De novo truncating variants in the AHDC1 gene encoding the AT-hook DNA-binding motif-containing protein 1 are associated with intellectual disability and developmental delay.

Whole-exome sequencing (WES) represents a significant breakthrough in clinical genetics, and identifies a genetic etiology in up to 30% of cases of intellectual disability (ID). Using WES, we identified seven unrelated patients with a similar clinical phenotype of severe intellectual disability or neurodevelopmental delay who were all heterozygous for de novo truncating variants in the AT-hook DNA-binding motif-containing protein 1 (AHDC1). The patients were all minimally verbal or nonverbal and had variable neurological problems including spastic quadriplegia, ataxia, nystagmus, seizures, autism, and self-injurious behaviors. Additional common clinical features include dysmorphic facial features and feeding difficulties associated with failure to thrive and short stature. The AHDC1 gene has only one coding exon, and the protein contains conserved regions including AT-hook motifs and a PDZ binding domain. We postulate that all seven variants detected in these patients result in a truncated protein missing critical functional domains, disrupting interactions with other proteins important for brain development. Our study demonstrates that truncating variants in AHDC1 are associated with ID and are primarily associated with a neurodevelopmental phenotype.

INTRODUCTION

Intellectual disability (ID) is present in 3% of children (Petersen et al. 1998; Shevell et al. 2003; Michelson et al. 2011). Up to 50% of moderate-to-severe ID cases are due to genetic causes (McLaren and Bryson 1987). Cytogenetically detectable chromosome aberrations account for almost 15% of cases (Leonard and Wen 2002). X-linked genetic defects are estimated to account for ∼10% of males with ID (Ropers 2010) and associated with more than 100 X-linked genes (Musante and Ropers 2014). Autosomal-dominant de novo mutations have been found as an important cause of ID in sporadic patients and have a low recurrence risk (Veltman and Brunner 2012), whereas the role of inherited variants in unaffected parents is less well understood. However, autosomal-recessive ID has a high recurrence risk and is the most common type of ID in consanguineous families (Musante and Ropers 2014), although only approximately 30 loci and 10 genes have been identified to date (Afroze and Chaudhry 2013). During the past decade, significant progress has been made to elucidate the genetic causes of ID. However, up to 60% of individuals with ID still do not have an identified etiology (Rauch et al. 2006). The FORGE (Finding of Rare Disease Genes) Canada Consortium project has identified disease-causing variants for 146 rare disorders including ID (Beaulieu et al. 2014). The Deciphering Developmental Disorders (DDD) study is also trying to identify the genetic causes for ID and other related developmental disorders and has recruited more than 12,600 families from the United Kingdom and Ireland for genomic analyses.

Genetic evaluation of ID routinely includes a chromosomal analysis/microarray and fragile X testing (Michelson et al. 2011; Flore and Milunsky 2012). If this initial diagnostic workup is unrevealing and if there are no specific clinical features to suggest a diagnosis, whole-exome sequencing (WES) is increasingly utilized as the next diagnostic test because of the tremendous genetic heterogeneity and incomplete knowledge of all the genes causing ID (Yang et al. 2013, 2014), The yield of WES for patients with ID is 27%–33% in large clinical series (Yang et al. 2013, 2014).

Here we describe novel truncating de novo variants in the AHDC1 gene associated with ID, identified using WES. AHDC1, located on Chromosome 1p36.11, encodes the 1603 amino acid protein AT-hook DNA-binding motif–containing protein 1. AHDC1 has previously been implicated as a rare cause of ID (Xia et al. 2014), and all five published variants are in the coding exon. Four de novo truncating mutations are associated with syndromic expressive language delay, hypotonia, and sleep apnea (Xia et al. 2014). A de novo missense variant was identified in 1 out of 53 sporadic cases of schizophrenia by exome sequencing (Guipponi et al. 2014). Recently, a de novo balanced translocation with a breakpoint in AHDC1 intron 1 that disrupts the 5′ untranslated region of AHDC1 was identified in a 5-yr-old boy with developmental delay and ID (Quintero-Rivera et al. 2015). To better understand the association of AHDC1 with ID and characterize the clinical phenotype of this new syndrome, we report our findings of de novo variants in seven patients from unrelated families with a similar clinical phenotype of severe intellectual disability or neurodevelopmental delay.

RESULTS

Exome sequencing was performed on 2157 cases with ID or developmental delay. From these cases, there were seven proband-parent trios with de novo AHDC1 variants. These seven cases produced an average of ∼9.5 Gb of sequence per sample. Mean coverage of captured regions was ∼115× per sample, with >97% covered with at least 10× coverage, an average of >92% of base call quality of Q30 or greater, and an overall average mean quality score of >Q35 (Supplemental Table 1). Initial filtering identified 271 genes (382 unique sequence changes) of interest across the seven families when considering all possible modes of inheritance. After manual filtering of variants with a minor allele frequency of >1% in reference populations, known disease association, pattern of inheritance, and similarity of clinical phenotype, there was one common gene of interest. In each case AHDC1 was the leading candidate because there was a novel, de novo loss of function variant, and loss of function alleles were absent from any public population database (6000 individuals of European and African American ancestry from the NHLBI Exome Variant Server), the Database of Single Nucleotide Polymorphisms (dbSNP), as well as our internal database including more than 6000 exomes of reportedly unaffected individuals. Only one nonsense variant p.Arg925Ter, located at Chr1: 27875854 G>A, has been seen in 1 out of 65,748 alleles in Exome Aggregation Consortium (ExAC), but with a low GQ score of ∼40, suggesting that this may not be present. Other variants in other genes were eliminated from consideration because there was a single variant in a gene observed in asymptomatic parents or because it was de novo but not predicted to alter protein function.

All seven de novo AHDC1 variants are either novel frameshift or nonsense variants that are predicted to result in loss of protein function (Table 1). The seven patients from unrelated families range in age from 2 to 16 yr and have multiple neurobehavioral manifestations. All have significant developmental delay or intellectual disability and are either nonverbal or had limited ability for verbal expressive language (Table 1). Exact developmental quotient/intelligence quotients were not consistently determined with the same instruments, but the degree of impairment ranged from moderate to severe, based on standardized age appropriate measures of intelligence. All of the patients demonstrate problems with muscle tone and global developmental delays, with the age at walking ranging from 20 mo to 4 yr 9 mo. Five of the patients have ataxia. The majority of the patients have behavioral issues. Two patients have autism (based on clinical diagnosis using DSM IV or V criteria), two have self-injurious behavior, one is impulsive, one is aggressive, and one patient is described as being very affectionate. Four of the children have problems with sleep: one with sleep apnea, one a “restless sleeper,” one with trouble staying asleep, and another with trouble falling and staying asleep. One patient had two seizures that were not captured on EEG and another had an abnormal EEG with epileptiform activity. MRIs of the brain were abnormal in five patients, three of whom had a small or thin corpus callosum.

Table 1.

Clinical features of individuals with AHDC1 de novo predicted pathogenic variants

	Age	Gender	Variant	De novo/inherited	Developmental delay	Cognitive delay	Age at walking	Current speech	OFC (%)	Hypotonia	Dysmorphic	Growth or feeding issues	Brain abnormality	Other neurobehavioral problems	Other
Patient 1	2 yr	M	c.1945delGp.A649Pfs*83	De novo	+	N/A	No independent walking	No verbal speech	96	+	+	+	−	ASD	Omphalocele, cryptorchidism, exotropia, sleep apnea
												Short stature (1%)
Patient 2	7 yr	F	c.2529_ 2545del17p.D845Rfs*40	De novo	+	+	3 yr 10 mo	2–3 word sentences	85	+	+	+	+	Cerebral palsy with spastic quadriparesis	CVI, estotropia, nystagmus, blepharoptosis, scoliosis, restless sleeper, not toiled trained
												G-tube	Periventricular leukomalacia, thin CC, Subependymal cysts
Patient 3	5 yr	M	c.1881delGp.Q627Hfs*105	De novo	+	+ (IQ <50)	UKN	UKN	UKN	+	+	−	−	Ataxic gait, autism, epileptic activity on EEG	Sagittal craniosynostosis, cochlear nerve dysfunction, hypermobile joints
Patient 4	4 yr	F	c.1122dupCp.G375Rfs*3	De novo	+	+	20 mo	Single words	91	+	−	+	+	Wide based gait, nystagmus	Affectionate
												Feeding difficulties	Small CC, loss of posterior ventricular white matter
Patient 5	9 yr	M	c.3809delAp.Q1270Rfs*75	De novo	+	+	4 yr 9 mo	No verbal speech (had a maximum of 2–3 words at 2 yr	75–90	+	+	+	+	Ataxia, intermittent nystagmus, ptosis, mildly self-injurious and intermittent behavior problems, disrupted sleep,	Left central macular dystrophy, isolated complex IV deficiency on muscle biopsy (24% of control)
												FTT	Hypomyelination, thin CC, small brain stem
Patient 6	5 yr	M	c.2373_ 2374delTGp.C791Wfs*57	De novo	+	+	2 yr	Two word sentences	50	+	+	−	+	Ataxia, seizures, impulsive	Mild joint laxity
													Prominent CSF space—small arachnoid cyst vs. abnl cisterna magna
Patient 7	16 yr	F	c.1480 A>Tp.K494Ter	De novo	+	+	UKN	No verbal speech	50–75	−	+	+	+	Autism, aggressive, self-injurious, strabismus, developmental regression, increased pain tolerance	Club foot, pes cavus
												Feeding difficulties, short stature	Demylenation involving subcortical U fibers, retrocerebellar cyst
Totals					7/7	6/6				6/7	6/7	5/7	5/7	5/7 Ataxic, 2/7 autism

OFC, occipital frontal circumference; M, male; F, female; UKN, unknown; N/A, not applicable; ASD, atrial septal defect; CVI, cortical visual impairment; EEG, electroencephalogram; CC, corpus callosum; FTT, failure to thrive; CSF, cerebrospinal fluid.

In addition to the neurobehavioral features, there are other clinical manifestations. Six patients have dysmorphic facial features including bitemporal hollowing, brachycephaly, plagiocephaly, short nose, narrow mouth, highly arched palate, tented upper lip, and high, prominent forehead with prominent glabella (Fig. 1). Five patients had a history of growth problems including short stature or failure to thrive and feeding difficulties. Two of the patients have joint laxity. Five patients have ophthalmologic problems, including one with cupped optic disks with unilateral central macular dystrophy and nystagmus, one with nystagmus only, two with strabismus only, and one with nystagmus along with strabismus and cortical visual impairment.

Figure 1.

Facial characteristics of individuals with AHDC1 variants including a high, broad prominent forehead, hypertelorism, depressed nasal bridge, and thin upper lip.

DISCUSSION

Seven proband-parent trios with de novo AHDC1 variants were identified through whole-exome sequencing on 2157 cases with ID or developmental delay. The seven patients from unrelated families with a similar clinical phenotype of severe intellectual disability or neurodevelopmental delay, dysmorphic features, and problems with feeding and growth were all found to be heterozygous for novel de novo truncating variants in AHDC1. We report six novel predicted deleterious variants and one recurrent variant, (p.Cys791Trpfs*57), which was previously reported in two patients by Xia et al. (2014). Xia et al. has proposed that AHDC1-associated ID is due to a dominant-negative mechanism given the autosomal-dominant inheritance and the single coding exon of this gene. However, a de novo balanced translocation with a breakpoint in AHDC1 intron 1 that disrupts the 5′UTR of AHDC1 has been identified in a 5-yr-old male patient with developmental delay and ID (Quintero-Rivera et al. 2015). AHDC1 expression in lymphoblastoid cells was found to be reduced to 50% of wild-type levels, supporting haploinsufficiency as disease mechanism for AHDC1 mutations (Quintero-Rivera et al. 2015). Additionally, several large CNV deletions that include the AHDC1 gene have been described in pediatric patients with developmental delay and intellectual disability (Itsara et al. 2009; Cooper et al. 2011; Coe et al. 2014). However, due to the large size of the deletions, it is uncertain whether AHDC1 is the only gene contributing to the neurodevelopmental phenotype.

AHDC1 consists of seven exons with only one coding exon (exon 6). Five 5′ noncoding exons and one 3′ noncoding exon flank the coding exon. AHDC1 encodes the 1603 amino acid protein AT-hook DNA-binding motif-containing protein 1, which likely binds DNA (Fig. 2; Reeves and Nissen 1990). By aligning human AHDC1 to other protein orthologs, two conserved regions have been reported in AHDC1 (Xia et al. 2014). The first conserved region has two AT-hook motifs located at amino acids 396–408 and 544–556. AT-hook motifs are known as auxiliary protein motifs that cooperate with other DNA-binding activities and facilitate DNA structure changes (Aravind and Landsman 1998). The AT-hook binding domain in MECP2, the gene mutated in Rett syndrome, plays an important role in chromatin organization (Baker et al. 2013; Xu and Pozzo-Miller 2013). The second conserved region near the carboxyl terminus contains a carboxy-terminal PDZ binding domain consensus sequence that interacts with PDZ domain proteins (Shalaby et al. 2011). PDZ domain proteins regulate multiple biological processes such as transport of electrolyte and fluid (Seidler et al. 2009), ion channel signaling, and signal transduction systems by recognizing short amino acid motifs at the carboxyl termini of target proteins (Lee and Zheng 2010).

Figure 2.

AHDC1 gene organization and distribution of truncating mutations. AHDC1 consists of seven exons with only one coding (exon 6). Three mutations above the schematic protein were reported by Xia et al. (2014). Six mutations below the schematic protein are novel mutations reported in this study. The mutation in the box reported by Xia et al. was found in one of our probands.

The in vivo targets for AHDC1 binding are largely unknown. However, in vitro assays demonstrate that AHDC1 interacts with a number of other nuclear proteins (Chatr-Aryamontri et al. 2015), influencing transcriptional and epigenetic regulation during development (Lim et al. 2006; Vandamme et al. 2011), receptors in neurite growth, neuronal adhesion, and axonogenesis (Olah et al. 2011), and transporters for specific neuron proteins (Butland et al. 2014). AHDC1 is expressed throughout the brain (Ponten et al. 2008; Uhlen et al. 2015). In the mouse, Ahdc1 is expressed at embryonic day E11.5 and E16.5 in the developing brain (Quintero-Rivera et al. 2015), suggesting that AHDC1 may be involved in early brain development.

Four of our de novo AHDC1 variants and three previously reported variants (Xia et al. 2014) are predicted to cause protein truncation beyond the AT-hook motifs and to eliminate the PDZ binding domain consensus sequence. The other two remaining de novo AHDC1 variants (p.Lys494Ter and p.Gly375Argfs*3) lead to an even earlier stop of translation thus destroying both of the conserved regions. As these truncating variants are all located in the single coding exon of AHDC1, they may not result in nonsense mediated mRNA decay (Schoenberg and Maquat 2012). Although we did not conduct expression studies in our cohort, based on these observations, we speculate that AHDC1 variants may interrupt protein translation, leading to truncated proteins missing the functional domains, possibly disrupting the interaction with other proteins, such as PDZ domain proteins important to brain development and function. The phenotype in all seven patients we report is dominated by neurobehavioral manifestations including ID, autism spectrum disorder, sleep disturbances, gross motor delays, and ataxia, but also includes facial dysmorphism. Our series expands the phenotype of AHDC1 disorders to include behavioral phenotypes of aggression, self-injurious behavior. Other medical features include problems with growth, epilepsy, nystagmus, omphalocele, and clubfoot.

WES has become an important diagnostic tool to evaluate patients with developmental delay and ID, especially in patients with nonspecific clinical features. Our study demonstrates that truncating de novo variants in AHDC1 are a rare cause of severe ID and are associated with feeding difficulties and problems with growth, and a neurobehavioral phenotype is emerging that includes challenging behaviors such as self-injury, aggression, autism, ataxia, and disordered sleep. Variants in AHDC1 should be further investigated in patients with this neurobehavioral profile.

METHODS

Whole-Exome Sequencing

Genomic DNA from whole blood from the affected children and their parents was extracted. Exome sequencing was performed on exon targets isolated by capture using the Agilent SureSelect Human All Exon V4 (50 Mb) kit (Agilent Technologies). One microgram of DNA from blood specimen was sheared into 350- to 400-bp fragments, which were then repaired, ligated to adaptors, and purified for subsequent PCR amplification. Amplified products were then captured by biotinylated RNA library baits in solution following the manufacturer's instructions. Bound DNA was isolated with streptavidin-coated beads and re-amplified. The final isolated products were sequenced using the Illumina HiSeq 2000/2500 sequencing system with 100-bp paired-end reads (Illumina). DNA sequence was mapped to the published human genome build UCSC hg19/GRCh37 reference sequence. Targeted coding exons and splice junctions of known protein-coding RefSeq genes were assessed for average depth of coverage with a minimum depth of 10× required for inclusion in downstream analysis. Local realignment around insertion–deletion sites was performed using the Genome Analysis Toolkit v1.6. (DePristo et al. 2011). Variant calls were generated simultaneously on all sequenced family members using SAMtools v0.1.18. (Li et al. 2009). All coding exons and surrounding intron/exon boundaries were analyzed. Automated filtering removed common sequence changes (defined as >10% frequency present in 1000 Genomes database) and additional filtering focused on variants with an allele frequency <1%. The targeted coding exons and splice junctions of the known protein-coding RefSeq genes were assessed for the average depth of coverage and data quality threshold values. Whole-exome sequence data for all sequenced family members were analyzed using GeneDx's XomeAnalyzer (a variant annotation, filtering, and viewing interface for WES data), which includes nucleotide and amino acid annotations, population frequencies (NHLBI Exome Variant Server and 1000 Genomes databases), in silico prediction tools, amino acid conservation scores, and mutation references. Variants were filtered based on inheritance patterns, gene lists of interest, phenotypes, and population frequencies, as appropriate. Resources including the Human Gene Mutation Database (HGMD), 1000 Genomes database, NHLBI Exome Variant Server, OMIM, PubMed and ClinVar were used to evaluate genes and detected sequence changes of interest. The general assertion criteria for variant classification are publicly available on the GeneDx ClinVar submission page (http://www.ncbi.nlm.nih.gov/clinvar/submitters/26957/). Additional searches were performed using specific gene lists related to clinical features of the patients. Identified sequence changes of interest were confirmed in all family members by conventional di-deoxy DNA sequence analysis using an ABI 3730 (Life Technologies) and standard protocols with a new DNA preparation.

ADDITIONAL INFORMATION

Ethics Statement

The study was approved by the Institutional Review Board of Columbia University and written consent was obtained for collecting blood samples and sequencing from all study participants.

Database Deposition and Access

Whole-exome sequencing data is not publicly available because patient consent could not be obtained. The AHDC1 variants have been deposited in ClinVar (http://www.ncbi.nlm.nih.gov/clinvar/) under accession numbers SCV000243809–SCV000243815.

Acknowledgments

We thank the families for their generous contributions.

Author Contributions

H.Y. analyzed the data, drafted and critically reviewed the manuscript. G.D. analyzed the data and critically reviewed the manuscript. K.G.M. analyzed the data and critically reviewed the manuscript. K.R. generated and analyzed the data and critically reviewed the manuscript. M.T.C. analyzed the data and critically reviewed the manuscript. L.F.E. provided the clinical data and critically reviewed the manuscript. M.E.T. provided the clinical data and critically reviewed the manuscript. J.S. provided the clinical data and critically reviewed the manuscript. L.H.R. provided the clinical data and critically reviewed the manuscript. D.S. provided the clinical data and critically reviewed the manuscript. W.M. provided the clinical data and critically reviewed the manuscript. G.E. provided the clinical data and critically reviewed the manuscript. J.P. provided the clinical data and critically reviewed the manuscript. R.C. provided the clinical data and critically reviewed the manuscript. P.G.W. provided the clinical data and critically reviewed the manuscript. C.C. provided the clinical data and critically reviewed the manuscript. A.C. provided the clinical data and critically reviewed the manuscript. R.T. provided the clinical data and critically reviewed the manuscript. G.R. analyzed the data and critically reviewed the manuscript. P.V. conceived of the study, analyzed the data, drafted and critically reviewed the manuscript. W.K.C. conceived of the study, analyzed the data, drafted and critically reviewed the manuscript.

Funding

This work was supported in part by a grant from the Simons Foundation.

Competing Interest Statement

H.Y., G.D., K.G.M., K.R., M.T.C., G.R., and P.V. are employees of GeneDx. Wendy Chung is a consultant to BioReference Laboratories.