Author Affiliations: Center for Blistering Diseases, Department of Dermatology, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands.
Background Pemphigus foliaceus is an autoimmune skin disease characterized by subcorneal blistering and IgG antibodies directed against desmoglein 1. In the skin, these antibodies deposit intraepidermally. On rare occasions, an additional “lupus band” of granular depositions of IgG and complement is seen along the epidermal basement membrane zone. This combined pattern has been connected with a variant of pemphigus foliaceus named pemphigus erythematosus.
Observations We describe 3 pemphigus foliaceus cases in which phototherapy was administered after a misdiagnosis of psoriasis. This treatment resulted in a flare of skin lesions. Direct immunofluorescence of skin biopsy specimens that were obtained several weeks later demonstrated intraepidermal and granular basement membrane zone depositions. The basement membrane zone depositions consisted of IgG, complement, and the ectodomain of desmoglein 1 and were located below the lamina densa.
Conclusions High doses of UV light are likely to induce the cleaving of the desmoglein 1 ectodomain. In patients with pemphigus foliaceus, the circulating anti–desmoglein 1 antibodies precipitate this cleaved-off ectodomain along the basement membrane zone, resulting in a lupus band–like appearance. In pemphigus erythematosus, a similar mechanism may be active, which might explain the lupus-band phenomenon.
Pemphigus foliaceus (PF) is an autoimmune skin disease characterized by subcorneal blistering and intraepidermal deposition of IgG antibodies that bind the desmosomal cadherin desmoglein 1 (Dsg1). Additional deposition of IgG is occasionally present along the epidermal basement membrane zone (BMZ), and this combined pattern has been correlated with pemphigus erythematosus (PE). Pemphigus erythematosus was first described in 1926 by Senear and Usher1,2 as a condition with a lupuslike butterfly rash or severe seborrheic dermatitis, which they suggested was a combination of pemphigus vulgaris and lupus erythematosus (LE). When insights into the differences between pemphigus vulgaris and PF crystallized, PE was not classified with pemphigus vulgaris but considered an early or a nongeneralized form of PF.3 When immunofluorescence (IF) became a diagnostic tool, the association with LE revived. Chorzelski et al4 described a “lupus-band” deposition in sun-exposed skin areas of patients with PE, together with antinuclear antibodies (ANAs) as in LE. Later studies, however, showed less clinicopathological concurrency with LE because ANAs often appeared to be absent, and the overall significance of this finding became disputed as it emerged that ANAs were also present in a high percentage of the healthy population.5- 9 Although occasional cases of LE can present simultaneously with pemphigus, the gross findings in patients with PE do not meet the criteria for systemic LE as published by the American College of Rheumatology.10,11 Therefore, what is called PE today should be separated from the sparse cases of actual concurrent LE and pemphigus/PF. At present, basic dermatology textbooks consider PE to be a localized form of PF.12 The significance of the lupus-band phenomenon in this entity, however, remains unexplained, whereas as much as 60% of biopsy results reported in patients with PE are described as having coarse granular depositions of IgG and complement along the BMZ in addition to the pemphigus anti–cell surface (ACS) deposition.4,5 Therefore, this lupus band likely reflects a unique immunopathological aspect of PE.5 In the present study, we investigated the composition of lupus-band depositions that were present in the skin of 3 patients with PF who received extensive UV exposure through phototherapy.
An 80-year-old woman was admitted to our hospital with a 3-year history of generalized progressive erythematosquamous skin lesions with pustules and flaccid blisters. This condition had been diagnosed elsewhere as psoriasis pustulosa complicated by a secondary infection with Staphylococcus aureus. The patient had received several therapies, including methotrexate sodium, systemic erythromycin stearate, acitretin, and cyclosporine. Owing to methotrexate-related hepatotoxic effects and the insufficient effectiveness of the other therapies, the patient switched to a twice-weekly regimen of psoralen–UV-A therapy with 40 mg of methoxsalen. During psoralen–UV-A therapy, the skin lesions worsened, and therapy was stopped after 3 weeks. Results of a physical examination revealed suberythroderma consisting of confluent and scattered red macules with scales and purulent crusts. A malar distribution was present on her face (Figure 1A). Multiple erosions and flaccid blisters were seen, and findings were positive for the Nikolsky sign. The mucous membranes were not involved. Results of the histopathological examination revealed subcorneal blisters. Direct IF microscopy of lesional and nonlesional skin specimens (1 biopsy specimen each, taken from the upper leg) showed intraepidermal ACS depositions of IgG and complement factor C3c in the lesional skin and additional coarse granular IgG and C3c depositions along the BMZ in the nonlesional skin. Indirect IF microscopy on monkey esophagus revealed ACS IgG antibodies, and retrospective enzyme-linked immunosorbent assay analysis demonstrated anti-Dsg1 antibodies. Blood tests yielded negative findings for ANA, antiextractable nuclear antigen, anti–double-stranded DNA, anti–Sjögren syndrome antigen A, antismooth muscle, and antistriated muscle antibodies.
Figure 1. Representative skin lesions in 2 patients. A, Patient 1. The typical pemphigus erythematosus facial butterfly eruption. B, Patient 2. Extensive lesions 2 months after ending UV-B therapy.
A 76-year-old woman with a 2-month history of generalized cutaneous blistering was admitted to our hospital. Six months earlier, she had developed itching plaques all over her body and scalp with the exception of her legs. This condition was diagnosed elsewhere as psoriasis vulgaris; 2 months later, the patient was treated with UV-B therapy. After 2 more months, she developed painful blisters on the trunk and face, with a burning sensation resembling that of sunburn. The UV-B therapy was stopped, but the blistering progressed. Physical examination at our hospital revealed multiple crusts on the scalp, face, and lips, without involvement of the oral mucosa. In addition, significant erosive lesions were present on the neck, arms, and legs, with crusted erythema on the trunk (Figure 1B). Findings on the legs and dorsal trunk were positive for the Nikolsky sign. The overall presentation resembled that of toxic epidermal necrolysis, staphylococcal scalded skin syndrome, or PF. A skin biopsy specimen showed ulcerative and erosive inflammation and secondary impetigo with beginning reepithelialization. Direct IF microscopy revealed smooth intraepidermal ACS IgG with weaker C3c depositions. In addition, granular IgG, IgM, and C3c depositions were present along the BMZ in nonlesional and perilesional skin (biopsy specimens were taken from the arm and leg). Indirect IF microscopy on monkey esophagus showed circulating anti-ACS antibodies and retrospective enzyme-linked immunosorbent assay anti-Dsg1 antibodies. Circulating ANAs were detected at titers of 1:20, falling below the cutoff range, and thus findings were evaluated as negative.
A 68-year-old man was referred to us with a 2-year history of red scaly skin lesions starting in the medial corner of the right eyelid and progressing to his chest and back. This condition was initially diagnosed elsewhere as psoriasis. The patient was subsequently treated with methotrexate. Five weeks before our examination, the patient had been treated with psoralen–UV-A therapy. The next day generalized itching developed, followed by blistering on the whole body, including the scalp and extremities. Results of a physical examination revealed facial malar erythema and erythematous confluent macules with central erosions, excoriations, crusts, and flaccid blisters on the scalp, trunk, and extremities. Mucous membranes were not involved. Findings were positive for the Nikolsky sign at the blister margins but were negative on nonlesional skin. Skin biopsy specimens from the upper leg and back obtained 5 weeks after the start of psoralen–UV-A therapy showed a globally intact epidermis with what looked like a remainder of a blister in the corneal layer and subepidermal neutrophilic infiltrates surrounding the blood vessels. Direct IF microscopy of perilesional skin (taken from the leg) showed intraepidermal intercellular IgG, IgA, and C3c depositions. In addition, granular IgG, IgA, and C3c deposits were present along the BMZ. Indirect IF microscopy on monkey esophagus revealed the presence of circulating ACS antibodies, and anti-Dsg1 IgG antibodies were detected by enzyme-linked immunosorbent assay. Blood tests for ANAs were negative.
Despite a malar distribution of skin lesions in 2 of the 3 patients and the relation to UV exposure in all 3, the patients did not otherwise fulfill the American College of Rheumatology criteria for LE. Instead, the immunopathological findings fit the diagnosis of PF, whereas the malar involvement and the presence of granular BMZ depositions were as described for PE.4,7
Nine skin biopsy specimens had been stored from the 3 cases. The specimens were all obtained from UV-exposed sites because the patients had received whole-body UV therapy. All 9 biopsy specimens showed intraepidermal deposition of IgG, and 6 had additional BMZ deposition of IgG and complement C3c (Figure 2A shows the biopsy findings of all 3 patients). These BMZ deposits were not found in 51 biopsy specimens of 14 other patients with PF who had not received phototherapy (results not shown).
Figure 2. Immunofluorescence analysis of skin biopsies. A, “Lupus-band” IgG deposition in the skin specimens of the 3 patients with pemphigus foliaceus (PF). B, The deposited IgG (left, green) colocalizes with desmoglein 1 (Dsg1) (middle, red) in the skin of patients with PF. C, The process does not occur in skin from a patient with systemic lupus erythematosus (SLE). All images have the same original magnification ×40 μm. Pt indicates patient. Patients are identified by number in the “Report of Cases” section.
Because Dsg1 is the autoantigen in PF, we stained our biopsy specimens with monoclonal Dsg1-P23 that is directed to the ectodomain of Dsg1. The staining overlapped with the IgG and C3c depositions. Thus, Dsg1 was present not only in the intraepidermal deposits but also in the BMZ deposits (Figure 2B).We stained 4 biopsy specimens of different patients with LE with the anti-Dsg1 monoclonal antibody to use as control specimens.Patients with LE also have BMZ depositions of IgG (lupus band) but ANA instead of anti-Dsg1 antibodies. As expected, Dsg1 was absent from the BMZ depositions in the LE biopsy specimens; thus, the presence of Dsg1 in such a band is exclusive to PF (Figure 2C).
To confirm our observation, we stained with 3 other anti-Dsg1 monoclonal antibodies. The Dsg ectodomain–specific monoclonal antibody Dsg1-P124 colocalized with the BMZ IgG, whereas the Dsg endodomain antibodies (27B2 and DG3.10) did not, demonstrating the absence of the Dsg1 endomain in the lower BMZ deposits (Figure 3A and B). The somewhat higher clusters contain the Dsg1 endomain and ectodomain (Figure 3B, yellow clusters). These clusters are part of the normal intraepidermal granular IgG depositions in PF skin that consist of IgG, full-length Dsg1, and cytoplasmic protein plakoglobin15 and may look like BMZ deposits but are not because they are located in the basal cells (yellow dots in Figure 3B, right panel). To investigate whether plakoglobin was also present in the BMZ deposits, we double stained for plakoglobin and the Dsg1 ectodomain. Plakoglobin appeared to be present only in the epidermal deposits (Figure 3C, left) but not in the BMZ deposits that contained the Dsg1 ectodomain. Furthermore, the BMZ deposits did not contain other desmosomal cadherins, Dsg3, or desmocollin 1 or 3 (not shown).
Figure 3. Immunofluorescence analysis of the basement membrane zone (BMZ) deposits. A, The desmoglein 1 endodomain (Dsg1 en; green) is absent in the BMZ deposition. The white bar indicates 40 μm. B, Detail of part A showing that the lower BMZ deposits contain the Dsg1 ectodomain (Dsg1 ec; red) only, whereas the higher deposits in the basal cells contain the endodomain and ectodomain (yellow). The white bar indicates 10 μm. C, In addition, plakoglobin (PG; green) is absent in the BMZ deposition. The white bar indicates 40 μm. D, The Dsg1 ec (green) is found to partly overlap with type VII collagen (ColVII). The white bar indicates 40 μm. E, Immunoelectron microscopy showing Dsg1 ec deposition below the lamina densa. The upper half of the deposition is at the level of ColVII.13 The cryosection was stained with monoclonal Dsg1-P23 and the diaminobenzidine reaction product using the gold-substituted silver-intensified peroxidase reaction.14 The black bar indicates 2 μm; the blue dotted line, the position of the lamina densa.
The absence of the cytoplasmic protein plakoglobin and the endodomain of Dsg1 in the BMZ deposits suggested that the deposits were located outside the cells and were not connected with the cell membrane of the basal keratinocytes. To more precisely determine the location of the deposits, we performed double stainings with monoclonal antibodies to the Dsg1 ectodomain and the adhesion molecules type XVII collagen, laminin 332, and type VII collagen that map to different levels of the BMZ. Most of the deposits partly colocalized with type VII collagen (Figure 3D). Additional immunoelectron microscopy revealed that deposits were located at the level of type VII collagen and somewhat lower (Figure 3E).13,14
Previous case reports have demonstrated that UV irradiation can induce skin lesions in PE and PF.16- 18 Although BMZ deposition of complement is often seen in PF, BMZ deposition of IgG is uncommon. The 3 cases described herein suggest that UV light therapy can induce a lupus band in patients with PF.
In PE, the BMZ IgG deposition has previously been shown to be related to UV exposure. Giannetti7 demonstrated for patients with PE that BMZ depositions were specific to sun-exposed areas. He investigated biopsy specimens taken from lesional skin of the face and back of 5 patients. Where BMZ IgG deposition was present in 4 of 5 facial biopsy specimens, no depositions were seen in any biopsy specimens of the back.
The granular BMZ depositions in the skin of our patients treated with UV irradiation consisted of the Dsg1 ectodomain, IgG, and complement. The mechanism involved in the shedding of this Dsg1 ectodomain might include apoptosis. One study19 reports that Dsg1 is a target of the effector caspase 3 in UV-induced apoptosis of keratinocytes. This apoptotic proteolysis also involves additional metalloproteinase-dependent shedding of the 75-kDa Dsg1 ectodomain fragment. This fragment contains the extracellular 1 and 2 domains that harbor the epitopes recognized by the pathogenic autoantibodies in PF.20 Therefore, when these fragments diffuse into the dermal compartment, they can form immune complexes with the circulating anti-Dsg1 antibodies and deposit along the BMZ.
Although a limiting flaw of our study is that we could not examine biopsy specimens of patients with PE, the data obtained from our UV irradiation–induced BMZ depositions in the 3 patients with PF may well provide the first clue for unraveling the mechanism behind BMZ deposition in PE since its first demonstration by Chorzelski et al4 in 1968. We hypothesize that a UV-driven mechanism is also active in PE that releases Dsg1 fragments from the cell membrane, thus forming deposits along the BMZ. Because this does not happen in ordinary PF, patients with PE somehow are predisposed to develop such pathophysiologic features. In this way, PE parallels discoid and subacute cutaneous LE, in which the lupus band is also present in sun-exposed lesional skin but not in sun-protected nonlesional skin.21 Pemphigus erythematosus and both LE forms are autoimmune diseases but with different autoantibodies, that is, anti-Dsg1 and ANA, respectively. In discoid LE and subacute cutaneous LE, characterization of the antigen involved in the lupus band so far has been unsuccessful, but the favored view is that the IgG reacts with nuclear and cytoplasmic antigens that are slowly released from damaged, possibly apoptotic, keratinocytes.22 Similarly, in PE, the IgG might react with the Dsg1 fragments.
Future studies of the composition of the lupus band deposition in the skin of patients with PE should confirm whether this hypothesis is correct.
Correspondence: Hendri H. Pas, PhD, Center for Blistering Diseases, Department of Dermatology, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, the Netherlands (firstname.lastname@example.org).
Accepted for Publication: May 6, 2012.
Published Online: July 16, 2012. doi:10.1001 /archdermatol.2012.1896
Author Contributions: Drs Oktarina, Poot, Jonkman, and Pas had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Oktarina and Pas. Acquisition of data: Oktarina, Poot, Kramer, Jonkman, and Pas. Analysis and interpretation of data: Oktarina, Poot, Kramer, Diercks, Jonkman, and Pas. Drafting of the manuscript: Oktarina, Poot, and Pas. Critical revision of the manuscript for important intellectual content: Oktarina, Poot, Kramer, Diercks, Jonkman, and Pas. Obtained funding: Oktarina, Jonkman, and Pas. Administrative, technical, and material support: Oktarina, Poot, Kramer, Diercks, and Pas. Study supervision: Jonkman and Pas.
Financial Disclosure: None reported.
Oktarina DAM, Poot AM, Kramer D, Diercks GFH, Jonkman MF, Pas HH. The IgG “Lupus-Band” Deposition Pattern of Pemphigus ErythematosusAssociation With the Desmoglein 1 Ectodomain as Revealed by 3 Cases. Arch Dermatol. 2012;148(10):1173-1178. doi:10.1001/archdermatol.2012.1896