Hypohidrotic ectodermal dysplasia (HED) is a congenital disorder characterized by abnormal morphogenesis of structures of ectodermal origin that may result in varying degrees of ectodermal defects such as hypodontia or anodontia, hypotrichiosis or alopecia, hypohidrosis or anhidrosis as well as nail abnormalities [1]. Conical teeth are a telltale sign in the HED patients, however other symptoms may vary between different individuals, even within the same family. The reduced or absent ability to regulate body temperature by sweating may lead to life-threatening hyperthermia, especially in very young children with an undiagnosed disorder.

The most prevalent form of HED is inherited as an X-linked condition [2] but autosomal dominant [3] and autosomal recessive [4] forms of the disorder are also found, albeit at a much lower frequency. According to the literature the most distinguishing feature for the HED variants appears to be the respective inheritance pattern [5].

HED may result from defects in any of three interacting proteins; ectodysplasin, EDAR or EDARADD. Ectodysplasin [6] is a soluble ligand protein that interacts with the transmembrane receptor EDAR, ectodysplasin 1 anhidrotic receptor [7–10] that is a death domain containing member of the tumor-necrosis factor receptor family. The intracellular part of EDAR subsequently binds to the adapter protein EDARADD (EDAR-associated death domain [11]) that links the receptor to downstream NF-κB signaling pathways [12].

Mutations in the gene coding for ectodysplasin, EDA situated on chromosome Xq12-q13.1, were the first to be found linked to HED [13]. The two most common splice isoforms of EDA (EDA-A1 and EDA-A2) bind the two distinct receptors EDAR and XEDAR respectively. They display different functions, where EDA-A1 appears to be essential for several skin appendages, and EDA-A2 probably is responsible for developmental timing and completion [14]. Later, both autosomal dominant and autosomal recessive variants of HED were found to be caused by different mutations in the EDAR gene on chromosome 2q11-q13 [15]. Autosomal recessive HED was also found to result from a mutation in the EDARADD gene on chromosome 1q42.2-q43 [16]. Interestingly, HED caused by mutations in either EDA or EDAR has been detected in several other species, including cattle [17], dogs [18] and mice [9, 19, 20]. Here we report the identification of a non-sense mutation in the EDAR gene in two large Swedish families with autosomal dominant (AD) HED.

In this report we have studied two presumed unrelated Swedish AD HED families stemming from the same small geographical region, and have identified an identical C to T mutation in exon 12 of the EDAR gene present in patients from both families. The mutation changes an arginine CGA codon into a TGA nonsense codon, thus introducing a truncation of the cytoplasmic portion of the EDAR protein at amino acid position 358. Thereby the p.Arg358X mutation leads to complete abolishment of the death domain region, amino acid 367 to amino acid 431, of the EDAR protein.

Two additional genetic variants were also detected in the PCR fragment that contained exon 12 of the EDAR gene; one polymorphism was a silent change, a T to C, just 15 nucleotides upstream of the p.Arg358X mutation, the other polymorphism, a T to G change, was located in the untranslated region downstream of the native stop codon of the EDAR gene. None of these polymorphisms appear to be associated with the AD HED disease, since the first variant was found both in a few healthy and in affected individuals and the second variant was only detected in two related healthy individuals, a parent and child. The most frequent sequence variant in both polymorphic loci, T, was found in homozygotic form at position c.1056 and position c.1389 in both healthy and affected individuals. Interestingly, the first polymorphism, c.1056 T to C was also detected in a previous study [15], but in that material the C allele was predominant and not the T allele as in our study. Likewise, in the published EDAR sequence [26] the nucleotide at position c.1389 is given as G, whereas in our material the G allele was only present as one copy each in two related individuals and all other individuals were homozygous for the T allele.

No known relationship exists between the two Swedish AD HED families, but since both families carry the same p.Arg358X mutation it is tempting to speculate that there is a previously unidentified relationship, especially when considering that the families originate from the same restricted geographical area.

All individuals affected by AD HED demonstrated absence of permanent teeth, albeit the actual number of missing teeth differed between the patients. Apart from the obligate odontological findings, other ectodermal symptoms were non-specific and, if present, generally mild. This is in contrast to a recently described family with AD HED where the affected individuals appear to have more pronounced symptoms overall [27]. Even though the actual mutation has not been identified in that family the disease gene maps to the same chromosomal region as the EDAR gene and it is very plausible that a mutation in the EDAR gene is the culprit in that AD HED family also.

The variable degree of expression of symptoms in individuals with AD HED is interesting and can probably be explained to some extent by different causative mutations. However, the variability of symptoms in teeth and other organs in individuals carrying the same EDAR mutation is intriguing and invites speculation to the cause. At least three different scenarios can be envisioned, two directly related to the nature of the EDAR mutation and one involving other gene products. The dominant negative action of the p.Arg358X mutation in AD HED could result from prevention of proper trimerization of the TNFR-like EDAR protein due to the lack of an intact death domain. Functional EDAR trimer complexes are required for accurate interactions with EDARADD and subsequent intracellular signaling [16]. Since EDAR trimers in AD HED patients most likely are formed by random combinations of normal and mutant variants of the protein, the proportion of functional EDAR homo-trimers in different individuals during embryogenesis and later developmental stages could vary. This variation could lead to a more or less severe clinical outcome due to chance deviations in the level of proper cell signaling in the ectoderm. An alternative explanation for the dominant-negative effect of the p.Arg358X mutation might be that the mutation leads to an altered subcellular location of the protein as compared to the non-mutated protein in analogy to what has been observed in studies of the orthologous mutated and wild-type Edar protein in mouse cell cultures [28]. A third scenario might be dissimilarities in the interactions between the mutant EDAR protein and different non-pathogenic variants of the ectodysplasin and EDARADD proteins.

This is the first study corroborating the p.Arg358X mutation in the EDAR gene and it supports the notion that EDAR is one of the key proteins in the development of HED. Thus, our report strengthens the role of this particular mutation in the etiology of autosomal dominant HED and confirms the importance of EDAR in the development of HED.