Background Overgrowth conditions are a heterogeneous group of disorders characterised by increased growth and variable features, including macrocephaly, distinctive facial appearance and various degrees of learning difficulties and intellectual disability. Among them, Sotos and Weaver syndromes are clinically well defined and due to heterozygous mutations in NSD1 and EZH2, respectively. NSD1 and EZH2 are both histone-modifying enzymes. These two epigenetic writers catalyse two specific post-translational modifications of histones: methylation of histone 3 lysine 36 (H3K36) and lysine 27 (H3K27). We postulated that mutations in writers of these two chromatin marks could cause overgrowth conditions, resembling Sotos or Weaver syndromes, in patients with no NSD1 or EZH2 abnormalities.
Methods We analysed the coding sequences of 14 H3K27 methylation-related genes and eight H3K36 methylation-related genes using a targeted next-generation sequencing approach in three Sotos, 11 ‘Sotos-like’ and two Weaver syndrome patients.
Results We identified two heterozygous mutations in the SETD2 gene in two patients with ‘Sotos-like’ syndrome: one missense p.Leu1815Trp de novo mutation in a boy and one nonsense p.Gln274* mutation in an adopted girl. SETD2 is non-redundantly responsible for H3K36 trimethylation. The two probands shared similar clinical features, including postnatal overgrowth, macrocephaly, obesity, speech delay and advanced carpal ossification.
Conclusions Our results illustrate the power of targeted next-generation sequencing to identify rare disease-causing variants. We provide a compelling argument for Sotos and Sotos-like syndromes as epigenetic diseases caused by loss-of-function mutations of epigenetic writers of the H3K36 histone mark.
- Clinical genetics
- Molecular genetics
Statistics from Altmetric.com
Sotos syndrome (SoS) (Mendelian Inheritance in Man (MIM) 117550), also known as cerebral gigantism, was first described by Sotos et al in 1964.1 Although most cases are sporadic, several reports of autosomal-dominant inheritance have been described. This overgrowth syndrome is characterised by excessive growth during childhood, macrocephaly, distinctive facial appearance and various degrees of learning difficulty. The typical craniofacial features include macro-dolichocephaly, a broad forehead with a receding hairline, a prominent chin and down-slanting of the palpebral fissures.2 SoS can be associated with other features like advanced bone age, neonatal jaundice, hypotonia, seizures, cardiac defects and genitourinary anomalies.3 Since its first description,1 hundreds of SoS cases have been reported in the literature. The phenotypic spectrum of these patients is usually broad, varying from a classical SoS phenotype to patients exhibiting only a few Sotos syndrome features (Sotos-like syndrome).4–6 Weaver syndrome (WS; MIM 277590) is seen less commonly than SoS and shows significant phenotypic overlap with SoS.7 WS comprises prenatal and postnatal overgrowth, a typical craniofacial appearance (micrognathia with a deep horizontal chin crease), deep set nails, camptodactyly and advanced carpal osseous maturation.8 Carpal bone development is advanced over the rest of the hand in WS, whereas in SoS carpal bone development it is at or behind that of the rest of the hand.
In 2002, the genetic origin of SoS was elucidated.9 In a patient with SoS, the 5q35 translocation breakpoint of a de novo balanced chromosomal translocation was shown to disrupt the NSD1 gene (nuclear receptor binding SET-domain containing gene 1) (MIM 606681). The NSD1 protein belongs to histone ‘writer’ proteins. NSD1 is a SET (Su(var)3–9, Enhancer of Zeste and Trithorax) domain histone methyltransferase that primarily targets nucleosomal histone H3 lysine 36 (H3K36).10 Histone lysine methylation signalling is a principal chromatin regulatory mechanism that influences fundamental nuclear processes linked to downstream biological functions by methyllysine-binding proteins.11 Lysine (K) residues can accept up to three methyl groups to form mono-, di- and trimethylated derivatives. H3K36 dimethylation (H3K36me2), mediated by NSD1, is regarded as an activating chromatin mark.12 Since this discovery, reports have indicated that approximately 60–80% of the clinically diagnosed patients with SoS harbour heterozygous NSD1 intragenic mutations or microdeletions.2 ,6 ,13 Therefore, there are a substantial number of patients suspected of SoS but without a molecular explanation.
Molecular overlap has been demonstrated between SoS and WS conditions. NSD1 mutations account for some WS cases.6 However, trio-based whole-exome sequencing in three NSD1-negative WS families recently identified de novo mutations in the EZH2 gene (enhancer of zeste homologue 2) (MIM 601573).14 EZH2 mutations were then identified in 3/29 individuals with Weaver-like phenotypes (A.S.A. Cohen et al, 2013, 63rd ASHG Annual Meeting, abstract). The EZH2 protein partners with SUZ12 and EED to form the minimal polycomb repressive complex 2 (PRC2).15 This complex catalyses the trimethylation of lysine 27 of histone H3 (H3K27me3), and EZH2 itself forms the catalytic subunit for this reaction. EZH2 forms a key component of molecular machinery that shuts off transcription of loci to which trimethylated H3K27 is bound.
Constitutional heterozygous mutations in NSD1 and EZH2 cause the two overlapping overgrowth syndromes SoS and WS. NSD1 and EZH2 are both epigenetic writers that catalyse two specific post-translational modifications of histones: methylation of histone 3 lysine 36 (H3K36) and lysine 27 (H3K27), respectively. We postulated that mutations in other genes coding writers of these two chromatin marks could cause overgrowth conditions resembling Sotos or Weaver syndromes in patients with no NSD1 or EZH2 abnormalities. Due to its low cost and high throughput, targeted next-generation sequencing (NGS) delivered a step change in the ability to simultaneously sequence multiple genes. We analysed the coding sequences of 14 H3K27 methylation-related genes and eight H3K36 methylation-related genes in 16 patients with SoS, Sotos-like and WS, using a targeted NGS approach.
Subjects and methods
A total of 16 patients were included in the study, namely, 3 SoS, 11 ‘Sotos-like’ patients and 2 WS patients. All presented overgrowth, macrocephaly and learning disability. They were regularly followed (once a year) and had repeated bone age assessment at various ages. Three patients were considered as typical Sotos patients as they fulfilled the diagnostic criteria defined by Cole and Hughes: facial gestalt, overgrowth >2 SD, advanced bone age and macrocephaly >2 SD.2 ,16 Eleven patients were considered as Sotos-like patients as they presented with less specific or absence of facial gestalt but fulfilled the other diagnostic criteria. Finally, two patients were considered as Weaver syndrome patients. They both presented with the suggestive facial gestalt, overgrowth, macrocephaly, deep set nails, camptodactyly and accelerated carpal maturation. All patients were regularly followed in the French Department of Medical Genetics and phenotypically scored by clinical geneticists. The study was approved by the local ethics committee. Informed consent was obtained from all patients and/or parents. In all cases, routine G banding and R banding chromosome analyses showed a normal karyotype. A high-resolution array-comparative genomic hybridisation (array-CGH; Agilent technologies, Palo Alto, California, USA) was also performed in all index cases and excluded any large genomic rearrangements. Blood samples from probands and their parents were obtained and genomic DNA was isolated from EDTA anticoagulated using a Nucleon kit (Amersham, UK) according to the manufacturer's instructions. The NSD1 gene was previously screened in all patients, and no NSD1 mutation was found.
Targeted next-generation sequencing
The coding sequences of 14 H3K27 methylation-related genes (AEBP2, EED, EZH2, EZH1, HDAC2, JARID2, PCL1, PCL2, PCL3, RBBP4, RBBP7, SIRT1, SUZ12, and UTX)15 and 8 H3K36 methylation-related genes (NSD1, NDS2, NSD3, SETD2, SETD3, ASH1L, SETMAR and SMYD2)17 were analysed using a targeted NGS approach. Experiments were performed in the NGS platform of the Cochin hospital, Paris (Assistance Publique—Hôpitaux de Paris, France). The custom primers panel targeting the 22 genes was designed using the AmpliSeq Designer (Lifetechnologies, Saint-Aubin, France). Twenty nanograms of genomic DNA are amplified to generate the library using the Ion AmpliSeq Library Kit V.2.0 (Lifetechnologies). NGS libraries preparation was performed using the Ion AmpliSeq Library Kit V.2.0 (Lifetechnologies) according to the manufacturer’s instructions (Ion AmpliSeq Library Preparation, Publication Part Number MAN0006735, Revision 5.0, July 2013, Lifetechnologies). The amplified libraries were purified using Agencourt AMPure XP beads (Beckman Coulter, Brea, California, USA). Prior to library pooling and sequencing sample preparation, amplified libraries were validated and quantified using the 2100 Bioanalyzer microfluidic platform (Agilent Technologies, Santa Clara, California, USA). Emulsion PCR was performed using the Ion OneTouch Instrument (Lifetechnologies). Enrichment of the template-positive Ion OneTouch 200 ion sphere particles (ISPs, containing clonally amplified DNA) PCR was performed using the Ion OneTouch ES (Lifetechnologies) according to the manufacturer's procedures. An ISP quality control was then performed using a QubitR 2.0 Fluorometer. The ISP Quality Control assay on the Qubit 2.0 Fluorometer labelled the ISPs 200 with two different fluorophores: Alexa Fluor 488 and Alexa Fluor 647. The probe labelled Alexa Fluor 488 annealed to all of the ISPs present while the probe labelled Alexa Fluor 647 to only the ISPs with extended templates. The ratio of the Alexa Fluor 488 fluorescence (all ISPs present) to the Alexa Fluor 647 fluorescence (templated ISPs) yielded the % templated ISPs. The template-positive ISPs were loaded on Ion 316 chips and sequenced with an Ion Personal Genome Machine (PGM) System (Lifetechnologies). Data Ion Torrent reads were collected by the Ion Torrent Suite software V.3.6.2, which also sorted the data according to the barcodes. Data collected on the PGM were collated and reanalysed using the Torrent Suite 3.6.2 using FASTQ files from the Ion Torrent Browser.
Sequence alignment and extraction of single-nucleotide polymorphisms (SNPs) and short insertions/deletions (indels) were performed using the Variant Caller plugin on the Ion Torrent Browser and DNA sequences visualised using the Integrated Genomics Viewer (IGV, V.2.3) from Broad Institute (Cambridge, Massachusetts, USA). The NextGENe software V.2.3.3 (Softgenetics, State College, Pennsylvania, USA) was also used for sequence alignment, extraction of SNPs and short indels, and their visualisation and annotation. In brief, major calling parameters were chosen as follows: minimum allele frequency ≥10% for both SNPs and In/Del (short insertions/deletions), minimum sequencing depth ≥6X for SNPs and ≥15X for In/Del, and minimum sequencing depth on either strand ≥5X for In/Del. A NGS bioinformatics analysis was also performed for single and multiexon deletions/duplications identification. In this method, quantitative values were obtained from the number of reads for each amplicon of each sample, extracted using the Coverage Analysis plugin on the Ion Torrent Browser V.3.6.2 (Lifetechnologies). Read number for each separated amplicon was normalised by dividing each amplicon read number by the total of amplicon read numbers of a control gene from the same sample. Normalised read number obtained for each amplicon of a sample was then divided by the average normalised read number of control samples for the corresponding amplicon. Copy number ratios of <0.7 and >1.3 were considered deleted and duplicated, respectively.
Point mutations detected by targeted NGS were confirmed using Sanger DNA sequencing analysis performed on the corresponding exon only. Mutational screening was performed using bidirectional DNA sequencing of the purified PCR products with the ABI Big Dye terminator sequencing kit (Applied Biosystems) on an ABI Prism 3130 automatic DNA sequencer (Applied Biosystems). Sequences were aligned with Seqscape analysis software V.2.5 (Applied Biosystems). The primer oligonucleotide sequences and PCR conditions are available upon request.
Two heterozygous mutations in the SETD2 gene were identified in two patients with ‘Sotos-like’ syndrome. In proband 1, we identified a heterozygous c.820C>T (p.Gln274*) nonsense variant in the SETD2 gene ((Su(var)3–9, enhancer of zeste (E(z)) and trithorax (trx)) domain-containing protein 2) (MIM 612778) (RefSeq NM_014159.6). We also identified a heterozygous missense variant c.5444T>G (p.Leu1815Trp) in SETD2 in proband 2. None of the two variants were reported in dbSNP,18 the 1000 Genomes Project19 or in the Exome Variant Server. We confirmed the two mutations using Sanger sequencing. Proband 1 carrying a nonsense SETD2 mutation (c.820C>T; p.Gln274*) was adopted and her biological parents could not be tested. We confirmed that the SETD2 variant c.5444T>G (p.Leu1815Trp) found in proband 2 was not found in either of his parents, indicating that it was a de novo mutation (figure 1).
Proband 1 was adopted when she was 4 months old. Her birth weight was 2850 g (25–50th centile), length was 50 cm (50th centile) and occipitofrontal circumference (OFC) was 34.5 cm (25–50th centile). Her development was considered as normal between 4 months and 3 years of age. She walked at the age of 14 months. Speech was slightly delayed. She was first seen at the genetic clinic at 7 years and 8 months of age for speech delay, overgrowth and overweight. Her weight was 75 kg (>97th centile), BMI was 23, height was 137 cm (>97th centile) and OFC was 57 cm (>97th centile). The clinical examination showed prominent forehead with high frontal hairline, downward slanting palpebral fissures, prominent mandible, long and large hands and feet. She had frequent ear and renal infections without any associated malformations. Ophthalmological evaluation, EEG, metabolic screen in blood and urine, endocrinologic screen were all normal. At 7 years and 8 months, CT scan showed mild ventricular dilatation. Her full-scale IQ score was 91, with a verbal IQ of 99 and a performance IQ of 80. Chromosome analysis showed a 46, XX karyotype, and CGH array gave a normal result. At a follow-up evaluation at 12 years and 2 months, weight was 84 kg (>97th centile), height 164 cm (>97th centile) and OFC was 60 cm. At 12 years and 2 months, MRI of the brain disclosed nodular and punctiform hypersignals at the anterior parts of corona radiate and in the centrum semi-ovale. Radiographs of lumbar and sacral vertebral column and pelvis were normal. Hand films revealed major advance in carpal and phalangeal bone age with a bone age of 10 years and 11 months then 16 years and 4 months at chronological ages of 7 years and 8 months and 13 years and 7 months, respectively. She had been ‘slow’ in school with signs of attention deficit and impaired fine motor skills. She also had temper tantrums and could be aggressive during such outbursts. She had received intensive support from her family and then completed basic secondary school. At 17 years, she was referred for obesity, hirsutism and menstrual irregularity. Hormones assay and pelvic ultrasound indicated polycystic ovary syndrome. Her weight was 90.8 kg and height 163.5 cm. She complained of pain and swelling of the wrists, the ankles and the knees. Despite extensive evaluation, the aetiology of those pains could not be identified. At the last evaluation, she was 23 years of age. Her weight was 108 kg (>97th centile), height 164.4 cm, BMI 40.2 and OFC 60 cm (>97th centile). Facial features include a high forehead with bilateral temporal retraction, long face, malar hypoplasia, triangular and massive chin, long nose (table 1 and figure 1). She was expecting a job in a sheltered environment because of her slowness and her lack of autonomy.
Proband 2 was the fourth child of unrelated French parents. Mother's height was 154 cm and father's height was 174 cm. He was born at term, and birth parameters were 51 cm (length), 3600 g (weight) and 36 cm (OFC). He was able to walk at 12 months of age, but speech delay was noted and had school difficulties. He was followed for Hashimoto’s thyroiditis also known in her mother and maternal grandmother. Puberty began at around 12 years, along with an excessive weight gain. He was first seen at the genetic clinic at 14 years of age for developmental delay with stature and weight advance. Weight was 78.5 kg (+4 SD), height 174.5 cm (+2.8 SD) and head circumference 60 cm (+4 SD). Minor facial features were noted including a long face (table 1 and figure 1). He had slurred speech and was slow in speech and gestures. He had an accelerated carpal osseous maturation (15 at 14 years of age). Array CGH analysis was normal as well fragile-X research. At the last evaluation, he was 26 years old. Weight was 94 kg (+4.5 SD), height 177 cm (+0.5 SD) and OFC 62 cm (+4 SD). He had multiple nevi, marked hirsutism in the back, mild facial features (long face, pointed chin), astigmatism and behavioural troubles characterised by shyness and low sociability. He was still extremely slow in the daily life. He was poorly performing at school and had then difficulties to find work.
We report here heterozygous loss-of-function mutations in the SETD2 gene in two patients with a consistent overgrowth phenotype characterised by postnatal overgrowth (+2.5 SD), macrocephaly (+4 SD) and obesity (>+4 SD) in the course of the disease, minor facial features (long face and pointed chin) and advanced carpal ossification. Both presented with speech delay, slowness and low sociability, leading to difficulties in their professional integration.
SETD2 was discovered in 1998 and was shown to function as a histone methyltransferase via the conserved SET domain. The SET domain was first recognised as a conserved feature in chromatin-associated proteins and a number of SET domain-containing proteins have since been characterised as histone methyltransferases.20 ,21 SETD2 is non-redundantly responsible for all trimethylation of lysine 36 of histone H3.22 ,23 Homozygous disruption of Setd2 in mice resulted in embryonic lethality with severe defects in blood vessel development.24 One of the two SETD2 mutations introduces a premature stop codon in SETD2 mRNA (proband 1, c.820C>T, p.Gln274*), and in the other case (proband 2), the mutation could result in a loss-of-function allele (c.5444T>G, p.Leu1815Trp). The affected Leu1815 codon is evolutionarily conserved according to phyloP (score=0.998). A positive phyloP score is interpreted as a signature of evolutionary conservation, which is consistent with functional importance.25 The in silico analysis of the p.Leu1815Trp variant using PolyPhen-2 classified it as probably damaging (score=0.98), and the Sorting Intolerant From Tolerant (SIFT) software26 classified it as damaging (score=0).
Our results illustrate the power of targeted NGS to identify rare disease-causing variants. Remaining mutation-negative patients (14 cases) could present mosaic mutations undetectable in blood. Unfortunately, no additional tissues other than blood could be collected from the affected cases in our study. The remaining 14 negative index cases should now be analysed with a whole exome NGS approach to search for causal mutations in other loci in the genome. This exome sequencing approach has proven to be a useful and relevant method for the identification of disease-causing genes.27
Our data identify heterozygous mutations in SETD2 (located at 3p21.31) in two patients with Sotos-like syndrome. Interestingly, involvement of region 3p21 in the development of a Sotos-like syndrome was suggested nearly 15 years ago. An apparently balanced translocation t(3;6)(p21;p21) was found in a 6-year-old boy with mental retardation, postnatal overgrowth and facial dysmorphism.28 Another report published in 1992 described a non-smoking female with Sotos-like syndrome (including excessive growth during childhood, accelerated osseous maturation, developmental delay, incoordination) who died of small cell lung cancer at the unusually young age of 22.29 Tumour cells showed a loss of heterozygosity of markers at region 3p2l.
Our data provide a compelling argument for Sotos and Sotos-like syndromes as epigenetic diseases caused by loss-of-function mutations of epigenetic writers (methylase) of the H3K36 histone mark. Recently, mutations in the DNA methyltransferase gene DNMT3A (encoding DNA (cytosine-5)-methyltransferase 3A) have also been shown to cause an overgrowth syndrome with intellectual disability.30 Interestingly, DNMT3A has been described to be a H3K36 epigenetic mark reader. Emerging technologies to interrogate the epigenome may demonstrate how H3K36 methylation dysregulation contributes to overgrowth phenotypes. Post-translational modification of histones by methylation is an important and widespread type of chromatin modification that is known to influence biological processes in the context of development and cellular responses.31 Widely described to be associated with active chromatin and continuing transcription,32 ,33 H3K36 methylation has also been implicated in alternative splicing, DNA replication and repair, and DNA methylation.17 ,20 ,34 ,35 The causality of each of these processes in the SoS phenotype remains to be explored. Further functional and mutational analyses will be of interest to extend and illuminate these observations.
The phenotypic similarities between the two syndromes caused by NSD1 and SETD2 alterations may be explained by the common H3K36 writer properties of these two proteins. Several factors probably contribute to how mutations in SETD2 or NSD1 cause different symptoms. In particular, NSD1 has been described to catalyse H3K36 dimethylation,10 ,11 ,12 whereas SETD2 is non-redundantly responsible for H3K36 trimethylation.22 ,23 However, the specific phenotype caused by SETD2 alterations remains to be further characterised in larger cohorts in order to document a well-defined nosologic entity.
Defects in the genes that maintain the levels of H3K36 methylation have also been identified at the somatic level in several cancer types. Each NSD family member behaves as an oncogene in multiple cancers,17 and translocations in NSD1 can lead to the development of acute leukaemia.36 Inactivation of SETD2 was found to be a common event in clear cell renal cell carcinoma with loss or decrease of H3K36me3 mark.37 SETD2 mutations were also described in high-grade gliomas and synovial sarcomas.38 ,39 Moreover, downregulation of SETD2 at transcriptional and protein levels was observed in breast cancer.40 A somatic variant p.Leu1815Phe in SETD2 has recently been described in lung adenocarcinoma and affects the same codon than the constitutional variant p.Leu1815Trp found in probant 2 in the present study.41 Data from long-term follow-up of individuals with Sotos syndrome (including SETD2 mutated patients) will precise the potential neoplastic complications of this rare disorder. The accumulating data on epigenetic abnormalities in cancer have led to the emerging realisation that many of the mediators of acetyl and methyl histone are susceptible to inhibition by small molecules.42 The hope is that these advances in oncology could benefit to a rare disease like SoS.
1000 Genomes Project, http://www.1000genomes.org/
NHLBI Exome Sequencing Project (ESP) Exome Variant Server, http://evs.gs.washington.edu/EVS/
Online Mendelian Inheritance in man (OMIM), http://www.omim.org
SWISS-Model Server, http://swissmodel.expasy.org/
UCSC Genome Browser, http://genome.ucsc.edu
The authors gratefully acknowledge the generosity of the families in providing samples and clinical details for this study.
EP and VC-D contributed equally.
Contributors All authors of this manuscript fulfil the criteria of authorship. AL and IL performed sequencing. Clinical studies are from VC-D, VM, CF, SO, FG, DL and RT. VC-D, MV and EP designed the study and wrote the manuscript.
Competing interests None.
Patient consent Obtained.
Ethics approval Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
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.