IL-17 A is functionally relevant and a potential therapeutic target in bullous pemphigoid

IL-17A has been identified as key regulatory molecule in several autoimmune and chronic inflammatory diseases followed by the successful use of anti-IL-17 therapy, e.g. in ankylosing spondylitis and psoriasis. Bullous pemphigoid (BP) is the most frequent autoimmune blistering disease with a high need for more specific, effective and safe treatment options. The aim of this study was to clarify the pathophysiological importance of IL-17A in BP. We found elevated numbers of IL-17A+ CD4+ lymphocytes in the peripheral blood of BP patients and identified CD3+ cells as major source of IL-17A in early BP skin lesions. IL17A and related genes were upregulated in BP skin and exome sequencing of 51 BP patients revealed mutations in twelve IL-17-related genes in 18 patients. We have subsequently found several lines of evidence suggesting a significant role of IL-17A in the BP pathogenesis: (i) IL-17A activated human neutrophils in vitro, (ii) inhibition of dermal-epidermal separation in cryosections of human skin incubated with anti-BP180 IgG and subsequently with anti-IL-17A IgG-treated leukocytes, (iii) close correlation of serum IL-17A levels and diseases activity in a mouse model of BP, (iv) IL17A-deficient mice were protected against autoantibody-induced BP, and (v) pharmacological inhibition of lL-17A reduced the induction of BP in mice. Our data give evidence for a pivotal role of IL-17A in the pathophysiology of BP and advocate IL-17A inhibition as potential novel treatment for this disease.


Introduction
Bullous pemphigoid (BP) is a chronic subepidermal blistering autoimmune disease of the elderly associated with autoantibodies against two structural components of the dermal-epidermal junction, BP180 (type XVII collagen, Col17) and BP230 [1]. In BP, binding of anti-Col17 antibodies to the dermal-epidermal junction is a prerequisite for blister formation as shown in various experimental models [2][3][4][5][6]. These data are supported by the close correlation of anti-Col17 antibody serum levels with disease activity in BP patients [2,[7][8][9][10][11]. The activation of complement and the subsequent influx of inflammatory cells into the upper dermis are crucial for the dermal-epidermal splitting that is finally mediated by reactive oxygen species and proteases such as matrixmetalloproteinase-9 (MMP-9) and neutrophil elastase [6,[12][13][14][15]. The mechanisms and key mediators that lead to the infiltration of inflammatory cells and orchestrate the inflammatory milieu in the upper dermis are still largely elusive.
The aim of the present study was to (i) identify the cellular source of IL-17A in BP skin, (ii) analyze IL17A and related genes as risk factors for BP, and (iii) explore the functional relevance of IL-17A in experimental models of BP.

Human material
BP patients (n = 15) were diagnosed based on a compatible clinical picture, linear deposits of IgG and/or C3 at the dermal-epidermal junction by direct IF microscopy of a perilesional skin biopsy, and circulating autoantibodies by indirect IF microscopy on human salt-split skin (epidermal binding) and BP180 NC16A ELISA (Euroimmun, Lübeck, Germany). Whole blood and perilesional skin biopsies were taken at the time of diagnosis before therapy was initiated.
Whole blood was taken from patients with non-inflammatory dermatoses (mostly non-melanoma skin cancer) aged above 75 years (n = 12). Perilesional skin was used from patients with other inflammatory dermatoses (n = 7). Skin biopsies were stored at −80°C. Peripheral blood cells were assayed on the same day when blood was taken. DNA from the peripheral blood of 51 BP patients diagnosed by the same criteria as outlined above, was isolated following the manufacturer's instruction (Qiagen, Hilden, Germany). DNA samples were stored at −80°C until analyzed. The study was performed in accordance with the Declaration of Helsinki and was approved by the ethics committee of the University of Lübeck (08-156, 11-234, 12-178, 15-051). Blood from healthy volunteers for the in vitro and ex vivo studies was used on the same day. Latter studies were approved by the ethics committee of the University of Lübeck (09-140).

Cryosection assay
Separation of the dermal-epidermal junction in cryosections of normal human skin was evaluated using an ex vivo model as previously described [31][32][33]. Briefly, normal human skin was incubated with BP patient and normal human serum, respectively, for 1 h at 37°C. Human leukocytes purified by a sedimentation gradient with dextran 500 (Carl Roth) were added together with 100 μg, 200 μg and 300 μg neutralizing anti-IL-17A antibody and an isotype control (Novartis), respectively, and incubated for 3 h at 37°C before intense washing with PBS. Skin sections were stained with hematoxylin and eosin (H&E), and dermalepidermal separation was evaluated microscopically (Keyence).

Mice
C57BL/6J mice were bred and housed at 12-h light-dark cycle at the experimental animal facility in the University of Lübeck. IL17A −/mice were kindly provided by Yoichiro Iwakura (Tokyo University of Science, Tokyo, Japan) and bred at the experimental animal facility of the Christian-Albrechts-Universität zu Kiel. All injections and bleedings were performed on eight to twelve-week old mice narcotized by intraperitoneal (i.p.) injection of a mixture of ketamine (100 μg/g) and xylazine (15 μg/g). Animal experiments were approved by the Animal Care and Use Committee of Schleswig-Holstein (Kiel, Germany; 21-2/ 11) and performed by certified personnel.

Antibody transfer-induced BP in adult mice
Experimental murine BP was performed as previously described [5]. Briefly, affinity-purified rabbit anti-murine BP180 IgG was injected subcutaneously into the neck of 6-to 8-week-old mice every other day over a period of 12 days at individual doses of 10 mg IgG/injection/ mouse. At the time of IgG injections, mice were weighed and examined for their general condition and evidence of cutaneous lesions (i.e., erythema, blisters, erosions, and crusts). The percentage of affected body surface area was determined on days 4, 8, and 12. Cutaneous lesions were scored as involvement of the skin surface as previously described, assigning a certain percentage of body surface to particular body parts [5]. Calculating the total percentage of affected body surface area resulted in an overall clinical score. At day 12, the mice were sacrificed, sera and biopsies were taken and stored at −20°C until used. The anti-IL-17A antibody (clone 17F3, BioXCell, West Lebanon, USA) and IgG1 isotype control antibody (clone MOPC-21, BioXCell) were injected at a dose of 200μg/injection as detailed below. Serum levels of IL-17A were determined by BioPlex (BioRad).

Statistics
The data are presented as the median or mean ± standard deviation. Statistical calculations were performed using GraphPad Prism version 6 (GraphPad Software Inc, San Diego, USA). For comparison of more than two groups Two-way ANOVA with Sidak's multiple comparisons test was used. When data were nonparametric Mann-Whitney-U test or Kruskal-Wallis test were applied followed by a Dunn posthoc or Bonferroni t-test for multiple comparisons. Pearson product moment correlation was used to test for correlations. In all tests, a p-value of 0.05 was considered to be statistically significant. The gene regulatory network was generated using NetworkAnalyst web server with "A list of proteins and genes" module [34]. In this analysis, significant genes identified by RT-PCR were provided as input. STRING database within web server with the medium threshold of 400 was used as background database for searching interacting partners [35]. We only considered zero degree or direct interaction. The output "SIF" files were visualized using Cytoscape software [36].
3.2. mRNA levels of IL-17A and related mediators are upregulated in perilesional skin of BP patients mRNA levels of IL-17A and selected related cytokines, chemokines, and regulatory factors (n = 32) were analyzed in perilesional skin biopsies from patients with BP and other inflammatory dermatoses, respectively (Table S2). Eighteen genes including IL-17A were significantly upregulated while one gene, CD8, was significantly less expressed in perilesional skin of BP patients compared to controls ( Fig. 2A; Table S1). When the gene regulatory network was visualized IL-17A and related cytokines, e.g., IL-22 and the neutrophil-active chemokine CXCL1 were localized in the center (Fig. 2b). Perilesional skin biopsies of 7 BP patients were stained with antibodies against IL-17A and CD3, CD4, CD68, tryptase, and myeloperoxidase (Fig. 3A). 41% of CD3 + cells, 61% of CD4 + cells, 45% of monocytes/macrophages, 45% of mast cells, and 35% of neutrophils showed IL-17A reactivity (Fig. 3B). Since on some cells, IL-17A staining only appeared on the cell surface, only those cells were counted as IL-17A producers that showed intracellular staining for IL-17A as determined by confocal microscopy (Fig. 3A). When the IL-17A producing cells of a certain cell population were analyzed in relation to the total number of IL-17A producing cells, CD3 + cells appeared to be the main source of IL-17A representing 60% of the IL-17A producers followed by A-D, When CD4 + cells, CD8 + cells, neutrophils, and monocytes of the peripheral blood from treatmentnaïve BP patients and age-and sex-matched controls were analyzed by flow cytometry for their potential to produce IL-17A, significantly more IL-17A + CD4 + cells were found in BP patients compared to controls (A). No difference in the number of IL-17A producing cells between BP patients and controls were seen for CD8 + cells, (B), neutrophils (C), and monocytes (D). ***, p < 0.001. CD68 + monocytes/macrophages (18%) and neutrophils (15%) (Fig. 3C).

BP is associated with mutations in IL17A and related genes
To evaluate whether dysregulation of the IL-17 pathway in BP might be caused by genetic predisposition, we performed whole exome sequencing of 51 patients. 157 genes of the IL-17 pathway, transcription factors of IL-17, IL-17 inhibitors, and related cytokines were analyzed for germline mutations, single nucleotide variants as well as small insertion and deletions (Table S3). Using this approach in 18 (35%) of the 51 BP patients, mutations in at least one of 12 IL-17-related genes were found (Fig. 4, Table S4). Only non-synonymous mutations having a CADD score > 5 as well as known rare mutations (minor allele frequency [MAF] < 0.01) are reported (Fig. 4, Table S4).

IL-17A activates human neutrophils in vitro
When neutrophils derived from peripheral blood of healthy volunteers (n = 3) were incubated with recombinant human IL-17A (500 ng/ ml for 1 h) expression of CD62L on neutrophils was significantly decreased compared with unstimulated cells indicating a higher activation of IL-17A-treated cells (Figs. S2A and B). As expected, treatment of neutrophils with PMA also led to a significant reduced CD62L expression compared to untreated cells (Figs. S2A and C).

Incubation of cryosections of normal human skin with anti-Col17
IgG and subsequently with leukocytes from healthy volunteers pretreated with an isotype control antibody resulted in dermal-epidermal splitting (Fig. 5D). When leukocytes were pre-treated with 100 μg anti-IL-17A antibody dermal-epidermal split formation occurred (Fig. 5A), whereas pre-treatment with 200 μg and 300 μg prevented splitting (Fig. 5 B, C.).

In the antibody transfer mouse model of bullous pemphigoid, IL-17A serum levels correlate with diseases severity
When sera of wildtype mice injected with anti-murine Col17 IgG (n = 34) were assayed for IL-17A, IL-17A levels significantly correlated with the extent of lesion formation in mice (r = 0.531; p = 0.001; Fig.  S3).

IL-17A −/mice are protected from the pathogenic effect of anti-Cil17 IgG
When IL-17A −/-(n = 8) and wildtype mice (n = 11) were injected with anti-Col17 IgG, wildtype mice developed significantly more erythema and erosions on the neck and ears compared to IL-17A −/animals that were largely protected (day 4, p = 0.007; day 8, p = 0.0003; day 12, p= < 0.0001) (Fig. 6 A, B, E, F, I, J). No difference was seen in the deposition of IgG and C3 (data for C3 not shown) at the dermalepidermal junction by direct IF microscopy of perilesional skin biopsies (Fig. 6C, G). While a dense inflammatory infiltrate was seen in the upper dermis of both IL-17A −/and wildtype mice, dermal-epidermal splitting was only found in wildtype mice (Fig. 6 D, H).

Discussion
BP is the most frequent autoimmune blistering disease in Central Europe and North America with a prevalence of about 21.000 patients in Germany [37][38][39]. The only demonstrably effective therapies are systemic or superpotent topical corticosteroids that are, however, associated with considerable adverse events [40]. The number of severe adverse events and the increased mortality of BP patients are, at least in part, due to the corticosteroid treatment [1,41]. While recent randomized controlled clinical trials suggested some effect of doxycycline and dapsone [42,43], more specific, effective, and safe treatment options for BP are lacking.
IL-17A is a pleiotropic cytokine with particular significance at epithelial barrier sites including the skin and oral cavity. It has the ability to induce the expression of proinflammatory cytokines and chemokines in various stromal and innate immune cells [44,45]. IL-17A is of particular importance for the initiation of the immune response by mediating the influx of neutrophils and monocytes and keeping these cells at the site of inflammation [45]. In addition to its role in host defense against microorganisms the importance of IL-17A in autoimmunity and chronic inflammatory disorders such as inflammatory bowel disease and psoriasis has previously been highlighted [45][46][47].
Here, we explored the cellular sources and functional relevance of IL-17A and related inflammatory mediators in the pathophysiology of BP and probed IL-17A inhibition in the treatment of this disease. In the peripheral blood of BP patients, CD4 + lymphocytes, but not CD8 + cells, neutrophils, and monocytes generated IL-17A. Peripheral CD4 + cells from BP patients also produced elevated levels of IL-22. These data indicate that inflammation in BP is not confined to the skin but also involves circulating Th17 cells. To address the IL-17 network in the skin, mRNA levels of IL-17A and related inflammatory mediators in the perilesional skin of BP patients were compared to levels in patients with other inflammatory dermatoses. Upregulation of IL17A, IL17RC, CCR6, and IL23R point to an activated IL-17 network at the site of blister formation. IL-17A is a strong inducer of CXCL1 and CXCL2 expression [44] and the finding of upregulated CXCL1 and CXCL2, encoding powerful chemoattractants for neutrophils, and upregulated ELANE, encoding for neutrophil elastase, validate previous data about the importance of neutrophils and neutrophil elastase in the pathogenesis of BP [13,[48][49][50].
The central role of IL-17A, IL-22, and CXCL1 in the dermal immune network of BP was also highlighted by the gene regulatory network of significantly regulated genes in BP patient skin. Of interest, although not in the focus of this study, increased mRNA levels of anti-inflammatory mediators, such as IL-10 were found in BP skin indicating that mechanisms of resolution of inflammation maybe in place.
To identify the major producer of IL-17A in early skin lesions, the most abundant cell populations that are known to express IL-17A, i.e. neutrophils, macrophages, CD3 + lymphocytes, and mast cells were analyzed for IL-17A expression. By confocal microscopy, we excluded cells that showed IL-17A staining on the cell surface [51] assuming that in these cells, IL-17A was bound to the ubiquitously expressed IL-17RA/ IL-17RC e.g. on different T cell subsets and neutrophils [52,53]. By this method, we found CD3 + lymphocytes to be the major source of IL-17A in BP skin accounting for 60% of the IL-17A-producers. Since here, early BP lesions without split formation were analyzed and cells bearing IL-17A on their surface were not counted our data may reconcile contradictory findings of previous studies that described lymphocytes [54], Th17 cells [55,56], and neutrophils [48] as major source of IL-17 in BP skin.
Since BP is a relatively rare disease with an incidence of about 14-42 new patients/million/year in Central Europe and the UK [1], we decided to perform exome sequencing of 51 BP patients to detect rare mutations associated with BP. One third of the patients showed rare, but mostly severe mutations in the coding regions of 12 IL-17-related genes. While none of the mutations was homozygous, they could have a modulatory effect on the IL-17 pathway activity. The most prevalent genes were related to NFκB signaling, such as IKBKB, NFKB1 and MUC5B. Incorrect activation of NFκB is associated with inflammatory diseases and constant inhibition of NFκB leads to incorrect immune cell development or delayed cell growth. MUC5B was most commonly mutated in our patients. It has been shown to play a role in the airway defense and its deficiency leads to chronic infection [57]. Further studies will need to validate these findings in other BP patient cohorts and analyze whether mutations in IL-17-related genes are associated with certain clinical or immunopathological characteristics.
We then explored the functional relevance of IL-17A for BP in five different approaches. We showed that (i) IL-17A activated normal human neutrophils in vitro, (ii) the dermal-epidermal splitting induced by anti-Col17 IgG was abrogated by incubation of leukocytes with anti-IL17A antibody, and (iii) serum levels of IL-17A closely correlate with the extent of skin lesion in experimental BP in adult mice. Our findings are in line with previous reports on the increased secretion of MMP-9 from neutrophils by IL-17A and CXCL10 [48,58]. MMP-9-deficient mice Fig. 4. Affected IL-17-related genes as identified by exome sequencing. The heatmap depicts rare and severe mutations related to IL-17 signaling and regulation found by whole exome sequencing in a cohort of 51 bullous pemphigoid patients. Heatmap colors correspond to the largest variant allele frequency (VAF) of the dominantly occurring mutation within one gene per patient. The two leftmost columns denote the corresponding population frequency (MAF, minor allele frequency) and the severity of the mutation as quantified by the combined annotation dependent depletion (CADD) score. The MAF score has been extracted from the maximum allele frequency of several population frequency databases, including the 1000 Genomes Project, ESP6500, ExAC, CG46. Only mutations with a CADD score of at least 5 were selected. The score quantifies the deleteriousness of variants in the human genome and is based on multiple annotations such as allelic diversity, pathogenicity, or experimentally measured regulatory effects. Rows and columns have been hierarchically clustered by complete linkage. Annotation of all mutations is detailed in Table S4. were resistant to anti-Col17 IgG-induced skin lesions and CXCL10 expression in macrophages and keratinocytes was shown to be induced by IL-17A [58][59][60].
In additional two approaches (iv), and (v), further evidence for the importance of IL-17A in the pathophysiology of BP was generated. IL-17A −/mice were nearly completely protected against the otherwise pathogenic effect of anti-Col17 IgG and anti-Col17 IgG-injected mice showed significantly reduced skin lesions when treated with an anti-IL-17A antibody compared to animals that received the isotype control antibody. Latter finding suggested pharmacological inhibition of IL-17A being a rational treatment approach in BP. Further studies in experimental murine BP will aim at identifying the main target cell of IL-17A in skin lesions to further explore how IL-17A exerts its effect in BP.
The presented data indicate that IL-17A has a functional role in BP and suggest anti-IL-17A therapy being a rational and promising therapeutic approach in patients with severe BP.
Conflicts of interest D.T. has received honoraria or fees for serving on advisory boards, as a speaker, as a consultant for Lilly, Leo-Pharma, Novartis and UCB, Fig. 6. In experimental murine bullous pemphigoid, IL17A −/− mice are protected and pharmacological inhibition of IL-17A significantly reduces skin lesions. A-J, When IL-17A −/-(A-D) and wildtype mice (E-H) were injected with anti-Col17 IgG (white triangles), wildtype animals developed erythema, erosions, and crusts on neck and ears (A, B), while IL-17A −/mice hardly showed any skin lesions (E, F). Disease activity, as quantified by the affected body surface area, was significantly lower on day 4, 8, and 10 as well as during the entire experimental period in IL-17A −/mice compared to wildtype animals (I, J). In both strains, no difference in IgG deposits at the dermal-epidermal junction was observed (C, G). By lesional histopathology a dense inflammatory infiltrate was seen in the upper dermis in both strains, while subepidermal splitting was only present in wildtype animals (E, H). K-N, when wildtype mice injected with anti-Col17 IgG (white triangles) were treated with anti-IL-17A and isotype control antibody, respectively (black triangles), anti-IL-17A-treated mice showed significantly less skin lesions compared to control antibody-treated mice. Two independent experiments are shown in K and L as well as in M and N. *p < 0.05, **p < 0.01, ***p < 0.001. and grants from Novartis. D.Z. received honoraria from UCB and travel fees from Novartis and UCB. R.J.L. and E.S. received honoraria and travel fees from Novartis as well as honoraria from UCB.