What’s new in bullous pemphigoid


Hideyuki Ujiie, M.D., Department of Dermatology, Hokkaido University Graduate School of Medicine, N.15 W.7, Kita-ku, Sapporo 060-8638, Japan. Email: h-ujiie@med.hokudai.ac.jp


Bullous pemphigoid (BP) is the most common autoimmune blistering disease. BP patients have autoantibodies against type XVII collagen (COL17, also called BP180 or BPAG2), a type II transmembrane protein that spans the lamina lucida and projects into the lamina densa of the epidermal basement membrane. The non-collagenous 16A domain of COL17 is considered to contain pathogenic epitopes of BP. The transfer of immunoglobulin (Ig)G from BP patients fails to cause blisters on mouse skin probably due to differences between humans and mice in the amino acid sequence of NC16A pathogenic epitope of COL17. Passive transfer of rabbit IgG antibodies against the murine homolog of human COL17 NC16A triggers immune reactions to COL17 in mice, including complement activation, mast cell degranulation and neutrophilic infiltration, resulting in dermal–epidermal separation. Recent studies using COL17-humanized mice that express human COL17 but lack murine COL17 were the first to demonstrate the pathogenicity of anti-COL17 human BP IgG autoantibodies in vivo. These new findings provide a greater understanding of BP pathomechanisms and facilitate the development of novel specific and efficient therapeutic strategies for BP.


Bullous pemphigoid (BP) is an autoimmune disorder induced by autoantibodies against the components of the skin basement membrane zone. The most common autoimmune blistering disease,1,2 it is most prevalent in the elderly but can also appear in younger people. Clinically, tense blisters, erosions and crusts with itchy urticarial plaques and erythema develop on the entire body (Fig. 1a). Mucosal surfaces can be also affected (Fig. 1b). Histological examination of biopsied lesional skin samples reveals subepidermal blisters with inflammatory infiltration of eosinophils, and lymphocytes in the dermis (Fig. 1c). In addition to these clinical and histopathological findings, immunofluorescence (IF) analysis is required for the diagnosis of BP. Direct IF of the lesional skin shows linear in vivo deposition of immunoglobulin (Ig)G and complement C3 at the dermal–epidermal junction (DEJ) (Fig. 1d). Indirect IF using the patient’s sera reveals linear deposition of IgG and C3 at the DEJ of normal human skin. When 1 mol/L sodium chloride split skin is used as the substrate, circulating autoantibodies from the patient’s serum deposit on the roof side of the artificial blister at the DEJ (Fig. 2). This distinguishes BP from other autoimmune bullous diseases, such as epidermolysis bullosa acquisita and bullous systemic lupus erythematosus, in which deposition of autoantibodies is at the floor of the blister. Autoantibodies in BP patients usually react to proteins of 180 kDa (Fig. 3) or 230 kDa in epidermal extractions of normal human skin. The 230-kDa protein called BP230 or BPAG1 was originally identified as the major antigen of BP.3,4 BP230 is a cytoplasmic component of hemidesmosomes that belongs to the plakin family; it promotes the linkage of keratin intermediate filaments to hemidesmosomes.5 More than 80% of BP patients show auto-reactivity against BP230 antigen by enzyme-linked immunosorbent assay (ELISA) and western blot.6,7 At present, it is unclear whether anti-BP230 autoantibodies in BP patients directly contribute to blister formation or whether they are just by-products of epitope-spreading associated with disease extension; several studies point to the pathogenicity of BP230.8,9 Meanwhile, the autoantibodies against the 180-kDa protein that is present in the DEJ of the skin are considered to trigger the inflammatory process, resulting in the disruption of dermal–epidermal connection. This review focuses on recent progress in our understanding of the pathogenesis of BP.

Figure 1.

 Typical clinical, histological and direct immunofluorescence findings of bullous pemphigoid. (a) Here, tense blisters, erosions and crusts developed in itchy edematous erythema on the chest. (b) Blisters and erosions on the buccal mucosa. (c) Histopathological observation in a skin specimen taken from a bulla. Subepidermal blister formation with dermal inflammatory cell infiltration mainly of eosinophils and lymphocytes. (d) Direct immunofluorescence analysis of lesional skin, showing linear deposition of immunoglobulin G at the dermal–epidermal junction.

Figure 2.

 Indirect immunofluorescence analysis using 1 mol/L sodium chloride split skin as a substrate, showing linear deposition of immunoglobulin G on the roof side of the separation at the dermal–epidermal junction.

Figure 3.

 Serum obtained from bullous pemphigoid (BP) patient reacts to the 180-kDa protein in an epidermal extraction of normal human skin.

Major Pathogenic Autoantigen: Type XVII collagen (COL17)

Bullous pemphigoid patients have circulating IgG autoantibodies against the hemidesmosomal antigen of 180-kDa, type XVII collagen (COL17, also called BP180 or BPAG2). COL17 is a type II transmembrane protein that spans the lamina lucida and projects into the lamina densa of the epidermal basement membrane zone (BMZ) (Fig. 4).10–14 The extracellular domain of COL17 has been proven to have at least one loop structure in the lamina densa in vivo (Fig. 4).15 This structure is considered to contribute to dermal–epidermal adhesion by intertwining with other basement membrane components.15 The extracellular region of COL17 contains 15 collagen domains separated from one another by non-collagenous domains (Fig. 5).11 The non-collagenous 16A domain (NC16A), located at the membrane-proximal region of COL17, is considered to be the major pathogenic epitope for BP (Fig. 5).16,17 Epitope mapping using various fragments of COL17 has shown that sera from most BP patients recognize NC16A.16 ELISA analysis using recombinant COL17 NC16A demonstrated that 94–96% of BP sera reacts to COL17 NC16A peptides.18,19 In addition, several major antigenic sites have been identified within the N-terminal 45 amino acid stretch of NC16A, while the remaining 28 amino acids of NC16A were shown to have no BP-associated epitopes.17 Several studies have confirmed that the serum levels of anti-COL17 NC16A autoantibodies correlate with the disease severity of BP.20,21 In addition to these clinical observations, the pathogenicity of anti-COL17 NC16A IgG antibodies from BP patients was shown by in vitro study. Sitaru et al.22 demonstrated that patients’ IgG purified against recombinant hCOL17 NC16A caused dermal–epidermal separation in cryosections of human skin when the skin was co-incubated with leukocytes from healthy volunteers. They also demonstrated that IgG depleted of reactivity to NC16A lost its blister-inducing capacity.

Figure 4.

 Schematic representation of the type XVII collagen (COL17) molecule in vivo. COL17 is a type II transmembrane protein that spans the lamina lucida and projects into the lamina densa of the epidermal basement membrane zone. The extracellular domain of COL17 has at least one loop structure in the lamina densa in vivo.

Figure 5.

 Schematic of the molecular organization of type XVII collagen (COL17). The extracellular region of COL17 involves 15 collagen domains separated from one another by non-collagenous domains. The non-collagenous 16A (NC16A) domain, located at the membrane-proximal region of COL17, is considered to be the major pathogenic epitope for bullous pemphigoid (underline).

An ELISA kit using a recombinant NC16A protein (Medical and Biological Laboratories, Nagoya, Japan) is commercially available to measure the levels of anti-COL17 NC16A autoantibodies in BP patient sera. The COL17 NC16A ELISA can be used to follow disease activities.23 High index value of COL17 NC16A ELISA at the time of cessation of therapy seems to correlate with the relapse of BP within the following year.24 Thus, the COL17 NC16A ELISA is used not only as an easy method for diagnosing BP but also as a useful tool for evaluating the disease activity of BP.

Autoantibodies against the C-terminal regions of the COL17 ectodomain are also detected in a portion of BP sera. Hofmann et al.25 demonstrated that 47% of the 116 BP sera recognized the C-terminal regions of the extracellular domain of COL17. They mentioned that the presence of autoantibodies against both N- and C-terminal portions of the COL17 ectodomain was associated with the clinical feature of mucosal lesions in BP patients. Autoantibodies bound to the C-terminal region of COL17 have been regarded as relating to the clinical features of cicatricial pemphigoid.13 At the moment, the pathogenic role of autoantibodies against the C-terminal region of COL17 remains unclear.

Experimental animal models for bp

Passive transfer of IgG autoantibodies from BP patients fails to induce a BP-like phenotype in animals,26,27 despite the fact that IgG purified from patients with pemphigus can induce acantholysis in neonatal mice.28 The failure in demonstrating the pathogenicity of autoantibodies in BP patients through a passive transfer animal model is probably due to the differences between human and animals in the amino acid sequence of the COL17 pathogenic epitope. Alternatively, the potential pathogenic role of COL17 has been demonstrated in an experimental mouse passive transfer model using rabbit IgG antibodies against the murine homolog of human COL17 NC16A (murine COL17 NC14A).29 In this study, rabbit polyclonal antibodies against murine COL17 (mCOL17) NC14A were induced by injection of recombinant mCOL17 NC14A protein and then passively transferred into neonatal BALB/c mice.29 The injected mice demonstrated skin fragility associated with in situ deposition of rabbit IgG and mouse C3 at the DEJ of their skin, and dermal–epidermal separation with an extensive inflammatory cell infiltration, which correspond to the clinical, histological and immunopathological features of BP.29 Using this experimental BP model, Liu et al.30revealed that subepidermal blister formation in their neonatal mouse model depends on complement activation, mast cell degranulation31 and neutrophil infiltration.32

To further understand BP pathomechanisms, especially the pathogenic roles of anti-human COL17 NC16A antibodies, Olasz et al.33 generated a novel transgenic (Tg) mouse expressing human COL17 (hCOL17) under the control of a keratin 14 promoter. These hCOL17 Tg mice displayed a normal phenotype and expressed hCOL17 in the mouse epidermal basement membrane with appropriate immunoreactivity and localization, as confirmed by light, IF and immunoelectronmicroscopy. They demonstrated that wild-type (Wt) mice grafted with hCOL17 Tg skin developed high titers of anti-hCOL17 IgG, and lost the Tg skin grafts associated with the deposition of IgG and C3 at the epidermal basement membrane as well as with neutrophil-rich leukocytic infiltration and subepidermal blisters like BP.33 These findings clearly show that the anti-hCOL17 IgG antibodies elicited by Tg skin grafting have pathogenicity against Tg skin that expresses hCOL17 antigens. They also demonstrate that major histocompatibility (MHC) class II−/− mice grafted with Tg skin develop neither anti-hCOL17 IgG nor graft loss, indicating that MHC II and CD4+ T-cell interactions are crucial for these responses. Furthermore, they demonstrate that blockade of CD40L, one of the most critical co-stimulatory molecules for B-cell proliferation and IgG production by B cells, prevents anti-hCOL17 IgG production and graft loss in Wt mice with hCOL17 Tg skin graft.34

These studies strongly support the hypothesis that anti-COL17 NC16A IgG autoantibodies in BP patients have a pathogenic activity. However, this concept has not yet been directly demonstrated using anti-COL17 NC16A IgG autoantibodies prepared from BP patients in vivo. In 2007, our group succeeded in confirming this hypothesis by using the unique technique of “humanization of autoantigen’’.35 First, we generated murine Col17-knockout (mCol17−/−) mice that developed blisters and erosions on the skin, symptoms that reproduce the human disease non-Herlitz epidermolysis bullosa, which is caused by null mutations in the COL17A1 gene. Then, we crossed hCOL17+/+ Tg mice with the heterozygous mCol17+/− mice. Mice that carried both the heterozygous null mutation of mCol17 and the transgene of hCOL17 (hCOL17+/−, mCol17+/−) were bred to generate COL17-humanized (hCOL17+/+, mCol17−/−) mice (Fig. 6). COL17-humanized mice lack mCol17 but express hCOL17 (Fig. 6). They showed no apparent clinical phenotype and were able to deliver COL17-humanized mice pups by mating with COL17-humanized mice parents. Neonatal COL17-humanized mice that were passively transferred with either total IgG or the IgG affinity-purified against hCOL17 NC16A from BP patients developed diffuse erythema and showed epidermal separation by gentle skin friction at 48 h after the transfer (Fig. 7). Lesional skin specimens demonstrated dermal–epidermal separation and inflammatory cell infiltration of neutrophils and lymphocytes. Direct IF analysis revealed linear deposition of human IgG at the DEJ (Fig. 8), which simulates the human BP phenotype. This passive-transfer neonatal mouse model was the first to directly show the pathogenicity of anti-COL17 NC16A IgG autoantibodies in BP patients.

Figure 6.

 The method for generating type XVII collagen (COL17)-humanized mice. By crossing heterozygous murine Col17 knockout (mCol17+/−) mice with human COL17 transgenic (hCOL17+/+, mCol17+/+) mice, hemizygous hCOL17 transgenic, heterozygous mCol17 knockout (hCOL17+/−, mCol17+/−) mice are generated. Then, they are bred to generate COL17-humanized (hCOL17+/+, mCol17−/−) mice that express hCOL17 but not mCOL17.

Figure 7.

 A passive-transfer bullous pemphigoid model using the type XVII collagen (COL17)-humanized mouse. The control COL17-humanized mouse that was passively transferred with normal human immunoglobulin G demonstrates no skin fragility (upper). The neonatal COL17-humanized mouse that was passively transferred with immunoglobulin G affinity-purified against human COL17 (hCOL17) NC16A from bullous pemphigoid patients shows epidermal detachment by gentle skin friction at 48 h after transfer (lower).

Figure 8.

 A passive-transfer bullous pemphigoid model using the type XVII collagen (COL17)-humanized mouse. (a) Lesional skin specimen demonstrates dermal–epidermal separation and infiltration of inflammatory cells, including neutrophils and lymphocytes. (b) Direct immunofluorescence analysis reveals linear deposition of human immunoglobulin G along the dermal–epidermal junction.

This COL17-humanized BP mouse model is quite valuable not only for analyzing the pathomechanisms of BP, but also for evaluating disease-specific therapies by blocking autoimmune reactions against hCOL17 molecules. To investigate this, we executed a therapeutic trial using recombinant peptides of hCOL17 NC16A domains. After the transfer of total IgG from BP patients, COL17-humanized mice were treated with a recombinant peptide composed of the full length of the NC16A domain. The mice treated in this way showed markedly reduced blister formation. This further attests to the potential of “epitope decoy” therapy in treatments for antibody-mediated autoimmune diseases such as BP. In addition, we recently developed a novel therapeutic strategy of treating BP by using the recombinant Fab fraction of IgG monoclonal antibodies to block the complement activation that would otherwise be induced by pathogenic autoantibodies.36 In BP, complement activation is considered to be critical to blister formation.30 To block autoantibody-induced complement activation in BP, we used phage display to generate recombinant Fab fragments of IgG monoclonal antibodies against hCOL17 NC16A from antibody repertoires of BP patients. Two recombinant Fab fragments, Fab-B4 and Fab-19, showed marked ability to inhibit the binding of BP autoantibodies and to inhibit subsequent complement activation in vitro. In the in vivo experiments using the COL17-humanized BP mouse model, these recombinant Fab fragments protected mice against BP autoantibody-induced blistering disease. Thus, the use of engineered Fabs to block pathogenic epitopes appears to demonstrate efficacy and may lead to disease-specific treatments for BP.

Using this COL17-humanized mouse, we developed another novel neonatal BP mouse model, which was naturally induced by maternally transferred antibodies.37 Heterozygous mCol17-deficient (mCol17+/−) female mice were crossed with COL17-humanized (hCOL17+/+, mCol17−/−) male mice. It was expected that 50% of the pups would show the heterozygous genotype of hCOL17+/−, mCol17−/− (Fig. 9). Then, mCOL17+/− female mice were immunized by hCOL17 Tg skin grafting and mated with COL17-humanized (hCOL17+/+, mCol17−/−) male mice. As we expected, the circulating anti-hCOL17 IgG antibodies that were produced by the immunized mother were transferred into their neonates through not only placenta but also milk, and half of the neonates whose skin expressed human but not murine COL17 (hCOL17+/−, mCol17−/−) developed blisters associated with the histological and immunological features that are seen in BP patients (Fig. 10). In addition, to examine the pathogenic role of anti-hCOL17 IgG antibodies produced by immunized mCOL17+/− female mice, total IgG with or without immunoadsorption using recombinant hCOL17 NC16A peptides was injected into the neonatal COL17-humanized (hCOL17+/+, mCol17−/−) mice. Purified total IgG without immunoadsorption induced skin fragility in the injected neonatal mice, whereas IgG with immunoadsorption did not cause phenotypic changes. These findings clearly show that IgG antibodies against hCOL17 NC16A play a major role in inducing blister formation. This novel experimental system has some advantages: (i) the model does not require the technically difficult injection procedure; (ii) the pathogenic IgG persists for longer than in conventional passive transfer models; and (iii) the immune reaction totally depends on the murine immune system, even though the antigen is hCOL17.

Figure 9.

 The method for generating the neonatal bullous pemphigoid model induced by maternally transferred antibodies. Heterozygous murine type XVII collagen knockout (mCol17+/−) mice are immunized against human COL17 (hCOL17) by grafting hCOL17-expressing transgenic mouse skin. The immunized female mice are crossed with COL17-humanized (hCOL17+/+, mCol17−/−) male mice. Fifty percent of the neonates express only hCOL17 in the skin (hCOL17+/−, mCol17−/−), and those are expected to develop the phenotype of bullous pemphigoid.

Figure 10.

 A neonatal bullous pemphigoid model induced by maternally transferred antibodies. (a) The neonatal mice develop spontaneous small blisters on the trunk (arrowheads). (b) Histological findings of blistering lesions on the tail. Subepidermal separation associated with neutrophilic infiltration is seen. (c,d) Direct immunofluorescence analysis demonstrates in vivo linear deposition of mouse immunoglobulin G (c) and mouse C3 (d) at the dermal–epidermal junction 1 week after birth.

Liu et al.38 recently developed a new BP mouse model using a humanized mouse in which the murine genomic region encoding the mCol17 NC14A domain was replaced with the homologous hCOL17 NC16A sequence. They demonstrated that COL17 NC16A-humanized mice injected with anti-COL17 NC16A IgG autoantibodies from BP patients develop BP-like subepidermal blisters, and the immune responses requires complement activation, mast cell activation and neutrophil recruitment. These results further support the idea that the anti-hCOL17 NC16A IgG antibodies present in BP patients are the major pathogenic autoantibodies.

Ige autoantibodies in bp

Immunoglobulin G is not the only response in BP. It is known that there is also an IgE response.39 The early urticarial phase of the eruptions seen in BP seems to be associated with IgE, based on the common knowledge of IgE-mediated degranulation of mast cells in allergic forms of urticaria.40 In clinical terms, total IgE levels are elevated in 70% of untreated BP patients and IgE autoantibodies against COL17 are detected in 86% of untreated BP patients.41 Iwata et al.42 reported that the presence of IgE autoantibodies against COL17 was associated with a severe form of BP, and BP patients with IgE anti-COL17 antibodies required a longer period of treatment for remission, a higher dose of prednisolone, and more intensive treatments for remission. These findings suggest that IgE autoantibodies against COL17 are associated with BP pathogenesis and disease activity. Dresow et al.43 recently reported that IgE autoantibodies in BP patients recognize not only COL17 NC16A, but also additional regions of the intracellular domain of COL17, indicating the possibility of intramolecular epitope spreading during disease progression.

Interestingly, the passive transfer models for BP using only IgG antibodies do not totally replicate human BP because eosinophil infiltration, a characteristic finding of human BP, is not detected in those models.29,35 Zone et al.44 has successfully induced the itchy erythematous lesions in engrafted human skin on severe combined immunodeficiency (SCID) mice using IgE antibodies against LABD97, which is a component of the shed ectodomain of hCOL17. They developed an IgE hybridoma to the LABD97 antigen, and this hybridoma was injected s.c. in SCID mice with engrafted human skin. IgE antibodies against LABD97 were produced by the injected hybridoma in vivo, and they bound to the BMZ of the engrafted human skin in a linear pattern. Human skin grafts developed erythema at days 8–11 after the s.c. injection of IgE hybridoma. At day 21, all the injected mice developed severe eosinophilic infiltration and mast cell degranulation within the grafts and most of them developed histological, but not clinically detectable, basement membrane blisters. This new BP model induced by IgE antibodies reproduces the clinical and histological findings of human BP, including the initial itchy urticarial lesion, intense infiltration of eosinophils, mast cell degranulation and dermal–epidermal separation. These findings indicate that IgE autoantibodies may play an important role in the pathogenesis of BP.

Fairley et al.45 developed another experimental system for investigating the pathogenesis of IgE autoantibodies in BP patients. They isolated total IgE from the sera of BP patients and injected it into human skin grafted on athymic nude mice. Elevated erythematous plaques similar to the early lesions seen in human BP developed in all the human skin grafts at 24 h after injection of the BP IgE autoantibodies. Histological examination of the lesions revealed the engorgement of blood vessels and a dermal inflammatory infiltrate composed of neutrophils, eosinophils and degranulated mast cells. Higher doses of BP IgE autoantibodies induced histological dermal–epidermal separation in the grafts. This study has also provided direct evidence of a pathogenic role for IgE autoantibodies in BP. More recently, Fairley et al.46 reported a case of steroid-unresponsive BP that was successfully treated with omalizumab, a humanized monoclonal antibody that inhibits IgE binding to the high-affinity receptor FcεRI. Although a randomized control study will be required to confirm the efficacy of anti-IgE therapies, these results suggest that IgE autoantibodies could be a new therapeutic target in BP.


Recent studies using animal models have made considerable progress toward our understanding of the pathogenesis of BP. Passive transfer studies using anti-COL17 IgG or IgE antibodies have revealed not only the pathogenicity of these antibodies, but also the subsequent immune responses, such as complement activation, mast cell degranulation and infiltration of inflammatory cells, including of neutrophils and/or eosinophils. These studies can provide basic knowledge for the development of more specific and efficient therapies for BP. To further advance our knowledge of BP pathomechanisms and to develop novel therapeutic strategies for BP, the establishment of an active, stable disease model is anticipated.