Received 7/3/97 Accepted 7/10/97

Regardless which BMZ antigen is targeted by autoantibodies, the histopathological findings of all autoimmune subepidermal blistering diseases are similar (42). Although the inflammtory cell infiltrate may provide clues in the diagnosis of certain diseases, such as eosinophilic infiltrate in the diagnosis of BP, they are not always reliable and cannot be used as definitive diagnostic criteria. Therefore, more specific techniques are required for accurate diagnoses.

As depicted in Figure 1, the skin BMZ can be chemically separated at the middle lamina lucida by either 1 M NaCl or 20 mM Na-EDTA (43). These techniques can be applied to direct immunofluorescence and indirect immunofluorescence in the detection of the relative binding sites of in vivo-bound and circulating autoantibodies, respectively. Autoantibodies bound to the epidermal roof of the chemically separated skin section indicate the target antigens are located in the hemidesmosome/upper lamina lucida area, whereas autoantibodies bound to the dermal floor indicate the target antigens are located in the lower lamina lucida, lamina densa, or sub-lamina densa areas. Figure 2 demonstrates the circulating IgG autoantibodies from a patient with anti-laminin cicatricial pemphigoid (a), a patient with bullous pemphigoid (b), and a patient with epidermolysis bullosa acquisita (c) bound to the dermal floor, the epidermal roof, and the dermal floor of chemically separated skin sections, respectively. (E=epidermis; D=dermis; arrowheads indicate the binding sites of the autoantibodies)

Figure 2. Indirect immunofluorescence microscopy on chemically separated skin sections showing the autoantibodies from patient with anti-laminin cicatricial pemphigoid (left), patient with BP (middle), and patient with epidermolysis bullosa acquisita (right) labeling the antigenic sites (arrowheads) at dermal floor, epidermal roof, and dermal floor, respectively.

Since immunofluorescence microscopy can only detect the relative location of the autoantibody binding sites (target antigens), more accurate diagnostic technique is required to delineate the fine location of the target antigens. Immunoelectron microscopy combines ultrastructural definition and immunolocalization of the target antigen, and therefore can provide a more detailed localization of the target antigens (44). The target antigens can be visualized by immunoperoxidase or immunogold techniques. A direct method can be used to detect the binding site of in vivo-bound autoantibodies, whereas an indirect method is used to detect the binding site of circulating autoantibodies. Figure 3 illustrates the immunoperoxidase particles (arrows) localizing the target antigen of the IgG autoantibodies from a patient with epidermolysis bullosa acquisita to the lamina densa and sub-lamina densa areas. (E=epidermis; D=dermis)

Figure 3.

The most definitive diagnosis can be obtained by determining the specific target antigens of the patients’ autoantibodies. The determination of target antigens can be achieved by immunoblotting, immunoprecipitaion, or ELISA (2-6,23,45-46). Figure 4 demonstrates the immunoblotting studies showing IgG autoantibodies from a patient with epidermolysis bullosa acquisita (middel lane) that bound to the 290-kDa type VII collagen (arrows), compared to the negative labeling by sera from two individuals who do not have the disease (left and right lanes). The presence of circulating autoantibodies, however, are required to determine the specific target antigens. Thus, the application of these techniques is limited in cases with detectable circulating autoantibodies.

Bullous pemphigoid is the most common and the most thoroughly studied autoimmune subepidermal blistering disease, seen predominantly in the elderly people. The average age of onset is about 65. Mucosal involvement is rare and quite limited. The typical clinical phenotype is a tense blister occurring on the torso and extremities, with a predilection on the flexure areas. The typical histopathologic finding consists of a subepidermal blister with a prominent mast cell and eosinophil infiltrate (42). IgG is deposited linearly at the BMZ of the patient’s peri-lesional skin biopsy specimen. Direct and indirect immunofluorescence microscopy detect in vivo-bound and circulating IgG binding to the epidermal roof of chemically separated skin section (43, Figure 2b). The majority of bullous pemphigoid patients have circulating autoantibodies that recognize a 230-kDa and/or a 180-kDa epidermal protein (BPAg1 and BPAg2 respectively) (2,3). Passive transfer of rabbit antibody raised against the extracellular non-collagenous (NC16A) domain of the mouse BPAg2 to neonatal BALB/c mice induces skin blisters similar to that of human BP blisters (47).

However, prominent mast cell and eosinophil infiltration, the typical histopathologic features characterizing the human BP lesion, is absent in the passive transfer experiment conducted in mice (47).

On the other hand, eotaxin, a newly identified epidermal cell-produced eosinophil-specific chemoattractant, is found to be upregulated in human BP lesional epidermis (48). Furthermore, pro-inflammatory cytokines, IL-1a and TNFa, shown to be potent stimulants for epidermal cell eotaxin production, are produced by mast cells (48-49). Thus, eotaxin may be responsible for recruiting eosinophils to the lesions of BP. Once migrated near the BMZ, eosinophils can release at the BP blister site, a potent protease, MMP-9 (a 92-kD gelatinase), that is capable of cleaving the BPAg2, and may be critical for the BP blister formation (50).

Linear IgA bullous dermatosis is a clinically distinct subepidermal blistering disease commonly seen in children and young adults (4,42,51). One of the striking clinical features is the grouping of blisters in a circular manner that resembles a "string of pearl" configuration (42). This disease is commonly associated with a significant pruritus. Histopathologically, a subepidermal blister is seen with prominent neutrophilic infiltration mostly at the papillary dermis, the so called "papillary microabscess". IgA is linearly deposited at the perilesional BMZ of patients’ skin biopsy specimen. In a minority of patients, circulating IgA autoantibodies that bind to the epidermal side of chemically separated skin section are detected (4,51). Some patients, in addition to IgA, also have IgG class autoantibodies against the skin BMZ (51). Although initially the target antigen of the patients’ IgA autoantibodies was thought to be a unique protein located at the hemidesmosome/upper lamina lucida areas (4), now, it appears that the target antigen is an epitope within the BPAg2 (52). It is not known why the immune system, instead of IgG class, initiates the formation of IgA class autoantibodies to target the same protein.

Cicatricial pemphigoid is a putative autoimmune disease clinically characterized by chronic inflammation and scarring at the mucous membranes, especially the oral and ocular mucosae. Sometimes, minimal skin involvement is observed (53). The histologic and direct immunofluorescence findings in cicatricial pemphigoid is quite similar to that of BP (42, 53). Within the clinical entity of cicatricial pemphigoid, different subsets can be classified according to the specific target antigens recognized by the patients’ autoantibodies. A subset of patients with cicatricial pemphigoid have autoantibodies recognizing the BP antigens, particularly the BP180 (BPAg2) (53-54). Further studies indicated that the autoantibodies from a subset of cicatricial pemphigoid patients recognized epitopes on BPAg2 that are different from the epitopes recognized by BP patients (54). The cicatricial pemphigoid epitopes are located beneath the lower lamina lucida or lamina densa area, so that direct or indirect immunofluorescence studies performed on chemically separated skin sections detect the in vivo-bound or circulating autoantibodies of this subset of patients at the dermal side. Thus, the location of cicatricial pemphigoid antigenic epitopes is distinct from that of the bullous pemphigoid epitopes which are located close to the hemidesmosome area (54). This deeper location of the cicatricial pemphigoid antigenic epitope may be a factor that influences the scarring nature of the disease. A second subset of cicatricial pemphigoid with clinical disease restricted to ocular mucosa is characterized by the absence of circulating autoantibodies and the presence of in vivo-bound fibrin deposition at the patients’ BMZ (53).

The third subset of cicatricial pemphigoid patients were reported to have IgG autoantibodies targeting a lower lamina lucida component epiligrin (55). Subsequently, epiligrin was identified as a member of the laminin family, and was named laminin-5. This protein was also previously known as kalinin, BM600, and nicein (56-57). Laminin-5, a heterotrimer of a₃, b₃, and g₂ chains, shares homology with another member of laminin family, laminin-6 (heterotrimer of a₃, b₁, g₁), which is also located at the skin BMZ (14,17). Autoantibodies from a subset of patients with cicatricial pemphigoid recognized the a₃chain of laminin-5 and thus also recognized laminin-6 (23,57). As a result, both laminin-6 and laminin-5 become target antigens for this subset of cicatricial pemphigoid (23). Therefore, the appropriate name for this subset of cicatricial pemphigoid should be anti-laminin cicatricial pemphigoid (23). Since both laminin-6 and laminin-5 are localized to the lower lamina lucida areas, autoantibodies from these patients bind to the dermal side of chemically separated skin section (Figure 2a) and localize to the lower lamina lucida by immunoelectron microscopy (23,55).

Anti-p105 pemphigoid is a rare IgG-mediated subepidermal blistering disease first reported in 1993 (6), and subsequently, in another patient in 1994 (58). Since the disease is rarely observed, a typical clinical phenotype has not been characterized. The two cases reported thus far showed different clinical phenotype with one case being similar to toxic epidermal necrolysis or pemphigus vulgaris (6,59) and other case being similar to bullous pemphigoid (58). Histopathologically, a subepidermal blister is present predominantly with a neutrophilic infiltration, resembling the histology of linear IgA bullous dermatosis. IgG, and not IgA, is deposited linearly at the BMZ of patient’s perilesional skin biopsy specimen. Direct or indirect immunofluorescence microsopy detect in vivo-bound and circulating IgG binding to the dermal floor of chemically separated skin. These patients also had IgG circulating autoantibodies that recognized a lower lamina lucida-located 105-kDa protein (P105) synthesized and secreted by epidermal keratinocytes and dermal fibroblasts (6).

Epidermolysis bullosa acquisita is an IgG-mediated subepidermal blistering disease with a distinct clinical phenotype. A common clinical feature is chronic inflammtion, blister formation, and scarring, predominantly at the extensor surfaces of the skin: elbows, knees, ankles, buttocks, and dorsal aspect of the hands (42). However, generalized blistering formation occurs in a minority of patients. Some patients have severe mucosal lesions. Histopathologically, epidermolysis bullosa acquisita is a subepidermal blister with a mixed inflammatory cell infiltrate (42). A wide band of IgG is deposited linearly at the patient’s skin BMZ when direct immunofluorescence is performed on the peri-lesional skin biopsy specimen. Direct or indirect immunofluorescence detect IgG autoantibodies that bind to the dermal floor of the chemically separated skin section (43, Figure 2c). Immunoelectron microscopy delineates the IgG binding site to be at the lamina densa and sub-lamina densa area (5, Figure 3). Approximately 50% of the patients have detectable circulating IgG autoantibodies that recognize a 290-kDa skin BMZ protein (5, Figure 4) synthesized and secreted by epidermal keratinocytes and dermal fibroblasts. This 290-kDa epidermolysis bullosa acquisita antigen was subsequently identified as type VII collagen (5). Further studies showed that the patients’s autoantibodies recognized four major immunodominant domains located at the non-collagenous (NC1) domain of type VII collagen, also located at the N-terminus of the protein (60).

Figure 4. Immunoblotting detects the target antigen (a 290-kDa protein, arrow) recognized by autoantibodies from a patient with epidermolysis bullosa acquisita (middle), whereas sera from two individuals without the disease (left and right) do not recognize this protein.