Prevalence of specific anti-skin autoantibodies in a cohort of patients with inherited epidermolysis bullosa

Dystrophic EB (DEB) forms, also known as dermolytic forms, are characterized by tense blisters and erosions that heal with extensive, either atrophic or hypertrophic, scarring. The scarring tendency reflects the fact that blister formation occurs below the level of the lamina densa of the BMZ of the papillary derma. The spectrum of clinical severity in DEB is, however, highly variable, with milder forms of the disease showing a limited tendency to blistering or nail involvement and severe forms where the gastrointestinal tract, particularly the oesophagus, is affected by blistering and scarring. All forms result from mutations within type VII collagen gene, COL7A1 [10].

The type VII collagen, the main constituent of the anchoring fibrils, is composed of a triple-helical domain flanked by a large 145 kDa non-collagenous amino-terminal (NC-1) domain and a small 34 kDa non-collagenous carboxyl-terminal (NC-2) domain [11, 12]. In patients with DEB, ultrastructural analysis of skin demonstrates the presence of altered anchoring fibrils, reduced in number or completely absent in the different forms [13].

In addition to inherited forms of EB, there is an acquired form of epidermolysis bullosa (EBA), which is caused by an auto-reactive response against collagen VII. Patients with EBA have both tissue-bound and circulating IgG autoantibodies against type VII collagen [14, 15]. Specifically, circulating autoantibodies recognize epitopes within the amino-terminal NC-1 domain of type VII collagen [16, 17]; only a small number of EBA patients have been described with autoantibodies reactivity against the NC-2 domain [18]. The routine screening tests for the diagnosis of EBA are: direct and indirect immunofluorescence microscopy; in particular, the salt-split skin method. Patients with EBA have immune deposits on the dermal side of salt-split skin, whereas in bullous pemphigoid (BP) deposits are on the epidermal side. However, this method does not distinguish EBA from anti-laminin 332 mucous membrane pemphigoid (MMP) or anti-laminin c1 pemphigoid as these diseases also show immune deposition on the dermal side of salt-split skin. EBA sera recognize the 290 kDa type VII collagen protein in immunoblotting studies; in fact, this reactivity confirms the diagnosis of EBA. However, immunoblotting studies are difficult to carry out because they are time-consuming and technically demanding. Advances in the identification of target antigens and the subsequent development of an increasing number of sensitive and specific assays for the detection of circulating autoantibodies, including recombinant forms of the target antigens and ELISA, allow serological diagnosis in patients with autoimmune skin disorders. Several ELISA systems using recombinant fragments of BP180, BP230, desmoglein1, desmoglein3, have become commercially available and are highly valuable diagnostic tools [19]. ELISAs were also developed to detect autoantibodies in EBA using different recombinant proteins of the NC-1 domain of type VII collagen [20, 21]. Previous reports demonstrate that these tests showed high sensitivity and specificity, and in some clinical studies serial titres of anti-type collagen VII from each patient with EBA reflected the disease activity [22‐24]. Experimental studies provided evidence that autoantibodies against type VII collagen are pathogenic. When rabbit and human antibodies were passively transferred into mice, they induced subepidermal blister formation and nail dystrophy [25, 26]. The active mouse model of EBA was also established by immunizing mice with the recombinant NC-1 domain of murine type VII collagen [27].

Although DEB has been shown to result from mutations in COL7A1, relatively little is known about the mechanistic consequences of the mutations in relation to blistering. To facilitate our understanding of DEB, we have evaluated the presence of anti-skin autoantibodies in a wider cohort of patients suffering from DEB and ascertained whether they correlated with the disease severity.

Serum samples were obtained from 27 patients with inherited forms of EB, seen at the Dermatological Department of the University Hospital in Bari between January 2010 and December 2012. All patients examined had typical clinical features of EB with a variable phenotype as well as immunofluorescence mapping and transmission electron microscopy findings consistent with EB.

As many as 17 patients had recessive dystrophic EB (RDEB), mean age 21.2 years, range 2-50, 4 male and 13 female; 10 patients had EB simplex (EBS), mean age 17 years, range 1-56, 3 male and 7 female. In all patients, disease severity was evaluated through the Birmingham Epidermolysis Bullosa Severity score (BEBS), which includes 11 items: area of damaged skin, involvement of nails, mouth, eyes, larynx and oesophagus, scarring of hands, skin cancer, chronic wounds, alopecia and nutritional compromise. Area was allocated 50 points, and the 10 other items 5 points each, giving a maximum score of 100 [28]. The BEBS score was performed by a dermatology specialist at blood sample collection.

Diseases control sera were obtained from 41 patients, including 20 patients with pemphigus vulgaris (PV) (mean age 59 years, range 37-85, 6 males, 14 females) and from 21 patients with bullous pemphigoid (BP) (mean age 74 years, range 61-90, 9 males, 12 females).

Control sera were obtained from 20 healthy subjects (mean age 47 years, range 36-65, 6 males and 14 females). All sera were stored at -20°C until assayed.

Patients and control subjects gave written consent to participate in this study, and the Ethics Committee of the hospital approved this study.

The detection of circulating anti-skin autoantibodies was performed in all samples with an Indirect Immunofluorescence (IIF) method (Euroimmun, Lubeck, Germany) and with the currently commercially available ELISA methods (MBL, Nagoya, Japan). Anti-type VII collagen ELISA kits were provided free of charge by the manufacturer. All assays were performed at the Laboratory of Clinical Pathology in the University Hospital of Bari.

Slides were made of frozen primate oesophagus section. IIF assays were performed at a serum dilution of 1:10 in phosphate buffered saline (pH 7.3).

Two different fluorescence patterns can be observed on primate esophagus sections: a) anti-intercellular substance (ICS) and anti-basal membrane zone (BMZ). ICS antibodies, which are specific for pemphigus, are directed against prickle cell desmosomes and react with surface antigens of keratinocytes; on tissue sections they display a granular fluorescence of the intercellular matter in the whole stratum spinosum. BMZ autoantibodies, which are specific for pemphigoid diseases or epidermolysis bullosa acquisita, are directed against epidermal basal membrane and generate a fine linear colouring between the stratum basale and the connective tissue.

Samples were scored positive if a fluorescence reaction was observed at 1:10 sample dilution.

Autoantibodies detection was performed in all samples with the ELISA systems, each one employing different antigenic proteins coated on the wells of microplates. Assays were run according to the manufacturers’ instructions.

Patients and control sera, diluted 1:100, were added to each well of a microwell plate. For detection of anti-type VII collagen autoantibodies the microplates were coated with human recombinant NC-1 and NC-2 proteins; for detection of anti-BP180 and anti-BP230 autoantibodies, the microplates were coated with recombinant purified BP180-NC16A antigen and recombinant purified BP230-N and BP230-C antigens, respectively. Sera were incubated for 1 hour at RT. After washing to remove any unbound serum proteins, horseradish peroxidise-conjugated anti-human IgG was added and incubated for 1 hour at RT. After another washing step, the peroxidise substrate was added and incubated for 30 min at RT. The acid solution was then added to each well to terminate the enzyme reaction and stabilize the colour development. The value in each sample was obtained by comparing the optical density (OD) of the sample with the OD of the calibrator.

To assess the relationship between clinical features and autoantibodies titres, we divided inherited EB patients into EB simplex (EBS) patients and EB recessive dystrophic (RDEB) patients.

The Kruskal Wallis test was used to compare the titres of each autoantibody assayed by the ELISA methods in patients and in the control group.

The diagnostic sensitivity of ELISA for anti-type VII collagen autoantibodies was calculated in 17 patients with RDEB, and the specificity was calculated in 10 patients with EBS, in 41 control patients with other diseases and in 20 healthy subjects. The cut-off value for positivity was validated and optimised by Receiver Operating Characteristics (ROC).

Correlation analyses were performed by using the Pearson test.

MedCalc software (Mariakerke, Belgium) was used for statistical and ROC curve analysis.

In patients with EB, the IIF method did not show an anti-basal membrane zone (BMZ) pattern, as expected. In the present study, all samples were tested at 1:10, in accordance to the recommendations for qualitative IIF given by the manufacturer. However, as the use of only a single dilution may lead to blocking or masking effects in high-titered sera, thus causing false-negative results, sera were retested at 1:100 dilution. At this dilution, no prozone effects were noted, confirming that those sera were true negatives. Detection of antibodies by ELISA test allowed quantitative measurements of autoantibodies serum titres. Figure 1 shows that the mean concentration of anti-type VII collagen autoantibodies titres was statistically higher in patients with RDEB than in patients with EBS, in controls with other diseases and in healthy subjects (test Kruskal Wallis, P < 0.0001). Clinical and serological findings of patients with two different inherited types of EB are reported in Table 1.

The mean concentrations of anti-type VII collagen autoantibodies titres were statistically higher in patients with RDEB than in patients with EBS (P =0.0014). The mean concentrations of anti-BP180 and anti-BP230 autoantibodies were also statistically higher in patients with RDEB than in patients with EBS (P = 0.0028 and P = 0.0023, respectively).

The area under the curve (AUC) for anti-type VII collagen ELISA test was 0.931(CI, 0.850-0.976) and ROC analysis showed that anti-type VII collagen autoantibodies had a sensitivity of 88.2% (CI, 63.5 -98.2) and a specificity of 96.7% (CI, 88.6-99.5), using a cut-off value of 5 UA/mL (Figure 2). The prevalence of anti-type VII collagen autoantibodies was 88% in patients with RDEB and 10% in patients with EBS. The combined prevalence of anti-BP180 and anti-BP230 autoantibodies was 88% in patients with RDEB and 50% in patients with EBS.

Thus, in RDEB forms there was a higher prevalence of both anti-type VII collagen autoantibodies and autoantibodies against BMZ antigens, in particular BP180 and BP230 (chi-square, P < 0.0001; OR 79.95% IC 4.01-1575).

ELISA results for each autoantibody and BEBS score for each patient are reported in Table 2.

In all patients with RDEB, anti-type VII collagen autoantibodies titres significantly correlated with both anti-BP180 (slope < 0.0009, R² 0.53) (Figure 3A) and anti-BP230 (slope < 0.0001, R² 0.83) autoantibodies titres (Figure 3B).

Birmingham Epidermolysis Bullosa Severity score (BEBS) correlated with anti-type VII collagen autoantibodies titres (P = 0.0190) (Figure 4A) and also with anti-BP180 and anti-BP230, P = 0.0119 and P = 0.0097, respectively (Figure 4B and 4C).