Herpetiform pemphigus without anti-desmoglein 1/3 autoantibodies


Hiroko Nakashima, M.D., Department of Dermatology, Faculty of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. Email: nakashima-tky@umin.ac.jp


We report a patient with herpetiform pemphigus who was negative for autoantibodies to desmoglein (Dsg)1 or 3. He had erythemas with vesicles lining the margins on the trunk and extremities. Histopathology revealed intraepidermal blister with prominent eosinophil infiltration. Direct and indirect immunofluorescence demonstrated the presence of depositing and circulating immunoglobulin (Ig)G autoantibodies, but no IgA antibodies, to keratinocyte cell surface. Nonetheless, neither anti-Dsg1 nor Dsg3 antibodies were detected by enzyme-linked immunosorbent assay. Immunoblotting using human epidermal extracts also showed no reactivity with known intraepidermal or epidermal–dermal junctional substances. Immunoelectronmicroscopy revealed the reactivity on the portion of keratinocyte cell surface but not on the desmosomes. This case suggests that non-desmoglein antigen on keratinocyte cell surface can be targeted in some patients with this unusual variant of pemphigus.


Pemphigus is a group of autoimmune bullous diseases exhibiting intraepidermal blisters. Classical pemphigus consists of two major subtypes, pemphigus vulgaris (PV) and pemphigus foliaceus (PF). Detection of anti-keratinocyte cell surface autoantibodies is the most reliable diagnostic criterion for pemphigus. PV patients and PF patients essentially possess autoantibodies against desmogleins. The 130-kD PV antigen and the 160-kD PF antigen have been identified as desmoglein (Dsg)3 and Dsg1, respectively.1,2

Herpetiform pemphigus (HP) is a variant of pemphigus, which was established by Jablonska et al. in 1975.3 HP combines the clinical features that resemble dermatitis herpetiformis and immunological features of pemphigus. Patients with HP present with coalescent annular vesiculopapular lesions, which typically shows eosinophilic spongiosis without apparent acantholysis by light microscopy.4 The diagnosis of HP is based on detection of IgG anti-keratinocyte cell surface antibodies, both in vivo and in circulation. Ishii et al.5 have demonstrated that the target of IgG autoantibodies in patients with HP is also desmogleins. Anti-Dsg1 antibodies are detected in the majority, while Dsg3 is detected in fewer cases. However, the pathogenic role of non-desmoglein antigens in classical pemphigus or HP has been controversial.

In our case, the clinical features, histological features and findings of immunofluorescence analysis all supported the diagnosis of HP, whereas his serum did not have reactivity against Dsg1 or Dsg3.

Case Report

A 73-year-old male developed vesicular and erythematous skin lesions on the palms, soles and inguinal area. He visited a dermatological clinic, where he was diagnosed with palmoplantar pustulosis. Despite treatment with topical corticosteroid and vitamin D3 ointment, the skin lesions spread to the trunk, extremities and face. Three months later, he was referred to our hospital. His previous history included tuberculosis, hypertension and colon polyps. His family history was unremarkable.

On admission, he had pruritic edematous erythemas with vesicles on the trunk and extremities (Fig. 1). Vesicles and small blisters lined the margins of the erythemas. No mucosal involvement of the oral cavity was present. Laboratory tests showed normal blood cell counts and blood chemistry except for slightly increased white blood cell count (9.3 ×103/mL; normal 4.0–7.0). During the evaluation, the gastric fiberscopy revealed the presence of an ulcerated tumor with a diameter of 5 cm located on the esophageal wall close to the cardia. Mucosal biopsy from the tumor revealed squamous cell carcinoma. No lymph node or remote metastasis was found.

Figure 1.

 Clinical appearances showing vesicles and papules in the periphery of annular region erythematous on the trunk (a) and the lower extremity (b).

Histological examinations of a skin biopsy of an erythema with vesicles revealed prominent infiltration of mainly eosinophils into the epidermis and intraepidermal blister, showing eosinophilic spongiosis (Fig. 2a,b). There were also eosinophil infiltrations around vessels in the superficial dermis. Direct immunofluorescence (IF) revealed intensive keratinocyte cell surface deposits of IgG (Fig. 2c) and C3 (Fig. 2d) in the epidermis. Indirect IF using human skin sections showed IgG anti-keratinocyte cell surface antibodies at a titer of more than ×320 (Fig. 2e). Cell surface staining of epidermis by direct and indirect IF was observed in all layers with a stronger staining in the upper layers. IgA deposits or IgA antibodies were not detected by direct or indirect IF.

Figure 2.

 Histopathology of a bulla with erythema showing infiltrations of mainly eosinophils into the epidermis and around vessels in the superficial dermis, and intraepidermal blister (a). Histopathology of an erythema demonstrating intraepidermal eosinophilic infiltration and bulla filled with eosinophils, characteristic of eosinophilic spongiosis (b). (Hematoxylin–eosin stain, original magnifications: [a] ×40; [b] ×160.) Cell surface deposits of IgG (c) and C3 (d) by direct IF. Indirect IF of the substrate of human skin sections, showing anti-IgG antibodies at a titer of ×160 (e).

Anti-Dsg1 or anti-Dsg3 antibodies were completely negative by enzyme-linked immunosorbent assay (ELISA) using recombinant Dsg1 or Dsg3 throughout the course (both <5 index). Immunoblotting using normal epidermal or dermal extracts, recombinant NC16a domain of BP180 or HaCaT cell culture medium showed no reactivity with known intraepidermal or epidermal–dermal junctional antigens. Immunoelectronmicroscopy by post-embedding methods with colloidal gold particles revealed that gold particles were observed on the extracellular regions other than the desmosomes (Fig. 3).

Figure 3.

 Immunoelectronmicroscopy by post-embedding methods with colloidal gold staining. The arrow and the capital D indicate gold particles with a diameter of 5 nm and the desmosomes on the cell surface, respectively.

We diagnosed him as HP from clinical, histological and immunological findings. Differential diagnoses includes paraneoplastic pemphigus, which is proposed by Anhalt et al.6 However, this is unlikely because he had no mucosal involvement and no serum reactivity against plakin protein family, and because he had a solid carcinoma, not hematological malignancy.

Although oral minocycline was given against intensive eosinophil infiltration, the skin lesions developed over his entire body surface. Diaminodiphenylsulfone was not effective, either. With consideration of his skin condition intolerable at operation, oral prednisolone (PSL) 30 mg/day and chemoradiotherapy against his esophageal cancer were started. With p.o. PSL administration, the skin lesions initially improved. However, the eruption relapsed 1 week after tapering to 25 mg/day. PSL 40 mg/day showed gradual improvement. After chemoradiotherapy for 1 month, subtotal esophagectomy was performed. After the surgery, he exhibited a remarkable improvement of the skin conditions. Oral PSL was tapered gradually to 12.5 mg/day with no recurrence. While indirect IF analysis during the preoperative period continued to show anti-cell surface IgG antibodies at a titer of more than ×320 in spite of p.o. PSL administration, the postoperative titer rapidly decreased to ×40.


Herpetiform pemphigus is a distinct entity of pemphigus, which presents clinical features similar to dermatitis herpetiformis and is associated with IgG autoantibodies against keratinocyte cell surface in the epidermis by immunohistochemical analysis. In many cases, pruritic annular erythema is the initial lesion, and subsequently vesicles and blisters line the margins of the erythema. The mucous membrane is usually intact. Histological findings include eosinophilic spongiosis, intraepidermal blisters at variable levels and frequent absence of apparent acantholysis. Depositing and circulating IgG autoantibodies against keratinocyte cell surfaces are demonstrated by direct and indirect IF. Diagnosis of HP depends on these clinical and immunohistological features.3,7

Herpetiform pemphigus is sometimes associated with clinical features resembling PV or PF. Some patients also shift to PV or PF.7,8 Recently, Ishii et al. have demonstrated in a study of a large series of HP patients that the target autoantigen is Dsg1 and Dsg3. They reported that 16 of 20 cases in HP serum samples contained anti-Dsg1 antibodies, and four cases contained anti-Dsg3 antibodies.9 Regarding the cause of the different clinical presentation between HP and classical pemphigus in spite of the same target antigens, Ishii et al.9 suggested that autoantibodies in HP may recognize a functionally less important part of the molecule and therefore are not able to induce loss of cell adhesion. These autoantibodies may be sufficient to cause some inflammatory processes through complement activation or induction of cytokine release by keratinocytes, leading to intercellular edema and eosinophilic spongiosis.

Previous reports examining autoantigens in HP patients have demonstrated that they have IgG autoantibodies to Dsg1 and/or Dsg3.10–14 In our case, clinical and histological features were typical for HP and anti-cell surface antibodies were detected at high titer by direct and indirect IF, although anti-Dsg1/3 antibodies were not detected by ELISA or immunoblotting. Furthermore, it was demonstrated by immunoelectronmicroscopy that the location of the binding sites of the autoantibodies in the human epidermis was the extracellular regions other than the desmosomes (Fig. 3). This suggests that the autoantibodies react with non-desmosomal cell surface proteins. Since highly-sensitive ELISA detecting anti-Dsg antibodies became available,15 there has been no report about HP without anti-Dsg antibodies. While the antigen has yet to be identified, this case suggests that the target of his pathogenic autoantibody may be a non-desmoglein protein located on the keratinocyte cell surface. Taking into consideration the rapid clinical improvement and decrease of the indirect IF titer after the surgery, it is possible that the presence of esophageal carcinoma has a role in the production of pathogenic antigens. The relationship between autoimmune bullous diseases and internal malignancies is contentious. In Japan, Ogawa et al.16 reported that association ratios of internal malignancies were significantly higher than those of the controls aged over 70 years in both pemphigus and bullous pemphigoid. The pathomechanism is still unknown. The expression of foreign tumor antigens may be cross-reacted to epidermal antigens. Alternatively, this reaction may be related to the mechanism of epitope spreading.

It has been controversial whether pathogenic autoantibodies against non-Dsg epidermal antigens exist in pemphigus. Nugyen et al.17 demonstrated that acetylcholine receptor-α9 regulating keratinocyte adhesion is one of the target antigens, although its pathogenic role remains unclear. In our case, one can postulate that antibodies against non-Dsg antigen (may be cholinergic receptor or not) had poor pathogenicity to cause acantholysis and could not develop blisters even in high titers of autoantibodies. Nonetheless, as with the above-mentioned hypothesis of Ishii et al., these autoantibodies may be sufficient to cause inflammatory processes. This may lead to the characteristic clinical feature of HP rather than classical pemphigus. However, these hypotheses still need validation because there are few reports of similar cases. Further investigations are necessary to clarify the existence and pathology of intraepidermal bullous dermatosis with anti-non-Dsg antibody including our case.