If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Dermatology, PEDEGO Research Unit, University of Oulu, Oulu, FinlandDepartment of Pathology, Cancer and Translational Medicine Research Unit, University of Oulu, Oulu, FinlandMedical Research Center Oulu, Oulu University Hospital and University of Oulu, Oulu, Finland
Department of Dermatology, PEDEGO Research Unit, University of Oulu, Oulu, FinlandMedical Research Center Oulu, Oulu University Hospital and University of Oulu, Oulu, Finland
Department of Dermatology, PEDEGO Research Unit, University of Oulu, Oulu, FinlandMedical Research Center Oulu, Oulu University Hospital and University of Oulu, Oulu, Finland
Department of Dermatology, PEDEGO Research Unit, University of Oulu, Oulu, FinlandMedical Research Center Oulu, Oulu University Hospital and University of Oulu, Oulu, Finland
The deletion of exon 18 from Col17a1 in transgenic ΔNC14A mice results in the absence of the NC14A domain. NC14A corresponds to the human NC16A domain, the immunodominant epitope in bullous pemphigoid. Before the age of 1 year, 84% of ΔNC14A mice have developed severe itch and skin erosion. Further characterization of mice with mutated CoLXVII (Bp180) revealed acanthosis; subepidermal blistering; and inflammatory cell infiltrates, especially neutrophils, eosinophils, and mast cells in the lesional skin. Direct immunofluorescence analysis detected linear complement C3, IgG, and/or IgA deposition in the dermo‒epidermal junction of symptomatic ΔNC14A mice. Elevated gene expression of IL-17‒associated cytokines was detected in the lesional skin. An increased proportion of dendritic cells, myeloid-derived suppressor cells, and NK cells and a decrease of T cells were found in both the spleen and lymph nodes of symptomatic ΔNC14A mice. The proportions of B cells and regulatory T cells were increased in lymph nodes. An 8-week treatment with an anti‒IL-17A decreased the expression of Il6, Il23a, and Cxcl1 in the nonlesional skin. Our results suggest that the absence of the NC14A domain of CoLXVII in mice causes an autoimmune response against the cutaneous basement membrane and manifests as an IL-17‒associated inflammation in the skin.
COLXVII (also known as BP180) is a transmembrane protein expressed in basal keratinocytes in the skin. CoLXVII is a pivotal constituent of the hemidesmosomes, which bind the epidermis to the underlying basement membrane (
). Alterations in the COL17A1 gene result in one subtype of junctional epidermolysis bullosa (EB), the symptoms of which include skin and mucosal blistering, skin atrophy, graying, and loss of hair (
Autoantibodies against COLXVII/BP180 are found in patients with bullous pemphigoid (BP), which typically manifests as large, tense blisters on the erythematous, urticarial, or eczematous skin and severe pruritus in elderly people. In BP, IgG autoantibodies mainly target the extracellular NC16A domain of the COLXVII polypeptide but also affect other regions (
). In addition, IgA and IgE anti-COLXVII antibodies can be found. The gold standard for BP diagnosis is linear IgG and/or complement C3 (C3) deposition in the dermo‒epidermal junction (DEJ) detected by direct immunofluorescence assay. Histologically, BP is characterized by subepidermal blister formation and eosinophilic and neutrophilic infiltrate in the dermis and blister fluid. BP is treated with topical and systemic corticosteroids and other immunosuppressive drugs (
We have developed a transgenic mouse line, ΔNC14A, in which the deletion of exon 18 from Col17a1 results in the absence of the NC14A domain, which corresponds to the human NC16A domain. The ΔNC14A mice appear normal at birth, but during aging, most of them develop BP-resembling clinical and histological signs: severe itch, skin erosions, and ulceration with DEJ-targeting autoantibodies (
). To expand our understanding of the role of COLXVII in the structural integrity of the basement membrane and in the pathogenesis of blistering skin disease, we have investigated cutaneous inflammation and cell populations in the spleen and draining lymph nodes (LNs) of symptomatic ΔNC14A mice, focusing on cytokines/chemokines and inflammatory cells previously associated with BP. In addition, we have tested anti‒IL-17A treatment to ameliorate the symptoms of ΔNC14A mice.
Results
Cutaneous symptoms and signs of inflammation in ΔNC14A mice
Since the generation of the mouse line, a total of 388 ΔNC14A mice have been followed until the humane endpoint or the age of 1 year. Of these, 94% of the female (193 of 206) and 72% of the male mice (131 of 182) have developed visible symptoms, including swollen eyelids and ears; crusted skin; and lesions on the head, neck, snout, chin, and frontal body areas. A total of 62% of the females and 47% of males became symptomatic before the age of 5 months. ΔNC14A mice groom, shake, and scratch themselves intensely assumably owing to itch.
In symptomatic ΔNC14A (S-ΔNC14A) mice, the earliest and the most prominent clinical symptoms were seen in the ears. Ear thickness was two-fold (P < 0.001) greater in S-ΔNC14A mice (n = 23, mean ± SEM = 0.41 ± 0.026 mm) than in asymptomatic ΔNC14A (AS-ΔNC14A) (n = 7, mean ± SEM = 0.21 ± 0.024 mm) or wild-type (WT) (n = 12, mean ± SEM = 0.19 ± 0.002 mm) mice. In the histological studies, small crusts and microblistering at the DEJ, slightly acanthotic epidermis (Figure 1a), and dermal inflammatory cell infiltrates with numerous eosinophils (Figure 1b) were detected in the ears of S-ΔNC14A mice. Samples from more inflamed dorsal skin showed heavy scarring in the upper dermis, together with epidermal acanthosis. Mast cells were detected in the lower dermis in WT (median = 7; range = 5‒11 per high-power field) and AS-ΔNC14A (median = 13; range = 5‒21 per high-power field) samples. In S-ΔNC14A mice, smaller and often degranulated mast cells were found in the upper dermis (Figure 1c and d), and the count was increased both on lesional dorsal skin (median = 38; range = 22‒80 per high-power field) and in the ears (median = 40; range = 14‒60 per high-power field). T cells and myeloperoxidase-positive neutrophils were detected near the lesions (Figure 1e‒h).
Figure 1Histological characteristics of S-ΔNC14A mice developing itch and skin symptoms. (a) Thickened epidermis and (b) microblisters and inflammatory cell infiltrate with numerous eosinophils (arrowheads) are seen in H&E-counterstained swollen ear samples. (c) In the inflamed dorsal skin, Giemsa-stained mast cells (arrows) were often degranulated and mainly found in the upper dermis (enlargement shown in d). (e) Numerous myeloperoxidase-positive neutrophils were detected in the lesional skin of S-ΔNC14A mice. In addition, (f) CD3-positive and (g) CD4-positive T cells were found in the upper dermis, and (h) few CD8-positive T cells were found in the lower dermis. Bars = 100 μm for a and c and 50 μm for b and d‒h. S-ΔNC14A; symptomatic ΔNC14A.
Direct immunofluorescence analysis of ear samples was performed for 17 S-ΔNC14A mice. All these showed strong linear C3 staining at the DEJ (Figure 2a). Linear IgG fluorescence was detected in the DEJ or on the roof of the microblisters in 11 of the S-ΔNC14A samples, and of these, nine were positive for IgG1, five for IgG2b, and three for IgG3 (Figure 2a and Supplementary Table S1). Two S-ΔNC14A mice were IgG negative but positive for IgG subclass‒specific antibodies (Supplementary Table S1). All S-ΔNC14A mice were IgG2c negative, and WT samples were negative for all IgG subclasses. A strong linear IgA deposition was detected in the DEJ in 10 S-ΔNC14A mice (Figure 2a). IgE-positive cells were detected in the dermis of 14 of 17 S-ΔNC14A mice ears, whereas WT samples were negative (Figure 2a). Circulating autoantibodies were analyzed with indirect immunofluorescence, which showed linear IgG staining in the DEJ in 12 of 17 S-ΔNC14A mice and a positive IgA signal in the sera of three (Figure 2b and Supplementary Table S1). In addition, S-ΔNC14A mice had significantly increased plasma IgE levels (mean ± SEM = 172 ± 49 ng/ml) compared with WT mice (mean ± SEM = 0.9 ± 0.4 ng/ml) (P = 0.001) (Supplementary Figure S1). We also measured thymic stromal lymphopoietin (TSLP) levels in skin lysates because it has been previously associated with BP (
). We did not find differences in TSLP protein levels between study groups (Supplementary Figure S2).
Figure 2Ig subclasses and C3 were detected in the same location as COLXVII in S-ΔNC14A mouse skin. (a) Frozen ear sections from WT and S-ΔNC14A mice were immunostained with COLXVII, C3, and Ig subclass‒specific antibodies. A weak discontinuous granular DEJ staining of C3 in WT samples was considered negative as in S-ΔNC14A samples; the C3 staining was strong and linear in the DEJ. IgE-positive cells were detected in the dermis of S-ΔNC14A mice. (b) WT mouse tail sections were used to detect circulating IgG- and IgA-class autoantibodies. In WT serum, scattered staining in the DEJ was detected in a few mice, whereas in S-ΔNC14A, both IgG and IgA stainings were stronger and linear. Arrows and dashed lines depict the DEJ. Bars = 50 μm. C3, complement C3; DEJ, dermo‒epidermal junction; IIF, indirect immunofluorescence; S-ΔNC14A; symptomatic ΔNC14A; WT, wild-type.
). Elevated mRNA levels of Il1b, Il17a, Il23a, Il36a, Il36g, Tnf, S100a8, S100a9, and Cxcl2 were detected in the lesional dorsal skin of S-ΔNC14A mice compared with those in the nonlesional skin of S-ΔNC14A or WT mice. In addition, Il6 and Cxcl1 were increased in the lesional S-ΔNC14A skin compared with those in WT skin (Table 1 and Supplementary Figure S3). The levels of Il36g, Tnf, Cxcl2, and S100a9 were also higher in the nonlesional skin of S-ΔNC14A mice than in WT skin. The level of Tgfb1 was lower in S-ΔNC14A nonlesional skin than in WT skin and higher in the lesional skin than in nonlesional skin of S-ΔNC14A mice. Tslp expression level was slightly lower in the lesional skin of S-ΔNC14A mice than in the nonlesional skin of S-ΔNC14A or WT mice (Table 1). No differences were detected between groups in the expression of Il22 or Il12p40. Taken together, these results suggest that the IL-23/IL-17 axis is activated in the skin of S-ΔNC14A mice. Therefore, we analyzed IL-17 in more detail at the protein level. In immunostainings, IL-17A was detected in tryptase-positive mast cells and myeloperoxidase-positive neutrophils from the S-ΔNC14A skin (Figure 3a and b). Owing to the lack of functional pair of antibodies, we performed immunostaining of consecutive skin sections and detected numerous CD3-positive cells (Figure 3c) in the same area as for smaller IL-17A‒positive cells (Figure 3b), suggesting that there are some IL-17A‒expressing T cells in the S-ΔNC14A skin. ELISA assays revealed that the plasma concentration of IL-17A was greater in S-ΔNC14A mice (mean ± SEM = 1,111 ± 186 pg/ml) than in WT mice (mean ± SEM = 153 ± 82 pg/ml) (P < 0.0001) (Supplementary Figure S4).
Table 1Fold Differences in the Levels of Inflammatory Markers’ mRNA Expression as Compared between Specified Sample Groups
Target
Untreated
Anti‒IL-17A Treated
Anti‒IL-17A Treated Versus Untreated S-ΔNC14A Mice
↑ denotes upregulated, and ↓ denotes downregulated. Statistical significances were determined by (i) Mann‒Whitney U test and (ii) Wilcoxon signed-rank test. Expression levels were analyzed by qPCR as described in the Supplementary Materials and Methods. Fold levels were calculated by comparing the arithmetic means of the specified groups. Statistically significant (P < 0.05) results are depicted in bold.
Figure 3IL-17A was detected in the lesional skin of symptomatic ΔNC14A mice. (a) IL-17A (green) was partly colocalized to the same cells (arrows and insets) with mast cell tryptase (red) and (b) myeloperoxidase (red) (neutrophils). IL-17A positivity was also detected in smaller inflammatory cells. (c) Numerous CD3-positive cells (green) were detected in the same area in consecutive skin sections. Bars = 100 μm in a (20 μm, inset), 50 μm in b (20 μm, inset), and 50 μm in c.
S-ΔNC14A mice have significantly enlarged spleens (spleen [mg]/body weight [g] ratio, mean ± SEM = 6.4 ± 0.32) compared with the WT mice, whereas AS-ΔNC14A mice have a spleen weight comparable with WT mice (ratios 3.8 ± 0.22 and 3.7 ± 0.11, respectively) (Figure 4a). Immune cell populations in the spleen and draining LNs were analyzed using flow cytometry. Young AS-ΔNC14A mice had higher populations of CD11b+ myeloid cells in the spleen and of B cells (B220+) in LNs than WT mice. In contrast, T-cell populations, especially CD3+CD8+ cells, were reduced in AS-ΔNC14A LNs (Figure 4b and Supplementary Figure S5a). Despite the proportion of regulatory T cells (Tregs) (CD4+FoxP3+) being higher, the absolute numbers of Tregs did not increase because of the overall drop in T-cell population (Figure 4b). In general, no defect in T-cell development was observed in the thymus (data not shown) nor were there any significant changes in CD3+CD4+ cells (Figure 4b).
Figure 4Inflammatory cell populations in the spleen and lymph nodes change during skin inflammation. (a) S-ΔNC14A mice (untreated, n = 43; anti‒IL-17A treated, n = 10) have enlarged spleens compared with WT (n = 61) or AS-ΔNC14A (n = 25) mice. Immune cell populations in the spleen and lymph nodes of (b) WT mice (n = 4) and AS-ΔNC14A mice (n = 5) aged 7 weeks and (c) adult WT (n = 9) and S-ΔNC14A mice (n = 10) with or without anti‒IL17A treatment. Quantified data of flow cytometry analysis are shown. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 as determined by two-tailed Student’s t-test. AS-ΔNC14A, asymptomatic ΔNC14A; LN, lymph node; S-ΔNC14A, symptomatic ΔNC14A; WT, wild-type.
In S-ΔNC14A mice, the overall T-cell count in both spleen and LNs was reduced. In S-ΔNC14A mice, the proportions of CD3+CD4+ and CD3+CD8+ cells decreased both in the spleen and LNs (Figure 4c and Supplementary Figure S5b). The proportion of Tregs was increased, and CD11b‒Gr1+ decreased in LNs, whereas CD4‒IFNg+ cells increased in the spleen, and CD4+IFNg+ cells decreased in the LNs. In addition, the proportions of CD11b+Gr1+ myeloid-derived suppressor cells, CD11b+ myeloid cells, CD11b+CD11c+ dendritic cells (DC), and NK cells were increased in both the spleen and LNs of S-ΔNC14A mice compared with those of the WT mice (Figure 4c). The proportion of B cells was reduced in the spleen but increased in the LNs of S-ΔNC14A mice (Figure 4c).
Symptoms of ΔNC14A mice are restrained with anti‒IL-17A treatment
Finally, 10 S-ΔNC14A mice were treated with anti‒IL-17A for 8 weeks. As a control group, we closely monitored 10 untreated S-ΔNC14A mice. We were able to maintain only two untreated mice for 8 weeks, whereas the others reached the humane endpoint within 2–7 weeks. The relative skin score of anti‒IL-17A‒treated mice remained stable during the 8-week surveillance (Figure 5a). At termination, the treated mice had less ulcerated skin, and the area of lesional skin was smaller than in the untreated mice (Figure 5b and c). Histological as well as direct immunofluorescence and indirect immunofluorescence findings in the skin of the anti‒IL-17A‒treated mice were similar to those of the untreated mice (Supplementary Table S1). The plasma IL-17A levels (mean ± SEM = 1,290 ± 333 pg/ml, P = 0.776) (Supplementary Figure S4) and IgE levels (mean ± SEM = 125 ± 38 ng/ml, P = 0.615) (Supplementary Figure S1) in the treated mice were comparable with those in the untreated mice, as were the inflammatory cell populations in the spleen and LNs (Figure 4c). When the nonlesional skin of the anti‒IL17A‒treated mice was compared with the untreated nonlesional skin, decreased expression levels of Il6, Il23a, and Cxcl1 were detected in treated mice (Table 1 and Supplementary Figure S3). In the lesional skin, no differences were found between treated and untreated mice.
Figure 5Anti‒IL-17A treatment stopped the disease progression in S-ΔNC14A mice. Ten S-ΔNC14A mice were treated with anti‒IL-17A for 8 weeks. As a control group, we monitored closely 10 untreated S-ΔNC14A mice, of whom two reached the humane endpoint within 2 weeks, two reached within 3 weeks, two reached within 4 weeks, one reached within 5 weeks, and two reached within 7 weeks. Only two untreated mice were acceptable to maintain until the 8th week. (a) Affected skin area was compared with the timepoint when visible symptoms were first detected (mean ± SEM). (b) The area of lesions and crusted skin increased during the surveillance of untreated S-ΔNC14A mice. (c) Clinical appearance of S-ΔNC14A mice remained stable during the 8 weeks of anti‒IL-17A treatment. ∗P < 0.05 and ∗∗P < 0.01 as determined by Mann‒Whitney U test. S-ΔNC14A, symptomatic ΔNC14A.
Deletion of exon 18 from Col17a1 in mice results in a cutaneous inflammatory response that includes linear DEJ deposition of C3, IgG, and IgA as well as infiltrates of eosinophils, neutrophils, T cells, and mast cells. In this study, we have also described changes in cytokine profile in the skin of S-ΔNC14A mice. The most important finding was the upregulation of IL-17‒associated cytokines, which play a central role in the pathogenesis of several inflammatory skin diseases in humans, including BP (
IL-23/IL-17 axis activates IL-1β-associated inflammasome in macrophages and generates an auto-inflammatory response in a subgroup of patients with bullous pemphigoid.
). In addition, we detected an increase in the populations of myeloid cells, especially myeloid-derived suppressor cells and DCs, among splenocytes. These changes were observed in young AS-ΔNC14A mice but became more prominent in the spleen and draining LNs of S-ΔNC14A mice, indicating that these cells may contribute to the skin inflammation. Because Tregs have an important role in maintaining the anergy of autoreactive lymphocytes (
), the increase of Tregs and B cells with simultaneous decrease of CD8+ T cells in the LNs of AS-ΔNC14A mice suggests an imbalance of immunity that might cause the break of immune tolerance in the aging ΔNC14A mice. Interestingly, Tgfb was downregulated in the nonlesional skin of S-ΔNC14A mice. TGFβ is involved both in proinflammatory (T helper 17) and immunosuppressive (Tregs) responses (
). Whether the upregulation of IL-17A and the relative increase of Tregs in LNs are related to decreased Tgfb levels remains to be investigated.
Another mouse model, named ΔNC16A, has a slightly larger deletion from Col17a1 than our ΔNC14A line. It shares several features of our ΔNC14A mouse: the development of pruritus and skin lesions with inflammatory cell infiltrate, especially myeloid-derived suppressor cells and eosinophils, and the increase in IgE levels as well as certain cytokines such as IL-1β, CXCL1, and CXCL2 (
). In contrast to ΔNC16A mice, we have detected enlarged spleens only in S-ΔNC14A mice. In addition, whereas ΔNC16A mice had increased levels of TSLP and unchanged Tnf mRNA levels (
), S-ΔNC14A mice had similar TSLP protein levels and increased Tnf mRNA levels compared with WT mice. Different phenotypes of these closely related mouse lines may result from distinct deletions from the Col17a1 gene. ΔNC14A mice have a deletion of only one exon, resulting in the deletion of 72 of the 76 amino acids from the NC14A domain, whereas ΔNC16A mice have a deletion of two exons (
), we also found increased mRNA levels of Cxcl1 and Cxcl2 in the lesional skin of S-ΔNC14A mice. The current hypothesis is that mast cells residing near blood vessels initiate neutrophil recruitment from circulation by releasing CXCL1 and CXCL2 (
). In addition, in studies of passive transfer BP mice, granzyme B from basophils and/or mast cells degraded hemidesmosomal proteins, including COLXVII, inducing keratinocytes to secrete CXCL2 and resulting in the recruitment of neutrophils (
). An abundance of neutrophils along with C3 deposition and degranulated mast cells in S-ΔNC14A skin suggest that mast cells activated by complement have directly or indirectly attracted neutrophils on the site of inflammation.
Although cutaneous inflammation is detected in both ΔNC14A and ΔNC16A (
) mouse lines, linear DEJ depositions of C3 and IgG/IgA are present only in our S-ΔNC14A mice, IgG1 and IgG2b being the most common IgG subclasses. In mice, antigen-specific mechanisms drive an Ig-class switch to IgG1 (
), which is the Ig subclass that binds most potently to FcRn, a major histocompatibility complex‒like receptor with a major role in antigen presentation (
). In a humanized BP mouse model, anti-NC16A antibodies cause C3 deposition at the DEJ, dermal neutrophil infiltrate, and degranulation of mast cells, which together result in subepidermal blistering (
Human IgG1 monoclonal antibody against human Collagen 17 noncollagenous 16A domain induces blisters via complement activation in experimental bullous pemphigoid model.
). Circulating IgG autoantibodies targeting the DEJ were detected in 70% of the sera of our S-ΔNC14A mice. We have previously shown that autoantibodies bound to the epidermal side of the salt-induced blister and detected 180-kDa collagenase-sensitive protein in the keratinocyte extract, suggesting the target being COLXVII. Only a portion of the sera of S-ΔNC14A mice was found positive in immunoblotting, suggesting alternating titers and affinities of autoantibodies (
). The development of autoantibodies against DEJ in ΔNC14A mice is unexpected because there are only a few case reports of patients with EB having developed circulating anti-skin antibodies, including anti-BP180 (
). To the best of our knowledge, there is only one reported case of a patient with junctional EB with a LAMB3 alteration displaying linear C3 and IgG deposition at the DEJ as detected by direct immunofluorescence (
). Perhaps the disrupted structure of the basement membrane in the ΔNC14A mice skin is prone to the exposal of neoepitopes, leading to the development of autoantibodies during aging.
ΔNC16A and probably also ΔNC14A mice manifest a partial but not clean T helper 2 type inflammatory response illustrated by upregulation of IL-13 but not IL-4 in ΔNC16A mice (
). Our results suggest an IL-17/IL-23 response in ΔNC14A mice, but so far, this has not been studied in detail with ΔNC16A model. In the S-ΔNC14A lesional skin, Il1b, Il6, Tnf, Il23a, and Il17a were upregulated. Similar findings have also been reported in BP perilesional skin (
). IL-1β is involved in the initial formation of blisters in BP and is found in blister fluid where its quantity correlates with increased amounts of IL-17 and IL-23 (
IL-23/IL-17 axis activates IL-1β-associated inflammasome in macrophages and generates an auto-inflammatory response in a subgroup of patients with bullous pemphigoid.
). S-ΔNC14A mice had elevated IL-17A plasma levels, and in the lesional skin, IL-17A was localized in mast cells, neutrophils, and smaller inflammatory cells, likely T cells because numerous CD3-positive cells with similar size and shape as those of IL-17A‒positive cells were detected at the same area. Similarly, in BP skin, IL-17A was found in mast cells, neutrophils, and monocytes/macrophages as well as in T cells (
). Although IL-23 primarily induces IL-17 production in T helper 17 cells, both IL-23 and IL-1β induce IL-17 secretion in other cell types, for example, psoriatic skin mast cells. Therefore DCs producing IL-23 may induce mast cells and neutrophils to secrete IL-17 (
). Upregulation of Tnf in lesional skin together with the increased populations of DCs, B cells, and Tregs in the draining LNs implies the stimulation of antigen presentation in S-ΔNC14A mice. Further supporting IL-17‒associated inflammation, the expression of the IL-17‒inducible genes Il36a, Il36g, S100a8, and S100a9 were upregulated in the lesional skin of S-ΔNC14A mice. All these cytokines attract myeloid cells, thereby promoting inflammation (
Previously, anti‒IL-17A treatment of WT mice reduced anti-COLXVII‒induced skin lesions, and Il17a‒/‒ mice did not develop pathogenic skin symptoms after the injection of anti-COLXVII antibodies (
). Anti‒IL-17A treatment stopped the disease progression in S-ΔNC14A mice, allowing us to maintain the mice longer than without treatment. Anti‒IL-17A treatment reduced the neutrophil-attracting and -associated cytokines Il6, Il23a, and Cxcl1 in the nonlesional skin of treated mice but did not change plasma levels of IL-17A. However, because anti‒IL-17A was administered subcutaneously to the mice, plasma levels do not necessarily reflect the effectiveness of the therapy. For example, secukinumab, human anti‒IL-17A, was shown to rapidly distribute to psoriatic lesions after subcutaneous administration and improve the skin status, whereas serum levels of secukinumab were lower than in healthy subjects (
). Although IL-17A was not cleared from the circulation of S-ΔNC14A mice, the antibody bound to IL-17A may have inhibited its interaction with IL-17 receptor and further signaling.
Our current data show that in mice, the lack of the NC14A domain of COLXVII, which corresponds with the immunodominant epitope in BP, paradoxically causes an autoimmune response against the DEJ, IL-17‒associated inflammation, and expansion of inflammatory cell populations in the spleen and LNs. To the best of our knowledge, such inflammatory responses have not been reported in patients with junctional EB carrying alterations to the COL17A1 gene. In addition, our findings that ΔNC14A mice develop DEJ-targeting antibodies that detect immunoprecipitated COLXVII protein (
) and that prominent skin lesions are developing in adult mice represent more the clinical and histological signs of patients with BP than those of patients with junctional EB. Perhaps the structural modifications of COLXVII and the DEJ in ΔNC14A mice somehow reflect the stage of BP in which the autoimmune reaction results in the destruction of the NC16A domain and the decrease of COLXVII/BP180, resulting in further blister formation. Altogether, ΔNC14A mice provide a feasible model for studying the pathological events that may mimic human BP after the loss of hemidesmosome integrity.
Materials and Methods
Animal experiments
The generation of the ΔNC14A mouse line has previously been described (
). Briefly, mice of a C57BL/6JOlaHSd background have exon 18 deleted from the Col17a1 gene, resulting in the absence of 72 amino acids (ΔAla498–Asn569) of the COLXVII NC14A domain. The University of Oulu Animal Ethics Committee and the Southern Finland Regional State Administrative Agency (ESAVI/10181/04.10.07/2017) approved the animal experiments described in this study. The study followed the animal care principles and experimental procedures according to Finnish legislation, the European Convention for the protection of vertebrate animals used for experimental and other scientific purposes (ETS 123), and EU Directive 2010/63/EU. A total of 50 μg neutralizing anti-mouse IL-17A antibody (BP0173, BioXCell, Lebanon, NH) was administered subcutaneously to symptomatic female mice aged 10–14 weeks once a week for 8 weeks. The area of the affected skin was evaluated twice per week using skin scores (
). Briefly, each body part of the mouse was assigned for a certain percentage (ears, 5%; eyes, 1%; snout and chin, 5%; head and neck, 9%; forelegs, 10%; hind legs, 20%; tail, 10%; and trunk, 40%), and the percentage of each area affected by skin lesions (erythema, crusts, erosions) was evaluated, resulting in overall skin score. Untreated mice were sacrificed when the humane endpoint was reached, typically within 2–3 weeks after the detection of symptoms, that is, scratching, skin lesions, or erosions. WT littermates were collected at the same age as for untreated mice.
Additional methods
Histology; immunostainings; and IL-17A, TSLP, and IgE ELISA as well as gene expression analysis by qPCR and flow cytometric analysis of immune cell populations in the spleen and LNs are described in the Supplementary Materials and Methods.
Data availability statement
No datasets were generated or analyzed during this study.
KT has received educational grants from Sanofi Genzyme and honoraria from Abbvie, Novartis, Sanofi Genzyme, Janssen-Cilag, Bristol-Myers Squibb, and UCB Pharma for consulting and/or speaking. The remaining authors state no conflict of interest.
Acknowledgments
We thank Anja Mattila and Erja Tomperi for their excellent technical assistance. This work was carried out with the support of The Oulu Laboratory Animal Centre Research Infrastructure, University of Oulu, Finland. Steve Smith is acknowledged for careful proofreading and language editing of the manuscript. This study was financially supported by grants from thel Academy of Finland (grant numbers 294738 and 330822 to KT and 325965 to ZJC); the Sigrid Juselius Foundation; the University of Oulu Graduate School; and the Medical Research Center, Oulu University Hospital (Oulu, Finland).
Formalin-fixed and paraffin-embedded ear and skin sections of wild-type and ΔNC14A mice (both lesional and nonlesional skin) were stained with H&E and Giemsa for morphological analysis. Mast cells were calculated using a ×40 high-power field objective at the point of the highest density of the infiltrate. For immunostainings, deparaffinized and rehydrated sections were boiled in Tris/EDTA or citrate buffer (Supplementary Table S2) and incubated with anti‒IL-17A, anti-myeloperoxidase, anti‒mast cell tryptase, or CD3 antibodies and were then further incubated in Alexa Fluor 488‒conjugated anti-rabbit or Alexa Fluor 568 anti-mouse antibodies (Supplementary Table S2). Alternatively, rabbit-specific HRP/DAB (ABC) Detection IHC kit (Abcam, Cambridge, United Kingdom) was used to visualize bound antibodies. Frozen ear sections (6 wild-type and 17 symptomatic ΔNC14A mice) were blocked with 1% BSA (in PBS), and Ig subclasses were detected with specific antibodies conjugated with fluorescent labels. Anti‒complement C3 and anti-COLXVII (carboxy-terminal specific) antibodies were followed with anti-rabbit IgG Alexa Fluor 488 as a secondary antibody. Indirect immunofluorescence was performed on frozen sections of wild-type mouse tail with 1:50 dilution of mouse serum and Alexa Fluor 488‒conjugated anti-mouse IgG or FITC-conjugated anti-mouse IgA (Supplementary Table S2). H&E and Giemsa sections were captured with a Leica DM3000 microscope with ×20 objective and fluorescent images with a Carl Zeiss AxioImager D2 with ×20 objective and appropriate filter sets.
ELISA
Mouse blood samples were collected in terminal anesthesia, and plasma samples were stored at ‒70 °C. Dilutions of 1:10 were used to run the ELISA for mouse IL-17A (BioLegend, San Diego, CA), and 1:500 dilutions were used for mouse IgE (BioLegend) according to the manufacturer’s instructions. All the samples were analyzed in duplicates.
The amount of thymic stromal lymphopoietin protein was measured from skin lysates. Frozen skin samples were pulverized. After the addition of PBS (Corning, Manassas, VA) with 2 mM EDTA and protease inhibitors (Sigma-Aldrich, St. Louis, MO), samples were homogenized in a TissueLyser (Qiagen, Hilden, Germany) and centrifuged at 20,000g for 20 minutes at +4 °C. A total of 20 μg of total protein was used for thymic stromal lymphopoietin ELISA (BioLegend). Samples were analyzed in duplicates.
Differences between groups were analyzed using Kruskal‒Wallis and Mann‒Whitney U-tests.
qPCR
Skin samples from lesional and nonlesional dorsal areas were collected after termination. Freshly frozen mouse skin samples were pulverized and shredded in a TissueLyser (Qiagen), and total RNA was isolated with the RNeasy Plus Universal kit (Qiagen) according to the manufacturer’s instructions. A total of 0.5 μg total RNA was used to reverse transcribe cDNA with the High-Capacity cDNA Reverse Transcription kit (Life Technologies/Thermo Fisher Scientific, Waltham, MA). A pooled cDNA sample, including all the samples, was used as a control. The expression levels were analyzed with validated PrimePCR primer pairs (Supplementary Table S3), SsoAdvanced Universal SYBR green supermix, and CFX connect qPCR machinery and software (all from Bio-Rad Laboratories, Hercules, CA). Hprt and B2m were used as reference genes because their stability values were below the recommended values for heterogenous samples (coefficient variance <0.5, mean value <1). Statistical analyses were conducted with IBM Statistical Package for the Social Sciences software for Windows, version 24.0 (IBM, Armonk, NY). Differences in gene expression between wild-type and nonlesional and lesional skin of anti‒IL-17A‒treated and ‒untreated symptomatic ΔNC14A mice groups were analyzed using the Kruskal‒Wallis test and pairwise with the Mann‒Whitney U test. Statistical significances between dependent nonlesional and lesional groups were analyzed with the Wilcoxon signed-rank test. P < 0.05 was considered statistically significant.
Isolation of splenocytes and lymph node cells and flow cytometric analysis
Spleen, thymus, axillary, inguinal, and brachial lymph nodes were collected after termination of the mice. The tissues were crushed in PBS, and cells were collected by centrifugation. Red blood cells were lysed by incubation in 150 mM ammonium chloride, 1 mM sodium bicarbonate, and 0.1 mM EDTA (pH = 7.2) for 5 minutes (room temperature). The washed cells were stored in liquid nitrogen until analysis. Cells were stained with antibodies against the following surface markers: CD3, CD4, CD8, CD44, CD62L, CD11b, Gr1, B220, NK1.1 (BD Biosciences, Franklin Lakes, NJ), and CD11c (eBioscience, San Diego, CA). Intracellular staining of Foxp3 (BD Biosciences) was performed using a Foxp3 Staining kit (eBioscience) according to the manufacturer’s instruction. The stained cells were analyzed using an LSR Fortessa flow cytometer (BD Biosciences), and data were analyzed using FlowJo software (version 10, Tree Star, Ashland, OR). P-values were calculated using the Student’s t-test.
Supplementary Figure S1Plasma IgE levels were increased in untreated and anti‒IL-17A‒treated S-ΔNC14A mice compared with those in WT littermates. IgE concentrations were measured by ELISA. Bars represent mean values, and whiskers depict SEM. Statistical significances were determined by Kruskal‒Wallis and Mann‒Whitney U tests. ∗∗P < 0.01 and ∗∗∗P < 0.001. S-ΔNC14A, symptomatic ΔNC14A; WT, wild-type.
Supplementary Figure S2TSLP protein levels are comparable in the nonlesional and lesional skin of S-ΔNC14A and WT mice. Anti‒IL-17A administration did not change the skin TSLP levels. The levels were measured by ELISA from skin samples collected after the termination. Bars depict mean values, and whiskers indicate SEM. S-ΔNC14A, symptomatic ΔNC14A; TSLP, thymic stromal lymphopoietin; WT, wild-type.
Supplementary Figure S3Various cytokine levels (relative normalized mRNA expression) were analyzed by qPCR from the dorsal skin of WT mice (n = 10) and S-ΔNC14A mice with (n = 10) or without (n = 15) anti‒IL-17A treatment. Statistical significances were determined using the Mann‒Whitney U test between independent groups and the Wilcoxon signed-rank test between dependent groups (nonlesional vs. lesional skin of the same mice). ∗∗∗P < 0.001, ∗∗P < 0.01, and ∗P < 0.05. The black line depicts the arithmetic mean. S-ΔNC14A, symptomatic ΔNC14A; WT, wild-type.
Supplementary Figure S4S-ΔNC14A mice have increased circulating levels of IL-17A. Plasma levels of IL-17A were analyzed by ELISA from untreated (n = 17) and anti‒IL-17A‒treated (n = 10) S-ΔNC14A mice as well as WT littermates (n = 10) after termination. Bars depict mean values, and whiskers indicate SEM. ∗∗∗P < 0.001 as determined by Mann‒Whitney U test. S-ΔNC14A, symptomatic ΔNC14A; WT, wild-type.
Supplementary Figure S5Representative flow cytometry plots of immune cell populations in the spleen and lymph nodes of (a) WT and AS-ΔNC14A mice aged 7 weeks and (b) adult WT and S-ΔNC14A mice that received or did not receive anti‒IL17A treatment. AS-ΔNC14A, asymptomatic ΔNC14A; S-ΔNC14A, symptomatic ΔNC14A; SSC, side scatter; WT, wild-type.
Supplementary Table S1Circulating (Serum) and Skin-Resident Igs and Complement C3 Detected by Indirect and Direct Immunofluorescence
Mouse
Skin
Serum
IgG
IgG1
IgG2b
IgG2c
IgG3
IgA
IgE
C3
IgG
IgA
WT-1
‒
‒
‒
‒
‒
‒
‒
‒
‒
‒
WT-2
‒
‒
‒
‒
‒
‒
‒
‒
‒
+
WT-3
‒
‒
‒
‒
‒
+
‒
‒
+
‒
WT-4
‒
‒
‒
‒
‒
‒
‒
‒
‒
‒
WT-5
‒
‒
‒
‒
‒
‒
‒
‒
‒
‒
WT-6
‒
‒
‒
‒
‒
‒
‒
‒
‒
‒
ΔNC14A-1
‒
‒
‒
‒
+
+
+
+
+
‒
ΔNC14A-2
+
‒
‒
‒
+
‒
+
+
+
‒
ΔNC14A-3
‒
+
+
‒
‒
+
+
+
+
+
ΔNC14A-4
‒
‒
‒
‒
‒
+
‒
+
+
‒
ΔNC14A-5
+
+
+
‒
‒
‒
+
+
+
‒
ΔNC14A-6
+
+
+
‒
‒
‒
‒
+
+
‒
ΔNC14A-7
‒
+
‒
‒
‒
‒
‒
+
+
‒
ΔNC14A-8
+
+
‒
‒
‒
+
+
+
‒
‒
ΔNC14A-9
+
+
+
‒
‒
‒
+
+
+
‒
ΔNC14A-10
‒
‒
‒
‒
‒
‒
+
+
+
‒
ΔNC14A-11
+
+
‒
‒
‒
+
+
+
+
‒
ΔNC14A-12
‒
‒
‒
‒
‒
‒
+
+
‒
‒
ΔNC14A-13
+
+
‒
‒
+
+
+
+
‒
+
ΔNC14A-14
+
‒
‒
‒
‒
+
+
+
+
‒
ΔNC14A-15
+
+
+
‒
‒
+
+
+
+
+
ΔNC14A-16
+
‒
‒
‒
‒
+
+
+
‒
‒
ΔNC14A-17
+
‒
‒
‒
‒
+
+
+
‒
‒
Anti‒IL-17A treated
Start/End
Start/End
Start/End
Start/End
Start/End
Start/End
Start/End
Start/End
End
End
ΔNC14A-1
+/+
+/‒
‒/‒
‒/‒
‒/‒
+/+
+
+/+
+
‒
ΔNC14A-2
‒/+
‒/‒
‒/‒
‒/‒
‒/‒
+/+
+
+/+
+
‒
ΔNC14A-3
‒/‒
‒/‒
‒/‒
‒/‒
‒/‒
+/+
+
+/+
‒
‒
ΔNC14A-4
‒/‒
‒/‒
‒/‒
‒/‒
‒/+
‒/+
+
+/+
+
‒
ΔNC14A-5
+/+
‒/‒
‒/‒
‒/‒
‒/‒
‒/‒
+
+/+
‒
‒
ΔNC14A-6
+/+
‒/+
‒/‒
‒/‒
+/‒
+/+
+
+/+
‒
‒
ΔNC14A-7
‒/‒
‒/‒
‒/‒
‒/‒
‒/‒
‒/‒
+/‒
+/+
‒
‒
ΔNC14A-8
‒/‒
‒/‒
‒/‒
‒/‒
‒/‒
‒/+
+/+
+/+
‒
‒
ΔNC14A-9
+/+
‒/‒
‒/‒
‒/‒
‒/‒
‒/‒
+/+
+/+
‒
‒
ΔNC14A-10
+/+
+/‒
‒/‒
‒/‒
‒/‒
+/‒
+/+
+/+
+
+
Abbreviation: WT, wild-type.
Skin-resident antibodies were analyzed from anti‒IL-17A‒treated mice both before the initiation (start) and at the end of the treatment. Staining was judged as positive if signal was detected in dermo‒epidermal junction. For IgE, none of the samples showed positivity at the dermo‒epidermal junction, but samples containing IgE-positive cells in the dermis were depicted as positive. ‒ denotes negative, and + denotes positive.
IL-23/IL-17 axis activates IL-1β-associated inflammasome in macrophages and generates an auto-inflammatory response in a subgroup of patients with bullous pemphigoid.
Human IgG1 monoclonal antibody against human Collagen 17 noncollagenous 16A domain induces blisters via complement activation in experimental bullous pemphigoid model.
01777-8&id=gr1.jpg){kind=link}
01777-8&id=gr2.jpg){kind=link}
01777-8&id=gr3.jpg){kind=link}
01777-8&id=gr4.jpg){kind=link}
01777-8&id=gr5.jpg){kind=link}
01777-8&id=fx1.jpg){kind=link}
01777-8&id=fx2.jpg){kind=link}
01777-8&id=fx3.jpg){kind=link}
01777-8&id=fx4.jpg){kind=link}
01777-8&id=fx5.jpg){kind=link}