Viewpoint 1

The pathophysiology of many dermatological diseases is complex. While some of these diseases are complicated by the involvement of multiple unidentified factors, others only appear to be complex because we do not understand the underlying principles and logic. Once their logic emerges, one can appreciate the beauty of nature and its elegant order. Fifty years ago, pemphigus was viewed as a complex disease. In 1964, Beutner and Jordon (1) discovered circulating immunoglobulin G (IgG) autoantibodies against the keratinocyte cell surface in the sera of pemphigus vulgaris (PV) patients, which opened up the modern history of pemphigus. In the late 1970s and early 1980s, the autoantibodies in pemphigus were shown to be pathogenic in that they induced blister formation in skin organ cultures (2,3) and following a passive transfer of IgG from pemphigus patients to neonatal mice (4). In the mid- and late 1980s, the target antigens of pemphigus were identified by immunochemical methods, such as immunoprecipitation and immunoblotting, as the 160- and 130-kDa glycoproteins for pemphigus foliaceus (PF) and PV, respectively (5–9). In the early 1990s, the isolation of cDNA using the technology of molecular biology revealed the 160- and 130-kDa antigens as desmoglein (Dsg)1 and Dsg3, respectively, which are cadherin-type cell–cell adhesion molecules that are expressed in the skin and mucous membranes (10–12). These findings, as well as the findings discussed below, have clarified the basic pathophysiology of pemphigus, and the clinical phenotype and localization of blister formation are logically explained by the anti-Dsg antibody profile and the expression pattern of Dsg isoforms in the skin and mucous membranes (Dsg compensation theory) (13–15). There are always exceptions to the general rule. However, the exceptions do not necessarily validate an alternative theory, supported by a minimal amount of solid evidence (16–19). A simple disease does not have to be explained by a complex theory. Compelling lines of evidence, collected not only from patient serology studies, but also from experiments with mouse models, genetic abrasion, Dsg1-specific protease and specific monoclonal or single-chain antibodies, support the theory that autoimmunity to Dsg is the principal cause of pemphigus pathogenesis. The following representative data corroborate this established theory (20). Anti-Dsg1 IgG autoantibodies have been found in patients with PF, but not in normal individuals (9,21–25). Anti-Dsg3 IgG autoantibodies have been found in patients with PV, but not in normal individuals (8,9,12,24–27). When monitored in individual patients, the titres of serum anti-Dsg1 and anti-Dsg3 IgG autoantibodies, as measured by indirect immunofluorescence or enzyme-linked immunosorbent assay, generally correlate with disease activity (24,28,29). Patients with mucosal dominant-type PV have only anti-Dsg3 IgG autoantibodies, while patients with mucocutaneous-type PV have anti-Dsg3 IgG and anti-Dsg1 IgG autoantibodies (30–33). Pemphigus subtype transition is accompanied by changes in the anti-Dsg antibody profiles. For example, when patients show clinical transition from PF to PV, anti-Dsg1 IgG is detected in the PF stage and anti-Dsg3 IgG is detected in the PV stage (34–38). IgG prepared from PF sera induced blisters with typical PF histology when passively transferred to neonatal mice (39). Removal of anti-Dsg1 IgG autoantibodies by immunoadsorption with baculovirus-expressed recombinant extracellular domain of Dsg1 (rDsg1) abolished the blister-forming activity of PF sera (40,41). Purified anti-Dsg1 IgG on rDsg1 from PF sera induced blisters with the typical histology (40). IgG prepared from PV sera induced blisters with typical histology when passively transferred to neonatal mice (4). Removal of anti-Dsg3 IgG by immunoadsorption with rDsg3 abolished the pathogenic activity of PV sera (26). Anti-Dsg3 and anti-Dsg1 IgG autoantibodies were necessary for efficient PV blister formation in the skin of neonatal mice (13). IgG prepared from the sera of patients with paraneoplastic pemphigus (PNP), which is another subtype of pemphigus that occurs with neoplasm, induces blisters when passively transferred to neonatal mice (42). Removal of anti-Dsg3 and anti-Dsg1 IgG autoantibodies by immunoadsorption abolished the blister-forming activity of the PNP sera (43). Specific inactivation of Dsg3 by genetic ablation of the DSG3 gene in mice resulted in a strikingly similar phenotype to that of mucosal dominant-type PV patients who carry anti-Dsg3 IgG autoantibodies alone (44). The exfoliative toxins (ETs), ETA, ETB and ETD, which are produced by Staphylococcus aureus, specifically cleave Dsg1 and induce blisters with identical histology to the PF blisters in humans and mice (45–47). Furthermore, patients with bullous impetigo caused by ETs and patients with PF show similar clinical manifestations (48). The anti-Dsg3 IgG mouse monoclonal antibody AK23 induces virtually the same PV phenotype as that of PV model mice or Dsg3−/− mice, with typical histology (49,50). The epitope of AK23 has been mapped to the N-terminal adhesive interface of Dsg3, which in functional terms is the most important part of the molecule. Single-chain monoclonal anti-Dsg antibodies isolated from a PV patient by phage display induced blisters that were histologically similar to those seen in pemphigus patients (51). The Dsg compensation theory is still under development. Anti-Dsg IgG antibodies do not necessarily possess equal pathogenic strength; some antibodies are more potent than others in blister-forming ability (50,51). Thus, the pathogenic potential of a given antibody constitutes an additional factor for explaining pemphigus cases that do not fall into classical forms of pemphigus (35,52–55). Cross-reactivity among Dsg isoforms also has to be considered. A subset of the anti-Dsg1 IgG antibodies cross-reacts with another isoform of Dsg, Dsg4, although the Dsg4/Dsg1-cross-reactive IgG has no significant pathogenic effect (56). Single-chain antibodies that react with both Dsg1 and Dsg3 have been isolated and shown to induce PV blisters in neonatal mice (51). We believe that the Dsg compensation theory will continue to provide a solid basis for a better understanding of the pathophysiological mechanism of blister formation in pemphigus.