Bullous pemphigoid (BP), by far the most common autoimmune blistering disease, is induced by autoantibodies against the structural components of the dermal–epidermal junction (DEJ) (1). In most cases, the disease develops cryptically (2). The suggested causes of BP include silicosis (3), psoralen and ultraviolet A therapy (4), infections (5), physical or chemical insults (6–8), certain fruits (9), and medications (10, 11). However, the validation of these factors in the pathogenesis of BP remains be established. BP mainly affects the older age group of both sexes, or those 70 years old and above, but it can also affect infants, children, and adolescents (1, 12). This disease mainly involves the skin but occasionally the eyes, mouth, and genitals (1, 2). The cutaneous manifestations of BP are polymorphic and can be classified into three groups, namely classical BP, non-bullous cutaneous pemphigoid, and various rare variants (13, 14). Classical BP is clinically characterized by large (1–3 cm), tense, serous, or hemorrhagic blisters that appear on erythematous, urticarial, or eczematous lesions and even on apparently normal skin (1, 13). The biopsied lesions exhibit subepidermal splitting or blisters, which is the hallmark of BP, with dense inflammatory infiltration of eosinophils, basophils, neutrophils, lymphocytes, and mast cells in the dermis (1). Immunofluorescence analysis is necessary for the diagnosis of BP (15). Direct immunofluorescence is the most sensitive method for BP diagnosis, in which the lesion shows linear deposition of immunoglobulin G (IgG), C3 complement, and even IgE at the DEJ (16–18). Indirect immunofluorescence using the patient’s sera and a substrate, especially salt-split skin, reveals a linear deposition of IgG along the roof of the artificial split (18).

One typical serologic characteristic of BP is the presence of circulating autoantibodies, which are mostly against BP180 (collagen XVII) and BP230 (15, 19, 20). BP180 is a 180 kDa transmembrane glycoprotein with a 16th non-collagenous (NC16A) domain, which is the immunodominant part in BP (14). BP230 is an intracellular constituent of the hemidesmosomal plaque and belongs to the spectraplakin family (20, 21). The autoantibodies reported in BP include IgG and IgE (1, 22). Usually, IgG autoantibodies to BP180 are the ones first to be detected, and then IgG autoantibodies to BP230 subsequently appear (23). IgE antibodies to BP230 can also be detected in the blood of BP patients (24). Given the existence of autoantibodies, there have been commercially available enzyme-linked immunosorbent assay (ELISA) kits that target BP180 and BP230 antibodies for BP auxiliary diagnosis (25, 26).

Due to the age group involved and the application of more sensitive and specific diagnostic assay systems, the reported BP morbidity has increased (14, 19, 27, 28). Moreover, for disease-specific factors, due to the concomitant occurrence of neurodegenerative disorders, use of higher doses of oral corticosteroids, and the propensity to malignancies and venous thromboembolism, BP mortality showed an increasing trend as well (19, 29–36). These findings suggested the contributory role of activation of blood coagulation in the pathogenesis of BP (35, 36). Presently, topical or systematic corticosteroids, with or without immunosuppressive agents, are still the mainstays for BP treatment (1, 14, 37, 38). Intravenous Ig has also been introduced as an alternative therapy for BP (39–41), however, its effectiveness is still questionable (42, 43). Therefore, it is of highly importance to discover new targets to reduce BP morbidity and mortality. Recently, increasing evidences show that autoimmune responses to BP180 are important in the initiation and evolution of BP (44). The binding of autoantibodies to BP180 is a central step for blister formation. Moreover, BP180 is associated with severe and extensive lesions that require higher dose of steroids, which is a key risk factor for death (14, 28, 45). The serum level of anti-BP180 NC16A autoantibody correlates with the more active and severe disease, as well as poorer prognosis (33, 46). We, thus, consider BP180 as the most important culprit in the pathogenesis of BP and focused this review on recently updated knowledge on BP180 and its autoantibodies in BP.

BP180 is a type II transmembrane protein with a cytosolic NH2 terminal and an extracellular COOH domain (47). The N-terminal domain, transmembranous stretch, and extracellular C-terminus have 466, 23, and 1,008 amino acids (aa) in length, respectively (48). The ectodomain contains 15 collagenous subdomains (COL1–COL15) interspersed by 16 non-collagenous sequences (NC1–NC16). The NC16A domain, a juxtamembranous linker region, appears to be biologically important, as it serves as the nucleus for the formation of a collagen-like triple helix (49, 50). The extracellular domain contains coiled-coil structures, which are physiologically shed from the cell surface by a disintegrin metalloproteinase (ADAM) (50). The ectodomain forms a loop structure as it spans the lamina lucida, extends to lamina densa, and then kinks back into the lamina lucida (49). BP180 contains multiple binding sites for hemidesmosome proteins, including the extracellular domains of integrin α6 and laminin-332 (laminin-5) and the cytoplasmic domains of integrin β4, plectin, and BP230 (20). The structure and location of BP180 indicate that it acts as a core anchor protein that connects the intracellular and extracellular hemidesmosomal proteins and plays a key role in the pathogenesis of BP.

Previous studies mainly focused on extracellular NC16A domain (aa residues 490–562), which is the main target of BP autoantibodies. The NC16A domain has seven antigenic sites, including NC16A1, NC16A1-3, NC16A1-5, NC16A2, NC16A2.5, NC16A3, and NC16A3-4 (51–53) (Figure 1). Among these sites, NC16A2 and NC16A2.5 are the major antigenic sites, which can be targeted by all IgG and IgE antibodies. However, recent studies have described additional autoantibody-binding domains of BP180, such as the intracellular domain (ICD) and ectodomain (44, 54). The ICD (aa 1–452) has five target sites, namely ICD A, ICD B, ICD C, ICD D, and ICD A-D, and a central region (aa 112–199) (Figure 1). A previously published study reported that out of 18 sera of BP patients, 16 reacted with recombinant ICDs and that most of the antibodies bind to the central portion (55). A great number of sera combined with at least one of the ICD regions. With regard to ectodomain, it has been reported that 7.8–47% of BP sera recognized the C-terminal regions of the ectodomain (54, 56). Further mapping identified the six regions outside of NC16A that were recognized by the sera of the patients: aa 809–1106, aa 1080–1107, aa 1280–1315, aa 1331–1404, aa 1365–1413, and aa 1048–1465 (11, 52, 54, 57). aa 809–1106 and aa 1080–1107 were at the midportion, whereas aa 1331–1404 and aa 1365–1413 were at the COOH-terminal (Figure 1). Other epitopes embracing more than one domain, such as aa 467–567, aa 490–812, and aa 490–1497, were also reported (11, 52). It has been suggested that the pattern of epitope recognition may influence the course of the disease (23). Therefore, the recognition of target regions within BP180 is substantial in understanding the disease initiation and clinical characteristics of BP.

The etiology of BP is complex, but the presence of autoantibodies was widely accepted as the sine qua non of the condition. Anti-BP180 autoantibodies also exist in healthy people, even though these antibodies are conformationally different from pathogenic ones; however, only those bound to skin basement membrane can induce BP—suggesting that autoantibodies in the healthy may not be pathological per se (58, 59). The autoantibodies may assume function of surveillance and self-tolerance (60). In pathologic conditions, self-tolerance of the autoantibodies is dysfunctional, thus leading to the production of a higher-level of autoantibodies that bind to skin basement membrane and give rise to the occurrence of BP. The development of BP suggests that there is a threshold or checkpoint in terms of autoantibody generation (61). It remains unclear why immune tolerance to BP180 is dysfunctional in some individuals. Previous study suggests that CD4+ CD25+ Foxp3+ regulatory T (Treg) cells play an indispensable role in maintaining self-tolerance and in suppressing excessive production of autoantibodies deleterious to the host (62–65).

The reduction of CD4+ CD25+ Foxp3+ Treg cells in BP, as induced by triggers that are variants of pre-existing genetic factors, such as HLA-BQB1*0301, CYP2D6, MT-ATP8, and so on, leads to the breakage of self-tolerance, followed by the increase in autoreactive Th2, Th1, and B cells that can recognize different domains of BP180 mediated by epitope spreading to produce different autoantibodies (14, 44, 59, 65–69). The pathogens can exacerbate the process by sensitizing B cells via binding to toll-like receptors. The autoreactive T cells can interact with autoreactive B cells via combinations of CD40L–CD40, B-cell activating factor–transmembrane activator and CAML interactor (TACI)/B-cell maturation antigen, and proliferation-inducing ligand–TACI to further break peripheral tolerance and induce Ig production and class switching (70–74) (Figure 2). Moreover, the reactivity of T and B cells that target the NH2-terminal portion of the BP180 ectodomain is associated with severe BP, whereas the crosstalk of T and B cells targeting the central portion of BP180 is more frequently recognized in limited BP (75). The exploration in gene therapy might provide clues to retrieve Treg-mediated tolerance and to hinder the production of autoantibodies in skin-grafted animals (76).

Previously, most studies pointed out that the NC16A might be the major pathogenic epitope in BP (47, 74). ELISA analysis using recombinant BP180 NC16A demonstrated that 22–100% of BP sera reacted to BP180 NC16A peptides and that autoantibodies targeting NC16A domain are associated with tense blisters, severe urticarial erythema, extensive lesions, and elevated eosinophils (45, 77). Therefore, there is a variety of autoantibody types that act on this domain and mediate various pathogenesis.

Recent studies reported that 59–82% of BP sera can recognize the ICD of BP180, while 7.8–49% of BP sera are reactive against the ectodomain of BP180 (54, 77, 122, 123). All autoantibodies, including IgG, IgE, and IgA, can target ICD; however, these autoantibodies bind to different sites (55, 114, 122, 123). The autoantibodies can penetrate live cells, reach their intracellular targets, and alter cellular functions (124) (Figure 4). The central region of BP180 ICD harbors binding sites that are critical for the interaction of BP180 with β4 subunit of the α6β4 integrin, which is vital for the incorporation of the protein into the hemidesmosome (49). Thus, it implicates that autoantibodies against BP180 ICD impair the interaction of BP180 with other molecular constituents of the hemidesmosome. Otherwise, the damaged basal keratinocyte induced by the binding of autoantibodies to BP180 ectodomain leads to the exposure of the ICD to the immune system, which is referred to as “epitope spreading” (125) (Figure 4).

In addition, the COOH-terminal region of the BP180 ectodomain is shown to be recognized by 47% of BP sera (56). IgG, IgE, and IgA autoantibodies can all bind to the terminal region (52, 54, 103, 122). The presence of autoantibodies against N- or C-terminal portions of the BP180 ectodomain is associated with the mucosal lesions in BP patients (56, 126). In addition, there are existing autoantibodies against the midportion of BP180; and these are associated with the occurrence of hemiplegia, clinical presentation of lack of erythema around the bullae, and histopathologic eosinophil infiltration inside and around subepidermal bullae (57). Other studies revealed that high levels of autoantibodies against C-terminal portions are associated with older age, administration of dipeptidyl peptidase-4 inhibitors before BP onset, and a positive response to moderate doses of oral prednisolone (11, 123). However, there is also a report refuting the association of autoantibodies with dipeptidyl peptidase-4 inhibitors (127). As BP180 extends from the cytoplasm of the basal keratinocyte to the lamina densa, it is presumed that the autoantibodies against this region might be responsible for the scarring phenotype observed in cicatricial pemphigoid patients (56) (Figure 4). The development of novel ELISA kits to detect the autoantibodies against the ectodomain, or even ICD, is beneficial in diagnosing BP without NC16A domain (56, 128).

More novel animal models have been recently constructed, thus making it possible to determine the role of different domains. One of the animal models is the ΔNC14A mice, which have BP180 NC14A replaced with the homologous human BP180 NC16A epitope cluster region (129). BP lesion develops in these ΔNC14A mice after passive transfer of BP IgG (129). The NC14A region can also be genetically deleted in C57BL/6 mice, which then have less amount of BP180 in skin but have normal ectodomain shedding (130). They spontaneously produce IgG and IgA autoantibodies against BP180 and present eosinophilic infiltrations, as well as the clinical features of pruritus and crusted erosions (130). Hence, the ΔNC14A mice may be an ideal experimental model for investigating the early clinical changes in BP. However, in the absence of NC16A domain, it is impossible to explore the detailed functions of anti-NC16A autoantibodies. It is also presumed that the pruritus and eosinophils are associated with the ectodomain. Therefore, the ΔNC14A mice may be utilized as a model for the exploration of autoantibodies acting on the ICD or on the ectodomain. However, the mechanisms involved remain to be confirmed. Another animal model is the COL17-humanized mice, which can express human BP180, and it is suitable for the analysis of the pathogenesis of BP in humans (131). The spontaneous production of high titers of anti-BP180 antibodies in blisters and erosions on erythematous skin lesions makes the observation of dynamic immune reactions possible. The pathogenicity of autoantibodies against ICD and ectodomain of BP180 remains unclear, and further studies are warranted. The development of novel ELISA system to detect such autoantibodies is necessary (77).

An IgM-mediated BP has been recently reported (132, 133). Direct immunofluorescence microscopy showed that linear deposition of IgM can be found at the DEJ of 6–22% of BP patients (17, 134, 135). However, the target of IgM autoantibodies is unknown, and immunoblotting with recombinant protein of BP180 C-terminal domain showed multiple non-specific bands (136). IgM is mainly associated with BP caused by lupus erythematosus (132); however, it is rarely associated with BP due to infections (137), macroglobulinemia (136, 138), and surgical factors (139). The presence of IgM autoantibodies seems to not influence the course or outcome of the disease; and the role of IgM autoantibodies in the pathophysiology of BP remains elusive.

Followed by various autoantibody-mediated inflammatory responses, the BP180 cleavage and depletion have been proposed as the terminal effect that causes reduced adhesion and blister formation. In vitro, the cleavage and shedding of BP180 ectodomain is an event related to detachment, migration, proliferation, differentiation, and wound healing of keratinocytes (50, 140–144). Generally, the cleaved ectodomain does not generate pathogenic epitopes. However, excessive cleavage, shedding, or depletion can lead to reduced adhesion and blister formation.

Bullous pemphigoid autoantibody-induced infiltration of mast cells, eosinophils, and neutrophils can lead to the production of various inflammatory factors and proteases that contribute to the induction of blister formation. Increased levels of IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-13, IL-17, IL-22, IL-23, IL-31, IL-36, interferon-γ, tumor necrosis factor (TNF)-α, transforming growth factor-β, RANTES (regulated on activation, normal T cell expressed and secreted), monocyte chemotactic protein 1, interferon gamma-induced protein 10, and C-C chemokine ligand (CCL) 17 have been detected in skin lesions, serum, or blister fluid of BP patients (14, 19, 97, 145–150). In addition, C-C chemokine receptor 3 ligands, such as CCL 11, CCL13, CCL18, CCL26, and CCL28, have been shown to be increased in skin and/or sera of BP patients (43, 146, 151, 152). Increased levels of CCL1, CCL2, and chemokine C-X-C motif ligand-10 were detected in sera of BP patients (153, 154). Moreover, increasing data revealed their functional involvements in BP (97, 149, 151, 153, 155–158) (Figure 5A). The proteases produced by inflammatory cells are functionally involved as well (79, 159) (Figure 5B). The inflammatory cells can release mast cell protease (MCP)-4, matrix metalloproteinase (MMP)-9, neutrophil elastase (NE), plasmin, and eosinophil cationic protein (ECP), which cleave and degrade BP180, thus leading to dermal–epidermal separation and blister formation (20, 149, 157, 160–164) Pathogenic anti-BP180 IgG failed to induce subepidermal blistering in mice that were deficient in either NE or MMP-9 (89). MMP-9 can regulate NE activity by inactivating α1-proteinase inhibitor (α1-PI) (159). Furthermore, α1-PI serves as a chemoattractant for neutrophils once it is cleaved and exacerbates tissue damage (165). MMP-9 can also cleave BP180 into small tripeptides Pro-Gly-Pro, which significantly enhance neutrophil chemotaxis and NE release (149). These infiltrated cells also release IL-17, which significantly upregulates the production of MMP-9 and elastase in neutrophils (149, 166). The released IL-17 could, in turn, stimulate neutrophils to produce more IL-17 and form an amplified loop (167) (Figure 5A). Therefore, inflammatory factors and proteases induced by inflammatory cells play key roles in the cleavage and depletion of BP180, and targeting these inflammatory networks may be a promising therapeutic strategy in the treatment of BP.

However, BP180 cleavage may also occur in the absence of anti-BP180 autoantibodies (140). Such physiological cleavage is mediated by ADAMs (140). Our study further reveals that TNF-like weak inducer of apoptosis (TWEAK), which is a multifaceted cytokine that participates in various skin inflammatory responses, can exacerbate the BP180 reduction and keratinocyte adhesion (19). Moreover, the effect of TWEAK on BP180 cleavage involves the activation of extracellular signal-regulated kinase and nuclear factor-κB pathways as well as the downstream ADAMs, in which ADAM 8, 9, 10, 15, and 17 have been suggested to participate in BP180 cleavage or BP development (19, 168, 169). We also found high expression of MMP-9, ADAM9, ADAM10, and ADAM17 in BP lesions and in keratinocytes upon TWEAK/Fn14 activation (19). The upregulation of MMP-9 and ADAM10 is responsible for the shedding of membrane CD46, which further enhances BP180 NC16A IgG-mediated complement activation and blister formation (82). Therefore, the role of TWEAK in BP development can be mainly ascribed to the abnormally high expression of ADAMs and other proteases. By considering the absence or insignificant expression of TWEAK in noninvolved skin, we conclude that TWEAK likely plays a secondary inflammatory role rather than being a primary participant (19, 170). Further investigations are required to establish the clear-cut function of TWEAK in BP.

Considerable progress made by recent studies updated our understanding of BP pathogenesis. The availability of novel BP animal models provides important tools to further gain insights on the pathophysiology of the autoimmune disease. However, there is a limited progress regarding BP therapy. As BP180 is a molecule with multiple epitopes, a better insight on the mechanisms of immune responses induced by binding of autoantibodies to BP180 on different epitopes is crucial for the design of novel and more specific therapeutic strategies for this life-threatening autoimmune disorder (Table 1).

Bullous pemphigoid has been regarded as a well-characterized, organ-specific, mainly anti-BP180 autoantibody-mediated blistering skin disorder. Both IgG and IgE play vital roles in BP development via complement-dependent or -independent inflammatory pathways. However, the roles of IgA and IgM are still uncertain, and further investigation is needed. Knowledge of the BP180 target sites and of the interaction between BP180 and anti-BP180 autoantibodies is pivotal for the exploration of novel and more specific therapeutic methods so as to reduce BP morbidity and mortality. The translation of bench findings into bedside strategies for the treatment of this complex disease still remains to be a challenge. Although BP180-based therapy appears not to be close at hand yet, a better understanding of the role of BP180 would further approximate that to practice.

YL and YX conceived this paper. YL and LL wrote this manuscript. All the authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.