Background

Historically, conditions classified as ectodermal dysplasia (ED) have exhibited genetic and clinical heterogeneity. A unifying factor among these conditions is the genetic basis of developmental anomalies in tissues originating from the ectoderm [1]. The ectoderm gives rise to various structures, including the epidermis, central and peripheral nervous systems, placodes (including cranial placodes), and neural crest cells [2]. Early classification systems, such as those formulated by Dr. Newton Freire-Maia, categorized EDs based on their inheritance patterns and phenotypic characteristics. The Freire-Maia classification has significantly contributed to our understanding of these disorders and directed approaches to their management. The conditions were classified into two groups based on the extent of tissue involvement: group A, characterized by cases where at least two critical ectodermal tissues, such as hair, teeth, nails, eccrine and other glands, were affected, and group B, encompassing conditions that involved one of the aforementioned tissues along with at least one other ectodermal derivative [2].

Advancements in our understanding of the human genome have revealed molecular correlations among certain ED conditions beyond their traditional categorization based on phenotypes. The identification of causative genetic mutations in genes such as EDA, EDAR, EDARADD, TRAF6, and NFkB pathway-related genes has provided insights into the genetic foundation of EDs [3, 4]. These genetic mutations can yield comparable clinical phenotypes, and heterozygosity, compound heterozygosity, and homozygosity can contribute to these disorders. Almost 50% of historically classified EDs are now recognized to have causative genetic mutations encompassing those underlying more common ED conditions. Over the last decade, significant developments have been made in molecular methodologies that facilitate diagnostic differentiation for EDs [5]. Given the expanding therapeutic possibilities, these diagnoses have significant clinical value. Therefore, it is crucial to establish effective strategies for molecular diagnosis tailored to each patient. The choice between targeted panel sequencing (TPS) and whole exome sequencing (WES) should be guided by specific phenotypes observed in ED cases, emphasizing the need for a well-defined strategy.

Materials and methods

Patients and enrolment criteria

We enrolled 27 patients with ED between August 2018 and October 2022. We obtained 15 samples from 11 clinical sites throughout South Korea, which were sent to the coordinating center of the Korean Genetic Diagnosis Program for Rare Diseases (KGDP) Phase II, Molecular Diagnostics Laboratory, Seoul National University Hospital (SNUH) (https://www.ncbi.nlm.nih.gov/gtr/labs/320228/). The remaining 12 samples were obtained from patients with ED who were seen at the Seoul National University Hospital. The inclusion criteria encompassed the enrollment of individuals manifesting at least one of the symptoms associated with ectodermal dysplasia. All genetic tests were conducted at the Molecular Diagnostics Laboratory, and the results were interpreted by clinical pathologists with expertise in molecular diagnostics. All study procedures were approved by the Institutional Review Board of SNUH (IRB number: 2212-114-1389), and the study adhered to the Declaration of Helsinki for biomedical research involving human subjects.

Workflow of the mutation screening

Among the 27 patients, all 15 patients enrolled by KGDP underwent WES. Additionally, among the remaining 12 patients, those referred before September 2021 underwent TPS, while those referred after October 2021 underwent WES. Furthermore, two patients (ED4, ED26) who showed copy number variations (CNVs) by bioinformatics tool for global normalization (as suggested by NextGENe; SoftGenetics) underwent multiplex ligation-dependent probe amplification (MLPA).

Targeted panel sequencing (TPS)

Of the 27 patients, 4 (14.8%) underwent TPS to analyze representative genes (EDA, EDAR, EDARADD, LTBP3, MSX1, NFKBIA, PAX9, WNT10A) associated with ED. TPS was performed as described below. DNA was extracted from peripheral blood using a Chemagic 360 instrument (Perkin Elmer, Baesweiler, Germany). Library preparation was performed according to the SureSelect QXT target enrichment protocol (Agilent Technologies). Paired-end 150 bp sequencing was performed using the MiSeq Dx platform (Illumina, San Diego, California, USA). The generated sequencing data were then aligned to the human reference genome sequence (GRCh37/hg19), with subsequent identification and annotation of qualified variants using NextGENe V.2.4.0.1 (SoftGenetics). The detected variants were subjected to variant classification in accordance with the American College of Medical Genetics and Genomics (ACMG) guidelines [6]. Furthermore, to predict CNVs from the targeted NGS data, a bioinformatics tool for global normalization (as suggested by NextGENe; SoftGenetics) was employed for the log2 ratio calculation.

Whole exome sequencing

DNA was isolated from the peripheral blood of the remaining 23 patients (85.2%) for WES using a Chemagic 360 instrument (Perkin Elmer, Baesweiler, Germany). The extracted DNA was fragmented using a Covaris E220 focused ultrasonicator (Covaris, Woburn, MA, USA). DNA fragments were targeted and captured using the Agilent SureSelect All Exon V8 (Agilent Technologies, Santa Clara, CA, USA). A total of 500 ng of genomic DNA was used as the input. Library preparation was performed using the SureSelect XT Target Enrichment Protocol (Agilent Technologies). Paired-end 150 bp sequencing was conducted using the NovaSeq 6000 platform (Illumina). Bioinformatics processes from alignment to annotation were performed using NextGene (Version 2.4.0.1; Software Genetics, State College, PA, USA).

Multiplex ligation-dependent probe amplification (MLPA)

DNA denaturation, probe hybridization, ligation, and PCR of the ligated probes were conducted in accordance with the manufacturer’s instructions. The amplified products were subsequently analyzed using an ABI 3130xl capillary sequencer (Applied Biosystems, Foster City, CA, USA). The GeneMarker V.1.51 software (SoftGenetics) was employed for the determination of fragment length and copy number for each fragment.

Variant interpretation

All identified variants were categorized according to the five-tier system delineated by the ACMG, which encompasses pathogenic, likely pathogenic, variants of uncertain significance (VUS), likely benign, and benign classifications. Variants falling into the pathogenic or likely pathogenic categories were deemed clinically significant. Multiple databases, including HGMD (www.hgmd.org), ClinVar (www.ncbi.nlm.nih.gov/clinvar/intro/), and OMIM (https://www.omim.org/), were consulted for reference [6].

Statistical analysis

Categorical variables were presented as counts and percentages (%) and analyzed using the χ2 test. All statistical analyses were performed using the R V.4.0.0 (The R Foundation; www.r-project.org). Statistical significance was established for p-values less than 0.05.

Results

Descriptive analysis of mutation-positive ED cases

Descriptive statistics for the 20 positive cases are presented in Table 1. Hypotrichosis was observed in all but one case (ED16), and anhidrosis/hypohidrosis was present in all but two cases (ED16 and ED23). Hypodontia/anodontia was observed in all cases except three (ED6, 9, and 23). For ED23, accurate tooth assessment is challenging because of delayed dentition. Altogether cases carrying mutations in the EDA and EDAR genes were 16, accounting for 80% (16/20) of the total positive cases. Among these, EDA accounted for the majority (13 cases), and notably, CNVs were responsible for three cases (23.1%). The remaining genes in which mutations were detected included one case each of ERCC2, DSG4, LRP6, and LIPH.

Table 1 Descriptive statistics of positive ectodermal dysplasia cases (n = 20)
Full size table

Cases with EDA/EDAR mutations

EDA and EDAR mutations were detected in 16 cases, including 15 male patients and 1 female patient, who tested positive for EDAR. The average patient age was 15.5 years. In positive cases, mutation sites within EDA and EDAR were annotated in the functional domains. (Fig. 1). Missense mutations were localized toward the termini of the established furin and TNF domains, whereas loss-of-function (LOF) mutations were predominantly clustered within the COL domain. Concerning EDAR mutations, three alterations were detected in the death domain. Notably, a novel missense mutation in EDAR gene (NM_022336 c.1043T > C, p.L348P) was confirmed as a de novo mutation based on parental testing results (Supplementary Figure S1).

Fig. 1
figure 1

Genetic variants of EDA/EDAR in positive ectodermal dysplasia cases. Among 16 cases, 15 sequence variants were identified (p.G201Rfs*39 was detected in both ED10 and ED20). Three EDA CNVs are shown at the top left. Among the four frameshift mutations, three were located in the COL domain of the EDA gene, two missense mutations were in the furin domain, and three were at the end of the TNF domain. Three missense EDAR mutations were identified: two in the C-terminal end of the DD and one slightly upstream of the DD domain. COL, collagen-like domain; TNF, tumor necrosis factor; DD, death domain

Full size image

Cases with mutations in genes other than EDA/EDAR

ED6 had biallelic mutations on ERCC2. He is a 5-year-old boy who exhibited classic symptoms including hypotrichosis and hypohidrosis. Hypodontia was not prominent; instead, dysmorphic teeth were observed. Notably, he had a history of frequent febrile episodes, along with the presence of micropenis accompanied by underdeveloped scrotum, and adrenal hemorrhage during the neonatal period (Table 1). ERCC2 is associated with xeroderma pigmentosum group D (MIM: 278730) in an autosomal recessive manner. Moreover, biallelic mutations in ERCC2 can also lead to photosensitive trichothiodystrophy. Two mutations in ERCC2 were identified: a frameshift mutation and a missense mutation that was initially categorized as VUS. However, subsequent parental testing led to the segregation and reclassification of these two mutations as likely pathogenic, thereby confirming the genetic basis of the disease.

ED9 had a homozygous mutation, c.574T > C, p.Ser192Pro, in the DSG4 gene. DSG4 is the causal gene of autosomal recessive hypotrichosis 6 (MIM: 607903), which is inherited in an autosomal recessive manner. Interestingly, ED9 exhibited hypohidrosis, whereas typically DSG4-mutation positive patients show normal sweating.

ED16, a 4-year-old boy, presented with a normal complement of 20 deciduous teeth; however, routine X-rays revealed multiple missing permanent teeth (more than half absent). A de novo mutation in LRP6 was detected. This variant is a heterozygous pathogenic nonsense mutation (c.94 C > T, p.Arg32*) in the LRP6 gene. LRP6 is associated with the autosomal dominant tooth agenesis-selective 7 (MIM: 616724).

Lastly, ED23, a 7-year-old boy, exhibited hypotrichosis, although hypohidrosis was not observed. Hypodontia could not be confirmed due to delayed dentition. ED23 patient harbored a pathogenic mutation (c.736T > A, p.Cys246Ser, heterozygous) and a likely pathogenic frameshift variant (c.928del, p.Asp310Ilefs*4, heterozygous) in the LIPH gene. LIPH is associated with autosomal recessive hypotrichosis 7 (MIM: 604379).

Analysis of negative cases with ED symptoms

Table 2 displays 7 negative cases. Among these, all seven patients were male, with dental issues evident in six cases (85.7%), hypohidrosis/anhidrosis in two cases (28.6%), and hypotrichosis in three cases (42.9%). Furthermore, three patients exhibited symptoms unrelated to the ectodermal origin, such as optic neuropathy and failure to thrive. Novel VUS were identified in four patients. Specifically, WNT10A variants were observed in 2 cases, and EDA and KDF1 were present in 1 case. Except for the WNT10A c.511 C > T variant, the remaining three variants were predicted to be deleterious based on all three in silico prediction tools. The sequence variants identified in EDA and KDF1 are not annotated in the GnomAD database of human variations. Parental testing was only feasible for the KDF1 case, where the mother was found to be a carrier and had few permanent teeth.

Table 2 Descriptive statistics of negative ectodermal dysplasia cases (n = 7)
Full size table

Correlation between EDA/EDAR mutation detection and classic symptom presentation

Table 3 illustrates the mutation detection rates of EDA and EDAR based on the presence of hair, skin, or dental symptoms. Notably, of patients exhibiting all three symptoms, 94.1% had EDA/EDAR mutations; of those who did not show these three symptoms, none had EDA/EDAR mutations.

Table 3 Detection rate of EDA/EDAR mutations according to symptoms
Full size table

Discussion

To date, this is the largest molecular study conducted in a Korean population with suspected ED. The EDA and EDAR genes constitute the molecular basis of 74.1% of patients with ED. The detection of mutations in relatively novel genes, including DSG4, ERCC2, LIPH, and LRP6, which elucidate disorders characterized by only one or two classic symptoms of ED, underscores the efficacy of WES [7].

The number of genome-wide methods capable of diagnosing rare diseases has been increasing. However, to optimize the allocation of the limited funds available for clinical NGS diagnostics, it is essential to utilize existing resources in an efficient and economically viable manner [8]. Therefore, selecting an appropriate genetic test based on the phenotype of a rare disease is an efficient strategy. We demonstrated the presence of EDA and EDAR mutations in over 90% of patients exhibiting the three classical ectodermal symptoms.

In the case of one patient with VUS on EDA, the c.410 A > T variant was reported as a VUS with “insufficient evidence” in ClinVar. For EDA patients, this variant has been previously reported to be absent in the normal population and has been predicted to be deleterious by in silico prediction tools such as Sorting Intolerant From Tolerant (SIFT) [9], Polymorphism Phenotyping version 2 (PolyPhen-2) [10], and Combined Annotation-Dependent Depletion (CADD) [11]. Therefore, there is a possibility that it may be reclassified as likely pathogenic based on additional evidence, such as segregation study data, in the future.

However, in cases where classical ED symptoms are not fully exhibited, or when other tissue symptoms are present, such as in Group B, WES may be a favorable strategy. Furthermore, in cases of atypical ED, various genes have sporadically been observed to have mutations, and new genes continue to be discovered [12, 13]. Conversely, in cases where patients exhibit typical ED symptoms but test negative, techniques such as MLPA or WGS should be considered because of the possibility of mutations that are not easily detected by conventional NGS methods, such as CNVs or structural variations in EDA and EDAR.

Promising therapeutic avenues have emerged for patients with ED [14]. A drug-targeting strategy involving the neonatal Fc receptor has shown potential in addressing sweating deficiency associated with X-linked hypohidrotic ectodermal dysplasia, especially when administered during fetal development [15, 16]. Ongoing clinical trials of fetal therapy aim to validate the observed enhancements in male fetuses undergoing in utero treatment. Moreover, there are ongoing advancements in translational research on regenerative therapies for skin and corneal lesions using patient-derived stem cells, and dental interventions are being devised to enhance oral function in patients with EDs [5]. Thus, establishing diagnostic strategies for ED is becoming increasingly important. Our study sheds light on effective diagnostic strategies based on ED phenotype.

Conclusions

In conclusion, when performing molecular diagnostics for ED, it is recommended to choose targeted sequencing of EDA/EDAR mutations in cases with classical symptoms. In contrast, WES is considered an effective approach for cases lacking these typical symptoms.