although SIN3A missense variants affecting critical functional domain in SIN3A also appear to be associated with disease. Further studies of this nature are required to ascertain clinical correlation in this disorder. HICF-1009-003]. This study makes use of DECIPHER (http:// decipher.sanger.ac.uk), which is funded by the Wellcome Trust. See Nature PMID: 25533962 or www.ddduk.org/access.html for full acknowledgement. We would also like to thank all the families for consenting to this publication.
Author contributions MB and TK designed and supervised the study; MB was responsible for collection of data from UK cohort of patients whilst TK was responsible for data collection from Dutch and international cohort; SA, AD and RR performed all data collection and molecular interpretation of SIN3A variants; all authors contributed to clinical or molecular data collection and approval of submitted manuscript.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethics statement This study was performed in adherence to the principles set out in the Declaration of Helsinki. The DDD study has UK Research Ethics Committee approval (10/H0305/83, granted by the Cambridge South REC, and GEN/284/12 granted by the Republic of Ireland REC). Informed consent for publication of clinical and molecular data and inclusion of photos where applicable was obtained from all patients and/ or their guardians included in this study.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.