A novel heterozygous ZBTB18 missense mutation in a family with non-syndromic intellectual disability

Intellectual disability (ID) is a common neurodevelopmental disorder characterized by significantly impaired adaptive behavior and cognitive capacity. High throughput sequencing approaches have revealed the genetic etiologies for 25–50% of ID patients, while inherited genetic mutations were detected in <5% cases. Here, we investigated the genetic cause for non-syndromic ID in a Han Chinese family. Whole genome sequencing was performed on identical twin sisters diagnosed with ID, their respective children, and their asymptomatic parents. Data was filtered for rare variants, and in silico prediction tools were used to establish pathogenic alleles. Candidate mutations were validated by Sanger sequencing. In silico modeling was used to evaluate the mutation’s effects on the protein encoded by a candidate coding gene. A novel heterozygous variant in the ZBTB18 gene c.1323C>G (p.His441Gln) was identified. This variant co-segregated with affected individuals in an autosomal dominant pattern and was not detected in asymptomatic family members. Molecular studies reveal that a p.His441Gln substitution disrupts zinc binding within the second zinc finger and disrupts the capacity for ZBTB18 to bind DNA. This is the first report of an inherited ZBTB18 mutation for ID. This study further validates WGS for the accurate molecular diagnosis of ID.


Introduction
Intellectual disability (ID) is a generalized neurodevelopmental disorder characterized by substantial impairment in intellectual capacity as well as adaptive behaviors and affects 1 to 3% of children [1,2].ID can be diagnosed in isolation or in combination with congenital malformations, as well as with additional neurological features such as epilepsy, sensory impairment, and autism spectrum disorder (ASD).The severity of ID can range from mild, moderate, severe, and profound [3,4].Individuals living with an ID diagnosis face significant medical, financial, and psychological challenges.The burden of ID is considerable and falls upon families, the community, and on the health care systems [5].
ID can be affected by genetic and environmental factors, of which 25-50% of ID patients are caused by genetic factors [6].With the rapid development of molecular biology technologies, next-generation sequencing (NGS) approaches for clinical diagnostic applications, such as whole-exome sequencing (WES) and wholegenome sequencing (WGS), have facilitated a broad and high-resolution discovery platform to characterize ID genes and their causal mutations [7][8][9][10][11][12][13][14][15][16].To date, over 700 genes and 130 rare CNVs have been identified as putative genetic causes of X-linked, autosomal-dominant as well as autosomal-recessive ID and ID-associated disorders [2,17].Notably, inherited genetic forms of ID are rare in occurrence, detected in approximately 5% of cases [18].
ZBTB18 gene is located on the q arm of the 1-chromosome (1q44) and encodes a transcriptional repressor protein which is essential to the development of the mammalian brain [19].The ZBTB18 protein has a BTB/ POZ and four zinc finger domains which are highly conserved (>95%) between human and mice [20].The animal experiments in mice show that ZBTB18 deficiency can result in the deficiency of cortical and cerebellar neurons' growth, differentiation, and maturation [21].In humans, the first ZBTB18 mutation has been reported in one patient with non-syndromic ID [7].Subsequently, ZBTB18 mutations have been identified in patients with syndromic ID [22][23][24][25][26][27] and 1q43q44 microdeletion syndrome [25].To date, over 60 mutations in the ZBTB18 gene have been described in the HGMD database (professional 2022.1).The mutation types include missense/ nonsense mutations, small deletions, small insertions, small indels, and gross deletions.
In this study, we performed whole genome sequencing of family members with non-syndromic ID to identify a novel mutation to the ZBTB18 gene that shows an autosomal-dominant pattern of inheritance.

Subjects
The study was approved by the Ethics Committee of West China Second University Hospital, Sichuan University (No: 2015011).Written informed consent was obtained from all the subjects or their guardians when they were enrolled in the study.Furthermore, all methods were carried out in accordance with the Declaration of Helsinki and relevant guidelines and regulations.
A non-consanguineous family from the Sichuan province in China with two generations of individuals with ID was recruited (Fig. 1).Two affected (III:1, III:2) members received a full clinical evaluation.The blood samples of four affected subjects (II:2, II:3, III:1, III:2) and the parents of the affected twins (I:1, I:2) were collected for analysis.

Clinical features
Four individuals in this family have mild ID.The proband (III:1) was a 7-year-old boy.A language impairment screening assessment and magnetic resonance imaging (MRI) scan of the brain were performed at 31 months of age.The global level of language development was remarkably lower than his peers.Specifically, the abilities of speech, listening, visual expression, and whole language were equivalent to the level of 9, 12-13, 12-13, and 11-12 months old, respectively.Brain MRI revealed dysplasia of the corpus callosum.At age 6 years, a brain MRI showed enlarged posterior horns of both lateral ventricles, abnormal corpus callosum, and small hippocampus.Electroencephalogram (EEG) examination showed the subject had an abnormal EEG but was not consistent with clinical epileptiform.
For the case of subject III: Clinical diagnostic data was not available for subjects II:2 and II:3.However, according to the description of their parents, both twins often developed fever with bouts of convulsions after they were weaned at 1 year old, and this phenomenon occurred 3-4 times.At a later age, convulsions occurred when body temperature was documented as low-grade fever (37.5 °C), and this was diagnosed as epilepsy.Anti-convulsion medication Diazepam was effective to reduce the incidence of epileptic episodes.As adults, both II:2 and II:3 are restricted to ordinary housework, such as cooking, washing dishes, and sweeping the floor.Both are unable to perform more complicated tasks alone, such as shopping and paying for goods and handling money.

DNA extraction and whole genome sequencing
Genomic DNA was isolated from peripheral blood leukocytes and collected from participants using a DNA QIAamp mini kit (Qiagen, Hilden, Germany) according to instructions of the manufacturer.WGS was performed on samples from 4 affected patients (II:2, II:3, III:1, III:2) and 2 normal family members (I:1, I:2).Sequencing was carried out with a BGI-seq 500 with 50 bp paired-end reads (BGI-Shenzhen).

Bioinformatics analysis
Sequencing data was analyzed using the SOAPnuke package [28] to remove adapter sequences and low-quality reads, following which reads were mapped to the human genome reference (UCSC GRCh37/hg19) through the Burrows-Wheeler aligner (BWA-MEM, version 0.7.10) [29].Variant calling was performed using the genome analysis tool kit (GATK, version 3.3) [30].Variant effect predictor (VEP) was used to annotate and classify all the variants [31].We then screened variants based on their frequency in public and internal databases (1000 genome, GnomAD, and data not shown) [32], retaining only variants with a minor allele frequency (MAF) <0.005.Variants were subsequently filtered based on the ID inheritance model of the pedigree.Finally, pathogenic as well as likely pathogenic variants were identified using Sift [33], PolyPhen-2 [34], PROVEAN [35], MutationTaster [36], and CADD [37].

Sanger sequence validation
Primers for polymerase chain reaction (PCR) validation of sequence reads were designed using Primer 5.0.Putative candidate variants were verified by Sanger sequencing to exclude false positive variants.The six family members (affected individuals II:2, II:3, III:1, III:2; unaffected individuals I:1, I:2) were sequenced by bidirectional Sanger sequencing to determine co-segregation of the candidate mutations.PCR and sequencing primers are available upon request.

In silico modeling
Schrodinger Suite 2018-3 was used for in silico modeling.The H441Q mutation was introduced into the homology model of the ZBTB18-E1 protein-DNA complex [38] using Maestro.The mutated residues, as well as residues within 6.0 Å of the mutation, were energy minimized using Prime.

Mutation detection
To identify the causative variants in the ID family, we performed WGS.High-quality data was obtained, with mean coverage in excess of 90%, and with average read depth at > ×40 (Table 1).
Bioinformatics analysis identified a de novo mutation in both II:2 and II:3 defined as c.1323C>G, in reference to RefSeq transcript NM_205768.2,located within Exon2 of the ZBTB18 gene.This variant segregated perfectly in affected family members and was not detected in unaffected individuals when tested.In addition, Sanger sequencing was also to confirm co-segregation of the heterozygous variant in these affected members in the pedigree (Fig. 2).This variant has not been reported in the HGMD [39], gnomAD, or ClinVar [40] databases, suggesting it is a novel ZBTB18 missense variant.In addition to the c.1323C>G variants in the ZBTB18 gene, several variants in some other genes, including CFTR, TRBV10-1, ANKRD18A, MUC6, TAS2R19, NCAN gene, and variants in the intergenic region, were also identified in four patients, but not in two normal members, as shown in table S1.
The ZBTB18 gene encodes a transcription factor protein that binds DNA in a sequence-specific manner to influence gene expression during fetal nervous system development [21,[41][42][43][44].The c.1323C>G mutation detected in this family pedigree results in an amino acid substitution from histidine to glutamine at position 441 (p.His441Gln) of the polypeptide sequence.This missense variant is possible pathogenicity, based on the prediction results of software SIFT, PolyPhen-2, PROVEAN, MutationTaster, and CADD (Table 2).Sequence alignment with vertebrate orthologues revealed codon 441 of ZBTB18 to be highly conserved from humans to frogs to fish (Fig. 3), further suggesting a mutation at this codon may have an impact on the function of the polypeptide.Moreover, we performed molecular modeling to find that His441 is a zinc-coordinating residue within the second zinc finger motif of ZBTB18 [38].As such, an H441Q substitution likely disrupts zinc binding that is critical to the formation of this zinc finger protein folding  motif that, in turn, alters the structural stability of ZBTB18 and affects its capacity to bind DNA (Fig. 4).Given the putative role of His441 in protein folding, it is plausible that a p.His441Gln substitution could also alter steady-state levels of the ZBTB18 protein variant, leading to cellular dysfunction.

Discussion
Here, we report a novel heterozygous missense mutation (c.1323C>G, p.His441Gln) to the ZBTB18 (NM_205768.2) gene in a Han Chinese family that segregates with ID in an autosomal dominant fashion.The variant was detected as a de novo mutation in a monozygotic twin female pair, but not their parents.Each child of the twin females inherited this variant, and they were born with brain developmental disorder and ID.To our knowledge, this is the first report of an inherited ZBTB18 mutation for ID.
It was noticed that in our WGS analysis, several variants in some genes, including ZBTB18, CFTR, TRBV10-1, ANKRD18A, MUC6, TAS2R19, NCAN gene, and variants in intergenic region were identified.CFTR gene encodes a member of the ATP-binding cassette (ABC) transporter superfamily.The encoded protein functions as a chloride channel, making it unique among members of this protein family, and controls ion and water secretion and absorption in epithelial tissues.Mutations in the CFTR gene usually cause pulmonary cystic fibrosis [45] and congenital bilateral absence of vas deference [46].ANKRD18A mutation might cause thrombocytopaenia [47].MUC6 gene encodes a member of the mucin protein family.Mucins are high molecular weight glycoproteins produced by many epithelial tissues and play an important role in epithelial cells, usually associated with the risk of gastric cancer [48].NCAN gene encodes neurocan, a chondroitin sulfate proteoglycan thought to be involved in the modulation of cell adhesion and migration.Mutations in CFTR gene might be associated with the risk of autism spectrum disorder [49] and schizophrenia [50].As for TRBV10-1 and TAS2R19, there are few reports on functional studies and disease association.Because there have many literatures reporting the association between ZBTB gene mutation and intellectual disability, but there have few literatures on the association between CFTR, TRBV10-1, ANKRD18A, MUC6, TAS2R19, and NCAN gene mutations and intellectual disability, so ZBTB18 gene was considered as a candidate gene for subsequent analysis.
ZBTB18 is essential to the development of the mammalian brain [19].The expression and function of ZBTB18 in neural stem cells and their postmitotic progeny are crucial to the formation and differentiation of appropriate numbers of neurons and astrocytes within the fetal brain [41].During embryonic cerebral cortex development, ZBTB18 mediates the multipolar-to-bipolar transition of migrating cortical neurons by negative inhibition of the expression downstream target genes such as Neurog2 and Rnd2 [43,44].Moreover, ZBTB18 is essential for the growth and organization of the cerebellum and regulates the development of both GABAergic and glutamatergic neurons [21].It has been reported that ZBTB18 acts to repress the expression of pax6, ngn2, and neuroD1, since expression of these three sequential proneurogenic genes can signal intermediate neurogenic progenitors (INP) within the embryonic mouse cerebellum to differentiate into neurons and initiate migration [42].Given such critical roles for ZBTB18 in mammalian brain development, it is perhaps unsurprising that mutations to this gene lead to ID and human brain disorder.
The ZBTB18 gene encodes a transcriptional repressor protein comprising an N-terminal BTB (broad complex tramtrack bric-brac) domain mediating protein-protein interaction, as well as four Cys 2 His 2 (C2H2) type zinc fingers for DNA-binding within its C-terminus, respectively.In humans, genetic mutations to ZBTB18, including copy number variation, microdeletions, microduplications, and single nucleotide variants (SNVs), are associated with structural brain abnormalities, neuronal migration disorder, and ID [7,9,11,12,[22][23][24][25][26][27][51][52][53][54][55][56][57][58].Interestingly, while disease-associated nonsense and frameshift mutations are documented across the coding sequence, the overwhelming majority of disease-associated missense mutations are mapped to the C-terminal, C2H2 zinc finger domain, while a small proportion of such variants map to the BTB domain.Within the coding sequence of the C2H2 zinc finger domain, the majority of disease-associated ZBTB18 missense variants are clustered map to the second, third, and fourth zinc fingers, corresponding with the importance of these three motifs for DNA binding [38].Missense variation to residues within zinc fingers has been shown to impair the capacity of ZBTB18 to bind DNA and disrupt its function as a site-specific transcriptional repressor [24,38].In addition, disease-associated ZBTB18 variants could have a dominantnegative effect by disrupting the DNA binding by circulating wild-type protein in the presence of the mutant protein [24].Thus, growing evidence indicates that alterations to DNA binding and disruptions to transcriptional regulation could represent pathological mechanisms through which ZBTB18 missense variants cause disease in humans [38].
Different pathogenic variants in ZBTB18 have caused clinical consequences with an extensive and highly variable phenotypic impact.The phenotypes of patients with relatively sufficient clinical data and point ZBTB18 mutations were compared in Table 3.All patients presented with developmental delay in varying degrees and speech delay [7,9,[22][23][24][25][26][27], as shown in our family pedigree.Among 17 patients who underwent MRI scans, corpus callosum abnormality was observed in 10 patients, including one patient (III:1) in this study [9,24,25,27].Also, two patients in our pedigree (II:2, II:3) exhibited epilepsy, a feature observed in 6 of the 17 previously reported patients [24][25][26][27].In addition, microcephaly, hypotonia, and dysmorphic facial features were described in 4 patients [9,22,23,27], 9 patients [9,[24][25][26][27], and 14 patients [22,24,26,27], respectively.Different mutation types have different effects on the protein and, therefore, result in different disease phenotypes.A milder ID phenotype in our study and one previous study [7] is possibly due to only incomplete haploinsufficiency associated with the missense mutation compared to complete haploinsufficiency due to nonsense mutations in other studies [23,26].However, there was no clear genotype-phenotype correlation with regard to the position of the variant or type of the variant.It was worth noting that people who carried the same ZBTB18 mutation might have different clinical phenotypes.Among 4 patients who harbored the c.1323C>G mutation in our study, 2 patients presented with epilepsy, and 1 patient presented with corpus callosum abnormality.Besides, in previous studies, two patients carried the same c.1391G>A mutation, one presented with hypotonia [27], while the other did not [25].This phenotypic discordance may be due to incomplete penetrance [59].
The novel missense mutation c.1323C>G (p.His441Gln) of ZBTB18 is anticipated to be pathogenic based on the following evidence.Firstly, this variant is not detected in the general population (gnomAD) nor the National Center for Biotechnology Information (NCBI) single nucleotide polymorphism database (dbSNP) [60].Secondly, the mutation leads to the substitution of a highly conserved amino acid (His441) with basic properties, to glutamine that features polar, neutral side chains.Furthermore, five independent bioinformatics algorithms, including SIFT, PolyPhen-2, PROVEAN, MutationTaster, and CADD, described this variant as possible pathogenic.Thirdly, our in silico modeling studies reveal that His441 is critical for zinc binding within the second zinc-finger motif.However, functional studies are further required to confirm the potential effect of the mutation on the protein.As all four zinc fingers are essential for ZBTB18 to bind DNA [38,61], a His441Gln mutation could destabilize the protein, disrupt DNA sequence-specific binding, alter transcriptional regulation in cells, or all of the above.Indeed, recent studies of a p.Asn461Ser mutation detected in a child with ID [24] demonstrated that such a variant not only resulted in reduced steady-state levels of the mutant protein, but also compromised sequence-specific DNA binding and transcriptional repression [38].
The accurate detection of disease-causing mutations in subjects with ID is necessary to provide appropriate genetic counseling, as well as to inform potential future gene-specific therapies to alleviate clinical symptoms.Given the extensive genetic heterogeneity of ID, a genome-wide diagnostic approach is crucial to be able to detect all types of causative genetic variants, whether such variants are detected in coding or noncoding loci.Given the ever-improving cost-effectiveness of WGS in clinical diagnostic sequencing, it is increasingly being used to detect causative mutations [8,13,14,16].Our study demonstrates that WGS is efficient for detecting genetic causes of hereditary ID.   (continued) [25] [26] [23] [24] [22] [7]

Conclusion
This is the first report of an inherited ZBTB18 mutation for ID.Our finding of a c.1323C>G (p.His441Gln) mutation to ZBTB18 expands the spectrum of mutations that cause inherited forms of ID and demonstrates that WGS is highly efficient to provide a precise genetic diagnosis.

Table 1
Average depth and coverage of WGS.(BGI-Shenzhen, the permission was obtained from the creators)

Table 2
Prediction results of the mutation

Table 3
The clinical phenotypes of patients with point ZBTB18