Eight Mutations of Three Genes (EDA, EDAR, and WNT10A) Identified in Seven Hypohidrotic Ectodermal Dysplasia Patients

Hypohidrotic ectodermal dysplasia (HED) is characterized by abnormal development of the teeth, hair, and sweat glands. Ectodysplasin A (EDA), Ectodysplasin A receptor (EDAR), and EDAR-associated death domain (EDARADD) are candidate genes for HED, but the relationship between WNT10A and HED has not yet been validated. In this study, we included patients who presented at least two of the three ectodermal dysplasia features. The four genes were analyzed in seven HED patients by PCR and Sanger sequencing. Five EDA and one EDAR heterozygous mutations were identified in families 1–6. Two WNT10A heterozygous mutations were identified in family 7 as a compound heterozygote. c.662G>A (p.Gly221Asp) in EDA and c.354T>G (p.Tyr118*) in WNT10A are novel mutations. Bioinformatics analyses results confirmed the pathogenicity of the two novel mutations. In family 7, we also identified two single-nucleotide polymorphisms (SNPs) that were predicted to affect the splicing of EDAR. Analysis of the patient’s total RNA revealed normal splicing of EDAR. This ascertained that the compound heterozygous WNT10A mutations are the genetic defects that led to the onset of HED. Our data revealed the genetic basis of seven HED patients and expended the mutational spectrum. Interestingly, we confirmed WNT10A as a candidate gene of HED and we propose WNT10A to be tested in EDA-negative HED patients.

There are three genes associated with HED that need to be confirmed. In 2012, Wisniewski and Trzeciak reported a patient with HED caused by c.252delG mutation in EDA2R [10]. In the same year, they reported another patient with HED caused by c.1074_1081del mutation in TRAF6 [11]. In 2010, Cluzeau et al. reported 10 patients with HED caused by mutations in WNT10A [12]. WNT10A mutations accounted for 16% of ED patients in Cluzeau's cohort [12].
In this study, we aimed to reveal the genetic basis of seven HED patients and test if WNT10A mutations contribute to the onset of HED in our patients. EDA, EDAR, EDARADD, and WNT10A genes were analyzed. We identified eight mutations, of which two are novel. We provided new evidence that WNT10A is a promising candidate gene of HED.

Subjects
This study consisted of seven unrelated HED families. Two families (families 6 and 7) had been tested for EDA mutation in our previous study, but no pathogenic mutation was found [5]. The other five families (families 1-5) were newly recruited ones. All of the patients were diagnosed as HED according to the criteria of having at least two of the three ectodermal abnormalities of teeth, hair, and sweat glands. Fifty healthy individuals were recruited as a normal control to rule out genetic polymorphism. This study was approved by the Ethical Review Committee of Guanghua School and Hospital of Stomatology, Sun Yat-sen University (ERC-[2013]-9, date of approval: 1 December 2013). Informed consent was obtained, and the Declaration of Helsinki was followed.

Mutation Detection of EDA, EDAR, EDARADD, and WNT10A
We collected about a 4-mL quantity of peripheral blood from each individual. Genomic DNA was extracted with QiaAmp Kit (Qiagen, Düsseldorf, Germany). Primers covering exons and flanking intronic sequences of EDAR, EDARADD, and WNT10A were designed using Oligo 6.0 (Molecular Biology Insights, Colorado Springs, CO, USA). These primer sequences were available in Tables S1-S3. The primer sequences of EDA were reported in previous work [5]. PCR was performed, and the products were sequenced with an ABI 3730XL genetic analyzer (Applied Biosystems, Foster City, CA, USA). The sequencing results were analyzed with Sequence Scanner Software v1.0 (Applied Biosystems, Foster City, CA, USA). Mutation nomenclature was used, with +1 corresponding to the A of the ATG translation initiation codon of the reference sequence NM_001399.4(EDA), NM_022336.3 (EDAR), and NM_025216.2 (WNT10A).

Clinical Report
The pedigrees are available in Figure 1. The probands of families 2, 3, 5, and 7 presented the triad of hypodontia, hypotrichosis, and hypohidrosis. In family 3, the proband displayed conically-shaped teeth. His grandfather also had these clinical features. Interestingly, the proband of family 7 did not have any missing primary tooth congenitally, while 15 teeth were missing a geneogenous in the permanent dentition ( Figure 2). His hair was sparse in his early childhood, according to a report by his mother. As he grew up, his hair became thicker. When he was 11 years old and came to our clinic, we found his hair was not significantly different from that of an average person. The probands of families 1 and 4 manifested hypodontia and hypotrichosis. The proband of family 6 suffered from hypodontia and hypohidrosis, similarly to his mother. The number of the missing teeth in primary and permanent dentition is listed in Table 1. (More clinical information is available in Figure S1 and Table S4.) We used the structure of protein Wnt-8 (PDB ID 4F0A; X-ray, resolution 3.25Å) as a homology model to perform structural modeling with Swiss PdbViewer v4.1 [22]. PyMol v1.5.0.3 (The PyMOL Molecular Graphics System, Version 1.8 Schrödinger, LLC.) was used to visualize the 3D structure of the wild-type Wnt-10a protein and mutated Wnt-10a protein.

Clinical Report
The pedigrees are available in Figure 1. The probands of families 2, 3, 5, and 7 presented the triad of hypodontia, hypotrichosis, and hypohidrosis. In family 3, the proband displayed conicallyshaped teeth. His grandfather also had these clinical features. Interestingly, the proband of family 7 did not have any missing primary tooth congenitally, while 15 teeth were missing a geneogenous in the permanent dentition ( Figure 2). His hair was sparse in his early childhood, according to a report by his mother. As he grew up, his hair became thicker. When he was 11 years old and came to our clinic, we found his hair was not significantly different from that of an average person. The probands of families 1 and 4 manifested hypodontia and hypotrichosis. The proband of family 6 suffered from hypodontia and hypohidrosis, similarly to his mother. The number of the missing teeth in primary and permanent dentition is listed in Table 1. (More clinical information is available in Figure S1 and Table S4.)  (p.Thr338Met) and c.895G>A (p.Gly299Ser). In family 6, Ectodysplasin A receptor (EDAR) mutation c.1259G>A (p.Arg420Gln) was identified. WNT10A mutation c.354T>G (p.Tyr118*) and c.637G>A (p.Gly213Ser) were identified in family 7. Black arrows point to the probands. Red arrows point to the mutations. "+" indicates that the blood sample is available for genetic analysis in this study. In family 1, the sequence chromatogram showed reverse complemented sequence. The remaining sequence chromatograms showed forward sequence.

Figure 2.
Panoramic radiograph of proband in family 7. He did not have any missing primary teeth congenitally (#51 and #61 had replaced by #11 and #21). However, 15 teeth were missing in the permanent dentition.

Genetic Findings of the EDA, EDAR, EDARADD, and WNT10A
We identified eight mutations in seven families ( Figure 1; Table 1). Two of the eight mutations were novel mutations. The two novel mutations were not carried by healthy volunteers and were not reported by ExAC, 1000 Genomes, PubMed, or the Human Gene Mutation Database (HGMD) public version (accession date: 3 August 2016) [23]. Families 1-5 harbored heterozygous mutations in EDA. In family 6, we identified one EDAR mutation. In family 7, two WNT10A mutations and two EDAR single-nucleotide polymorphisms (SNPs) were found.

Family 1
A novel mutation, c.662G>A (p.Gly221Asp) in EDA gene, is identified in family 1. It was located at the collagen-like domain (Figure 3), and led to the replacement of glycine by aspartic acid at amino Figure 2. Panoramic radiograph of proband in family 7. He did not have any missing primary teeth congenitally (#51 and #61 had replaced by #11 and #21). However, 15 teeth were missing in the permanent dentition. We identified eight mutations in seven families ( Figure 1; Table 1). Two of the eight mutations were novel mutations. The two novel mutations were not carried by healthy volunteers and were not reported by ExAC, 1000 Genomes, PubMed, or the Human Gene Mutation Database (HGMD) public version (accession date: 3 August 2016) [23]. Families 1-5 harbored heterozygous mutations in EDA. In family 6, we identified one EDAR mutation. In family 7, two WNT10A mutations and two EDAR single-nucleotide polymorphisms (SNPs) were found.

Family 1
A novel mutation, c.662G>A (p.Gly221Asp) in EDA gene, is identified in family 1. It was located at the collagen-like domain (Figure 3), and led to the replacement of glycine by aspartic acid at amino acid residue 221. A cross-species alignment of protein sequences showed that p.Gly221 was evolutionarily conserved (Figure 4). SIFT, PolyPhen2, Mutation Taster, FATHMM, and PROVEAN predicted the mutational effect to be damaging, probably damaging, disease-causing, damaging and deleterious, respectively.

Family 7
Identified in family 7, c.354T>G (p.Tyr118*) mutation in the WNT10A gene was carried by the patient and his father, while c.637G>A (p.Gly213Ser) mutation was shared by the patient and his mother. The novel c.354T>G mutation truncated 300 amino acids from the C terminus of the Wnt-10a protein by inducing a termination codon at amino acid position 118. The mutated protein was predicted to undergo nonsense-mediated mRNA decay (NMD) by Mutation Taster [17]. Even if it was not decayed by NMD, the mutated Wnt-10a protein was unlikely to have any function with the majority of the protein structure lost ( Figure 5). c.637G>A mutation has been reported to cause oligodontia with minor signs of ectodermal dysplasia in autosomal recessive inheritance [27]. This is in accordance with our case.
There were two more variants, c.723G>A (p.Glu241Glu) and c.813T>C (p.Asp271Asp) in the EDAR gene, identified in the patient of family 7. Mutation Taster predicted the variants were disease causing mutations for they may change the splice site of EDAR. However, we considered them as SNPs for the following three reasons. Firstly, splicing analysis of the patient's total RNA showed normal splicing of EDAR ( Figure 6). Secondly, these two variants are on the same chromosome that was inherited from his mother, who showed no signs of HED. Thirdly, the homozygous individuals were recorded in 1000 Genomes and ExAC database. acid residue 221. A cross-species alignment of protein sequences showed that p.Gly221 was evolutionarily conserved (Figure 4). SIFT, PolyPhen2, Mutation Taster, FATHMM, and PROVEAN predicted the mutational effect to be damaging, probably damaging, disease-causing, damaging and deleterious, respectively.

Family 7
Identified in family 7, c.354T>G (p.Tyr118*) mutation in the WNT10A gene was carried by the patient and his father, while c.637G>A (p.Gly213Ser) mutation was shared by the patient and his mother. The novel c.354T>G mutation truncated 300 amino acids from the C terminus of the Wnt-10a protein by inducing a termination codon at amino acid position 118. The mutated protein was predicted to undergo nonsense-mediated mRNA decay (NMD) by Mutation Taster [17]. Even if it was not decayed by NMD, the mutated Wnt-10a protein was unlikely to have any function with the majority of the protein structure lost ( Figure 5). c.637G>A mutation has been reported to cause oligodontia with minor signs of ectodermal dysplasia in autosomal recessive inheritance [27]. This is in accordance with our case.
There were two more variants, c.723G>A (p.Glu241Glu) and c.813T>C (p.Asp271Asp) in the EDAR gene, identified in the patient of family 7. Mutation Taster predicted the variants were disease causing mutations for they may change the splice site of EDAR. However, we considered them as SNPs for the following three reasons. Firstly, splicing analysis of the patient's total RNA showed normal splicing of EDAR ( Figure 6). Secondly, these two variants are on the same chromosome that was inherited from his mother, who showed no signs of HED. Thirdly, the homozygous individuals were recorded in 1000 Genomes and ExAC database.

Family 7
Identified in family 7, c.354T>G (p.Tyr118*) mutation in the WNT10A gene was carried by the patient and his father, while c.637G>A (p.Gly213Ser) mutation was shared by the patient and his mother. The novel c.354T>G mutation truncated 300 amino acids from the C terminus of the Wnt-10a protein by inducing a termination codon at amino acid position 118. The mutated protein was predicted to undergo nonsense-mediated mRNA decay (NMD) by Mutation Taster [17]. Even if it was not decayed by NMD, the mutated Wnt-10a protein was unlikely to have any function with the majority of the protein structure lost ( Figure 5). c.637G>A mutation has been reported to cause oligodontia with minor signs of ectodermal dysplasia in autosomal recessive inheritance [27]. This is in accordance with our case.
There were two more variants, c.723G>A (p.Glu241Glu) and c.813T>C (p.Asp271Asp) in the EDAR gene, identified in the patient of family 7. Mutation Taster predicted the variants were disease causing mutations for they may change the splice site of EDAR. However, we considered them as SNPs for the following three reasons. Firstly, splicing analysis of the patient's total RNA showed normal splicing of EDAR ( Figure 6). Secondly, these two variants are on the same chromosome that was inherited from his mother, who showed no signs of HED. Thirdly, the homozygous individuals were recorded in 1000 Genomes and ExAC database.  EDAR gene of probands in families 6 and 7 was sequenced. Both probands carried a heterozygous C allele of SNP rs3827760.

Discussion
Although Cluzeau et al. reported 10 HED patients harboring WNT10A mutations [12], the disease-causing relationship between WNT10A mutations and HED was not confirmed by other scientists. Based on the information from OMIM (National Center for Biotechnology Information, Bethesda, MD, USA) and the latest update on ectodermal dysplasia clinical classification [3], WNT10A is not recommended as a candidate gene for HED. In this study, we reported a HED patient caused by compound heterozygous mutations in WNT10A. This is new evidence ascertaining WNT10A as a candidate gene for HED. With the newly-identified WNT10A-related HED patient in our study, we think this is a solid fact that WNT10A is a candidate gene for HED. We suggest WNT10A to be routinely tested in EDA-negative HED patients.
WNT10A mutation was reported to cause odonto-onycho-dermal dysplasia (OODD), a form of ED characterized by severe oligodontia, nail dystrophy, palmoplantar hyperkeratosis, and hyperhidrosis [28]. Schopf-Schulz-Passarge syndrome (SSPS) is a form of ED similar to OODD but distinguished by eyelid cysts [29]. Bohring et al. reported WNT10A mutations can cause not only OODD, but also SSPS and non-syndromic oligodontia [30]. Variability of phenotypes was shown to be inter-and intra-family [31][32]. This could be partly explained by the effect of some genetic variants.  EDAR gene of probands in families 6 and 7 was sequenced. Both probands carried a heterozygous C allele of SNP rs3827760.

Discussion
Although Cluzeau et al. reported 10 HED patients harboring WNT10A mutations [12], the disease-causing relationship between WNT10A mutations and HED was not confirmed by other scientists. Based on the information from OMIM (National Center for Biotechnology Information, Bethesda, MD, USA) and the latest update on ectodermal dysplasia clinical classification [3], WNT10A is not recommended as a candidate gene for HED. In this study, we reported a HED patient caused by compound heterozygous mutations in WNT10A. This is new evidence ascertaining WNT10A as a candidate gene for HED. With the newly-identified WNT10A-related HED patient in our study, we think this is a solid fact that WNT10A is a candidate gene for HED. We suggest WNT10A to be routinely tested in EDA-negative HED patients.
WNT10A mutation was reported to cause odonto-onycho-dermal dysplasia (OODD), a form of ED characterized by severe oligodontia, nail dystrophy, palmoplantar hyperkeratosis, and hyperhidrosis [28]. Schopf-Schulz-Passarge syndrome (SSPS) is a form of ED similar to OODD but distinguished by eyelid cysts [29]. Bohring et al. reported WNT10A mutations can cause not only OODD, but also SSPS and non-syndromic oligodontia [30]. Variability of phenotypes was shown to be inter-and intra-family [31][32]. This could be partly explained by the effect of some genetic variants. EDAR gene of probands in families 6 and 7 was sequenced. Both probands carried a heterozygous C allele of SNP rs3827760.

Discussion
Although Cluzeau et al. reported 10 HED patients harboring WNT10A mutations [12], the disease-causing relationship between WNT10A mutations and HED was not confirmed by other scientists. Based on the information from OMIM (National Center for Biotechnology Information, Bethesda, MD, USA) and the latest update on ectodermal dysplasia clinical classification [3], WNT10A is not recommended as a candidate gene for HED. In this study, we reported a HED patient caused by compound heterozygous mutations in WNT10A. This is new evidence ascertaining WNT10A as a candidate gene for HED. With the newly-identified WNT10A-related HED patient in our study, we think this is a solid fact that WNT10A is a candidate gene for HED. We suggest WNT10A to be routinely tested in EDA-negative HED patients.
WNT10A mutation was reported to cause odonto-onycho-dermal dysplasia (OODD), a form of ED characterized by severe oligodontia, nail dystrophy, palmoplantar hyperkeratosis, and hyperhidrosis [28]. Schopf-Schulz-Passarge syndrome (SSPS) is a form of ED similar to OODD but distinguished by eyelid cysts [29]. Bohring et al. reported WNT10A mutations can cause not only OODD, but also SSPS and non-syndromic oligodontia [30]. Variability of phenotypes was shown to be inter-and intra-family [31,32]. This could be partly explained by the effect of some genetic variants. The SNP rs3827760 (c.1109T>C, p.Val370Ala) has been associated with increased hair thickness [33]. In this study, we sequenced the EDAR gene of the probands in families 6 and 7, and both of them carried the SNP. Their phenotype of hair was consistent with the genotype. Lately, OODD and SSPS have been proven to be the same disease with variable symptoms caused by WNT10A mutations [3,31,32]. The clinical feature of HED is milder than OODD but more serious than oligodontia. We consider OODD, HED and oligodontia represented the full clinical manifestations of WNT10A-related ED. It is not rare that mutations of the same gene lead to different phenotypes. For example, mutations in the EDA gene are the cause of some XLHED and non-syndromic oligodontia in patients, and the two diseases are considered to be the same disease with different degrees of severity [5,34].
In family 7, we identified a novel mutation c.354T>G (p.Tyr118*) and a known mutation c.637G>A (p.Gly213Ser) in WNT10A gene. The novel mutation c.354T>G was predicted to be decayed by NMD or truncate a major part of the protein. It is a loss-of-function mutation [35,36]. Interestingly, the most prevalent mutation c.321C>A (p.Cys107*) and c.682T>A (p.Phe228Ile) did not show up in our study. By reviewing the literature, we found that the two prevalent mutations were frequently reported in European, but not in Asian, subjects [30,[37][38][39][40]. Population frequency from the ExAC database also showed that the two mutations are most popular in European but are not carried by Asian subjects [41]. In contrast, although c.637G>A (p.Gly213Ser) was first reported in Europe, it is far more prevalent in Asia [41]. This information should be noted in genetic counseling and prenatal diagnosis. Surprisingly, the ExAC database recorded four c.637G>A (p.Gly213Ser) homozygotes in Asian and 14 c.682T>A (p.Phe228Ile) homozygotes in European subjects. We cannot inquire whether the 18 homozygotes have ectodermal defects. We consider ExAC is a useful database for disease studies, and it will be better if medical records of the subjects are available.
The novel EDA mutation, c.662G>A (p.Gly221Asp), is located in the collagen-like domain (Collagen) of Ectodysplasin A protein (Figure 3). Collagen functioned as a connector for the trimerization of Ectodysplasin A [42]. Schneider et al. conducted functional analyses of collagen, and revealed that point mutations in collagen can disrupt the trimerization of Ectodysplasin A [42]. Collagen is composed of 19 repeating peptide triplets of glycine-X-Y with a single interruption [42,43]. Intriguingly, the disease-causing mutations in Collagen have a strong bias to glycine. Twelve (86%) of the 14 mutations in Collagen are mutated at glycine [44]. This phenomenon highlights the important role of glycine in maintaining the structure and function of Collagen. As c.662G>A mutation leads to a glycine being replaced by an aspartic acid in Collagen, the mutated protein probably cannot accomplish trimerization.
In a previous study, we reported four HED patients harbored EDA mutations [5]. With the new data in this study, EDA mutations were detected in nine of 11 patients (82%). This is in accordance with other studies, which showed a detection rate of 57%-88% [42,[45][46][47]. We found mutations in only one patient for WNT10A and EDAR genes. This is not large enough to give a prevalence. Other studies showed that WNT10A or EDAR was mutated in 9%-16% of HED patients [12,46,47]. Reviewing the distribution of EDA mutations in our cohort, we found that all of the mutations were located in furin sites, Collagen and tumor necrosis factor (TNF) domains, except c.741G>A (p.Gln247Gln) (Figure 3). Mutation c.741G>A may also affect the functional domain because it is located before the TNF domain and predicted to alter the splicing of EDA.
ED is a group of diseases with great heterogeneity. To date, only 75 EDs are possible to connect a specific gene or chromosomal region [3]. The genetic diagnoses of the remaining 88 EDs are not possible to obtain. This puts us at a disadvantage because we may not treat the patient effectively if we do not know the genetic defect and the mechanism of the disease. The USA, Europe, and China have launched a precision medicine initiative [48,49]. Our study is an effort to achieve precise diagnosis and eventually to practice precision medicine.

Conclusions
We conducted a genetic study in seven HED families. Two novel and six known mutations were identified in the EDA, EDAR, and WNT10A genes. Bioinformatics analysis, structural modeling, or splicing analysis proved the pathogenicity of the mutations. This study revealed the genetic basis of seven HED patients and expanded the mutational spectrum. New evidence ascertaining WNT10A as a candidate gene for HED was provided.