Article Text

The p63 gene in EEC and other syndromes
1. H G Brunner,
2. B C J Hamel,
3. H van Bokhoven
1. University Medical Centre, Department of Human Genetics 417, Geert Grooteplein 16, 6525 GA Nijmegen, The Netherlands
1. Correspondence to:  Dr H G Brunner, University Medical Centre, Department of Human Genetics 417, Geert Grooteplein 16, 6525 GA Nijmegen, The Netherlands;  H.Brunner{at}ANTRG.AZN.NL

## Abstract

Several autosomal dominantly inherited human syndromes have recently been shown to result from mutations in the p63 gene. These syndromes have various combinations of limb malformations fitting the split hand-split foot spectrum, orofacial clefting, and ectodermal dysplasia. The p63 syndrome family includes the EEC syndrome, AEC syndrome, ADULT syndrome, limb-mammary syndrome, and non-syndromic split hand/foot malformation. The pattern of heterozygous mutations is distinct for each of these syndromes. The functional effects on the p63 proteins also vary between syndromes. In all of these syndromes, the mutation appears to have both dominant negative and gain of function effects rather than causing a simple loss of function.

• EEC syndrome
• p63 gene
• EEC, ectrodactyly, ectodermal dysplasia, facial clefting
• SHFM, split hand/foot malformation
• AEC, ankyloblepharon, ectodermal dysplasia, and clefting
• SAM, sterile alpha motif
• LMS, limb-mammary syndrome
• CHANDS, curly hair, ankyloblepharon, and nail dysplasia syndrome
• CLPED1, cleft lip/palate, ectodermal dysplasia syndrome
• AER, apical ectodermal ridge

## EEC SYNDROME

The EEC syndrome (MIM 129900) is characterised by the triad of ectrodactyly, ectodermal dysplasia, and facial clefting. A number of associated anomalies are frequently found, among which are lacrimal tract anomalies, urogenital anomalies, and conductive hearing loss.1–3 The EEC syndrome is relatively common with over 200 cases published. It is well known for having both variable expressivity and reduced penetrance. A comparison of interfamilial and intrafamilial variability in expressivity found significantly larger interfamilial variability, suggesting that more than one gene or allele might be involved.3

### p63 gene mutations in EEC syndrome

p63 gene mutations account for most and possibly all cases of classical EEC syndrome (fig 1). An extended analysis of EEC syndrome patients showed heterozygous mutations in 40 of 43 unrelated families. This indicates that p63 is the major if not the only gene mutated in EEC syndrome.4 All but one of these mutations give rise to amino acid substitutions in the DNA binding domain common to all known p63 isoforms. The arginine codons 204, 227, 279, 280, and 304 are most frequently mutated. Mutations involving just these five amino acids account for 75% of all EEC syndrome cases. A single frameshift mutation was detected in exon 13 in a patient with EEC syndrome characterised by cleft lip and palate, ectodermal dysplasia, split hand-split foot malformation, and mammary gland aplasia.5 Although the phenotype of this patient most closely resembles EEC syndrome, the mammary gland aplasia would also be consistent with a diagnosis of limb-mammary syndrome (LMS).6 Interestingly, two typical limb-mammary syndrome patients have also been found to have frameshift mutations affecting the C-terminal region of the gene.

Figure 1

p63 gene structure and transcripts. The top panel is a diagram of the intron-exon structure, showing the two transcription initiation sites and alternative splicing routes, yielding at least six different p63 isotypes. The structure of the TAp63γ isotypes is most similar to p53, both containing a transactivation domain (TA), DNA binding domain, and tetramerisation domain (ISO). The α isotypes contain an additional C-terminal SAM domain.

Different factors may contribute to this highly specific distribution of p63 mutations in EEC syndrome. It is possible that the mutations involve nucleotides that are highly mutable. The fact that mutations commonly affect CpG sites is consistent with this notion. However, it is more likely that the restricted mutation spectrum in EEC syndrome reflects a specific pathogenetic mechanism. This possibility is supported by the finding that for each of the frequently mutated amino acids a number of different missense mutations occurred, such as R204W/Q, R279C/H/Q, R280C/H/S, and R304W/Q. All of these mutations were associated with EEC syndrome, whereas mutations affecting other domains of the p63 protein yield phenotypically distinguishable syndromes, such as AEC or LMS. This suggests that, although mutations can occur at many different sites along the p63 gene, only those affecting specific amino acids in the DNA binding domain of the molecule will yield an EEC phenotype.

### The EEC syndrome community

Several syndromes have been described which share features with EEC (table 1). Some of these are now known to result from mutations of the P63 gene and are thereby allelic to classical EEC syndrome in 3q27. In others, allelism has been excluded, and for yet others this is unknown.

Table 1

Dominantly inherited syndromes that share similarities with EEC syndrome

A large kindred has been reported in which a syndrome segregated that combines ectrodactyly with cleft palate, without cleft lip or the ectodermal features that occur with the EEC syndrome.7 The term ectrodactyly-cleft palate syndrome was suggested (MIM 129830).

The AEC syndrome (MIM 106260) of ankyloblepharon, ectodermal dysplasia, and clefting was originally described by Hay and Wells.8 It shares the ectodermal dysplasia and clefting with EEC syndrome. However, the ectodermal involvement is much more severe in AEC syndrome. Severe scalp dermatitis is common. In addition, the clefting when present in EEC syndrome is always cleft lip (with or without cleft palate), whereas it can be cleft palate only in AEC. Limb involvement in AEC syndrome is minimal or absent.

Popliteal pterygium syndrome (MIM 119500) may resemble AEC syndrome since it has ankyloblepharon filiforme, filiform bands between the jaws, lip pits, and cleft palate with or without cleft lip.9 Many patients have syndactyly. The characteristic popliteal pterygium is not present in all cases. A pathognomonic sign is a typical, triangular overgrowth of skin over the nail of the big toe.2 Popliteal pterygium syndrome may be allelic to van der Woude syndrome in 1q32.10

The limb-mammary syndrome (MIM 603543) was originally described in a large Dutch kindred with split hand-split foot malformation (ectrodactyly) and aplasia or hypoplasia of the mammary gland and nipple. Some patients had complete absence of breast tissue (amastia).6 Although a number of features were shared with the EEC syndrome, the frequent mammary gland abnormalities and the fact that ectodermal involvement occurred in only a small minority of patients clearly indicated a separate syndrome. In addition, clefting occurred, but unlike EEC syndrome it was cleft palate and not cleft lip.

Another EEC-like condition (MIM 103285) was described in a single large German kindred.11 Features included ectrodactyly, syndactyly, fingernail and toenail dysplasia, hypoplastic breasts and nipples, intense freckling, lacrimal duct atresia, frontal alopecia, primary hypodontia, and loss of permanent teeth. The absence of clefting in this multigenerational family, and the intense freckling were thought to set it apart from the EEC syndrome. The authors suggested the acronym ADULT syndrome for this condition to reflect acro-dermato-ungual-lacrimal-tooth involvement.

The combination of ectodermal dysplasia and clefting occurs in several other syndromes. One of these is the autosomal dominant Rapp-Hodgkin syndrome of wiry hair (pili torti or pili canaliculi), anhidrotic ectodermal dysplasia, and cleft lip and palate12 (MIM 129400). Because of overlapping features of EEC and Rapp-Hodgkin syndromes in a mother and child, it has been suggested that these are fundamentally the same disorder.13 This would seem to be supported by the findings of Rodini and Richieri-Costa,14 who reported on a Brazilian family with 11 affected patients over four generations. Clinical findings in different patients ranged from isolated trichodysplasia (sparse, brittle, and dry hair) to ectodermal dysplasia, cleft palate, tear duct anomaly, and minor limb anomalies.14 It has been suggested that AEC syndrome and Rapp-Hodgkin syndrome could represent the same entity.15 These authors described a child affected by ectodermal dysplasia associated with clefting, ankyloblepharon, severe scalp dermatitis, and the characteristic Rapp-Hodgkin facies.

Another dominantly inherited syndrome which shares features with EEC is the lacrimo-auriculo-dento-digital (LADD) syndrome (MIM 149730). LADD syndrome has aplasia or hypoplasia of the puncta with obstruction of the nasal lacrimal ducts, cup shaped pinnae with mixed hearing deficit, small and peg shaped lateral maxillary incisors, and mild enamel dysplasia. The digital features include fifth finger clinodactyly, radial ray abnormalities, and syndactyly.16 Clinical differentiation between LADD and EEC syndromes was difficult in a family in which the mother had a split hand/split foot deformity and the daughter a condition consistent with a diagnosis of LADD syndrome.17

Some recessively inherited syndromes share similarities with EEC syndrome, notably the Bowen-Armstrong syndrome of recessively inherited ectodermal dysplasia, and cleft lip and/or cleft palate. Some patients have congenital adhesions between the eyelids, cicatricial atrophy of the scalp, abnormal EEG, partial anodontia, genital hypoplasia, syndactyly, and delayed skeletal growth and maturation.18 This condition could be a separate syndrome or it could be the same as the CHANDS syndrome of curly hair, ankyloblepharon and nail dysplasia, which may further include alveolar synechiae (MIM 214350).19,20

Another recessively inherited syndrome consists of syndactyly, ectodermal dysplasia, and cleft lip/palate. This condition has been termed the cleft lip/palate, ectodermal dysplasia syndrome (CLPED1, MIM 225000) or Zlotogora-Ogur syndrome. The patients have sparse eyebrows and eyelashes, sparse, short and dry scalp hair, and hypodontia. Cleft lip/palate, syndactyly of fingers and toes, and onychodysplasia are often present. Some patients have mental retardation. CLPED1 was recently shown to result from mutations in the PVRL1 gene, encoding nectin-1, an immunoglobulin related transmembrane cell-cell adhesion molecule that is part of the NAP cell adhesion system.21

### Specific patterns of mutation of the p63 gene in other syndromes

Mutations of the p63 gene have been found in five human malformation conditions to date (fig 2). In addition to EEC, these are the AEC or Hay-Wells syndrome,8,22 limb-mammary syndrome,4,6 ADULT syndrome,11,23,24 and non-syndromic split hand-split foot malformation25 (SHFM, MIM 183600). These syndromes are characterised by limb abnormalities that fit the split hand-split foot spectrum, ectodermal dysplasias affecting the hair, teeth, nails, and sweat glands, absence of the mammary glands, and a range of other malformations of the facial skeleton, the urogenital system, and the eyes. The various malformations can be largely explained by assuming that the effect of the mutation is to disrupt normal ectoderm formation or function. This is probably also true for the limb malformations which are thought to reflect interference with normal formation of the apical ectodermal ridge (AER), which is the distal most part of the growing limb during embryogenesis. Combined p63 mutation data for these syndromes indicate extensive genotype-phenotype correlations, with each of these syndromes having a distinct pattern and type of mutations.4 This suggests that each of these mutations causes a specific perturbation of normal p63 gene function. Functional data are consistent with dominant negative as well as dominant gain or change of function defects (P Duijf, unpublished data).5 The p63 gene encodes at least six different protein isotypes through the use of alternative transcription start sites and alternative splicing at the 3 end of the gene.26

Figure 2

Pattern of p63 mutations in human syndromes. The p63 mutations identified in five different human disorders are indicated. Note the clustering of mutations in the DNA binding domain for EEC syndrome and in the SAM domain for AEC syndrome.

p63 gene mutations in Hay-Wells or AEC syndrome were initially detected in eight families.22 All were missense mutations in the SAM (sterile alpha motif) domain. These mutations are predicted to disrupt protein-protein interactions of the p63 protein with as yet unknown partners. A single family with AEC syndrome has been shown to have a mutation that results in a splicing defect that removes exon 11 and affects the C-terminus of both the alpha and beta isotypes of p63 (J Murray, personal communication). In vitro analysis of the functional effects of the mutations in the AEC patients indicated that such mutations have complex gain of function as well as loss and change of function effects.22

Initially, no p63 mutations were detected in the large Dutch LMS family that first linked this class of diseases to the p63 locus.6 However, two isolated, unrelated patients with an LMS phenotype were found to have frameshift mutations in exon 13 that resulted in truncations of the p63 protein.4 The location of these mutations may be significant, as LMS differs from EEC syndrome in at least three respects. The first is that mammary gland and nipple hypoplasia is a consistent feature of LMS and is only occasionally seen in EEC syndrome. The second is that LMS patients do not have hair and skin defects. The third is that whereas LMS patients have cleft palate or bifid uvula, those with EEC syndrome more often have cleft lip. Very recently the mutation in the original LMS kindred6 was found to reside in the N-terminal part of the gene, in exon 4 (P Duijf et al, unpublished data). This mutation differs from that in the two other LMS kindreds, but is also clearly different from the mutations seen in EEC syndrome.

ADULT (acro-dermato-ungual-lacrimal-tooth) syndrome differs from EEC syndrome by the absence of facial clefting. Instead, these patients show neurodermal signs, including exfoliative dermatitis of the digits and excessive freckling.11 Linkage studies in this large German family indicated allelism with EEC syndrome in 3q27.27 A p63 gene mutation was found in a small French kindred with ADULT syndrome by Amiel et al.23 This mutation is clearly different from the p63 gene mutations in EEC. The mutation is localised in exon 3, which is included only in the TA isotypes of p63 and causes an amino acid substitution (N6H) outside the DNA binding domain.23 Another ADULT syndrome mutation involving arginine 298 has now been found in the German kindred originally described by Propping and Zerres.24 Transactivation studies indicate that this latter mutation confers significant transactivation capacity to ΔN-p63gamma, which is otherwise inert in these assays.24 This unusual result raises that possibility that the ADULT syndrome mutation confers a novel gain of function to p63, although the precise molecular mechanism is still under investigation.

p63 gene mutations associated with non-syndromic split hand and foot malformation were first identified in two South African kindreds.25 Subsequent analysis of a larger group of 45 unrelated SHFM patients indicated that four out of 45 had a p63 mutation, suggesting that p63 mutations probably account for no more than 10% of non-syndromic SHFM.4 Four of the six p63 mutations detected in SHFM patients are not found in other syndromes: missense mutations K193E and K194E, splice site mutation IVS4-2A>C, and nonsense mutation Q634X. Two other mutations, R280C and R280H, both involving codon 280, have been found in SHFM as well as in EEC syndrome. Thus, there is partial overlap between the EEC and SHFM mutational spectra. This might perhaps have been expected from the considerable phenotypic variability that can be seen within and between EEC syndrome families. If this were the explanation, then some of the affected subjects in these kindreds should have shown mild EEC features. This does not seem to be the case. The two large South African SHFM families described by Ianakiev et al25 were carefully examined for signs of ectodermal dysplasia and none was found. Since these families are quite large, the fact that all family members with the R280H mutation had only SHFM and none of the other signs of EEC syndrome suggests specific modifier effects, possibly through interacting genes.

### The power of genotype-phenotype correlations

The finding that p63 mutations cause at least five different malformation conditions with a strong genotype-phenotype correlation suggests a number of topics for future research. First, these results validate the concept that human malformation syndromes can be grouped into syndrome families based on clinical similarities. This concept was first articulated for the skeletal dysplasias.28 Spranger's classification was based on the realisation that different skeletal dysplasias could be grouped on the basis of radiological similarities. The grouping together of achondroplasia, hypochondroplasia, and thanatophoric dysplasia has been fully vindicated by molecular studies. The same is true for diastrophic dysplasia and achondrogenesis 1b. Yet other examples are the type II collagenopathies that range from achondrogenesis II, via hypochondrogenesis, spondyloepiphyseal dysplasia congenita, and Kniest dysplasia, to Stickler arthro-ophthalmopathy and mild dominant spondyloarthropathy.29

This concept may be applicable in a more general sense. For syndromes that are clinically similar, we should assume a shared or at least related pathogenesis. As a case in point, one might consider the Stickler and Marshall syndromes. For many years their status as independent syndromes was debated.30,31 Molecular studies have since shown that both involve genes encoding component chains of type XI collagen protein. However, there is preferential involvement of the COL2A1 gene in Stickler syndrome and of the COL11A1 gene in Marshall syndrome.32 This not only illustrates that functionally related genes are involved in clinically similar syndromes, it also shows that even relatively small phenotypic differences may point to important differences at the molecular level. Our finding that many phenotypes are associated with p63 gene mutations would seem to support this notion.

Although superficially similar to EEC, the presence of pigmentary disturbances and the absence of clefting set the ADULT syndrome apart,11 and this is now confirmed by our recent demonstration of a specific gain of function mechanism.24 Our recent finding that the ADULT syndrome mutations in these families uncover a putative second transactivation domain underlines how pertinent clinical observations may help show aspects of a protein's function that would otherwise have been more difficult to detect.

Finally, the contrasting phenotypes associated with different classes of p63 mutations shows promise for our understanding of embryonic development in general. It is significant in this respect that EEC mutations but not AEC mutations cause limb involvement. Also, limb-mammary syndrome mutations cause mammary gland aplasia, whereas EEC and SHFM rarely or never do. Taking these clinical differences as leads, we can now start to assess which downstream targets of p63 are specifically perturbed by each of these classes of mutations. In doing so, we may be able to pinpoint some of the developmental pathways that are specifically involved in shaping the limbs, skin, and mammary gland during development. Clearly, molecular genetics will provide the answers, but only if the power of clinical genetics is exploited to pose the right questions.

## Acknowledgments

The authors are indebted to Pascal Duijf, Jacopo Celli, Ellen van Beusekom, Kaate Vanmolkot, Saskia van der Velde-Visser, and Bellinda van den Helm for their hard work in the Nijmegen Human Genetics Laboratory. We also thank clinicians world wide as well as the families themselves for graciously providing clinical data and DNA samples. We especially thank Fiorella Gurrieri, Giovanni Neri, John McGrath, Mike Bamshad, Andre Reis, Jeff Murray, Volker Doetsch, Klaus Zerres, Peter Propping, Rob de Waal, Arie Smits, and Gert Vriend for crucial contributions to the work and for sharing of unpublished data. This work was partly supported by NWO grant number 901-04-183 and by an Elisabeth Kinderziekenhuis grant to HGB and by NWO grant 901-02-232 to HvB.

View Abstract

## Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.