DEAR EDITOR, Dystrophic epidermolysis bullosa (DEB) is a blistering skin disease caused by mutations in the gene (COL7A1) encoding type VII collagen (C7). DEB can be inherited by either dominant (DDEB) or recessive (RDEB) mechanisms. RDEB results in severe wounds and scarring, and extracutaneous manifestations such as oesophageal strictures, chronic anaemia and pseudosyndactyly. DDEB is generally a milder form of the disease, with fewer and less severe blisters. 1 COL7A1 contains 118 exons. The central Gly-XY region is flanked by two noncollagenous (NC) regions: the amino-terminal NC1 domain (exons 1–28) and the carboxyl-terminal NC2 domain (exons 112–118). Patients with RDEB may either express the NC1 amino-terminal fragment of C7 (NC1+) or lack it (NC1 Ā). 2 The NC1 region has been reported to contain the major antigenic epitopes of C7, although immune responses to other C7 regions have been noted. 3 In the autoimmune disease epidermolysis bullosa acquisita, IgG autoantibodies bind to various epitopes in C7, causing dermoepidermal separation and blistering. 3 As a screening tool for C7 replacement trials, we investigated whether NC1 expression can be predicted using genetic testing. 2, 4 We proposed that subjects who do not express the NC1 domain (NC1 Ā) could be predicted by two loss-of-function mutations in exons 1–28. We compared genetic testing data with indirect immunofluorescence (IIF) using LH7. 2 antibody and Western blot techniques using FNC1 antibody (Fig. 1). Genetic testing predictions correlated with Western blot results for 21 of 21 (100%) samples (Tables 1 and S1; see Supporting Information). A total of 39 mutations were identified including nine previously unreported mutations (Table S1; see Supporting Information). IIF staining with LH7. 2 displayed significant variability, correlating with FNC1 Western blot results for 11 of 22 (50%) samples tested (Tables 1 and S1; see Supporting Information). LH7. 2 binds to an epitope between amino acids 36 (exon 2) and 153 (exon 4); the NC1 domain spans up to exon 28. 5 Therefore, a sample may show positive LH7. 2 staining even if the full NC1 domain is not produced, as seen in subjects 15 and 17. These results illustrate the variability in using IIF with LH7. 2 as a diagnostic tool for NC1 expression. Revertant mosaicism is another confounder to biopsy analysis. 6 Skin biopsies for epidermolysis bullosa (EB) diagnosis are obtained from uninjured, unscarred skin, which can be difficult to locate in older patients with EB, especially those with RDEB. It is possible that the sites selected for biopsy may actually be areas of undiagnosed revertant mosaicism, which would be positive for NC1 on IIF and Western blot but predicted to be NC1 Ā by genetic testing. Circulating antibodies to C7 were detected using two enzyme-linked immunosorbent assays (ELISAs). According to the Stanford ELISA (Data S1; see Supporting Information) seven of 26 had antibodies and 11 of 26 had antibodies according to the commercial ELISA (MBL International, Woburn, MA, USA). Only three subjects were positive according to both assays (Table 2). Most notably, all subjects were negative for IgG, IgM, IgA or complement deposits at the dermoepidermal junction by direct immunofluorescence (DIF) on their skin, providing evidence that these antibodies are not pathogenic, as has been previously reported by other groups. 7–9

Nine subjects had previously received synthetic allografts, which would expose the recipient to C7, although there have been no previous reports of immune response. 10 Only one subject (subject 18) was positive for C7 antibodies and had …