Background Moyamoya angiopathy (MMA) is characterised by a progressive stenosis of the terminal part of the internal carotid arteries and the development of abnormal collateral deep vessels. Its pathophysiology is unknown. MMA can be the sole manifestation of the disease (moyamoya disease) or be associated with various conditions (moyamoya syndrome) including some Mendelian diseases. We aimed to investigate the genetic basis of moyamoya using a whole exome sequencing (WES) approach conducted in sporadic cases without any overt symptom suggestive of a known Mendelian moyamoya syndrome.
Methods A WES was performed in four unrelated early-onset moyamoya sporadic cases and their parents (trios). Exome data were analysed under dominant de novo, autosomal recessive and X-linked hypotheses. A panel of 17 additional sporadic cases with early-onset moyamoya was available for mutation recurrence analysis.
Results We identified two germline de novo mutations in CBL in two out of the four trio probands, two girls presenting with an infancy-onset severe MMA. Both mutations were predicted to alter the ubiquitin ligase activity of the CBL protein that acts as a negative regulator of the RAS pathway. These two germline CBL mutations have previously been described in association with a developmental Noonan-like syndrome and susceptibility to juvenile myelomonocytic leukaemia (JMML). Notably, the two mutated girls never developed JMML and presented only subtle signs of RASopathy that did not lead to evoke this diagnosis during follow-up.
Conclusions These data suggest that CBL gene screening should be considered in early-onset moyamoya, even in the absence of obvious signs of RASopathy.
- CBL E3 Ubiquition Liqase
Statistics from Altmetric.com
Moyamoya angiopathy (MMA) is characterised by a progressive stenosis of the intracranial internal carotid arteries (ICA) and their bifurcation proximal branches, associated with the development of thin collateral vessels at the base of the brain.1 2 This is a rare condition, found in all ethnicities, but with a marked East-West gradient. In Japan, the prevalence is about 6–10/100 000.3 4 In Europe and USA, prevalence of moyamoya is estimated to be 10–20 times lower than in Japan.5 6 MMA leads to ischaemic and haemorrhagic stroke in children and adults, with an age of onset that follows a bimodal distribution, namely a first peak at about 5 years of age and a second peak in the fourth decade. Its pathophysiology is so far unknown. The term moyamoya disease (MMD) refers to isolated idiopathic MMA. Genetic factors have been suspected in MMD based on a strong ethnicity-related effect and a familial case prevalence close to 10%. To date, RNF213 gene is the sole susceptibility gene known in MMD, essentially in East Asian countries, where about 80% of patients with MMD carry the p.R4810K founder variant.7 8 However, RNF213 gene should be considered as a susceptibility gene rather than a Mendelian causative gene, given the very low penetrance of MMD in mutated patients. Indeed, in East Asian countries, only 1 out of 150 p.R4810K carriers develops MMD, and a large majority of the p.R4810K-mutated patients have inherited the variant from an unaffected parent.4 9 In contrast to MMD, the term moyamoya syndrome (MMS) refers to MMA associated with other neurological or extra-neurological manifestations. Several distinct Mendelian diseases are known to be responsible for MMS, most often with a low penetrance of the angiopathy, reflecting the genetic heterogeneity of moyamoya.10 Currently, the screening of the genes involved in this MMS is only performed in patients presenting suggestive symptoms of these Mendelian diseases.
In this study, we sought to investigate Mendelian causes of moyamoya using the powerful trio whole-exome sequencing (WES) approach in early-onset sporadic cases who did not present any overt systemic symptom suggestive of a specific aetiology.
We included a panel of 21 consecutive early-onset sporadic moyamoya cases of unknown aetiology, referred to the Department of Genetics at Lariboisière Hospital, Paris, to the French National Center for Pediatric Stroke, to the French National Center for Rare Vascular Diseases of the Brain and the Eye (CERVCO), and to the Department of Neurology at the Alfried Krupp Hospital, Essen. An early onset was defined by the occurrence of the first moyamoya symptoms before 15 years of age. DNA from both parents was available for four probands. Therefore, a total of four trios and 17 additional early-onset sporadic cases were included in the study. For each case, the diagnosis of moyamoya has been established according to the Research Committee on Spontaneous Occlusion of the Circle of Willis published guidelines.11 Detailed clinical interview and physical examination looking for features suggestive of known MMS was performed in each one as well as cerebral MRI; imaging of cervical, cerebral, renal and aortic arteries; blood pressure measurement; echocardiography and blood analysis including blood count; assessment of liver and kidney functions; coagulation tests; and haemoglobin electrophoresis. All parents provided their written informed consent for genetic analysis. Parents of proband 1 provided their written consent for clinical photographs.
Genomic DNA was isolated from peripheral blood leucocytes using the Wizard Genomic DNA Purification Kit. In addition, DNA was extracted from hair bulbs and buccal swabs of trio 1 proband using QIAamp DNA Investigator Kit and QIAamp DNA Blood Mini Kit (QIAGEN), respectively. WES was performed at IntegraGen platform (Evry, France) in the 21 probands and the two parents of each trio. Exomes were captured using the SureSelect Human All Exon V5 (trio 1 and 1 singleton) and V5 Clinical Research Exome kits (trios 2–4 and 16 singletons) of Agilent and were sequenced with an Illumina HiSeq2000 (paired end, 75 bp reads). Mapping and variant calling were performed with the CASAVA pipeline provided by Illumina. Reads were mapped to the GRCh37 build using ELAND software. Single-nucleotide substitutions and small insertion-deletion (indel) variants were annotated with an Integragen bioinformatics pipeline. Individual mean sequencing depth ranged from 62 to 113 reads.
In a first step, GUCY1A3, ACTA2 and BRCC3, three genes previously known to be involved in Mendelian MMS, were screened in all probands using WES data to exclude causative variants in those genes. Based on the sporadic nature of the disease in the four trios, we then interrogated the exome data under the assumption that the causative allele was dominant and had arisen de novo. We also performed WES analysis according to an autosomal recessive (AR) pattern of inheritance (homozygous or compound heterozygous variants) and an X-linked recessive pattern of inheritance in male probands. To limit the number of false positives, we used stringent quality criteria and restricted our analysis to variants with a high quality score (Q Phred score ≥30) and a coverage ≥20 X in probands (in all members of a family trio for de novo analysis) and excluded variants located in segmental duplications. A variant was considered de novo if the proband was heterozygous for a variant with a number of mutated reads ≥20% of the total number of reads with both parents homozygous for the reference allele (with a number of reference reads ≥95% of the total number of reads). We restricted our analysis to nonsense, missense substitutions, mutations in canonical splice sites and indel variants located in coding regions. The public databases dbSNP 144, 1000Genomes Phase 3, Exome Sequencing Project (ESP6500SIV2), ExAC (0.3) and ClinVar were screened for each coding variant detected. Variants with a minor allelic frequency (MAF) exceeding 0.01 in public databases were considered as polymorphisms and were excluded. Considering the rarity of the affection, we excluded X-chromosome variants with a MAF exceeding 10−4. Pathogenicity of candidate variants was predicted using Mutation Taster, SIFT and Polyphen 2 (HumVar) software. Missense variants were excluded if they were predicted benign by more than one out of these three software. Combined Annotation Dependent Depletion pathogenicity scores were also calculated for each variant. Standard PCR amplification and Sanger sequencing were performed in all members of a given trio for each candidate variant identified.
Mutations in genes showing candidate variants in the four trios were then searched for in the remaining 17 early-onset sporadic moyamoya cases using WES data of these patients. We extracted all nonsense, missense, splice-site and indel variants located in these candidate genes and excluded variants with a MAF ≥0.01 in the databases dbSNP, 1000Genomes, Exome Sequencing Project and ExAC databases as well as variants of poor quality (Q Phred score <30 and/or coverage <8X).
General characteristics of the four trios and of the 17 early-onset MMA singletons
Pedigrees of the four trios are shown in figure 1A. Age at moyamoya diagnosis of the four trio probands ranged from 21 months to 6.5 years. Parents of trio 1 proband were of Caucasian ancestry and parents of trio 2–4 probands originated from North Africa and Middle East. In the panel of 17 singletons, age of onset ranged from 1 year to 14 years, and these patients were Caucasian (n=13), North African (n=2), Korean (n=1) or Caribbean (n=1). Screening for currently known causes of acquired or inherited MMS was negative in the 21 probands. Causative variants in GUCY1A3, ACTA2 and BRCC3 genes, previously known to be involved in distinct Mendelian MMS, have been ruled out in the 21 probands. The entire coding DNA sequence of these genes was covered at ≥8X with both Agilent capture kits, except for the sixth exon of BRCC3, which was Sanger sequenced in all included male probands. Mean depth of coverage was respectively 145X for GUCY1A3, 135X for ACTA2 and 31X in male probands for BRCC3.
Identification of de novo CBL mutations in two trios
A total of four de novo coding variants located in three distinct genes (three missense and one splice-site variants) were found in three out of the four trio probands and were confirmed by Sanger sequencing (table 1). CBL was the sole gene showing a pathogenic de novo mutation in more than one trio proband (figure 1 and table 1). The heterozygous NM_005188.3:c.1111T>A substitution (rs267606706) leading to a p.Tyr371Asn substitution was found in trio 1 proband. Sanger sequencing of DNA extracted from hair bulbs and buccal swabs of trio 1 proband confirmed the germline nature of this mutation. The heterozygous NM_005188.3:c.1228-2A>G mutation (rs727504426), a canonical splice-site mutation which leads to a truncation of the enzymatic domain of the protein,12 was identified in trio 3 proband. No additional DNA source was available in trio 3 proband. Both variants were referred as variants causing a RASopathy in ClinVar database. A thorough analysis of exome sequencing data and additional Sanger sequencing of the first exon of CBL, which was insufficiently covered by WES, failed to identify any additional CBL mutation in the probands of trios 2 and 4. Analysis of WES data of the 17 additional sporadic patients with early-onset moyamoya did not show any candidate variant either in CBL or in the two additional genes showing de novo mutations in the trio analysis, namely SCEL and ZNF418.
Analysis of WES data according to the AR mode of inheritance did not identify any candidate gene in trios 1 and 3. Two candidate genes were detected in trios 2 and 4, FREM2 and SVEP1 (table 1). Importantly, none of these two candidate genes were located in the linkage regions previously identified in MMD (3p24-p26, 6q25.2, 8q23 and 12p12).13–15 Analysis of WES data according to the X-linked recessive mode identified two candidate hemizygous variants located in GLUD2 and ITGB1BP2 in trios 2 and 4 probands, both inherited from their unaffected mothers (table 1). Analysis of WES data of the 17 additional sporadic patients with early-onset moyamoya did not show any candidate variant in those four genes. Thus, CBL appears to be the sole gene showing recurrent candidate mutations in our study, both of them being a de novo mutation.
Clinical description of CBL-mutated probands
Trio 1 proband was born at term in 2003, following an uneventful pregnancy and a normal delivery. She was the second child of healthy unrelated Caucasian parents aged 34 years (father) and 38 years (mother) at the time of birth. Birth weight (BW) was 3770 g (+1 SD), birth length (BL) 49 cm (mean) and occipitofrontal circumference (OFC) 33.5 cm (−1 SD). Following a normal neonatal adaptation, she presented sucking difficulties, failure to thrive and hypotonia during infancy. Frequent airway infections were attributed to gastro-oesophageal reflux. After diet diversification, she presented a severe constipation requiring chronic use of laxatives. She presented mild motor and speech delay (walked independently at 26 months, vocabulary limited to 20 words at age 3 without word association). At 3 years of age, she presented a right hemiplegia, and MRI revealed both acute cerebral infarction and prior silent infarcts in the left middle cerebral artery (MCA) territory. Three months later, she developed a contralateral hemiplegia due to an acute ischaemic stroke in the right MCA territory. Conventional arteriography showed typical bilateral MMA consisting of a bilateral high-grade short stenosis of the supraclinoid ICAs, bilateral occlusion of proximal segments of the MCAs, and anterior cerebral arteries (ACAs) and bilateral basal collateral moyamoya network. Basilar artery and posterior cerebral arteries (PCAs) were normal and gave rise to leptomeningeal arterial-arterial anastomoses (figure 2). Global assessment included normal skin examination, and no dysmorphic features were noted at examination of face, neck, thorax and extremities. Revaluation after CBL mutation identification led to identify mild facial dysmorphisms: anteverted ear lobes, posteriorly rotated ears and hypertrophic lips (figure 3). Blood pressure measurement and echocardiography were normal. Renal arteries and aorta imaging did not show any other vascular abnormality, and funduscopy was normal. Repeated biological assessment showed prolonged prothrombin time (international normalised ratio 1.4) and elevated kaolin clotting time ratio (KCT ratio 1.6) without detected antiphospholipid (APL) antibodies. Antithrombotic treatment by aspirin was started and bilateral indirect revascularisation surgery with multiple burr holes was performed 2 months after the second symptomatic stroke. She did not experience further clinical or radiological stroke recurrence, with a 10-year follow-up. She had persistent oral dyspraxia, right hemiparesis and strabismus. Gastro-oesophageal reflux spontaneously resolved before the age of 2, and constipation normalised at 10 years of age. Coagulation assay normalised at 10 years of age. Leucocyte, red blood cell and platelet counts were normal during further follow-up, and neither hepatomegaly nor splenomegaly was observed. Spontaneous menarche occurred at 12 years and 6 months. At last evaluation, at 13 years of age, height was 154.5 cm (mean), weight 39.5 kg (−1 SD) and OFC 52.5 cm (mean). Mild psychomotor delay and dysexecutive syndrome were noted. She could read but had difficulties in writing. She followed an adapted academic course.
Trio 3 proband was born at term in 2003, following uneventful pregnancy and normal delivery. She was the first child of unrelated parents, with a 29-year-old mother of Moroccan origin and a 37-year-old father of Lebanese origin. Birth parameters were in the normal range (BW 3100 g (−0.5 SD), BL 49.5 cm (mean) and OFC 34 cm (mean)) as well as neonatal adaptation. During infancy, swallowing dysfunction and constipation requiring the regular use of laxatives were reported. She had normal psychomotor development and was able to walk independently at 12 months, but parents reported unstable gait. She started drooling at 18 months. At 21 months of age, she presented with focal status epilepticus and left hemiplegia revealing an acute infarction in the right MCA territory. A second stroke occurred 1 month later in the contralateral MCA territory, and MRI revealed several and bilateral silent infarcts in the MCA and ACA territories. Conventional arteriography showed bilateral MMA with a bilateral high-grade short stenosis of the supraclinoid ICAs and a bilateral occlusion of the proximal segments of the MCAs and ACAs. Bilateral collateral moyamoya network was of unusual localisation, arising from the ACAs and posterior territory and not from M1 segment of the MCAs as expected. Both PCAs were stenotic and gave rise to leptomeningeal arterio-arterial anastomoses. V4 segment of the left vertebral artery and the first segment of the basilar artery were dilated. Distal arteries were dysplastic (figure 4). Global assessment revealed mild hypertension, requiring antihypertensive monotherapy. Skin examination revealed four café-au-lait spots (three on the belly, one on the knee), and no dysmorphic feature suggestive of Noonan syndrome was noted at examination of face, neck, thorax and extremities. Transthoracic echocardiography and renal artery ultrasound were normal. Biological assessment revealed an APL syndrome, with persistent elevated anticardiolipin and anti-β2GP1 antibodies (respectively 58 units of IgG APL (UGPL), n<10 and 21 U/mL, n<10) and elevated KCT ratio (1.42). Lupus anticoagulant testing showed ambiguous results, with positive Rosner index and negative dRVVT ratio. She had no antinuclear antibodies on Farr testing. Antithrombotic treatment by aspirin was started and bilateral indirect revascularisation surgery with multiple burr holes was performed 1 month after the second symptomatic stroke. She did not experience further clinical or radiological stroke after 11 years of follow-up. Owing to her strokes, she kept mild quadriparesis, severe oral dyspraxia with major sialorrhea and cognitive impairment with attention deficit. She developed symptomatic focal epilepsy, requiring antiepileptic monotherapy until 9 years of age. Constipation and swallowing difficulties normalised during childhood. Blood pressure normalised and antihypertensive treatment could be stopped at 10 years of age. Leucocyte, red blood cell and platelet counts remained normal during follow-up, and she did not show hepatomegaly or splenomegaly. She had normal puberty onset (menarche at 12 years). At last evaluation, at 13 years of age, height was 160 cm (+2 SD), weight 46 kg (+1 SD) and OFC 55 cm (+1.5 SD). She was able to read a few words, had difficulties in writing and followed adapted academic course.
Our trio study led to the identification of de novo germline heterozygous CBL mutations in two unrelated cases presenting a bilateral severe early-onset MMA. The CBL gene encodes a RING finger E3 ubiquitin ligase that acts as a negative regulator of RAS signalling, a pathway involved in a large number of cellular processes such as cell survival, cell proliferation and development.16 The CBL protein controls proliferative signals via the ubiquitination of activated growth factor receptor tyrosine kinases, endocytosis and degradation of these receptors. In addition, CBL is a multiadaptor protein that controls intracellular trafficking. CBL protein contains a tyrosine kinase binding domain and a RING finger domain that mediates the E3 ubiquitin ligase activity. These two domains are separated by a linker sequence which is crucial for the ubiquitin ligase activity of the protein (figure 1B).17 Both CBL mutations observed in trio 1 and 3 probands affect the critical domains of the protein. The Tyr371 mutated in trio 1 proband is located in the linker domain. Previous functional studies showed that this highly conserved residue is required for the ubiquitin ligase activity of the protein through the maintenance of the tridimensional conformation of the linker domain.12 18 19 The c.1228-2A>G splice-site mutation found in trio 3 proband has been previously demonstrated to lead to a partial or complete exon 9 deletion.12 Thus, the mutant protein is predicted to lack the RING finger domain that mediates the E3 ubiquitin ligase activity.
Some heterozygous CBL germline mutations, including the two reported in this study, are known to cause a RASopathy within the Noonan syndrome phenotypic spectrum. This ‘Noonan-like syndrome’ related to CBL mutations is characterised by developmental anomalies, including dysmorphic features, and by an increased risk to develop myeloproliferative/myelodysplastic disorders, the most notable being juvenile myelomonocytic leukaemia (JMML).12 20 JMML has an early-infancy onset and a usually aggressive course. It is initiated by mutations in RAS genes (NRAS or KRAS) or RAS pathway regulators, including CBL in around 10% of cases,21–23 although only a small proportion of CBL mutation carriers develop JMML.24 In the vast majority of patients with JMML carrying a heterozygous germline CBL mutation, a second hit occurs in myeloid cells leading to the loss of the wild-type allele. In most cases, the loss of this wild-type CBL allele is the result of an acquired isodisomy.12 On the other side, the developmental anomalies described in CBL mutation carriers are characterised by impaired growth, developmental delay, cryptorchidism, cutaneous abnormalities and facial dysmorphic features.24 25 However, clinical manifestations are frequently mild and some asymptomatic carriers have been reported in some families.12 20 25 Pathogenic CBL mutations causing RASopathy usually affect the linker or the RING finger domain and lead to loss of ubiquitination of target proteins and upregulation of the RAS signalling pathway, thus conferring proliferative properties to cells.12 25 Of note, Tyr371, mutated in trio 1 proband, is one of the residues most commonly involved in CBL pathogenic mutations but the specific Tyr371Asn substitution identified in trio 1 proband appears to be rare since it was previously reported in only one patient with JMML and was associated to mild dysmorphic phenotype in this latter. The c.1228-2A>G mutation found in trio 3 proband was also previously reported in patients with JMML in whom developmental abnormalities are inconstant.12 25
An extensive review of the literature on CBL mutation carriers allowed us to identify only two patients with MMA. There were two major differences between these previously published patients and ours. First, they were both referred for a haematological condition strongly suggestive of a CBL mutation. One of them had a JMML, in addition to dysmorphic features suggestive of Noonan syndrome.26 The second one was referred for the association of splenomegaly, mild monocytosis and juvenile xanthogranuloma.22 Moreover, both of them, in addition to carry a germline heterozygous CBL mutation, had lost the CBL wild-type allele in haematopoietic cells in contrast to the two young girls reported herein, in which mutations were found in a heterozygous state in peripheral blood leucocytes. In the two young girls reported herein, the diagnosis of RASopathy was never evoked until the identification of these CBL mutations by exome sequencing, despite a regular follow-up in a specialised neuropaediatric unit. Importantly, they did not present any sign of haematological malignancy throughout a prolonged follow-up. In retrospect, some subtle signs could have been considered as part of ‘CBL syndrome’ in them. However, in the context of infant stroke, the distinction between some of the developmental signs of a RASopathy and neurological sequelae related to stroke appears quite difficult. In both cases, sucking and swallowing difficulties, frequently observed in context of orofacial dyspraxia following stroke involving insular regions, hypotonia, mild motor and speech delay, were attributed to ancient strokes in MCA territories that were seen on initial MRI. Other signs such as transitory coagulation defects without bleeding tendency in trio 1 proband—coagulation defects were reported so far in one CBL-mutated patient27—and café-au-lait spots in trio 3 proband were subtle and did not lead neuropaediatricians to consider a diagnosis of RASopathy. Indeed, no ‘red flag’ suggestive of a RASopathy has been identified by neuropaediatricians in both girls, in whom MMA was the hallmark manifestation at diagnosis. Thus, some CBL mutations leading to MMA as a main presentation are most likely overlooked in a neuropaediatric context.
These data suggest that CBL screening should be considered by neuropaediatricians in patients with early-onset MMA even in the absence of obvious developmental abnormalities, or malignant haematological manifestations suggestive of a RASopathy. Minor signs such as coagulation defects without bleeding tendency, café-au-lait spots, or developmental delay, even mild, are of great value and should particularly incite to search for a CBL mutation. Importantly, similarities were observed in the clinical presentation of these two cases, which may also help to orient diagnosis. Both presented a very-early-onset MMA, with a first symptomatic stroke before the age of 3 years, and they already had silent infarcts on initial imaging, suggesting an even earlier onset of the disease. The initial course of the disease was very severe with several symptomatic bilateral strokes recurring within very short delay and prompting rapid revascularisation surgery decision. Intriguingly, they also both presented unexplained early-onset digestive symptoms with severe but spontaneously resolving constipation.
The identification of these two de novo mutations in CBL, a major RAS pathway gene, strongly reinforces the involvement of this pathway in moyamoya pathophysiology. Indeed, neurofibromatosis type 1 and Noonan syndrome are classical causes of MMS. In addition, MMA has recently been reported in two patients presenting with a Noonan-like syndrome with loose anagen hair related to heterozygous mutations in the SHOC2 gene, which also belongs to the RAS pathway.28 Altogether, these data lead to consider the RAS pathway as a key pathway in the moyamoya pathophysiology. The pathophysiological links between the RAS pathway and some of the anomalies observed in CBL mutants, such as the presence of autoantibodies, and the role of these anomalies in the pathophysiology of MMA are however unclear. APL antibodies have been detected in proband 3. Propensity of RASopathy patients to present APL has been reported (although never reported in CBL-mutated patients to our knowledge).29 APL syndrome is commonly associated with thrombotic events and MMA.30 However, the development of this angiopathy cannot be attributed to the sole presence of APL antibodies. Indeed, APL antibodies have been reported only in a few Noonan and Noonan-related patients presenting with MMA.31 In the present study, APL have been detected only in one out of the two CBL-mutated probands, despite repeated screening. This does not exclude however the role of APL antibodies as a cofactor for the development of MMA.
In addition to CBL de novo mutations identification, our study led to identification of six additional candidate genes in trio 2 and 4 probands (table 1). Among them, ZNF418 and SVEP1, both mutated in trio 4 proband, are of particular interest based on available functional data. The ZNF418 gene is a transcriptional repressor acting on MAPK signalling pathway, and thus might exert a regulatory effect on RAS-MAPK pathway.32 The de novo mutation identified in trio 4 proband affects a DNA-binding C2H2 Zing finger domain of the protein and is predicted to be damaging. Trio 4 proband also carries biallelic mutations in SVEP1. SVEP1 encodes for a multidomain protein of extracellular matrix involved in cell adhesion in an integrin α9β1-dependant manner.33 It has been shown that SVEP1 could interfere with expression of adhesion molecules at the surface of HUVEC endothelial cells.34 A recent Genome Wide Association Study showed that a low-frequency missense variant of SVEP1 was associated with an increased risk of coronary disease and a higher blood pressure, suggesting that variants in this gene could have an impact on the vasculature.35 The two SVEP1 variants carried by trio 4 proband are both predicted to be of very high impact. However, exome analysis of the 17 additional early-onset sporadic MMA singletons failed to identify any recurrent mutation in these six additional candidate genes. The search for a recurrent mutation in either of these six genes would need to be carried out on a larger MMA cohort. Interestingly, in trio 2 proband, a rare coding variant was identified in a heterozygous state in the RNF213 susceptibility gene (data not shown). This variant was inherited from the unaffected father. Of note, variants of this susceptibility gene were also identified in seven out of the 17 singletons, each one of the mutated proband carrying a distinct variant (data not shown). However, further studies on large cohorts of patients and controls using gene-based burden tests would be needed for the interpretation of the possible pathogenic significance of RNF213 rare variants since rare variants in this gene are also present in controls.
In summary, this trio sequencing approach led us to identify de novo germline heterozygous CBL mutations in two unrelated cases presenting a bilateral severe early-onset MMA but no myeloproliferative/myelodysplastic syndrome, and only minor signs of RASopathy that did not lead to evoke this diagnosis despite a regular neuropaediatric follow-up in a university hospital. To our knowledge, they are the first cases of CBL causative mutations in early-onset MMA without myeloproliferative-myelodysplastic manifestations. Both mutations are predicted to alter the ubiquitin ligase activity of the protein. These data raise the question of a CBL screening in early-onset MMA of unknown aetiology, even in the absence of haematological malignancy and of obvious developmental abnormalities. The identification of a pathogenic CBL mutation in this context also raises questions regarding the haematological follow-up to be proposed to such patients. Additional work will be needed in the future to resolve this question.
The URL for data presented in this paper are the following:
The authors express their gratitude to the family members for participating to this study. They also thank Prof Christine Chomienne for fruitful discussions regarding haematological manifestations of CBL mutations, Thibault Coste, Mickaelle Corpechot, Jessica Hadjadj and Aurore Beauger for their help in molecular analyses and sample preparation. They also thank Florence Marchelli for her help in figure design.
Contributors SG, ET-L: study design. MK, FBrunelle, JA, SL, BD, CB, JCS, MK, DH: clinical investigation and sample collection. SG, LG, FBergametti, AD, MA: experimental work, preparation of samples and direct sequencing. SG, LG, ET-L: exome data interpretation and analysis. SG, MK, ET-L: first draft of the manuscript.
Funding The authors acknowledge support from INSERM funding to U1161 and from Fondation pourla Recherche Médicale (SG, PhD student grant).
Competing interests None declared.
Patient consent 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.