From ectodermal dysplasia to selective tooth agenesis


  • How to cite this article: Mues GI, Griggs R, Hartung AJ, Whelan G, Best LG, Srivastava AK, D'Souza R. 2009. From ectodermal dysplasia to selective tooth agenesis. Am J Med Genet Part A 149A:2037–2041.


The history and the lessons learned from hypohidrotic ectodermal dysplasia (HED) may serve as an example for the unraveling of the cause and pathogenesis of other ectodermal dysplasia syndromes by demonstrating that phenotypically identical syndromes (HED) can be caused by mutations in different genes (EDA, EDAR, EDARADD), that mutations in the same gene (EDA) can lead to different phenotypes (HED and selective tooth agenesis) and that mutations in genes further downstream in the same signaling pathway (NEMO) may modify the phenotype quite profoundly (incontinentia pigmenti (IP) and HED with immunodeficiency). But it also demonstrates that diligent phenotype characterization and classification is extremely helpful in uncovering the underlying genotype. We also present a new mutation in the EDA gene which causes selective tooth agenesis and demonstrates the phenotype variation that can be encountered in the ectodermal dysplasia syndrome (HED) with the highest prevalence worldwide. © 2009 Wiley-Liss, Inc.


The skin with its appendages is an important organ in all higher organisms. It constitutes the primary barrier delineating self from the environmental challenges. The development of such a versatile and complicated structure requires a highly coordinated interaction of several genetic signaling pathways within and between the ectodermal and underlying (ecto)-mesenchymal layers of embryonic tissues. Malfunctions in any part of this system or at any time point during the development of the ectodermal structures can lead to the variety of phenotypically distinctive entities which we classify as the ectodermal dysplasia (ED) syndromes.

The genetic basis of the most prevalent ED syndrome: Hypohidrotic ectodermal dysplasia (HED) was recognized even before the concepts of genes and inheritance patterns had been formulated. It was in 1875 that Charles Darwin described the peculiar inheritance pattern of HED which was later recognized as being a classical X-linked recessive Mendelian pedigree type. This observation helped to identify the causative gene “EDA” by limiting the search to the X-chromosome.


The EDA gene was located to Xq12–13 by Zonana et al. 1988 and the entire human gene was isolated and characterized 10 years later [Kere et al., 1996; Bayés et al., 1998]. It spans about 425 kb and consists of 12 exons, of which 8 are utilized to form the two major EDA proteins EDA-A1 and EDA-A2. EDA-A1 is the 391 amino acid-long, full-length form and binds the receptor EDAR while EDA-A2 is just two amino acids shorter and binds exclusively to a different receptor named XEDAR. Since EDA-A2 and XEDAR do not seem to play a role in any human syndrome and an intact XEDAR signaling pathway cannot rescue an interruption in the EDAR pathway we will focus exclusively on EDA-A1 and EDAR. The EDA/EDAR pair belongs to the tumor necrosis factor (TNF) ligand/receptor family of proteins. Both are membrane proteins, but EDA's carboxy-terminal end which consists of a collagen-like domain and a TNF-like domain, is released by proteolytic cleavage. The collagen domain aids in the formation of EDA homo-trimers and even larger protein complexes which seem to facilitate interaction with the receptor [Elomaa et al., 2001].

The structure and function of the EDA receptor EDAR is typical of other TNFRs in that it interacts with special adaptor proteins to activate the IKK complex (IKKα, IKKβ and IKKγ/NEMO) and ultimately the NFκB transcription factor. One of these adapter proteins, EDARADD (EDAR-associated death domain) is uniquely used in the EDA signaling pathway, the others called TRAFs (TNF receptor associated factors) are also employed by other TNFRs.


Many different mutations have been identified in EDA and shown to cause X-linked HED [Bayés et al., 1998; Monreal et al., 1998; Schneider et al., 2001]. Most of these mutations affect the three main functional domains of the EDA protein: The furin cleavage site, the collagen-like multimerization domain and the TNF homology domain which mediates binding to the downstream receptor. Most of the mutations are missense mutations, some are nonsense and frameshift-causing mutations. There are also several in-frame deletions in the collagen-like repeat domain and whole exon deletions. Attempts to uncover a correlation between types and/or location of mutations and severity of the phenotype have been unsuccessful.

The impact of several EDA missense mutations on different aspects of protein function has been thoroughly investigated [Schneider et al., 2001] and it could be demonstrated that either proteolytic processing, glycosylation, multimerization, solubility or receptor-binding can be affected by the different mutations. Most mutations were found to result in a total elimination of receptor signaling.


Most of the currently identified EDA mutations have been detected through ascertainment and screening of patients and families with classical signs and symptoms of HED which include significant hypoplasia of a variety of ectodermal appendages such as hair, teeth, glands and occasionally nails. Undiagnosed infant boys with the HED syndrome can die of hyperthermia during minor febrile infections due to absent or severely decreased sweating. Glandular hypoplasia also affects the respiratory tract and the middle ear where it may lead to recurrent infections. The eyebrows are missing, scalp hair is sparse and fine with early balding; most other body hair is also sparse except for facial hair. The dentition is variably affected from hypodontia to anodotia. On the average 22 permanent teeth are missing, excepting wisdom teeth. The upper central incisors and first molars are most often spared, but frequently they are misshapen [Lexner et al., 2007, 2008]. Facial appearance is unusual with frontal bossing and depressed nasal bridge.

Mutations in any part of the EDA gene can cause the HED syndrome with no apparent correlation between the location of the mutation and the severity of the phenotype. Often there is variability among affected males of the same family. Carrier females are mostly free of symptoms or show a much milder phenotype with only one or two missing teeth, slightly smaller teeth, sparse hair and a patchy distribution of loss of sweating along Blaschko lines, presumably due to lyonization. However there are a few females with severe hypodontia and other symptoms of full blown HED. Non-random X-chromosome inactivation can explain only some of these cases.


The pedigree of several families with a classical HED phenotype did not fit an X-linked inheritance pattern; instead they showed autosomal dominant or autosomal recessive patterns. Of great help in the explanation of this observation were the spontaneously occurring mouse mutants with ED. Mice like humans can be affected with HED. The mouse tabby phenotype is X-linked, corresponds to human X-linked HED and is caused by mutations of the EDA-analogous Ta gene [Srivastava et al., 1997]. Autosomal dominant and recessive mouse mutants had been located to two different loci called downless (dl) and crinkled (cr). The dl gene was isolated first and proved to be the EDA receptor while the mouse cr gene turned out to code for the receptor-associated adapter protein called EDARADD in humans. Both genes are located on autosomes, in humans on chromosome 2q12 and 1q42.2–3, respectively. Mutations in either gene can present with an autosomal dominant or recessive HED inheritance pattern [Monreal et al., 1999], but EDAR mutations seem to be more common than EDARADD mutations [Chassaing et al., 2006].

Rearrangements and mutations in a fourth member of the EDA/NFκB signaling pathway, the NEMO gene at Xq28, can cause an HED-like syndrome with additional manifestations. This syndrome is called incontinentia pigmenti (IP) [The International Incontinentia Pigmenti Consortium, 2000] and occurs mainly in heterozygous females with skewed X-inactivation. The condition is prenatally lethal for affected males; the few surviving males present with HED and severe immunodeficiency.


Intense research in the area of tooth development and agenesis has uncovered in recent years that mutations in several genes can lead to selective tooth agenesis. These genes include PAX9, MSX1, AXIN2 and many others which have not yet been identified. Surprisingly, the search for an X-linked tooth agenesis gene led several researchers to the EDA gene [Tao et al., 2006; Tarpey et al., 2007; Fan et al., 2008; Han et al., 2008; Li et al., 2008; Rasool et al., 2008]. The affected males and few females in these families showed only variable degrees of tooth agenesis without the systemic symptoms of HED and hence were never diagnosed with any ED syndrome. We have characterized an additional family with selective tooth agenesis as only manifestation of an EDA mutation which will be described here for the first time.


Several members of two different families in an American Indian community were found to have dental abnormalities, apparently inherited in an X-linked fashion. The dental defects manifested primarily as missing or poorly developed incisors and occasionally canines. Unfortunately, sufficient information to reconstruct accurate phenotype patterns for all affected individuals could not be obtained. The affected members of both families do not show typical characteristics of X-linked HED. There was no indication of hair abnormalities or reduced numbers of sweat pores in affected individuals. Analysis of the X-linked ectodysplasin-A gene in two affected males in one kindred revealed a novel missense alteration, c.494G>C (p.G165A), in both individuals. The carrier females were heterozygous for the alteration and showed a skewed X chromosome inactivation pattern. In the second kindred, a female with an affected son carried the identical alteration. The finding of an identical mutation in both families suggests that the mutation is probably inherited from a common ancestor, as both these families are members of the same regional community. Informed consent was obtained from all participating family members.


Dental examination showed abnormal tooth development and hypodontia in individuals V-3, -7, -11, -17 and IV-13 involving mainly the incisors, but also the morphology of canines and premolars to some extent (Figs. 1A,B and 2A–C). IV-13 who was in his 30s at the time of the examination, had some thinning of the lateral aspect of his eyebrows and scant hair on his arms and chest and showed a lack of permanent upper and lower lateral incisors. Additional dental loss to caries and other causes had left him with only 10 upper and 4 lower teeth. None of the cases showed significant hair abnormality. V-17 who died at the age of 10, was reported to have suffered from some unexplained episodes of fever, possibly caused by eccrine dysfunction. He also suffered from mental retardation, chronic seizures and minor anomalies, however if these symptoms were the sequel of hyperthermia or due to additional, unrelated problems cannot be determined with certainty. His surviving brother is 24 years old, has dental records indicating congenital hypodontia; but is otherwise in good health. None of the other individuals have a history of heat intolerance or abnormal eccrine function. Sweat pore impressions on two of these cases (V-7 and IV-13) show normal sweat pore density. Nails appear to be normal. A carrier female (IV-7) has long scalp hair of normal quality, no congenital dental abnormalities and only mild nail dystrophy in the form of periodic, transverse, white subungal bands.

Figure 1.

Ectodysplasin-A gene alteration in family K7150 (A) and family K7155 (B), identified in this study. Open symbols represent normal individuals, and filled squares represent affected males. All available samples in family K7150 and a carrier female in family K7155 were analyzed. X inactivation (Xi) data are indicated. WT, wild-type allele; MUT, mutant allele. C: DNA sequence chromatograms of ectodysplasin-A exon 3 showing a “G” to “C” alteration at nucleotide 494 (c.494G → C) in an affected individual (V-7) in family K7150, and a normal c.494G in his father (IV-7-1). The carrier mother (IV-7) and a carrier female (III-2) in family K7155 are heterozygous for the alteration. This alteration is predicted to cause a p.G165A missense change in ectodysplasin-A. D: Multiple protein alignment of partial ectodysplasin-A orthologs are shown. The amino acid residues that differ from the sequence of the human ectodysplasin-A are in red color. The highly conserved G165 residue is indicated with an arrow.

Figure 2.

Abnormal tooth appearance and significant hypodontia in kindreds K7150 and K7155. A: Complete absence of lower incisors in an affected male (K7150, V-3). B: Mild tri-lobed appearance of otherwise normal incisors in an obligate female carrier (K7150, IV-8). C: Panorex film showing dental abnormalities in an affected male (K7150, V-7). D: Panorex film of severe anodontia in an affected male (IV-3) in family K7155.


None of the family members reported heat intolerance, heat shock/exhaustion or recurrent febrile episodes as children. Although this family is of the same ethnic origin as Family K7150, there are no known common ancestors. Sweat pore analysis in IV-3, III-6 and -7 showed normal density of sweat pores. No hair abnormality was noted. Nails were also normal in all cases.

In case IV-3, a marked congenital hypodontia, particularly involving the incisors, was noted (Figs. 1A,B and 2D). Dental impressions of case IV-2 showed minor anomalies and hypodontia similar to IV-3. Case III-2, a carrier female, was normal on examination. There was no indication of mild nail dysplasia as noted in K7150, case IV-7.


All coding exons of the ectodysplasin-A gene were analyzed by PCR-amplification and sequencing. We identified a single nucleotide substitution, c.494G>C, in exon 3 (Fig. 1B). This missense mutation is predicted to cause the replacement of amino acid Glycine165 by Alanine. The change was present in the affected male but was absent in the normal father. The mother was heterozygous for the alteration and carried a wild-type and a mutant allele. We analyzed all other female samples in the family and confirmed their carrier status as expected (Fig. 1A). X inactivation studies showed highly skewed X inactivation (Xi: 94:6) in the carrier female IV-7 and moderately skewed X-inactivation (Xi: 87:13) in III-1. Subsequently, we also studied a female who was mother of an affected male in family K7155 and found that this female carried the identical alteration identified in family K7150. X inactivation studies showed moderately skewed X inactivation (Xi: 84:16). DNA samples on additional relatives in this family were not available.

There is a remote possibility that the identified mutation in these families is simply a normal polymorphism. However, we have not found this alteration in >250 control X chromosomes from normal individuals. Furthermore, the residue (G165) is highly conserved (Fig. 1C) and is located in the extracellular domain of ectodysplasin-1 between the furin cleavage site (residue 150–159) and the collagenous-like domain (residue 185–235) that is required for trimerization of the ectodysplasin-1 molecule. Further functional studies of the Gly165Ala replacement are warranted to characterize the pathogenic role of this mutation.


We are grateful to the patients and their family members for participation in this study. We thank Cindy Skinner for assistance in obtaining patients' samples. This study was supported, in part, by a grant from the South Carolina Department of Disabilities and Special Needs (SCDDSN).