Exclusion of EGFR, HRAS, DSP, JUP, CTNNB1, PLEC1, and EPPK1 as Functional Candidate Genes in 7 Families With Syndromic Diarrhoea

Abstract

Syndromic diarrhoea (SD) is a rare disease associating intractable diarrhoea and hair abnormalities. In an attempt to identify the gene causative for SD, we studied several functional candidate genes, based on their implication in overlapping phenotypes in mice (EGFR) or in humans (HRAS, JUP, DSP EPPK1, PLEC1, and CTNNB1) in 8 patients affected by SD. Except for EGFR and HRAS, all selected genes encode for cell adhesion proteins. Using direct sequencing or linkage analysis, we excluded all of the candidate genes as the disease-causing gene in our group of patients; however, the hypothesis of intercellular junctions defect in SD remains seductive.

Syndromic diarrhoea (SD) is a rare condition associating intractable diarrhoea in infancy with facial dysmorphism, hair abnormalities, intrauterine growth retardation, and immunological defects ^(1,2). Patients presenting with hepatic cirrhosis in addition to these symptoms have been diagnosed as tricho-hepato-enteric syndrome ⁽³⁾, but it is probably the same disease ⁽⁴⁾. In SD no histological anomalies are found by digestive exploration, and common etiological analyses, such as metabolic analysis, are negative.

Thanks to collaborative work among 4 pediatric units (Hôpital Robert Debré and Necker Enfants Malades, Paris; Hôpital des Enfants de Toulouse and Hôpital d'enfant de la Timone, Marseille), a cohort of 8 patients from 7 families has been characterised. Informed consent was obtained from the patients or their parents to participate in the study. All of the patients present with a highly homogeneous phenotype including symptoms of diarrhoea beginning in the first month of life and necessitating total and lifelong parenteral nutrition, hair abnormalities with woolly hair, facial dysmorphism, and immunodeficiency especially toward antigene response. Four patients belong to consanguineous families. Because all of the patients have normal standard karyotype, no putative disease-causing region has been pointed out.

In an attempt to identify the gene causative for SD, we selected a number of functional candidates genes, based on the phenotypes of knockout mice or clinical overlap with other human diseases, for sequencing or linkage analysis.

We first analysed EGFR encoding the epidermal growth factor receptor, a membrane-spanning glycoprotein, which has been implicated frequently in various cancer types. Description of egfr^−/− mice indicates several symptoms shared with SD, such as intrauterine growth retardation, intestinal dysfunction, and hair follicles anomalies ⁽⁵⁾. This led us to consider this gene as a candidate and to perform direct sequencing of affected individuals from each family. The complete coding region, including the flanking intronic sequences, was amplified by polymerase chain reaction; purified polymerase chain reaction products were subsequently sequenced using the Big Dye Terminator kit (Applied Biosystems, Foster City, CA) and analysed on an Applied Biosystems 3130 automated sequencer. No mutation, except previously described polymorphisms, was found either in exonic or intronic parts (Table 1).

TABLE 1: Nucleotidic variations found in coding regions of EGFR, HRAS, and JUP

Hair abnormalities are constant in SD and described as woolly hair with aspects of trichorrhexis nodosa in half of the cases. We thus consider genes whose mutations are described to cause similar defects (eg, HRAS, DSP, JUP) as serious candidates.

HRAS encodes a Rho-GTPase involved in the MAP-kinase pathway and has been found mutated in Costello syndrome, which associates mental retardation as well as skin and hair abnormalities ⁽⁶⁾. Sequencing of the 7 patients' DNA following the procedure described earlier only revealed known polymorphisms (Table 1).

JUP and DSP are genes encoding, respectively, for plakoglobin and desmoplakin, 2 proteins constitutive of the desmosome. The desmosome is an intercellular adhesive junction found in many tissues, especially epithelium and muscles, and provides mechanical strength for the tissues. The extracellular part of the desmosome is made of the transmembrane proteins cadherins, which bridge the space between adjacent cells. Desmoplakin constitutes the intracellular part of the desmosome and is attached both to the intermediate filaments of the cells via plakoglobin and to the cytoplasmic domains of cadherins (Fig. 1). Desmoplakin is a member of the plakins family and shares 2 subunits with the other members of this family: the plakin domain and the plakin repeat domain ⁽⁷⁾. Plakoglobin belongs to the catenins family and plays a role not only in desmosomes but also in adherens junctions, another type of cell–cell adhesion structure mediated by the linkage of cadherins complex to the actin cytoskeleton.

FIG. 1: Schematic view of a desmosome. D = desmoplakin; IF = intermediate filaments; C = cadherin; P = plakoglobin.

Both genes JUP and DSP have been incriminated in syndromes including woolly hair in association with cardiovascular and cutaneous symptoms, namely Naxos syndrome (JUP) and Carvajal syndrome (DSP) ^(8,9).

Regarding gene sizes (3.5 and 9.6 kb for coding regions), we decided to use a linkage analysis with flanking markers for JUP and DSP analysis. Five polymorphic markers (chr.6: 22AC 7499115, chr.6 15CA 7458911, chr.6 21GT 7745112, chr.17 18 GT 37198999, chr.17 18CA 37200387) were amplified by polymerase chain reaction and fluorescently labelled with Beckman Coulter (Fullerton, CA) dye D2. Pattern visualisation was performed using the Beckman Coulter CEQ 8000 Series Genetic Analysis System. The analysis of the segregation excluded the implication of DSP but was inconclusive for JUP. Consequently, the analysis was completed by direct sequencing of JUP following the same procedure as for EGFR sequencing and revealed only previously reported nonpathogenic mutations (Table 1).

We additionally investigated CTNNB1, EPPK1, and PLEC1 on the basis of their functional roles and their relation to JUP and DSP. CTNNB1 encodes for the β-catenin, a member of the catenin family sharing 70% identity with plakoglobin. β-Catenin plays 2 important roles in cells: It is a downstream signalling molecule in the Wnt pathway, an evolutionarily conserved pathway that plays a role in central cellular functions including cell proliferation, migration, and differentiation. It is also a component of the adherens junction that links the cadherins to the α-catenin. Moreover, β-catenin is expressed in hair and liver. A direct sequencing of CTNNB1 following previously described procedure revealed no mutation of the coding regions.

EPPK1 encodes for epiplakin, another protein in the plakin family (as is desmoplakin). EPPK1 has a peculiar structure, with only one 15-kb-long exon. This exon contains 13 subunits coding for 13 plakin repeat domains presenting a high nucleotide sequence homology, the last 5 being virtually identical ⁽¹⁰⁾. Moreover, at 60 kb downstream of EPPK1 is located PLEC1, another gene encoding a member of the plakin family. PLEC1 also shares nucleotide homology with EPPK1 (2 sequences of 350 bp long with 90% homology at the nucleotide level), which could lead to abnormal recombination and create pathogenic rearrangements between these 2 genes. Little is known about epiplakin function, but it is presumably involved in intermediate filaments linkage ⁽⁷⁾. The gene is highly expressed in digestive epithelium and liver ⁽¹¹⁾ and has never been related to a human disease.

To study the EPPK1/PLEC1 locus, we made a linkage analysis with SNP (rs11779128, rs1138597, rs35221054, rs12542653, rs3923388, rs3923387) and microsatellites (chr.8 16AC 145052042, chr.8 18 GT 145147530). The linkage between EPPK1 and the disease was excluded in 2 of the families, leading us to conclude that there is an absence of implication for EPPK1/PLEC1 locus in SD.

In conclusion, we excluded 7 functional candidate genes (EGFR, HRAS, DSP, JUP, CTNNB1, EPPK1, and PLEC1) as disease-causing genes in 8 patients with syndromic diarrhoea. The intractable diarrhoea of infancy is a rare condition, but with severe features. Up to now, the IPEX syndrome is the only one in this group characterised at the molecular level; it is related to the forkhead box P3 defect, a transcription factor involved in the immune response ⁽¹²⁾.

The importance of cell–cell junction integrity in the myocardium and in the skin has been underscored by the description of desmosome defects in cardiocutaneous syndromes such as Naxos and Carvajal syndromes ⁽¹³⁾. Such a major functional role in digestive epithelium is likely and could also be extended to other types of cellular junctions, such as adherens junctions and tight junctions. This is sustained by the recent description of tight junction defects in infectious diarrhoea and in chronic inflammatory bowel diseases ⁽¹⁴⁾. Even if JUP, CTNNB1, and DSP have been excluded as disease-causing genes in this study, the hypothesis of a cellular junction–related disease remains seductive in SD. Further analysis of other major proteins of cell adhesion, such as cadherins or α-catenin, may lead us to unveil the underlying defect in this group of pathologies.

Acknowledgment

We thank the patients and their families for their kind cooperation and willingness to participate.