Introduction

Epidermolysis bullosa (EB) comprises a group of cutaneous hereditary mechanobullous disorders that can be classified into three major categories, the simplex, the junctional, and the dystrophic forms, on the basis of the level of tissue separation within the basement membrane zone (BMZ; Fine et al. 2000). Dystrophic EB (DEB) is clinically characterized by mucocutaneous blistering in response to minor trauma, followed by scarring and nail dystrophy, in which patients exhibit tissue separation beneath the lamina densa at the level of the anchoring fibrils. It occurs as either an autosomal dominant (DDEB) or recessive (RDEB) trait, each form having a different specific clinical presentation and severity (Fine et al. 2000).

Both DDEB and RDEB are caused by mutations in the COL7A1 gene encoding type VII collagen, the major component of anchoring fibrils (Uitto et al. 1995; Fine et al. 2000). The most severe RDEB subtype, the Hallopeau–Siemens (HS) type, shows a complete lack of expression of type VII collagen, whereas some collagen expression is found in the non-Hallopeau–Siemens (nHS) type. Clinical features of DDEB are comparatively milder than those of RDEB. To date, several hundred pathogenic mutations within the collagenous and noncollagenous domains of type VII collagen gene have been identified in different forms of DEB (Christiano et al. 1995; Uitto et al. 1995; Shimizu et al. 1996; Pulkkinen and Uitto 1999; Whittock et al. 1999). Although particular molecular and phenotypic characteristics of DEB have been elucidated, we cannot always expect DEB clinical manifestations precisely from genetic information of COL7A1. Furthermore, no systematic study has thus far revealed detailed delineation of COL7A1 mutations in Japanese DEB patients apart from several recurrent COL7A1 mutations (Tamai et al. 1999; Murata et al. 2004).

In this study, we performed mutational analysis of 20 Japanese DEB families and have demonstrated the characteristic features of COL7A1 mutations in Japanese DEB patients.

Materials and methods

Subjects

Twenty unrelated Japanese DEB families, who had been referred to Hokkaido University Hospital’s special clinic for inherited skin disorders from January 2000 to December 2004, were studied (Table 1). DEB was at first clinically diagnosed and later confirmed by immunofluorescence antigen mapping that demonstrated tissue separation beneath the lamina densa. Clinical features and inheritance modes also helped to differentiate most, though not all, cases into recessive or dominant DEB subtypes. Immunofluorescence expression of type VII collagen was of significant diagnostic value in determining HS-RDEB and nHS-RDEB.

Table 1 Clinical phenotype, type VII collagen expression, and COL7A1 mutations in patients in this study
Full size table

Immunohistochemical analysis

Skin biopsies were taken from DEB patients and subjected to a routine immunofluorescence antigen mapping study (Shimizu et al. 1996). The specimens were embedded in OCT compound, and 10-μm thick sections were cut. The following monoclonal antibodies (mAbs) against BMZ components were used: mAbs HD1-121 for plectin; GoH3 and 3E1 (Chemicon International, CA, USA) for the α6 and β4 integrins, respectively; GB3 (Sera-lab, Cambridge, UK) for laminin 5; LH7.2 (Sigma, St. Louis, MO, USA) for type VII collagen; and S1193 and HDD20 for BPAG1 and BPA2, respectively. The antibodies GoH3, S1193, and HDD 20 were kind gifts from Dr. A. Sonnenberg of the Netherlands Cancer Institute. The antibody HD1-121 was also a kind gift from Dr. K. Owaribe of Nogoya University. The bound antibodies were detected with FITC-conjugated goat anti-mouse IgG antibody. In some cases, nuclei were counterstained with propidium iodide.

All DEB patients in this study were evaluated by several experienced dermatologists. This study was approved by the Ethical Committee at Hokkaido University Graduate School of Medicine. Informed consent was obtained from individual patients or their parents.

Mutation analysis

Genomic DNA was isolated from peripheral lymphocytes of patients and their families using standard procedures. COL7A1 segments including all 118 exons, all exon-intron borders, and the promoter region were amplified by PCR using pairs of oligonucleotide primers synthesized on the basis of intronic sequences according to the report by Christiano et al. (1997; GenBank numbers L02870 and L23982). The PCR products were examined on 2% agarose gel and subjected to direct automated nucleotide sequencing using the BigDye Terminator System (Applied Biosystems, Foster City, CA, USA).

Results and discussion

An increasing number of DEB mutations have elucidated some general genotype-phenotype correlations (Jarvikallio et al. 1997; Pulkkinen et al. 1999). DDEB patients basically harbor glycine substitution mutations within the collagenous domain on one COL7A1 allele, leading to disruptions in anchoring fibril assembly and relatively mild clinical features. On the other hand, patients with RDEB in its most severe form, the Hallopeau–Siemens variant (HS-RDEB), frequently have premature termination codon (PTC) mutations on both alleles. These mutations characteristically lead to nonsense-mediated mRNA decay that manifests as a complete absence of type VII collagen protein and total loss of anchoring fibrils. On the other hand, patients with the non-Hallopeau–Siemens variant (nHS-RDEB) show milder phenotype, and type VII collagen can be generally detected immunohistologically. This DEB subtype is caused by a combination of mutations such as PTC, missense, and splice-site mutations on both alleles.

The routine immunofluorescence antigen mapping study in a blister site showed that all BMZ antigens were located in the roof of the blister, indicating tissue separation beneath the lamina densa. Also, linear type VII collagen expression was found along the dermal epidermal junction in nHS-RDEB and DDEB patients, whereas HS-RDEB cases showed no expression (Table 1). We found retention of type VII collagen within epidermal keratinocytes in a DDEB (family 17, data not shown).

Examination of 40 alleles of 20 families (10 nHS-RDEB, six HS-RDEB, and four DDEB) identified 30 pathogenic COL7A1 mutations, including 16 novel mutations (Table 1). COL7A1 mutations of nHS-RDEB included five missense mutations [G1595R(4783G>A), G1815R(5443G>A), R1957Q(5870G>A), G2366C(7096G>T), C2875F(8627G>T)], five nonsense mutations [R236X(706C>T), R1340X(4018C>T), R1978X(5932C>T), Q2827X(8479C>T), E2857X(8569G>T)], one insertion-deletion mutation (5818delC), and four splice-site mutations (5604+2G>C, 6573+1G>C, 8109+2T>A, 8358+1G>T). HS-RDEB patients showed four nonsense mutations [R137X(409C>T), Q641X(1921C>T), R1683X(5047C>T), R2261X(6781C>T)] and four insertion-deletion mutations (434insGCAT, 1474del8, 5818delC, 6081insC). As predicted by previous DEB mutation reports, all combinations of PTC mutations caused HS-RDEB, while nHS-RDEB resulted from compound heterozygous COL7A1 mutations except for homozygous nonsense PTC/PTC mutations. Although we could not find positional effect of PTC mutations as suggested by the previous report (Tamai et al. 1999), final conclusions need further accumulation of RDEB patients with PTC mutations.

In DDEB patients, we identified three dominant glycine substitution mutations [G2034R (6100G>A), G2037E (6110G>A), G2064E (6191G>A)]. These glycine substitution mutations were previously reported, and, interestingly, the nucleotide changes were identical to those in previous reports (Kon et al. 1997; Rouan et al. 1998; Jonkman et al. 1999; Whittock et al. 1999; Lee et al. 2000). Glycine residues within the collagenous domain are critical for proper triple helix formation. Some COL7A1 glycine substitution mutations, which cause RDEB in association with a second mutation on the other allele, are silent in patients with a normal COL7A1 allele. In addition, heterozygous glycine substitution mutations can cause DDEB through dominant negative interference of the collagen triple helix. Although this study also identified both dominant and recessive glycine substitution mutations (Table 1), we could not clarify positional effect of glycine substitution on the inheritance mode. A single indel mutation, 8069del17insGA, was novel. The 17-nucleotide deletion from 8069 to 8084 with GA insertion resulted in a 15-nucleotide deletion within the collagenous domain, which failed to change an open reading frame of COL7A1 but interfered with the collagen triple helix (Gly-X-Y). This mutation causes a DDEB phenotype, probably in a dominant negative fashion.

We failed to detect one allelic mutation in families 8, 9, 10, and 15, and both allelic mutations in RDEB family 16 (Table 1). Thus, this study could demonstrate COL7A1 mutations in 30 out of 36 alleles that were expected to have COL7A1 mutations, and the resulting ratio of successful mutation detection was 83%. Similar, large-scale COL7A1 mutation reports using Italian patient data also failed to determine single allele mutations in 13 RDEB families out of a total of 49 families (Gardella et al. 2002). This suggests at least two possibilities: that pathogenic mutations lie in the other parts of the COL7A1 gene that were not examined in these studies, and that genes other than COL7A1 are responsible for the DEB phenotype.

Although COL7A1 mutations are generally family specific, some recurrent mutations have been reported in several populations: R2814X, R578R, and 7786delG in British patients (Mellerio et al. 1997); 2470insG in Mexican patients (Salas-Alanis et al. 2000); and 8441-14del21, 4783-1G>A, 497insA, and G1664A in Italian patients (Gardella et al. 2002). In Japanese patients, the mutations 5818delC (nine out of 50 cases: 18%), 6573+1G>C (6/50:12%), and E2857X (9/50:18%) are present only in individuals of Japanese ethnic origin (Tamai et al. 1999). The present study also detected 5818delC in four families, 6573+1G>C in two families, and E2857X in one family out of total of 20 unrelated families. Furthermore, the Q2827X mutation was found in two unrelated families, and this mutation should be regarded as a candidate recurrent Japanese COL7A1 mutation. However, 16 mutations were novel out of a total of 30 pathogenic DEB mutations identified, indicating that Japanese COL7A1 mutations are family specific. This result furthers our understanding of both the clinical and allelic heterogeneity displayed in Japanese DEB patients.