Natural Occurrence of Autoantibodies against Basement Membrane Proteins in Epidermolysis Bullosa

TO THE EDITOR Epidermolysis bullosa (EB) is a group of genetic blistering diseases characterized by lifelong trauma-induced blistering of the skin and mucosa and extracutaneous manifestations. Autoantibodies to a structural protein of the epidermal basement membrane zone (BMZ) such as dystonin (BP230), plectin, type XVII collagen (COL17/BP180), laminin-332, or type VII collagen (COL7) result in the same level of blister formation as in EB subtypes caused by mutations in their coding gene, such as in EB simplex (DST and PLEC), junctional EB (JEB) (COL17A1, LAMA3, LAMB3, or LAMC2), and dystrophic EB (DEB) (COL7A1) (Goletz et al., 2017; Has et al., 2020). The innate and adaptive immune systems are designed not to recognize the host’s own cells and proteins owing to natural immunological tolerance and negative selection of host-specific T lymphocytes in the central lymphatic organs. However, the lack of one of the proteins due to inherited mutations can interfere with this process. When the missing protein is introduced later in life, it can be recognized as dangerous, and an immune response can occur (Alberts, 2002; Siprashvili et al., 2016). Four previous publications presented results of serological tests, ELISA, and indirect immunofluorescence (IIF) on monkey esophagus in patients with EB (DEB and EB simplex) (Annicchiarico et al., 2015; Esposito et al., 2016; Tampoia et al., 2013; Woodley et al., 2014). Circulating antibodies against BMZ proteins were present in the serum, and the authors suggested the

TO THE EDITOR Epidermolysis bullosa (EB) is a group of genetic blistering diseases characterized by lifelong trauma-induced blistering of the skin and mucosa and extracutaneous manifestations. Autoantibodies to a structural protein of the epidermal basement membrane zone (BMZ) such as dystonin (BP230), plectin, type XVII collagen (COL17/BP180), laminin-332, or type VII collagen (COL7) result in the same level of blister formation as in EB subtypes caused by mutations in their coding gene, such as in EB simplex (DST and PLEC), junctional EB (JEB) (COL17A1, LAMA3, LAMB3, or LAMC2), and dystrophic EB (DEB) (COL7A1) (Goletz et al., 2017;Has et al., 2020). The innate and adaptive immune systems are designed not to recognize the host's own cells and proteins owing to natural immunological tolerance and negative selection of host-specific T lymphocytes in the central lymphatic organs. However, the lack of one of the proteins due to inherited mutations can interfere with this process. When the missing protein is introduced later in life, it can be recognized as dangerous, and an immune response can occur (Alberts, 2002;Siprashvili et al., 2016).
Four previous publications presented results of serological tests, ELISA, and indirect immunofluorescence (IIF) on monkey esophagus in patients with EB (DEB and EB simplex) (Annicchiarico et al., 2015;Esposito et al., 2016;Tampoia et al., 2013;Woodley et al., 2014). Circulating antibodies against BMZ proteins were present in the serum, and the authors suggested the need for further ex vivo experiments to assess their pathogenicity. Although three publications lacked direct immunofluorescence (DIF) on a skin biopsy specimen and IIF on salt split skin (SSS) for detection of tissue-bound autoantibodies, Woodley et al. (2014) additionally performed DIF and IIF on SSS in patients with DEB with a positive ELISA.
Treatment approaches for EB are being investigated, and progress has been recently made; however, they can be threatened by pre-existent circulating antibodies (Eichstadt et al., 2019;Gaucher et al., 2020;Hirsch et al., 2017). In these studies, their presence was assessed before transplantation only by ELISA, IIF, and western blot. In the study of Eichstadt et al. (2019), DIF was performed but only on the transplanted sites after the transplantation and not before. IgG deposition was found in one of the transplanted sites in one of the treated subjects at 3 months and 2 years after transplantation; however, circulating antibodies were only detectable at months 1 and 3 and until month 6 after transplantation. Therefore, they suggested that the humoral immune response was provoked by the transplantation site rather than that the circulating antibodies were preexisting. For the diagnosis of pemphigoid diseases, Meijer et al. (2019) recently proposed that DIF and IIF on SSS and not ELISA or blot are essential, and therefore, these techniques should be used to illustrate whether preexisting antibodies can bind to the skin (Schmidt and Zillikens, 2009). Because these data are missing in the literature, we have investigated skin biopsies and serum of 37 patients with EB with a wide variety of techniques, including DIF and IIF on SSS to assess the presence of circulating antibodies.
Of the 37 patients, 12 were affected with JEB due to mutations in LAMB3 and COL17A1, and 25 were affected with DEB due to mutations in COL7A1 (Table 1). A total of 10 of the 37 included patients had revertant mosaicism (6 with JEB and 4 with recessive DEB [RDEB]) (Supplementary Table S1), that is, healthy, natural skin patches due to correcting somatic mutations that occurred during embryo development or later in life (Pasmooij et al., 2012). We analyzed the already stored punch biopsies from 35 of the 37 patients. Serum samples from all the patients were obtained with permission from medical ethical committees in the Netherlands (University Medical Center Groningen 2013/317) and Spain (Code Hospital Universitario La Paz: PI1359 and PI1595). All patients or their parents provided written informed consent. Furthermore, 14 sera from 13 patients with severe burn wounds were used as the control for ELISA, blotting, and IIF. For a detailed methods description, see previous publications (Groth et al., 2011;Vodegel et al., 2004). The age of the patients at the time of biopsy and serum sampling varied from 0 to 61 years (Supplementary Table S1) for the patients with EB and from 6 to 86 years for the burn wound patients (Supplementary Table S2). DIF was performed on all available skin specimens (1-3 biopsies per patient) to detect human IgG and IgA. Furthermore, we performed IIF on two substrates, monkey esophagus, and SSS; keratinocyte footprint assay for laminin-332 (Giurdanella et al., 2020); and ELISA for COL17 (NC16A), BP230, and COL7. In addition, immunoblot was performed on keratinocyte cell extract to detect antibodies against BP230, COL17, and LAD-1 (Groningen, The Netherlands) (Pas, 2001), on dermal extract to detect antibodies against COL7, and on the recombinant C-terminus of laminin g1 (Lü beck, Germany) to detect antibodies against the p200 protein. ELISA for the NC16A domain of COL17, BP230, and COL7 were performed in two different laboratories in Groningen (The Netherlands) and in Lü beck (Germany) using MBL and Euroimmun kits, respectively. The most important finding of our study is that only two patients (2 of 35, 5.7%) showed linear binding of IgG along the BMZ in DIF. Both patients, #23 and #29, have severe RDEB and were negative for COL7 staining, although patient #29 also had a proven revertant patch. Patient #23 was negative for all serological tests except for one of the ELISA's for COL7 (Table 1). Because patient #23 was negative for the COL7 protein with LH7.2 in the skin, the IgG in the DIF was either not directed to COL7 or it is possible that the patient expresses small amounts of truncated COL7 protein to which the IgG is directed. Patient #29 DIF showed 3þ IgG staining to the BMZ in the revertant skin and negative in the mutant skin. Serology revealed dermal binding of IgG in SSS and positive anti-COL7 autoantibodies in both ELISAs, consistent with a diagnosis of EB acquisita (Figure 1 and Table 1). Both patients did not report any noticeable change of skin phenotype that would indicate the manifestation of EB acquisita, and in the case of patient #29, his revertant skin patch did not blister, even after inducing mechanical trauma (minimal skin rub test) (Figure 1b). This suggests that his general blistering was caused by RDEB and not by circulating autoantibodies as in a case published by Guerra et al. (2018), where EB acquisita occurred in a patient with a mild DEB phenotype.
In 22 of the 37 patients (59.5%), we found at least one positive serological test (Table 1), and in three other patients, we found at least one serological test that was doubtful, meaning that 67.5% (25/37) of our cohort had circulating antibodies against BMZ proteins. Interestingly, the proportion of patients with at least one positive or doubtful serological test was highest in the severe RDEB subgroup (83%, 14/18  A Gosty nski et al. In our cohort, we found positive Euroimmun ELISA for NC16A in 5 of 37 patients (13.5%) and positive MBP ELISA for NC16A in 12 of 37 patients (32.4%). van Beek et al. (2014) described that ELISA for NC16A in the elderly (aged >70 years) was positive in about 6.5% and 3.5% for Euroimmun and MBP ELISAs, respectively. This suggests more positive reactions in the patients with EB than in the published elderly group.
An important question is why circulating autoantibodies are found so frequently in patients with EB and especially in patients with severe RDEB, which do not seem to be clinically relevant. Esposito et al. (2016) showed that patients with EB and especially those with RDEB have higher levels of proinflammatory cytokines than the levels in the control population. In addition, a recent review by Huitema et al. (2021) states that there is evidence that patients with RDEB have an underlying immunity defect. The high number of positive ELISAs in patients with EB may thus be caused by exposure to self-antigens due to repeated skin damage combined with a chronic immunological response or underlying immunity defect because all serological tests were negative in 13 patients with severe acute burn used as controls (data not shown). However, the exact reason is still unknown and warrants further investigation.
To summarize, the clinical relevance of autoantibodies in EB is disputable, especially those detected by ELISA or immunoblot because in the majority of patients, no in vivo binding of antibodies could be shown. Furthermore, because more than half of our cohort had a positive serological test without apparent clinical meaning, an exclusion for the therapy trials for EB based on ELISA causes a risk of omitting possible candidates. We suggest DIF combined with IIF on SSS because these methods have a better diagnostic and prognostic value. However, in EB, the clinical significance of reactivity even in DIF and/or IIF on SSS remains uncertain.

Data availability statement
No datasets were generated or analyzed during this study.