Background: Oesophageal atresia (OA) and mandibulofacial dysostosis (MFD) are two congenital malformations for which the molecular bases of syndromic forms are being identified at a rapid rate. In particular, the EFTUD2 gene encoding a protein of the spliceosome complex has been found mutated in patients with MFD and microcephaly (MIM610536). Until now, no syndrome featuring both MFD and OA has been clearly delineated.
Results: We report on 10 cases presenting with MFD, eight of whom had OA, either due to de novo 17q21.31 deletions encompassing EFTUD2 and neighbouring genes or de novo heterozygous EFTUD2 loss-of-function mutations. No EFTUD2 deletions or mutations were found in a series of patients with isolated OA or isolated oculoauriculovertebral spectrum (OAVS).
Conclusions: These data exclude a contiguous gene syndrome for the association of MFD and OA, broaden the spectrum of clinical features ascribed to EFTUD2 haploinsufficiency, define a novel syndromic OA entity, and emphasise the necessity of mRNA maturation through the spliceosome complex for global growth and within specific regions of the embryo during development. Importantly, the majority of patients reported here with EFTUD2 lesions were previously diagnosed with Feingold or CHARGE syndromes or presented with OAVS plus OA, highlighting the variability of expression and the wide range of differential diagnoses.
- Mandibulofacial dysostosis
- esophageal atresia
- 17q21.31 deletion
Statistics from Altmetric.com
Oesophageal atresia (EA) with or without tracheoesophageal fistula (TEF) is the consequence of an abnormal septation of the foregut into oesophagus and trachea.1 EA incidence is estimated at about 1/3500 live births. The molecular bases for isolated EA are currently unknown and probably multifactorial. The known molecular bases for syndromic EA are heterozygous loss-of-function mutations or deletions of the SOX2, CHD7, MYCN and MIR17-92 genes in AEG, coloboma, heart malformation, atresia of the choanae, retardation of growth and/or development, genital anomalies and ear anomalies (CHARGE) and Feingold syndromes, respectively.2 ,3 Although far less frequent and specific than laryngeal cleft, TEF has also been reported in Opitz GBBB syndrome associated with a MID1 gene mutation.4
Mandibulofacial dysostosis (MFD) is the consequence of an abnormal development of the first and second branchial arches.5 The core phenotype is represented by malar and mandibular hypoplasia and dysplastic ears. Conductive hearing loss, lower lid anomalies and/or cleft palate are frequent associated features. Several genes have been identified as disease-causing in syndromic MFD, including TCOF1 (MIM 606847), POLR1D (MIM 613715) and POLR1C (MIM 610060) in Treacher-Collins (MIM 154500); DHODH (MIM126064) in Miller (MIM 263750); SF3B4 (MIM 605593) in Nager (MIM 154400); and ORC1, ORC4, ORC6, CDT1 and CDC6 genes (MIM 601902, 603056, 607213, 605525 and 602627) in Meier-Gorlin (MIM 224690) syndromes. Recently, loss-of-function mutations of the EFTUD2 gene on 17q21.31 have been reported in a syndromic form of MFD with microcephaly and intellectual deficiency (MIM 610536).6
Until now, no syndrome featuring both MFD and EA has been clearly delineated, although EA is reported in about 5% of patients diagnosed with oculoauriculovertebral spectrum (OAVS; also known as hemifacial microsomia or Goldenhar syndrome, MIM 164210).7 OAVS is phenotypically heterogeneous, but frequently includes dysplasia of the external ears, preauricular tags, epibulbar dermoids, upper eyelid coloboma, macrostomia with lateral cleft and vertebral anomalies. Facial features are more often unilateral and asymmetric when bilateral. Most cases of OAVS are sporadic, and although there is no consistent molecular basis, a number of chromosomal aberrations have been associated with OAVS (reviewed in Rooryck et al8). Importantly, one or more genes for EA possibly map to 17q21.3q24 (an interval harbouring EFTUD2) according to a small number of patients presenting multiple congenital anomalies-mental retardation with EA and de novo interstitial deletion at 17q (reviewed in Puusepp et al9).
We identified a 17q21.31 microdeletion by array-CGH in two unrelated patients presenting MFD and EA. Both deletions encompassed four genes, including EFTUD2. This prompted us to hypothesise either a contiguous gene syndrome or that a broader spectrum of congenital malformations could be ascribed to EFTUD2 loss-of-function mutations, given that EA was not previously reported in association with EFTUD2 lesions. We therefore studied the locus for deletions and/or mutations in patients classified as OAVS with or without EA and/or microcephaly, and in patients referred for molecular testing for possible Feingold or CHARGE syndromes and with no mutation identified in MYCN, MIR17-92 or CHD7 genes, respectively. We identified heterozygous mutations in EFTUD2 in eight cases including a familial case with vertical transmission, leading us to expand the clinical spectrum of malformations ascribed to EFTUD2 loss-of-function.
Material and methods
Detection of gene copy number was performed by array comparative genomic hybridisation (CGH) following the manufacturer's recommendations (Agilent, Agilent Technologies, Santa Clara, Ca) using 44 000 oligo probes approximately spaced at 35–40 kb intervals across the genome (Human Genome CGH microarray 44 K kit, Agilent). Commercial (Promega) or non-commercial female genomic DNA was used as reference in hybridisations which were extracted with Feature extraction software and analysed with the DNA-analytics software by applying an aberration detection method (ADM) 2 segmentation algorithm to identify chromosome aberrations. Copy-number gains and losses were determined using a threshold of 0.3 and −0.3, respectively. Aberrant signals obtained with three or more neighbouring oligonucleotides were considered indicative of genomic aberrations. For the Affymetrix GeneChip SNP6.0 genotyping array whole blood genomic DNA was processed as recommended by the manufacturer (Affymetrix).
The relative DNA copy numbers at the SNP/CNV loci were determined by comparison of the normalised array signal intensity data for the proband's DNA sample against the HapMap270 reference file provided by the manufacturer. Genotyping Console (Affymetrix) and Nexus Copy number v6.1 (BioDiscovery) were used to display and analyse the relative DNA copy number data.
All 27 EFTUD2 coding exons and their flanking intronic sequences (using NM_004247.3 and NG_032674.1 as the mRNA and genomic sequences of reference) were amplified by PCR using the FastStart Taq DNA Polymerase (Roche) in a total volume of 25 µl with the following conditions: 95°C for 4 min, 35 cycles of 95°C for 35 s/ 60°C for 30 s/ 72°C for 1 min, and 72°C for 7 min. Automatic sequencing was performed on a 3130 Genetic Analyser (Applied Biosystems) or a Ceq2000/8000 DNA sequencer (CEQ DTCS-Quick Start Kit, Beckman Coulter, Fullerton, CA, USA). All EFTUD2 primer sequences as well as those used for sequencing of CCDC103, FAM187A, CHD7, MYCN and MIR17-92 are available on request.
In situ hybridisation
A 532 bp fragment of Eftud2, spanning exons 22–26, was amplified from mouse retina cDNA, using the primers 5′-GAAAAGGCCTGTCCCTGGGTGTG (forward) and 5′-CAAAGTACGACTGGGATCTGCTG (reverse), and cloned into pCRII-TOPO (Invitrogen). Following HindIII or XhoI digestion, digoxigenin (Roche)-labelled antisense or sense probes were generated using T7 or SP6 polymerases, respectively. In situ hybridisation of probes to cryosections of E11.5 mouse embryos was performed according to standard protocols, using an alkaline phosphatase-conjugated α-digoxigenin antibody (Roche) followed by addition of nitro blue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl-phosphate (BCIP) (Roche) to visualise probe hybridisation.
We studied three groups of patients: (i) 17 cases with isolated EA (types A and C and unknown in 3, 7 and 7 cases, respectively); (ii) 19 OAVS cases presenting at least two of the following anomalies: facial asymmetry, ear dysplasia, preauricular tags, deafness, epibulbar dermoids, vertebral anomalies, with a cardiopathy in some cases but none presenting with microcephaly or EA; and (iii) 14 patients with MFD and EA and/or microcephaly with or without additional features. For this third group, the initial diagnoses were OAVS plus EA for three cases, VATER association for one case, CHARGE syndrome for seven cases and Feingold syndrome for three cases (table 2).
EFTUD2 sequencing revealed no mutation in patient groups i and ii. In the third group of patients, we identified 10 pathogenic EFTUD2 mutations or deletions and two novel variants of unclear significance. Of the 10 cases with pathogenic mutations, eight presented with EA as one component of the phenotype. The clinical presentations of the patients in whom EFTUD2 alterations were identified are summarised in table 2 and detailed below.
Case 1 is a 7-year-old girl referred to the genetics clinics at birth with a diagnosis of Goldenhar syndrome (ie, OAVS) and EA, the latter being surgically corrected soon after birth (figure 1A, table 2). She walked independently at 19 months and spoke her first words at 3 years of age. A bone-anchored implant was placed at 6 years of age. When reviewed at the age of 8 years, height and weight were on+2 SD and head circumference (HC) on -1 SD, and she was speaking in sentences and about to enter elementary school.
Cranial and abdominal ultrasounds, echocardiogram and vertebral x-rays were normal. Brain MRI under general anaesthesia was declined by the parents. A CT scan of the petrosal bones showed an ossicular malformation and a narrowing of the left outer auditory canal.
Chromosome examination revealed an apparently balanced de novo translocation: 46, XX, t(1;17)(q42.?2;q21.3?2). Further exploration by Agilent 44K array-CGH identified a deletion of at least 23kb (genome assembly NCBI36: chr17:40.318–40.341 Mbp) at 17q21.31 (figure 2A) encompassing EFTUD2 (5′ part), CCDC103, FAM187A and GFAP (3′ part).
Case 2, a boy, was born at 34/40 weeks gestation. He had EA associated with facial asymmetry, two left preauricular tags, cupped and low set ears (figure 2) and a pelvic kidney on the right side. Echo at birth showed persistent ductus arteriosus and patent foramen ovale. A diagnosis of VATER association was proposed. He experienced severe oesophageal dysmotility and gastro-oesophageal reflux subsequent to surgery for EA with no catch up for growth. Indeed, at 6.5 months (5 months corrected), weight and HC were below −3 SD. Cranial ultrasound showed no brain abnormality. Routine G banded karyotype showed an apparently normal male karyotype (46, XY) and an Affymetrix gene chips 6.0 SNP genotyping array identified a de novo deletion at 17q21.31 of approximately 34 kb (genome assembly NCBI36: chr17:40.319–40.353 Mbp, figure 2B) encompassing EFTUD2 (5′ part), CCDC103, FAM187A and GFAP.
Case 3, was diagnosed with EA type C at birth, and was referred to the genetic clinic at the age of 3 weeks for possible Goldenhar syndrome. Renal ultrasound and cardiac echocardiogram, vertebral x-rays and eye examination were normal. At 5 months, the patient had symmetric growth retardation with height, weight and HC at −3 SD. Standard karyotype showed normal chromosomes. Array-CGH (Agilent 44K) revealed two CNVs, each of them being inherited from one of the parents (paternal 100 kb 9p21.1 deletion and maternal 200 kb 8q24.11 duplication), and were considered variants unrelated to the patient's phenotype. EFTUD2 gene sequencing revealed a de novo heterozygous nucleotidic change affecting the donor splice site of intron 22, predicted in silico to abolish splicing (g.42931923C>T, c.2259+1G>A). This variant is absent from the dbSNP and 1000 genome databases.
Case 4, a female subject, was previously reported by de Pontual et al.3 She was first seen in the genetic clinic at 17 years of age for the exploration of mild intellectual disabilities (IQ of 73 at WISC III). On examination, Feingold syndrome was suspected due to the association of microcephaly, mild thenar hypoplasia and 2/3 syndactyly of toes. She also presented unilateral partial paralysis of the elevator of the angle of the mouth (figure 1C), high arched palate with no cleft, and asymmetric and dysplastic ear helices. Agilent 105 K array-CGH identified no pathological copy number variant. Direct sequencing and qPCR of the MYCN and MIR17HG genes were normal. Direct sequencing of the EFTUD2 gene identified a de novo nucleotide variant (c.2619-2621delinsGGTC) presumed to result in a frameshift and a premature stop codon (p.Phe874ValfsX11) in the absence of nonsense-mediated RNA decay.
Case 5 presented with the association of prenatal and postnatal microcephaly, EA, dysplastic ears, mixed hearing loss, epilepsy and intellectual disability requiring special education, and was referred for possible Feingold syndrome. The patient died at 12 years of sepsis following duodenal perforation. Array-CGH (Agilent 44K) identified no pathological copy number variant. Direct sequencing of EFTUD2 identified a nucleotide variant creating a premature stop codon (c.1705C>T; p.Arg569X) that arose de novo.
Case 6, a male subject, was referred for possible CHARGE syndrome and presented ante-natal and postnatal microcephaly, EA, dysplastic ears, conductive hearing loss and cleft palate (figure 1D). Intellectual disability required special needs education. Direct sequencing of CHD7 did not identify any pathological variant. Subsequently, EFTUD2 sequencing revealed a heterozygous nucleotidic change affecting the donor splice site of intron 8, predicted in silico to abolish splicing (c.619+1G>A). This variant is absent from the dbSNP and 1000 genome databases and arose de novo.
Case 7 was referred for possible CHARGE syndrome and presented with choanal atresia, EA, facial asymmetry, dysplastic ears, hearing loss, agenesis of the lateral semicircular canals and postnatal microcephaly. Direct sequencing of CHD7 identified no pathological variation. EFTUD2 direct sequencing revealed a heterozygous, de novo, nucleotidic change leading to a missense mutation (c.623A>G, p.H208R) absent from the dbSNP and 1000 genome databases and predicted to be damaging in silico. H208 falls within the switch II region of the EFTUD2 GTP-binding domain and is a putative guanine exchange factor interaction site (NCBI Conserved Domains Database entry cd04167). This histidine is conserved between human EFTUD2 and its yeast homologue Snu114p,10 and in a random mutagenesis screen in yeast, the equivalent mutation in Snu114p, H218R, resulted in a temperature-dependent growth defect,11 strongly supporting the pathogenicity of the EFTUD2 mutation in case 7.
Case 8 is in foster care and we have little information about the medical history of the family. The patient was referred for possible CHARGE syndrome and presented with choanal atresia, dysplastic ears, hearing loss and microcephaly. Direct sequencing of CHD7 identified no pathological variant. EFTUD2 direct sequencing revealed a heterozygous nucleotide variant leading to a premature stop codon (c.2198G>A; p.W733X). Parents were not available for testing.
Case 9 was referred for possible CHARGE and presented with prenatal and postnatal microcephaly, EA, asymmetric dysplastic ears, hearing loss and complex cardiac malformation with an atrial septal defect, ventricular septal defect and right pulmonary artery anomaly. No mutations were identified in CHD7. EFTUD2 direct sequencing revealed a heterozygous nucleotide variant in IVS 23 (c.2347+66A>G), predicted to create a donor splice site in silico and absent from the dbSNP and 1000 genome databases. The variant was inherited from the healthy father and therefore not considered disease causing. Blood from the paternal grandparents and other tissue from the father of case 9 are awaited for further study.
Case 10 was referred for possible CHARGE and presented with low birth weight, bilateral microtia and sensorineural hearing loss. The patient also presented with malar and mandibular hypoplasia. Growth has been in the normal range but the patient developed epilepsy at age 3 and has moderate intellectual disability. Direct sequencing of CHD7 identified no pathological variant. EFTUD2 direct sequencing revealed a heterozygous nucleotide change leading to a missense mutation (c.670G>A; p.G224R), absent from the dbSNP and 1000 genome databases and predicted to be damaging in silico. Parental DNA has not been available for study thus far. As for the residue H208, mutated in case 7, G224 falls within the 19 amino acid switch II region of the GTP-binding domain (NCBI Conserved Domains Database entry cd04167). G224 is adjacent to D223, which plays a role in salt bridge formation in related proteins,10 supporting the pathogenicity of the EFTUD2 mutation in case 10.
Case 11 was a female fetus for whom pregnancy was terminated at 29 weeks gestation due to the association of EA and delayed gyration. Necropsy reported EA type C, dysplastic ears, microretrognathia and microcephaly with no brain malformation (figure 1E). CHARGE syndrome was suspected in the fetus but no CHD7 mutation could be identified. A diagnosis of Feingold syndrome had been proposed for the mother, based on the association of microcephaly, mild intellectual disabilities and brachydactyly of fingers (figure 1F). Mixed hearing loss had been diagnosed during late childhood. Subsequently, EFTUD2 sequencing of DNA from the fetus revealed a heterozygous nucleotide change affecting the donor splice site of intron 27, predicted in silico to abolish splicing (c.2823+1del) and was absent from databases. The mutation was inherited from the mother in whom it arose de novo.
Case 12 was referred for possible CHARGE syndrome and presented with low birth weight, EA, congenital cardiac defect, unilateral microtia, facial asymmetry and unilateral thumb hypoplasia. Direct sequencing of CHD7 identified no pathological variant. EFTUD2 direct sequencing revealed a heterozygous 5 bp deletion in IVS12 (c.1058+3_1058+7del) absent from the dbSNP and 1000 genome databases and not predicted to alter splicing in silico. Unfortunately, parents are not available for testing and we are not able to conclude whether this variation is disease causing or not.
Finally, neither deletion nor mutation of EFTUD2 could be identified in two cases with EA and either MFD or microcephaly from our series of patients (patients AO24 and AO63 in table 2).
In situ hybridisation
To the best of our knowledge, the embryonic expression pattern of Eftud2 has not been previously published for any vertebrate species. We performed in situ hybridisation on cryosections of embryonic day 11.5 (E11.5) mice using digoxigenin-labelled Eftud2 antisense and sense probes. The antisense probe revealed expression of Eftud2 in a range of mesenchymal and epithelial tissues, while the sense control produced no appreciable staining (figure 3). Strong expression was observed in the mesenchyme of the limb buds and lung buds (figure 3A–C). Eftud2 expression was also noted in the trachea and oesophagus, mandibular mesenchyme, ventricular zone cells of the forebrain and the epithelium of the otic vesicle (figure 3G–I,M,O). Several of these Eftud2 expression sites correspond to tissues whose derivatives are affected in patients with EFTUD2 lesions.
EFTUD2 has recently been identified as the disease-causing gene in MFD with microcephaly (MFDM MIM 610536),6 ,12–16 with dysplastic ears, microcephaly and intellectual disabilities being hitherto consistent signs (15 cases with an EFTUD2 gene mutation or deletion, table 1). Other frequent features reported in the literature include preauricular skin tags, posterior cleft palate, auditory canal atresia, choanal atresia, congenital cardiac defects and radial ray anomalies (table 1). A high nasal bridge with bulbous tip and thick alae nasi and persistent epicanthic folds with ageing are also relatively common.
We report on 10 patients presenting MFD associated with EA and/or microcephaly and heterozygous loss-of-function of the EFTUD2 gene. In particular, two unrelated cases showed a heterozygous deletion encompassing EFTUD2 (5′ part), CCDC103, FAM187A and GFAP (3′ part) (figure 2 and table 2). Deletions encompassing EFTUD2, CCDC103 and/or FAM187A are not reported in the Database of Genomic Variants and there are no other entries with a CNV at this locus in the DECIPHER database besides these two patients. However, recurrent deletion at 17q21.3q24 has been reported in patients with EA (reviewed in Puusepp et al9). Under a contiguous gene syndrome model, GFAP was considered unlikely to explain EA since heterozygous gain-of-function mutations are disease-causing in Alexander disease (MIM 203450),17 and since Gfap null mice present no congenital malformation.18 Interestingly, the Stanford database SOURCE gene report for homo sapiens indicates CCDC103 expression in the trachea. However, it was recently identified as disease-causing in primary ciliary dyskinesia,19 a syndrome without tracheal or oesophageal abnormalities. Finally, nothing is known regarding the pattern of expression or function of FAM187A. Direct sequencing of these latter two genes in 17 isolated EA cases identified no mutation.
Despite the fact that EA was not previously reported in EFTUD2 mutation-positive cases,6 we hypothesised that EA may be a low-penetrance phenotype of EFTUD2 haploinsufficiency. We thus considered Feingold syndrome (MIM 164280, 614326), CHARGE syndrome (MIM 214800) and OAVS as differential diagnoses for MFD with microcephaly. Under this hypothesis, overlapping features with Feingold syndrome would include microcephaly, EA and limb anomalies, while overlapping features with CHARGE and OAVS would include ear dysplasia, facial asymmetry and, to a lesser extent, EA.7 Indeed, EA is reported in about 5% of patients with OAVS and about 15% of patients with CHARGE syndrome.7 ,20 ,21 Moreover, choanal atresia and inner ear malformations are additional overlapping features between MFD with microcephaly and CHARGE syndrome while preauricular tags, epibulbar dermoid and cervical spine anomalies overlap between MFD with microcephaly and OAVS (tables 1 and 2). An anomaly of the semicircular canals is reported in three EFTUD2 mutation-positive cases from our series (table 2) and should thus not be considered pathognomonic for CHARGE syndrome.
Heterozygous EFTUD2 truncating, splicing or missense mutations were identified in eight cases (table 2) and were found to be de novo in all sporadic cases for whom parents were available for testing (5/7), and in the mother of case 11 (table 2). Thus, we can conclude that EA is an additional malformation caused by EFTUD2 heterozygous loss-of-function. Importantly, EFTUD2 can now be regarded as a predisposing gene for EA in the del17q21.3-q24 syndrome (reviewed in Puusepp et al9). Four cases with EA/TEF and a deletion within this interval have been previously reported. Only one of the four deletions has been fine-mapped and is telomeric to EFTUD2.9 Another case associates EA and microcephaly without MFD and harbours a deletion probably telomeric to EFTUD2 at 17q22-q23.3, while the remaining two cases associate EA/TEF, microcephaly and MFD with a large interstitial 17q deletion encompassing EFTUD2.22–24 Therefore, additional EA-predisposing genes within the 17q21.3-q24 interval, further telomeric to EFTUD2, may exist.
Microcephaly is uncommon in CHARGE and OAVS and may not be a consistent feature in patients with EFTUD2 lesions (table 2). Thus, one could propose changing the name MFD with microcephaly to MFD Guion-Almeida type. Importantly, brain malformations have not yet been reported in patients with EFTUD2 gene mutation or deletion. Reviewing the literature, cases 1 and 2 in Sutphen et al7 likely represent further examples of MFD Guion-Almeida type (see table 2). This diagnosis could also be hypothesised in the familial case reported by Wieczorek et al,25 with the mother, one of her daughters and her son all affected (the son most convincingly so) (see table 2). In addition, all three individuals have zygomatic aplasia or hypoplasia, which has not been reported thus far in patients with EFTUD2 mutations, and should be more carefully looked for. Along these lines, EFTUD2 (and SF3B4, see below) should be tested in the patient reported by Ozkan et al26 and featuring Pierre Robin sequence, EA and radio-ulnar synostosis. However, EFTUD2 is unlikely to be involved in a large proportion of OAVS cases; indeed, our sequencing of 19 isolated OAVS cases identified no mutations. Of note, no chromosomal aberration concerning the 17q21 region was reported in a large series of OAVS cases submitted to CGH analysis.8
EFTUD2 encodes one of the major components of the spliceosome. The list of syndromes associated with mutations of genes involved in the mRNA maturation process carried out by the spliceosome is rapidly growing. Indeed, mutations in three genes encoding proteins of the major spliceosome complex have been reported in autosomal dominant retinitis pigmentosa.27–29 An autosomal recessive condition, Taybi-Linder syndrome (MIM 210710), was ascribed to RNU4ATAC (MIM 601428) gene mutations.30 RNU4ATAC encodes a small nuclear RNA component of the minor (U12-dependent) spliceosome complex. Finally, SF3B4 has been identified as the disease-causing gene in Nager syndrome, in which clinical overlap with MFD with microcephaly is striking.16 The latter point was illustrated by screening for EFTUD2 mutations in a series of Nager syndrome patients without an SF3B4 mutation; Bernier et al16 identified one EFTUD2 mutation in this series, although retrospectively it was concluded that the patient had microcephaly, suggesting that this is an important distinguishing feature. The spliceosome complex plays a key role in the regulation of gene expression through pre-mRNA splicing and in cell proliferation in vertebrates (reviewed in Konig et al31 and Johnson et al32). However, some of the spliceosome components also have other functions, possibly resulting in defects in specific regions of the developing embryo when mutated. Indeed, SF3B4 has been shown to regulate BMP-mediated chondrogenesis by binding directly to the BMPR-1A serine–threonine kinase receptor and reducing its amount at the cell surface.33 Disruption of this mechanism may underlie the radial ray anomalies observed in Nager syndrome.
We observed widespread expression of Eftud2 in embryonic mice, with certain regions displaying relatively more intense expression (figure 3). These sites included regions of mesenchymal outgrowth such as the distal limb bud and the lung bud; the ventricular zone neuroepithelium, which contains proliferative neuroprogenitors; and various epithelia of the developing respiratory and digestive systems. The observed pattern of Eftud2 expression correlates well with several of the developmental regions that are affected in patients with EFTUD2 lesions.
Altogether, we show that EFTUD2 loss-of-function leads to a wide range of congenital malformations and should be considered as a differential diagnosis for phenotypic variants of OAVS and CHARGE and Feingold syndromes. We also report EA as a common additional feature of MFD Guion-Almeida type and EFTUD2 as a novel EA predisposing gene at 17q.
We are grateful to the families for participating in the study.
CTG, FP, MH-E and JA contributed equally to this work.
Funding This work was supported by funding from E-Rare CRANIRARE (SL) and the Fondation pour la Recherche Médicale (JA).
Contributors All authors contributed to this work and have read and approved the manuscript.
Competing interests None.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
Ethics approval CCP Ile-de-France II.
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.