Biochimica et Biophysica Acta 1842 (2014) 1755–1761

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbad is

De novo GLI3 mutation in esophageal atresia: Reproducing thephenotypic spectrum of Gli3 defects in murine models

Lin Yang a,b,1, Chun Shen c,d,1, Mei Mei b, Guodong Zhan b, Yunke Zhao e, Huijun Wang a, Guoying Huang a,Zilong Qiu e, Weineng Lu f, Wenhao Zhou a,b,⁎a Key Laboratory of Birth Defects, Children's Hospital of Fudan University, Shanghai, Chinab Division of Neonatology, Children's Hospital of Fudan University, Chinac Division of Surgery, Children's Hospital of Fudan University, Shanghai, Chinad Key Laboratory of Neonatal Diseases, Ministry of Health, Children's Hospital of Fudan University, Chinae Institute of Neuroscience, Shanghai Institute of Biological Sciences, Chinese Academy of Sciences, Shanghai, Chinaf Division of Neonatology, Guangzhou Women and Children's Medical Center, Guangzhou, China

⁎ Corresponding author at: Division of NeonatologyUniversity, 399 Wan Yuan Road, Shanghai, 201102 China.

E-mail address: zwhchfu@126.com (W. Zhou).1 These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.bbadis.2014.05.0010925-4439/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 26 February 2014Received in revised form 29 April 2014Accepted 3 May 2014Available online 9 May 2014

Keywords:Esophageal atresiaGLI3 geneCopy number variationSonic hedgehog pathway

Esophageal atresia is a common and life-threatening birth defectwith a poorly understood etiology. In this study,we analyzed the sequence variants of coding regions for a set of esophageal atresia-related genes includingMYCN, SOX2, CHD7, GLI3, FGFR2 and PTEN for mutations using PCR-based target enrichment and next-generation sequencing in 27 patients with esophageal atresia. Genomic copy number variation analysis wasperformed using Affymetrix SNP 6.0. We found a de novo heterozygous mutation in the N-terminal region ofthe GLI3 gene (c.332 T N C, p.M111T) in a patient with esophageal atresia and hemivertebrae. The N-terminalregion (amino acids 1–397) of GLI3 contains the repressor domain, which interacts with SKI family proteins.Using the co-immunoprecipitation assay, we found that interaction of GLI3 with the SKI family protein SKILwas significantly compromised by the p.M111T mutation of GLI3. Thus far, all the identified mutations mappedwithin the repressor domain ofGLI3were nonsense and frame-shiftmutations. In this study, amissensemutationwas initially detected in this region. Our finding is thefirst to link thisGLI3 genemutationwith esophageal atresiain humans, which was previously suggested in an animal model.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Esophageal atresia (EA) is a developmental defect of the uppergastrointestinal tract in which the continuity between the upperand lower esophagus is lost. Esophageal atresia with or withouttracheo-esophageal fistula (TEF) is a rare malformation, occurringin approximately 1 in 3500 births [1]. Although the underlying mo-lecular mechanism remains obscure in most patients with EA/TEF,it has been associated with a wide spectrum of genetic syndromescaused by genetic alterations including mutations in single genesor chromosome aberrations [2].

Approximately 6% to 10% of syndromic EA/TEF is associated withchromosomal anomalies [3]. The majority is represented by trisomies(trisomy 13, 18, 21) [2,4] or microdeletions of 22q11 [5], 17q22–

, Children's Hospital of FudanTel./fax: +86 21 64931003.

17q23.3 [6] and 16q24.1 [7]. Meanwhile, mutations of MYCN, SOX2 andCDH7 have also been reported to be responsible for syndromic EA, suchas Feingold syndrome, AEG syndrome (Anophthalmia-esophageal–genital syndrome) and CHARGE association (coloboma, heart anomaly,choanal atresia, retardation, and genital and ear anomalies) [8]. Deletionsand mutations in the GLI3 and FGFR2 genes are responsible for Pallister–Hall syndrome and Apert syndrome, respectively, in which EA is infre-quently reported [9–11]. In rare cases, mutations in PTEN have beenreported in the VATER association (vertebral defects, anal atresia,tracheoesophageal fistula with esophageal atresia, and radial or renaldysplasia) [12].

In this study, we searched for sequence variants in the protein-coding regions of the human MYCN, SOX2, CHD7, GLI3, FGFR2 and PTENgenes by PCR-based target enrichment followed by next-generation se-quencing using the Ion Torrent Personal Genome Machine™ (PGMTM,Life Technologies, Carlsbad, CA). We also analyzed genomic copynumber variants (CNVs) in 27 patients with EA/TEF. We found a denovo heterozygous mutation in the N-terminal region of the GLI3 gene(c.332 T N C, p.M111T). Our finding is the first to link the GLI3 genemutation with esophageal atresia in humans, which was previouslysuggested in an animal model.

1756 L. Yang et al. / Biochimica et Biophysica Acta 1842 (2014) 1755–1761

2. Materials and methods

2.1. Subjects and DNA extraction

The subjects included 27 neonates with either isolated (6 neonates,22.2%) or syndromic (21 neonates, 77.8%) forms of EA/TEF (Table 1).Among these neonates, 14 (52%) were female and 13 (48%) weremale. Routine chromosome analysis was normal. The samples used inthis study were collected with appropriate informed consent and theapproval of the ethics committees of Children's Hospital, FudanUniversity.

Genomic DNA was extracted from whole blood using the QIAampDNA Blood Mini kit (Qiagen, Hilden, Germany). DNA concentrationwas measured using a Nano-Drop spectrophotometer (ND-1000,Thermo Fisher Scientific, Waltham, MA).

2.2. Candidate genes sequencing

Analysis of the protein-coding regions of the human MYCN, SOX2,CHD7, GLI3, FGFR2 and PTEN genes were performed using Ion TorrentPGMTM. The complete target genes have 43 amplicons (from 652 bp to7936 bp) covering all exons and at least 20 bp of all intron/splice sites.The library was prepared by following the instructions provided bythe manufacturers of kits for fragmentation (Ion Shear, Life Technolo-gies), adaptor and barcode ligation (Ion Xpress Barcode Adapters Kit,Life Technologies) and library quantification (Ion Library QuantificationKit, Life Technologies). We used the Ion OneTouch™ system (Life Tech-nologies) to clonally amplify pooled barcoded libraries on Ion Sphere™particles. Torrent Suite™ software was used to compare base calls.NextGENeTM software was used to read alignments and to call variantwith the human genomic reference hg19 (NCBI). The variants selectedfor further analysis met the following criteria: 1) The variant was

Table 1Summary of clinical features of all included patients. Note: PDA: patent ductus arteriosus, Phypertension, ASD: atrial septal defect, PLSVC: persistent left superior vena cava, TAPVD: total

PatientID

Gender Isolated/non-isolated Gastrointestinal Skeletal Heart

29 M Non-isolated – Polydactyly30 M Non-isolated – – VSD, PDA,107 M Isolated – –

208 M Non-isolated – – VSD, PDA,296 F Non-isolated – – CHD309 F Non-isolated – Polydactyly –

328 F Isolated – – –

329 M Non-isolated – Hemivertebrae –

332 F Non-isolated – – ASD, PDA343 F Non-isolated Anal atresia,

anorectal fistulaPolydactyly –

352 F Non-isolated Anal atresia,intestinal atresia

Spinalanomalies

PDA, PFO,

386 F Isolated – – –

396 F Non-isolated Anorectal fistula – ASD, PPHNPLSVC

399 F Isolated – – –

405 M Non-isolated Meckel'sdiverticulum

– VSD, PDA,

w24 M Non-isolated – – VSD, ASD,w43 M Non-isolated – – VSD, PDA,w44 M Non-isolated – – PDA, PFO,w46 F Non-isolated – Polydactyly –

611 M Non-isolated – – PDA, PFO612 F Non-isolated – – TAPVD, DO

pulmonar613 F Non-isolated – Hemivertebrae –

614 M Non-isolated – – VSD, PDA,

615 M Non-isolated Anal atresia – PDA, ASD617 M Isolated – – –

618 F Isolated – – –

619 M Non-isolated – – PDA, ASD

detected in sequence reads for both strands, 2) a minimum coverageof 10× was achieved, 3) the variant reads represented N20% of thesequence reads at a particular site, 4) the targeted region covered allexons and at least 20 bp of all intron/splice sites. The filtered variantswere then compared with dbSNP (http://www.ncbi.nlm.nih.gov/projects/SNP/). Novel Variants of unknown significance (VUS) were an-alyzed with in silico tools SIFT (http://sift.jcvi.org/) and PolyPhen2(http://genetics.bwh.harvard.edu/pph2/).

2.3. Validation by Sanger sequencing

The variants were validated by PCR followed by direct Sangersequencing using an automated sequencer (3500XL Genetic Analyzer,Applied Biosystems, Foster City, CA).

2.4. In vitro binding assay and co-immunoprecipitation

The human N-terminal GLI3 (397 amino acids) gene was obtainedfrom GenScript (Piscataway, NJ). The hGli3N sequence was insertedinto pRK5 with an HA tag to produce the WT hGli3N plasmid. ThehGli3N mutation was generated based on this construct. In brief,Gli3N-M111T was generated using KOD-Plus (Toyobo, Osaka, Japan)according to the manufacturer's instructions. Primers used in the pointmutation assay were For 5′-acggaccccaggaatggttacatgg-3′ and Rev 5′-ggcaaacaccgtcccgcggtacggc-3′. Antibodies used in this study wereagainst the HA epitope tag (#M20013, Abmart) and SKIL (#19218-1-AP, Proteintech). The plasmid expressing either HA-tagged wild-typeor p.M111T-mutant N-terminal GLI3 (397 amino acids) was transfectedinto 293 T cells.

For co-IP of 293 T cells, HEK293T cells were plated on 6-well plates,and transfectionwas performedwhen the cells reached 50% confluence.Lipofectamine2000 (Invitrogen, Carlsbad, CA)was used for transfection.

FO: patent foramen ovale, VSD: ventricular septal defect, PPHN: persistent pulmonaryanomalous pulmonary venous drainage, DORV: double outlet of right ventricle.

Urogenital Dysmorphic Additional anomalies

Hypospadia –

PFO Hypospadia – –

PPHN – – –

– Preauricular tags –

– –

– – –

– – –

– – –

– – –

PPHN – – –

– – –

, aortic coarctation, – – –

– – –

PFO – –

PDA, PPHN – – Diabetes mellitusPFO – – –

PLSVC – – –

– – –

– – –

RV, AVSD, PDA,y stenosis

– – –

– – –

PFO – – Coagulation dysfunction,hypothyroidism, hypokalemia

– – Hypoglycemia– – –

– – –

– – –

1757L. Yang et al. / Biochimica et Biophysica Acta 1842 (2014) 1755–1761

Then, 2 μg of HA-tagged hGli3N was added per well in 6-well plates.Cells were harvested 48 h after transfection. The cells were harvestedand lysed for 10 min at 4 °C in RIPA buffer. A total of 25% of the super-natant was saved for the input control, and the rest was incubatedwith 2 μg anti-HA agarose beads (Abmart, Beijing, China) overnight at4 °C. The immune complex was washed three times with the lysis buff-er, then boiled in 1× SDS loading buffer with 20 mM DTT and resolvedby 8% SDS-PAGE. The gel was transferred onto a PVDF membrane(Amersham Pharmacia Biotech, Uppsala, Sweden), and the membranewas blocked with 5% milk in TBST buffer for 1 h. The membrane wasthen incubated overnight at 4 °C with 1:1000 anti-SKIL antibodies,washed three times in TBST, then incubated for 2 h at room temperaturewith 1:3000 secondary antibodies, and after three timeswashing in TBST,the signalswere revealed byHRP reaction using the Supersignal chemilu-minescent substrate (Pierce, Rockford, IL). The co-immunoprecipitatedendogenous SKIL was normalized to the HA-tagged hGli3N. The datawere analyzed with Students' T-test with the GraphPad Prism software.

2.5. Whole-genome high-resolution SNP array analysis

Affymetrix Genome-Wide Human SNP Array 6.0 was used to per-form copy number variation analysis. The copy number variation dataof 2 patients (614, 617) were obtained from previous results with the244 K array comparative genomic hybridization (CGH) platform(Agilent, Santa Clara, CA). We used Affymetrix Genotyping Console(GTC) V.4.0 and ChAS V.2.0 software to call CNVs, with all the parameterthresholds set to default values. All annotated CNV genomic regionswere based on the hg19 (NCBI) human genome assembly. Commonvariants were evaluated using the Database of Genomic Variants(DGV) and were excluded from further analysis. Quantitative real-time PCR was used for validation of CNVs. The ALB and TBP geneswere used as controls as previously described [13,14]. Copy numbergain or loss was determined by the comparative CTmethod. The relativecopy number values for each CNV between 0.8 and 1.2 were consideredto be normal while values≥1.3 and≤0.7 were considered as evidencesof duplication and deletion, respectively.

3. Results

3.1. Candidate gene sequencing analysis

We pooled 10 cases of barcoded sample libraries on one 316™ chipfor sequencing using the Ion PGMTM. We obtained a total alignment ofbases from 416Mb to 478Mb, with an average mean depth of coverageof 375–421 Mb. We identified 156 variants from 27 probands (Supple-mentary Table 1) and an average of 5.78 variants per proband. Theseincluded 126 known single-nucleotide polymorphisms (SNPs), 11VUSs, and 19 variants were called on homopolymer sites. Variant callsfrom 5 noncoding homopolymer sites were filtered out to avoidhomopolymer-associated base-calling errors. Of the 11 VUSs, 9 werein coding regions (2were synonymous variants), and 2were in intronicregions. Sanger sequencing confirmed 5 of the 7 missense variants in

Table 2Summary of variants of unknown significance in 27 patients with EA/TEF.

Patient ID r Chromosomeposition

Gene CDS Chr Referencenucleotide

Coverage Hom

613 42187860 GLI3 2 7 T 300 Het328 61713068 CHD7 4 8 C 61 Het399 61654400 CHD7 1 8 T 629 Het611 61713055 CHD7 4 8 C 47 Het612 16082736 MYCN 1 2 G 100 Het29 123247625 FGFR2 12 10 T 1014 Het618 61769072 CHD7 33 8 C 221 Het208 42065790 GLI3 – 7 G 245 Het619 61734504 CHD7 – 8 G 189 Het

the coding regions (Table 2). Of the 5 confirmed variants, the variantof GLI3 (NM_000168.5, c.332 T N C, p.M111T) was predicted to be dam-aging in silico by both SIFT and PolyPhen2, and the variant in CHD7(NM_017780.3, c.2360C N T, p.787S N F) was predicted to be damagingin silico by PolyPhen2 but tolerated by SIFT. The p.M111T variant ofGLI3 is novel and de novo (Fig. 1A and D). Unfortunately, the parentswere not available to test the CHD7 p.787S N F variant.

The phenotypes of the patients with the GLI3 (p.M111T) and CHD7(p.S787F) variants were esophageal atresia with hemivertebrae andisolated esophageal atresia, respectively.

3.2. In vitro binding assay and co-immunoprecipitation

GLI3 is a Ski-interacting protein. The N-terminal region of GLI3(amino acids 1–397) contains the repressor domain responsible forinteracting with Ski [15,16]. To investigate whether the p.M111T muta-tion of GLI3 affects the interaction of GLI3 with SKI family proteins, weexamined the interaction between wild-type and p.M111T-mutantGLI3 with endogenous SKIL, a critical SKI family protein, using co-immunoprecipitation assay.

As shown in Fig. 1E, we found that the interaction between GLI3with SKIL was significantly compromised in the GLI3 p.M111T mutant(Fig. 1E). These data suggested that the p.M111T mutation affects thephysiological function of GLI3 by disrupting its interaction with SKIfamily proteins.

3.3. Copy number variation analysis

We identified a total of 969 CNVs in 27 probands. Among theseCNVs, 266 were mapped in the intronic regions and 693 CNVs were re-ported in DGV that were excluded for further analysis. Only nine CNVsfrom 8 probands were rare and mapped in the gene rich regions. Fourof them, ranging in size from 175.2 Kb to 382.2 Kb at 3p14.1, 4q21.22,10q24.2 and 17q24.3, were validated by qPCR (Table 3, Fig. 2). TheCNV in proband 386 was de novo, but the parents of the other three pa-tients were not available for testing. Clinical significance was evaluatedbased on the literature review and genome annotation of gene contentand functions. There was no evidence immediately supporting theclinical significance of these CNVs. Case W43 is a male patient withesophageal atresia IIIb and ventricular septal defect. A duplication wasfound at 3p14.1 with an estimated size of 382 kb, encompassing theSLC25A26 and LRIG1 genes. The 3p14 region is frequently detected invarious types of human cancers [17]. The LRIG1 gene has been proposedto interact with and counteract the effects of growth factor receptorssuch as EGFR [18], thereby functioning as a tumor suppressor[19]. Case 329 is a male patient with esophageal atresia IIIb andhemivertebrae. A duplication was found at 4q21.23 with an estimatedsize of 365 kb, encompassing the HNRNPD, HNRPDL, ENOPH1 andTMEM150C genes. Deletion of this region induces 4q21 deletion syn-drome, which is characterized by growth restriction and mental retar-dation [20]. Case 386 is a female patient with isolated esophagealatresia III. De novo microduplication was found at 10q24.2 with an

/Het Mutation call Amino acidchange

Interpretation SIFT PolyPhen2

c.332 T N C 111 M N T VUS 0.001 0.99c.2360C N T 787S N F VUS 0.224 0.964c.409 T N G 137S N A VUS 0.19 0.002c.2347C N T 783P N S VUS 0.053 0.069c.550G N T 184A N S VUS 1 0.301c.1599A N G 533C N C VUS – –

c.7233C N T 2411A N A VUS – –

IVS1242 + 8C N T Intron VUS – –

IVS2835 + 18G N C Intron VUS – –

Fig. 1. Identification of theGLI3mutation and functional analysis of GLI3 protein. A: The p.M111Tmutation ofGLI3 identified byNextGENeTM software. The FASTAfilewas derived from theIon Torrent Server. Nucleotides highlighted in blue indicate the mutation. B: Chromosomal location of GLI3. C: Schematic view of the GLI3 structure. The red vertical line indicates thep.M111T mutation. In the N terminal region, each black vertical line indicates a reported mutation from The Human Gene Mutation Database (HGMD, public database). ZFD is the ZincFinger Domain, PC is the Protein Cleavage site, and TA1 and TA2 are the two transactivator domains. Greig cephalopolysyndactyly syndrome (GCPS) is caused by mutations upstreamorwithin the zinc finger domain. Pallister–Hall syndrome (PHS) is caused bymutations affecting themiddle third of the protein. D: Sanger sequencing results confirmed the de novo, het-erozygous p.M111Tmutation; E: The in vitro translatedGLI3 (WT) or p.M111Tmutant and those that bound to SKILwere analyzed byWestern blotting and are indicated by a histogram inthe right panel. **: p b 0.005.

1758 L. Yang et al. / Biochimica et Biophysica Acta 1842 (2014) 1755–1761

estimated size of 175 kb, encompassing the PYROXD2, MIR1287, HPS1andHPSE2 genes. The 10q24 duplication has previously been implicatedin split hand footmalformation, but no patient with EA/TEF has been re-ported [21]. Case 614 is a male patient with esophageal atresia, ventric-ular septal defect, coagulation dysfunction, hypothyroidism andhypokalemia. A duplication was found at 17q24.3 with an estimatedsize of 249 kb, encompassing the SLC39A11 gene. Neither deletion/duplication of this region nor abnormality of the SLC39A11 gene hasbeen related to EA/TEF. Although chromosome anomalies have been re-ported in approximately 6%–10% of individuals with EA/TEF, none of thedetected CNVs are proposed to contribute towards gastrointestinalmalformations. Copy number variants may not be the common causein the investigated EA/TEF cohorts.

4. Discussion

The GLI3 gene encodes a zinc finger protein belonging to the GLIfamily and functions as a transcription factor that regulates the Sonichedgehog (SHH) signal transduction pathway [22]. The SHH-GLI path-way is one of many signaling pathways controlling the developmentof the primitive gut. The importance of SHH signaling in esophagealdevelopment was confirmed with Gli mutant mice, in which a 50% re-duction of Gli3 expression in Gli2–/– mice resulted in EA/TEF and severelung deficiencies [23]. These results suggest that mutations in the GLI3gene may be involved in human foregut malformations. However, to

Table 3Characteristics of detected rare copy number variants.

Patient ID Cytoband Chromosome start (hg19) Chromosome end (hg19)

W43 3p14.1 66415348 66797568329 4q21.22 83163011 83528606386 10q24.2 100092010 100267283614 17q24.3 70521861 70770996

our knowledge, no GLI3 gene structure aberrations or sequence varia-tions were detected in patients with isolated or syndromic EA/TEF.

Mutations in GLI3 are associated with two dominant human syn-dromes: Greig cephalopolysyndactyly syndrome (GCPS) and Pallister–Hall syndrome (PHS) [9]. The genotype–phenotype correlation of theGLI3 gene has been demonstrated based on mutation location and clas-sification. Mutations upstream of the zinc finger domain (the repressordomain) or within the zinc finger domain of the GLI3 gene are associat-ed with GCPS; mutations downstream of the zinc finger region includ-ing the protease cleavage site correlate with PHS (Fig. 1B and C). Asimilar phenomenon was observed in the mouse homologue [24]. Theprotein truncation mutations in GLI3 are associated with PHS, whereashaploinsufficiency of GLI3 induces GCPS [25,26]. These results stronglysuggest that these two allelic disorders have distinct modes of patho-genesis. In this study,we present a female patient (No. 613)with esoph-ageal atresia IIIa and hemivertebrae, carrying a c.332 T N C transitioncaused by a heterozygous mutation, p.M111T, in exon 3 of the GLI3gene. This mutation is de novo and is located in the N-terminal region.The N-terminal region of GLI3 contains the repressor domain [27].Thus far, all identified mutations mapped within the repressor domainwere nonsense mutations (T122X, L162X, and E236X) or frame-shiftmutations (M309fs, L346fs and E411fs), which should remove mostregions of the 1580 amino acid wild-type protein and induce GCPS byhaploinsufficiency [24,26,28,29]. However, we present a missense mu-tation, p.M111T, in this region. Using co-immunoprecipitation assays,

Size (kb) Event Inheritance Genes

382.22 Gain NA SLC25A26, LRIG1365.595 Gain NA HNRNPD, HNRPDL, ENOPH1, TMEM150C175.273 Gain de novo PYROXD2, MIR1287, HPS1, HPSE2249.135 Gain NA SLC39A11

Fig. 2. Copy number variation analysis and qPCR validation.Four copy number variation profiles detected by SNP array and CGH array in 4 patients and validated by real-time PCR. X-axis:genomicmapping; Y-axis: log2 copy-number values. A: CaseW43presented a duplication on 3p14.1with an estimated size of 382 kb; B: Case 329presented a duplication on4q21.23withan estimated size of 365 kb; C: Case 386 presented a duplication on 10q24.2 with an estimated size of 175 kb; D: Case 614 presented a duplication on 17q24.3 with an estimated size of249 kb.

1759L. Yang et al. / Biochimica et Biophysica Acta 1842 (2014) 1755–1761

1760 L. Yang et al. / Biochimica et Biophysica Acta 1842 (2014) 1755–1761

we found that the interaction of GLI3 with the SKI family protein SKIL issignificantly compromised by the p.M111T mutation of GLI3. This is amilder functional defect than haploinsufficiency; therefore, it maycontribute to a less severe and atypical phenotype of Pallister–Hallsyndrome.

PHS is a disorder that affects the development of many parts of thebody that is characterized by hypothalamic hamartoma, central andpostaxial polydactyly, imperforate anus, renal anomalies, and foregut-related anomalies. Although bifid epiglottis is a common finding inPHS, posterior laryngeal and tracheoesophageal cleft are uncommonin this disorder [30]. Likewise, hemivertebrae is not a major criterionfor diagnosis of PHS and is present only in a few cases [31,32]. We con-clude that the missense mutation in the repressor domain of the GLI3gene is associated with esophageal atresia and hemivertebrae, whichare partially consistent with PHS.

Zentner et al. [33] reviewed 254 individuals and found that approx-imately 19% of CHD7 mutation-positive CHARGE syndrome patientsalso present with EA/TEF. Case 328, a female patient with esophagealatresia III and patent foramen ovale, carried the transition c.2360C N T,which causes the heterozygous variant p.787S N F in exon 4 of CHD7(NM_017780.3). The p.787S N F residue is proximal to the chromatin or-ganizationmodifier (chromo) domain,which is a conserved region. Un-fortunately, the parents were not available to test for the CHD7 p.787S NF variant in this study. Although it remains to be confirmed, we predict-ed that this variant may contribute to EA/TEF.

In this study, we analyzed genomic copy number variants andscreened six esophageal atresia-related genes for protein-coding vari-ants. However, several limitations in our study may still be worthdiscussing and may be improved in the future. For example, severalother genes, such as EFTUD2 and HOXD13, which are known to causeEA, will be included in further studies. Secondly, the role of GLI3 in thepathogenesis of EA/TEF should be further investigated by extendingthe same analysis to a larger sample size and by conducting the studyin vivo.

In summary, we detected a de novo heterozygous mutation in theGLI3 gene in a patient with esophageal atresia and hemivertebrae. Thismutation affected the protein–protein interaction between GLI3 andSKIL, which suggests the functional relevance of the mutation. To ourknowledge, this is the first report that links a GLI3 gene mutation toesophageal atresia in humans, which has been suggested in studieswith animal models. Our findings suggest that esophageal atresiashould be added to the clinical spectrum of phenotypes caused bymutations in GLI3.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbadis.2014.05.001.

Conflict of interest

The authors have no conflicts of interest to disclose.

Funding source

This work was supported by grants from the Shanghai Munici-pal Commission of Health and Family Planning (2013-018), theScience and Technology Commission of Shanghai Municipality(11DZ1950302) and the Innovation Grant of Fudan University(EYF156020).

Acknowledgments

The authorswould like to express their gratitude to the participatingpatients and families and would also like to thank Dr. Yonghui Jiang forhis critical reading of the manuscript.

References

[1] C. Shaw-Smith, Oesophageal atresia, tracheo-oesophageal fistula, and the VACTERLassociation: review of genetics and epidemiology, J. Med. Genet. 43 (2006)545–554.

[2] J.F. Felix, D. Tibboel, A. de Klein, Chromosomal anomalies in the aetiology of oesoph-ageal atresia and tracheo-oesophageal fistula, Eur. J. Med. Genet. 50 (2007)163–175.

[3] D. Genevieve, L. de Pontual, J. Amiel, S. Lyonnet, Genetic factors in isolated andsyndromic esophageal atresia, J. Pediatr. Gastroenterol. Nutr. 52 (Suppl. 1) (2011)S6–S8.

[4] S. Bianca, M. Bianca, G. Ettore, Oesophageal atresia and Down syndrome, DownsSyndr. Res. Pract. 8 (2002) 29–30.

[5] M.C. Digilio, B. Marino, P. Bagolan, A. Giannotti, B. Dallapiccola, Microdeletion 22q11and oesophageal atresia, J. Med. Genet. 36 (1999) 137–139.

[6] A.J. Marsh, D. Wellesley, D. Burge, M. Ashton, C. Browne, N.R. Dennis, K. Temple,Interstitial deletion of chromosome 17 (del(17) (q22q23.3)) confirms a link withoesophageal atresia, J. Med. Genet. 37 (2000) 701–704.

[7] P. Stankiewicz, P. Sen, S.S. Bhatt, M. Storer, Z. Xia, B.A. Bejjani, Z. Ou, J. Wiszniewska,D.J. Driscoll, M.K. Maisenbacher, J. Bolivar, M. Bauer, E.H. Zackai, D. McDonald-McGinn, M.M. Nowaczyk, M. Murray, V. Hustead, K. Mascotti, R. Schultz, L. Hallam,D. McRae, A.G. Nicholson, R. Newbury, J. Durham-O'Donnell, G. Knight, U. Kini,T.H. Shaikh, V. Martin, M. Tyreman, I. Simonic, L. Willatt, J. Paterson, S. Mehta, D.Rajan, T. Fitzgerald, S. Gribble, E. Prigmore, A. Patel, L.G. Shaffer, N.P. Carter, S.W.Cheung, C. Langston, C. Shaw-Smith, Genomic and genic deletions of the FOX genecluster on 16q24.1 and inactivating mutations of FOXF1 cause alveolar capillarydysplasia and other malformations, Am. J. Hum. Genet. 84 (2009) 780–791.

[8] D. Genevieve, L. de Pontual, J. Amiel, S. Sarnacki, S. Lyonnet, An overview of isolatedand syndromic oesophageal atresia, Clin. Genet. 71 (2007) 392–399.

[9] J.J. Johnston, I. Olivos-Glander, C. Killoran, E. Elson, J.T. Turner, K.F. Peters, M.H.Abbott, D.J. Aughton, A.S. Aylsworth, M.J. Bamshad, C. Booth, C.J. Curry, A. David,M.B. Dinulos, D.B. Flannery, M.A. Fox, J.M. Graham, D.K. Grange, A.E. Guttmacher,M.C. Hannibal, W. Henn, R.C. Hennekam, L.B. Holmes, H.E. Hoyme, K.A. Leppig,A.E. Lin, P. Macleod, D.K. Manchester, C. Marcelis, L. Mazzanti, E. McCann, M.T.McDonald, N.J. Mendelsohn, J.B. Moeschler, B. Moghaddam, G. Neri, R. Newbury-Ecob, R.A. Pagon, J.A. Phillips, L.S. Sadler, J.M. Stoler, D. Tilstra, V.C. Walsh, E.H.Zackai, T.M. Zadeh, L. Brueton, G.C. Black, L.G. Biesecker, Molecular and clinicalanalyses of Greig cephalopolysyndactyly and Pallister–Hall syndromes: robustphenotype prediction from the type and position of GLI3 mutations, Am. J. Hum.Genet. 76 (2005) 609–622.

[10] T. Roscioli, D. Kennedy, J. Cui, B. Fonseca, G.F. Watson, J. Pereira, Y.G. Xie, D. Mowat,Pallister–Hall syndrome: unreported skeletal features of a GLI3 mutation, Am. J.Med. Genet. A 136A (2005) 390–394.

[11] L. Pelz, K. Unger, M. Radke, Esophageal stenosis in acrocephalosyndactyly type I, Am.J. Med. Genet. 53 (1994) 91.

[12] W. Reardon, X.P. Zhou, C. Eng, A novel germline mutation of the PTEN gene in apatient with macrocephaly, ventricular dilatation, and features of VATER associa-tion, J. Med. Genet. 38 (2001) 820–823.

[13] R.M. Linzmeier, T. Ganz, Human defensin gene copy number polymorphisms: com-prehensive analysis of independent variation in alpha- and beta-defensin regions at8p22–p23, Genomics 86 (2005) 423–430.

[14] M. Meins, J. Lehmann, F. Gerresheim, J. Herchenbach, M. Hagedorn, K. Hameister, J.T.Epplen, Submicroscopic duplication in Xq28 causes increased expression of theMECP2 gene in a boy with severe mental retardation and features of Rett syndrome,J. Med. Genet. 42 (2005) e12.

[15] P. Dai, H. Akimaru, Y. Tanaka, T. Maekawa, M. Nakafuku, S. Ishii, Sonic Hedgehog-induced activation of the Gli1 promoter is mediated by GLI3, J. Biol. Chem. 274(1999) 8143–8152.

[16] P. Dai, T. Shinagawa, T. Nomura, J. Harada, S.C. Kaul, R. Wadhwa, M.M. Khan, H.Akimaru, H. Sasaki, C. Colmenares, S. Ishii, Ski is involved in transcriptional regula-tion by the repressor and full-length forms of Gli3, Genes Dev. 16 (2002)2843–2848.

[17] I. Ljuslinder, B. Malmer, I. Golovleva, M. Thomasson, K. Grankvist, T. Hockenstrom, S.Emdin, Y. Jonsson, H. Hedman, R. Henriksson, Increased copy number at 3p14 inbreast cancer, Breast Cancer Res. 7 (2005) R719–R727.

[18] L. Chang, R. Shi, T. Yang, F. Li, G. Li, Y. Guo, B. Lang, W. Yang, Q. Zhuang, H. Xu, Res-toration of LRIG1 suppresses bladder cancer cell growth by directly targeting EGFRactivity, J. Exp. Clin. Cancer Res. 32 (2013) 101.

[19] W.M. Yang, Z.J. Yan, Z.Q. Ye, D.S. Guo, LRIG1, a candidate tumour-suppressor gene inhuman bladder cancer cell line BIU87, BJU Int. 98 (2006) 898–902.

[20] C. Bonnet, J. Andrieux, M. Beri-Dexheimer, B. Leheup, O. Boute, S. Manouvrier, B.Delobel, H. Copin, A. Receveur, M. Mathieu, G. Thiriez, C. Le Caignec, A. David, M.C.de Blois, V. Malan, A. Philippe, V. Cormier-Daire, L. Colleaux, E. Flori, H. Dollfus, V.Pelletier, C. Thauvin-Robinet, A. Masurel-Paulet, L. Faivre, M. Tardieu, N. Bahi-Buisson, P. Callier, F. Mugneret, P. Edery, P. Jonveaux, D. Sanlaville, Microdeletionat chromosome 4q21 defines a new emerging syndrome with marked growthrestriction, mental retardation and absent or severely delayed speech, J. Med.Genet. 47 (2010) 377–384.

[21] B.I. Dimitrov, T. de Ravel, J. Van Driessche, C. de Die-Smulders, A. Toutain, J.R.Vermeesch, J.P. Fryns, K. Devriendt, P. Debeer, Distal limb deficiencies, micrognathiasyndrome, and syndromic forms of split hand foot malformation (SHFM) arecaused by chromosome 10q genomic rearrangements, J. Med. Genet. 47 (2010)103–111.

[22] E.H. Villavicencio, D.O. Walterhouse, P.M. Iannaccone, The sonic hedgehog-patched-gli pathway in human development and disease, Am. J. Hum. Genet. 67 (2000)1047–1054.

1761L. Yang et al. / Biochimica et Biophysica Acta 1842 (2014) 1755–1761

[23] J. Motoyama, J. Liu, R. Mo, Q. Ding, M. Post, C.C. Hui, Essential function of Gli2and Gli3 in the formation of lung, trachea and oesophagus, Nat. Genet. 20 (1998)54–57.

[24] A. Jamsheer, A. Sowinska, T. Trzeciak, M. Jamsheer-Bratkowska, A. Geppert, A. Latos-Bielenska, Expanded mutational spectrum of the GLI3 gene substantiates genotype-phenotype correlations, J. Appl. Genet. 53 (2012) 415–422.

[25] S. Kang, J.M. Graham Jr., A.H. Olney, L.G. Biesecker, GLI3 frameshift mutations causeautosomal dominant Pallister–Hall syndrome, Nat. Genet. 15 (1997) 266–268.

[26] J.J. Johnston, I. Olivos-Glander, C. Killoran, E. Elson, J.T. Turner, K.F. Peters, M.H.Abbott, D.J. Aughton, A.S. Aylsworth, M.J. Bamshad, C. Booth, C.J. Curry, A. David,M.B. Dinulos, D.B. Flannery, M.A. Fox, J.M. Graham, D.K. Grange, A.E. Guttmacher,M.C. Hannibal, W. Henn, R.C. Hennekam, L.B. Holmes, H.E. Hoyme, K.A. Leppig, A.E. Lin, P. Macleod, D.K. Manchester, C. Marcelis, L. Mazzanti, E. McCann, M.T.McDonald, N.J. Mendelsohn, J.B. Moeschler, B. Moghaddam, G. Neri, R. Newbury-Ecob, R.A. Pagon, J.A. Phillips, L.S. Sadler, J.M. Stoler, D. Tilstra, C.M. Walsh Vockley,E.H. Zackai, T.M. Zadeh, L. Brueton, G.C. Black, L.G. Biesecker, Molecular and clinicalanalyses of Greig cephalopolysyndactyly and Pallister-Hall syndromes: robust phe-notype prediction from the type and position of GLI3 mutations, Am J Hum Genet76 (2005) 609–622.

[27] R. Tsanev, K. Vanatalu, J. Jarvet, R. Tanner, K. Laur, P. Tiigimagi, B.B. Kragelund, T.Osterlund, P. Kogerman, The transcriptional repressor domain of gli3 is intrinsicallydisordered, PLoS One 8 (2013) e76972.

[28] S. Driess, K. Freese, D. Bornholdt, A. Kobelt, W. Kress, G. Mortier, U. Radhakrishna,S.E. Antonarakis, A. Rauch, M. Suri, J.B. Verheij, H. Woerle, K.H. Grzeschik, M. Kalff-Suske, Gene symbol: GLI3. Disease: Greig cephalopolysyndactyly syndrome, Hum.Genet. 112 (2003) 103.

[29] M. Kalff-Suske, A. Wild, J. Topp, M. Wessling, E.M. Jacobsen, D. Bornholdt, H. Engel,H. Heuer, C.M. Aalfs, M.G. Ausems, R. Barone, A. Herzog, P. Heutink, T. Homfray, G.Gillessen-Kaesbach, R. Konig, J. Kunze, P. Meinecke, D. Muller, R. Rizzo, S. Strenge,A. Superti-Furga, K.H. Grzeschik, Point mutations throughout the GLI3 gene causeGreig cephalopolysyndactyly syndrome, Hum. Mol. Genet. 8 (1999) 1769–1777.

[30] F. Ondrey, A. Griffith, C. Van Waes, S. Rudy, K. Peters, L. McCullagh, L.G. Biesecker,Asymptomatic laryngeal malformations are common in patients with Pallister–Hall syndrome, Am. J. Med. Genet. 94 (2000) 64–67.

[31] F. Guimiot, P. Marcorelles, A. Aboura, G. Bonyhay, S. Patrier, F. Menez, V. Drouin-Garraud, V. Icowick, D. Eurin, C. Garel, H. Moirot, E. Verspyck, P. Saugier-Veber, T.Attie-Bitach, O. Picone, J.F. Oury, A. Verloes, A.L. Delezoide, A. Laquerriere, Giantdiencephalic harmartoma and related anomalies: a newly recognized entity distinctfrom the Pallister–Hall syndrome, Am. J. Med. Genet. A 149A (2009) 1108–1115.

[32] J.J. Johnston, J.C. Sapp, J.T. Turner, D. Amor, S. Aftimos, K.A. Aleck, M. Bocian, J.N.Bodurtha, G.F. Cox, C.J. Curry, R. Day, D. Donnai, M. Field, I. Fujiwara, M. Gabbett,M. Gal, J.M. Graham, P. Hedera, R.C. Hennekam, J.H. Hersh, R.J. Hopkin, H.Kayserili, A.M. Kidd, V. Kimonis, A.E. Lin, S.A. Lynch, M. Maisenbacher, S. Mansour,J. McGaughran, L. Mehta, H. Murphy, M. Raygada, N.H. Robin, A.F. Rope, K.N.Rosenbaum, G.B. Schaefer, A. Shealy, W. Smith, M. Soller, A. Sommer, H.J. Stalker,B. Steiner, M.J. Stephan, D. Tilstra, S. Tomkins, P. Trapane, A.C. Tsai, M.I. Van Allen,P.C. Vasudevan, B. Zabel, J. Zunich, G.C. Black, L.G. Biesecker, Molecular analysis ex-pands the spectrum of phenotypes associated with GLI3 mutations, Hum Mutat 31(2010) 1142–1154.

[33] G.E. Zentner, W.S. Layman, D.M. Martin, P.C. Scacheri, Molecular and phenotypicaspects of CHD7 mutation in CHARGE syndrome, Am. J. Med. Genet. A 152A(2010) 674–686.