Background We present a large Dutch family with seven males affected by a novel syndrome of X-linked intellectual disability, hypogonadism, gynaecomastia, truncal obesity, short stature and recognisable craniofacial manifestations resembling but not identical to Wilson-Turner syndrome. Seven female relatives show a much milder expression of the phenotype.
Methods and results We performed X chromosome exome (X-exome) sequencing in five individuals from this family and identified a novel intronic variant in the histone deacetylase 8 gene (HDAC8), c.164+5G>A, which disturbs the normal splicing of exon 2 resulting in exon skipping, and introduces a premature stop at the beginning of the histone deacetylase catalytic domain. The identified variant completely segregates in this family and was absent in 96 Dutch controls and available databases. Affected female carriers showed a notably skewed X-inactivation pattern in lymphocytes in which the mutated X-chromosome was completely inactivated.
Conclusions HDAC8 is a member of the protein family of histone deacetylases that play a major role in epigenetic gene silencing during development. HDAC8 specifically controls the patterning of the skull with the mouse HDAC8 knock-out showing craniofacial deformities of the skull. The present family provides the first evidence for involvement of HDAC8 in a syndromic form of intellectual disability.
Statistics from Altmetric.com
X-linked intellectual disability (XLID) comprises a genetically heterogeneous group of disorders with a prevalence as high as 1/600–1/1000 males.1 It is estimated that up to 200 genes might be responsible for XLID disorders.2 Approximately 80 genes on chromosome X have been already identified through positional cloning, candidate gene analysis and cytogenetic studies, and the introduction of exome sequencing strategies in the last 2 years increased the number to >90.3 This indicates that novel genes involved in XLID syndromes are still to be identified.
We characterised a five-generation Dutch family with seven males affected by a syndromic form of severe intellectual disability (ID) and seven females with a much milder phenotype (figure 1A). The main phenotypic characteristics in affected males are ID, truncal obesity, gynaecomastia, hypogonadism, short stature, small hands, and a typical face characterised by a small head, short ears, prominent supraorbital ridges, deep-set eyes, high malae, broad nasal tip, columella somewhat below the nasal alae, thin upper vermillion, and retrognathia (figure 1B, supplementary table S1). X-linked dominant inheritance was indicated by the absence of male-to-male transmission, and by a more severe phenotype in males as compared with the milder phenotype, comprising learning disorder and recognisable facial features, in obligate female carriers. In addition to the present syndromic form of XLID, mild isolated ID without an unusual phenotype was present in three individuals (figure 1A: II:1, III:1 and III:7). They had learning problems but did not show the recognisable facial dysmorphisms and/or additional clinical features. Female III:7 did not resemble the carrier females in this family. In particular, the high malae and broad nasal tip, noted in all carrier females, were absent in III:7.
The clinical phenotype overlaps with two known forms of XLID: Borjesson-Forssman-Lehmann syndrome (BFLS, MIM 301900) and Wilson-Turner syndrome (WTS, MIM 309585). BFLS is less likely due to the more coarse facial features and the large, fleshy ears in this entity.4 WTS has been described in a single family,5 and differs in a larger stature and head, larger chin, somewhat coarser face and tapering fingers. Females were not affected. Linkage analysis in the original family localised the WTS locus in the pericentromeric region of the X chromosome (Xp22.1–q22), between the DNA markers DXS426 and DSXS990, but the responsible gene has not yet been identified.5 The BFLS locus—that is, the PHF6 gene—is in the Xq26.3 region.
To characterise this family further, linkage with the published BFLS and WTS loci was tested by genotyping; 24 family members were genotyped for three markers in the Xq26–28 region spanning the PHF6 gene4 and six markers in the pericentromeric region of the X chromosome (Xp22.1–q22)5 (figure 1). Multipoint parametric linkage for a dominant X-linked model was performed with MERLIN package MINX. We assumed an allele frequency of 0.000001, and penetrances for the wild-type, the heterozygote and the homozygote or hemizygote of 0.000001, 0.99 and 0.99, respectively. Marker positions were taken from the Généthon map. In concordance with the initial clinical diagnosis, the BLFS region was excluded as the candidate region for the localisation of the gene underlying the disorder in this family, based on linkage analysis (data not shown). Also sequencing of the BFLS gene PHF64 failed to show a mutation in patient III:4. However, linkage was found using markers from the WTS critical region, with a maximum logarithm of odds (LOD) score of 4.93 at marker DXS983 (chrX: 69, 448 069). All affected individuals and obligate carriers showed completely shared haplotypes for the candidate WTS locus. In addition, due to a pericentromeric recombination in female III:7 who did not show the abnormal phenotype, the proximal border of the candidate region could be fine-mapped to Xq11–q22 (supplementary figure S1). This region extending from DXS983 to DXS986 contains 75 protein coding genes, 71 pseudogenes and 33 RNA coding genes.
Subsequently, we performed next-generation sequencing of all protein coding sequences on the X chromosome (X-exome) in five individuals from this family (three affected males (III:3; III:12; III:25), one affected female (II:10), and one healthy female (III:18)). In short, barcoded fragment libraries were equimolarily pooled and enriched using multiplexed targeted genomic enrichment6 with the Demo X-exome enrichment kit (Agilent Technologies, Santa Clara, CA, USA) and sequenced according to the SOLiD 3 Plus manual (Life Technologies, Carlsbad, CA, USA).. Raw sequencing data were mapped against the GRCh37/hg18 reference genome using a custom bioinformatic pipeline based on the BWA algorithm. Sequencing and mapping statistics are presented in supplementary table S2.
Candidate single nucleotide variants and small indels were called. The percentage cut-off for the non-reference allele (NRA) was set to 15% and the cut-off for minimum sequence coverage was set to 10 reads. Variants called in at least one out of the five individuals were evaluated in the other samples using less stringent parameters to exclude false negatives. Filtering of variants that fit an X-linked dominant inheritance model was performed using non-stringent criteria as follows: affected males (III:3, III:12, III:25) should be hemizygous for the candidate causal variant (>50% NRA), the affected female (II:10) should have at least one allele (15–90% NRA), and the healthy female (III:18) should not carry the variant (<5% NRA). Subsequently, all common and rare polymorphisms present in Ensembl 65 or our in-house X-exome database of approximately 100 samples were marked as known and not further considered for this study; remaining variants were considered to be novel and thus fulfil the criteria for an ultra-rare disease as presented here. All bioinformatic scripts and pipelines are available upon request. Sequencing data are available (Sequence Read Archive accession number ERP001237).
The analysis revealed two possible causal variants: ZNF81 c.376C>T (p.Ala3Val) and histone deacetylase 8 (HDAC8) c.164+5G>A (p?) (supplementary figure S2). Both variants were confirmed by Sanger sequencing to be hemizygous in the three affected males (III:3, III:12, III:25), heterozygous in the affected female (II:10), and not present in the healthy female (III:18) (figure 1C). Primer sequences are summarised in supplementary table S3. We performed Sanger sequencing in 96 control Dutch males and checked the variants in an online database of approximately 5000 exomes (Exome Variant Server, NHLBI Exome Sequencing Project, Seattle, Washington, USA (http://evs.gs.washington.edu/EVS/), January 2012). The ZNF81 gene is located close to but outside the fine-mapped candidate region in this family. Furthermore, the ZNF81 variant is present in 2/96 control males and 114/8148 exomes in databases. In addition, the variant was predicted to have no damaging effect on protein function (Polyphen 2, SIFT). This reduced the likelihood that ZNF81 is the causative gene.
HDAC8 c.164+5G>A was not present in controls or public databases. The mutation is located in the consensus sequence of the donor splice site of exon 2 (GERP score 3.8) and is predicted to negatively affect normal splicing. Additional segregation analysis in 32 individuals from this family confirmed segregation in all affected males, affected females and obligates carriers, and absence in non-affected individuals (figure 1A,B). Furthermore, HDAC8 is located within the fine-mapped region between markers DXS983 and DXS986 (supplementary figure S1). Therefore, the novel variant in the HDAC8 gene remains as the only candidate responsible for the syndromic form of XLID in this family. To date, no deletions involving the HDAC8 locus have been reported.
To test the effect on the structure of HDAC8 mRNA, total RNA was isolated from peripheral lymphocytes of patient III:9 and cDNA synthesis was performed using random hexamer primers (Roche, Basle, Switzerland). Subsequently, reverse transcriptase PCR (RT-PCR) with primers located in the flanking exons 1 and 3 (and 1 and 4) was performed. The patient was found to express a transcript lacking exon 2 sequences as compared to the control (figure 2A). As a result, translation in affected individuals will result in a premature stop (p.Ala38AspfsX3) (figure 2B) at the very beginning of the histone deacetylase catalytic domain (figure 2C).7 Although HDAC8 encodes several alternatively spliced products, skipping of exon 2 is predicted to result in a truncated protein in all known transcripts. In addition to the shorter product lacking exon 2, a normal fragment with very low intensity was visible in the patient. We confirmed the presence of this fragment with an RT-PCR with the exon 1 forward primer and an exon 2 specific reverse primer (figure 2A). Primer sequences are available in supplementary table S3. These results indicated that still some correctly spliced HDAC8 mRNA is expressed at least in circulating lymphocytes.
To evaluate affected female carriers in this family, we performed X chromosome inactivation analysis on DNA isolated from peripheral lymphocytes after amplification of the CAG repeat region at the nearby androgen receptor locus (supplementary table S3). To discriminate between methylated and unmethylated DNA, the PCR was performed on DNA with and without digestion with the methylation sensitive restriction enzyme CfoI. All female carriers of the HDAC8 mutation displayed a notably skewed X-inactivation pattern (skewing close to 100%) with the mutated allele inactivated (supplementary figure S3), which indicates a selection against cells with an active X chromosome with the mutated HDAC8 gene. The female phenotype features such as learning difficulties and facial characteristics may imply that a negative effect of the HDAC8 variant on target tissues takes place in a very early stage of development,8 and after the establishment of selective advantage for non-mutated cells development of the organism continues normally. It remains possible that the pattern of skewing found in the peripheral blood cells is not representative for other organs and tissues, for example, brain.4 In this family skewing was present in all affected females and absent in healthy females, confirming the X-linked condition and allowing it to be used as a diagnostic marker as is often possible in other XLID disorders.9
Next to negative selection, skewed X-inactivation patterns may have other causes such as structural abnormalities or variations in the XIST promoter region.10 This prompted us to perform karyotyping at a 500-band resolution, microdeletion analysis in 34 markers between DXS993 and DXS990, and mutation analysis by direct sequencing of the XIST minimal promoter region in individual III:4. No abnormalities were found (data not shown). Furthermore, some genes escape X-inactivation but this is not the case for HDAC8.11
HDACs are evolutionary conserved enzymes that remove acetyl groups from lysine residues of histones and other important cellular non-histone proteins.12 They play a major role in epigenetic gene silencing during development, for example, X-chromosome inactivation, cell differentiation and morphogenesis.13 Epigenetic chromatin remodelling and DNA modifications represent the central mechanisms for regulation of gene expression during brain development.14 Thus, based on the function of HDACs, increasing numbers of HDAC inhibitors are presently being developed for treatment of neuropsychiatric diseases.15
HDACs were already associated with disorders involving human brain development. Haploinsufficiency in HDAC4 causes brachydactyly-mental retardation syndrome (MIM 600430).16 Variants in the HDAC binding domain of TGIF cause holoprosencephaly-4 (MIM 142946), which also shows microcephaly and intellectual disability as part of the phenotype.17 BFLS shows a phenotypic overlap with the clinical characteristics in the present family, and is caused by a plant homeodomain (PHD)-like finger PHF6.4 This group of zinc-finger proteins is associated with chromatin remodelling and histone acetylation.18 Another member of the PHD finger family is PHF21A, a member of a HMG20B/HDAC complex that mediates repression of neuron specific genes through a cis-regulatory element.19
HDAC8 specifically controls patterning of the skull by repressing transcription factors in neural crest cells.20 Knock-out mice of HDAC8 have craniofacial deformities of the skull, show a distinct ossification defect with a wide foramen frontale and defects in the interparietal bone, reduction in body weight and size, and die early after birth. In relation to the genetic background in Hdac8 mutants the observed skull phenotype ranged from a small persistent anterior fontanelle to severe frontocranial dysplasia. Mice with a global deletion of Hdac8 are viable and show a runted phenotype with hypoplastic, dysmorphic skulls. In female mice heterozygous for the Hdac8-null, lethality was observed dependent on the extent of random X chromosome inactivation of the wild-type allele.20
Significant skull defects have not been identified at physical examination of the affected males in the family presented here. However, all males that carry the variant in HDAC8 show microcephaly and distinct facial deformities of the skull, especially prominent supraorbital ridges and high cheekbones. In comparison with the knock-out situation where mice died early after birth, this partial loss-of-function mutation might explain the non-lethal but still severe phenotype in male patients presented here.
In summary, we identified a splice site mutation in HDAC8 as causal of a possibly new XLID disorder. Since the difference between the phenotype in this family and WTS is small and HDAC8 is located in the original WTS locus, one can hypothesise that both are allelic, and with HDAC8 we identified the gene responsible for WTS syndrome. We cannot exclude the possibility, however, that both families represent XLID caused by different genes within the same chromosomal region.
We would like to express our thanks to the members of this family for their intensive cooperation and for photographic documentation. We thank Prof Dr Marc Timmers and Dr Martin Poot for valuable suggestions and discussions. We thank Mirjan Albring for performing the X-inactivation tests.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
▸ Additional supplementary figures and tables are published online only. To view this file please visit the journal online (http://dx.doi.org/10.1136/jmedgenet-2012-100921)
Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sector.
Competing interests None.
Patient consent Written informed consent was obtained from all study participants or their guardians and the study was approved at the recruiting centre.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All bioinformatic scripts and pipelines are available upon request. Sequencing data are available (Sequence Read Archive accession number ERP001237).
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.