- Email: [email protected]
Anti-epiligrin cicatricial pemphigoid and epidermolysis bullosa acquisita: Differentiation by use of indirect immunofluorescence microscopy
Anti-epiligrin cicatricial pemphigoid and epidermolysis bullosa acquisita: Differentiation by use of indirect immunofluorescence microscopy Robert M. Vodegel, MD,a Marcelus C. J. M. de Jong, PhD,a Hendri H. Pas, PhD,a Kim B. Yancey, MD,b and Marcel F. Jonkman, MD, PhDb Groningen, The Netherlands, and Milwaukee, Wisconsin Binding of autoantibodies to laminin 5 and type VII collagen causes anti-epiligrin cicatricial pemphigoid and epidermolysis bullosa acquisita, respectively. Differentiation between these two dermal-binding autoimmune bullous dermatoses is not yet possible by indirect immunofluorescence microscopy. In this study we tested whether two recently described immunofluorescence techniques, “knockout” skin substrate and fluorescent overlay antigen mapping, can differentiate between anti-epiligrin cicatricial pemphigoid and epidermolysis bullosa acquisita. A total of 10 sera were tested: 4 with antilaminin 5, and 6 with antitype VII collagen autoantibodies, as characterized by either immunoblot or immunoprecipitation analysis. Differentiation between anti-epiligrin cicatricial pemphigoid and epidermolysis bullosa acquisita was possible in all 10 sera by indirect immunofluorescence using either knockout skin substrate or fluorescent overlay antigen mapping technique. (J Am Acad Dermatol 2003;48:542-7.)
A
ntiepiligrin cicatricial pemphigoid (AECP) is a recently described1,2 subepidermal blistering disease characterized by IgG anti– basement membrane zone (BMZ) autoantibodies against laminin 5 (␣33␥2). Probably the disease is underreported3 as indicated by the incidence of CP (0.871.16/million4,5) and the finding that approximately 5% to 15% of patients with CP have IgG autoantibodies against laminin 5.6,7 A potential explanation for the presumed rarity of AECP may be grounded in the difficulty of distinguishing these patients from those with other forms of CP or mucosal predominant forms of epidermolysis bullosa acquisita (EBA) on the basis of routine clinical, histopathologic, and immunopathologic evaluations. It is particularly challenging to distinguish AECP from EBA because anti-BMZ autoantibodies in these patients both bind the dermal side of 1-mol/L salt-split skin.8 This distinction is significant given the recent demonstration From the Departments of Dermatology, Groningen University Hospitala and Medical College of Wisconsin.b Funding sources: Zon/MW. Conflict of interest: None identified. Accepted for publication September 22, 2002. Reprints not available from authors. Correspondence: Prof Dr Marcel F. Jonkman, MD, Department of Dermatology, Groningen University Hospital, PO Box 30.001, 9700 RB Groningen, the Netherlands. E-mail:m.f.jonkman@ derm.azg.nl. Copyright © 2003 by the American Academy of Dermatology, Inc. 0190-9622/2003/$30.00 ⫹ 0 doi:10.1067/mjd.2003.99
542
that in 10 of 35 patients with AECP, a solid cancer was found. The incidence is increased compared with the normal population.9 Although AECP and EBA can be differentiated by immunoelectron microscopy (immunoreactants are found above and below the lamina densa in patients with AECP and EBA, respectively1,2,10) this specialized investigative technique is expensive, time-consuming, and increasingly less available in many institutions. Immunoblotting cannot reliably be used to distinguish patients with AECP from those with EBA because approximately half of the latter lack circulating antiBMZ autoantibodies. Moreover, immunoblotting can yield false-negative findings as a result of loss of conformational epitopes in denatured preparations of such autoantigens. The diagnosis of AECP would be easier to make if it could be achieved with immunofluorescence microscopy techniques. We report the successful use of two indirect immunofluorescence microscopy (IIF) methods—IIF on antigen-deficient skin (“knockout” skin)11 and fluorescence overlay antigen mapping (FOAM)12 on 1-mol/L salt-split skin— for differentiation of AECP from EBA.
MATERIAL AND METHODS Patients Serum characterization. Serum samples were collected from 4 patients with AECP and 6 patients with mechanobullous IgG-mediated EBA. The sera had been characterized by immunoblotting, immunoprecipitation, or both as described previously.8,12
J AM ACAD DERMATOL VOLUME 48, NUMBER 4
The AECP sera reacted with the ␣3 subunit of laminin 5 in immunoblotting. The EBA sera reacted in immunoblot using dermal extract with a 290-kd antigen corresponding with type VII collagen. By IIF, using 1-mol/L salt-split skin,13 all 10 sera showed binding of IgG4 antibodies exclusively to the dermal side of the substrate at a titer of at least 40. Indirect immunofluorescence microscopy on knockout skin substrate. To assess the diagnostic value of skin substrates lacking defined antigens, we used frozen skin specimens lacking either laminin 5 or type VII collagen. These knockout substrates were obtained from (1) a patient with Herlitz junctional epidermolysis bullosa (JEB-H), who was homozygous for the LAMB3:R635X mutation; and (2) a patient with recessive dystrophic epidermolysis bullosa EB (Hallopeau-Siemens type), who was compound heterozygous for 2 null mutations in COL7A1 (M1I/353delG and 352-354insCCCCCTTGCAA). Characterization of these substrates by IIF using an extensive panel of monoclonal antibodies against epidermal BMZ antigens has been described elsewhere.14 The laminin 5– deficient skin did not bind the antilaminin 5 monoclonal antibody GB3,15 whereas the type VII collagen– deficient skin did not bind the antitype VII collagen monoclonal antibodies II-32 and LH7:2.16 For IIF, air-dried cryostat-cut sections of 4-m thickness were mounted on Polysine glass slides and encircled with a hydrophobic emulsion (PAP pen; Dako, Glostrup, Denmark). The sections were then overlaid with AECP and EBA serum samples diluted 1:4 to 1:8 in 0.01 mol/L phosphate-buffered saline (PBS) (pH 7.3). After 30 to 60 minutes incubation in a moist chamber at ambient temperature, followed by thorough washing with PBS, the sections were stained for 30 minutes with either Fc␥specific fluorescein isothiocyanate (FITC)-conjugated goat F(ab⬘)2 antihuman IgG (Protos Immunoresearch, Burlingame, Calif) or Alexa 488conjugated goat antihuman IgG (Molecular Probes Europe, Leiden, The Netherlands). Proper antibody dilutions were made in PBS supplemented with 5⫻ crystallized ovalbumin 2% wt/vol (Serva, Heidelberg, Germany). After a final wash in PBS, the sections were coverslipped under fresh PBS/glycerol (50% vol/vol) and examined with a fluorescence microscope (see below). Fluorescence overlay antigen mapping on salt-split skin. Originally, the computer-aided FOAM procedure was developed to distinguish BMZ immunoreactants in the skin of patients with bullous pemphigoid from those with EBA.12,17,18 For the purpose of this study, the FOAM technique was modified to enable topographic mapping of serum anti-
Vodegel et al 543
BMZ autoantibodies instead of in vivo bound antiBMZ antibodies.12 An essential modification involved the use of salt-split human skin as the test substrate.13 Fresh skin, obtained from breast reduction, was trimmed free of fat tissue and cut into samples of 4 to 6 mm using a biopsy punch. The skin samples were then incubated for 72 to 96 hours at 4°C in an excess of 1 mol/L NaCl solution supplemented with 1 mmol/L phenylmethylsulfonyl fluoride. When small splits appeared at the epidermal BMZ (as verified at regular time intervals by microscopic examination of cryostat-cut sections from parallel blocks) the skin samples were blotted on filter paper to remove excess moisture and stored at ⫺80°C. The level of cleavage in this substrate was always checked by immunofluorescence microscopy using monoclonal antibodies against type XVII collagen (1D119) and laminin 5 (GB3). Topographic mapping of AECP and EBA autoantibodies was performed on cryostat sections of 1 mol/L salt-split skin (see above). The mapping procedure included 3 consecutive immunostaining steps of 30 minutes each (see below) at ambient temperature, alternated by washing steps with 0.01 mol/L PBS for 30 minutes. The first staining step involved binding of rabbit antihuman type VII collagen (a gift of Dr L. Bruckner-Tuderman, Mu¨ nster, Germany). The second staining step was applied as a cocktail including (1) patient serum (1:4-1:20) and (2) rat IgG1a anti– heat shock protein73 (Stressgen Biotechnology, Victoria, British Columbia). The third staining step involved fluorescent detection of the antibodies bound in steps 1 and 2. For this purpose, we used a cocktail of (1) FITC-conjugated donkey anti-rat IgG (Jackson Immunoresearch, West Grove, Pa); (2) lissamine rhodamine sulfonyl chloride (LRSC)– conjugated donkey antirat IgG (Jackson Immunoresearch); (3) LRSC-conjugated donkey antirabbit IgG (Jackson Immunoresearch); and (4) FITC-conjugated mouse antihuman IgG4 (Zymed Laboratories, San Francisco, Calif). This staining procedure yielded: (1) red fluorescent type VII collagen, used as topographic EBMZ reference marker; (2) dual green and red fluorescent HSP73, used as epidermal geometric verification marker; and (3) green fluorescent human IgG4 of the respective AECP or EBA anti-BMZ autoantibodies. Detection of the IgG4 subclass instead of total human IgG had the advantage of producing much lower background staining in the dermal compartment. The specificity of the staining steps was carefully checked according to well-established principles (eg, replacement of the relevant antibodies by irrelevant ones and/or by staining reagents yielding purposely undesired specific fluorescence).
544 Vodegel et al
J AM ACAD DERMATOL APRIL 2003
Fig 1. IIF staining for serum IgG on knockout skin substrates lacking either laminin 5 (A and C) or type VII collagen (B and D). Antiepiligrin cicatricial pemphigoid serum shows faint (or absent) IgG binding in laminin 5– deficient skin (A), but clearly positive linear IgG binding to epidermal BMZ in type VII collagen– deficient skin (B). EBA serum shows opposite binding patterns: positive in laminin 5– deficient skin (C) and negative in type VII collagen– deficient skin (D). Note that antinuclear antibody in EBA serum binds to both knockout substrates.
The sections were examined with a microscope (Leica DMRA) equipped with two switchable light sources (XBO75W and HBO50W) and 8 switchable (and motor-driven) filter cubes for selective incident light fluorescence. For observation of FITC fluorescence (blue narrow-band excitation) we used the XBO75W lamp and a Chroma high Q filter set containing (1) exciter filter HQ 480/40; (2) dichroic beam splitter Q 505 LP; and (3) emitter filter HQ 535/50. For LRSC fluorescence (green excitation) we used the HBO50W lamp and a Chroma Texas red high Q filter set with (1) exciter HQ 560/55, (2) beam splitter Q 595 LP, and (3) emitter HQ 645/75. The microscopic detection of nonoverlapping red and green fluorescent spots at the BMZ in split skin was accomplished by blue-green excitation (HBO50W lamp) using a Chroma dual band filter set with (1) exciter F/TXRD X 495/20 and 570/20, (2)
beam splitter F/TXRD BS 510 and 590, and (3) emitter F/TXRD M 535/40 and 630/50. A PL Apo 40⫻/0.80 dry objective was used throughout this study. For digital analysis of FOAM patterns, pairs of red-and-green fluorescent images were acquired as separate 16-bit, gray-scale FTS images, using a cooled Pictor 416XT CCD camera (Meade, Irvine Calif). The acquisition of gray-scale fluorescent images at selective wavelength bands (see filter sets) has the advantage of being able to be converted into images with primary colors of red and green, without lack of color fidelity.12,18 Further digital processing of images involved (1) conversion of gray-scale image pairs into either pseudo-red or pseudo-green images, (2) overlay of the pseudo-red and green image pairs, and (3) correction of image shift (if present) by pixel shifting of the geometric verification marker.
J AM ACAD DERMATOL VOLUME 48, NUMBER 4
Vodegel et al 545
Fig 2. Fluorescence overlay antigen mapping on 1 mol/L salt-split skin illustrating staining patterns for AECP serum (A and B) and EBA serum (C and D). Both AECP (A) and EBA sera (C) contain IgG4 autoantibody that binds to dermal side of salt-split skin (green fluorescence). Overlay images (B and D) show distribution patterns of IgG4 anti-BMZ antibodies relative to red fluorescent type VII collagen (used as topographic reference marker in this study). AECP serum yields nonoverlap BMZ pattern (B) distinct from yellow overlap pattern of EBA serum sample (D). Note dual red and green fluorescent epidermal geometric verification marker, HSP73, shows overlap in both cases, indicating geometric fidelity of overlay images.
RESULTS Indirect immunofluorescence microscopy on knockout skin IIF with AECP sera showed negative or faint staining on JEB-H skin (the faint staining being a result of cross-reactivity of the autoantibodies to laminin 6 of the epidermal BMZ in laminin 5– deficient skin) (Fig 1, A). In contrast, AECP sera showed positive linear staining in type VII collagen– deficient skin (Fig 1, B). IIF with EBA sera showed linear staining along the epidermal BMZ in laminin 5– deficient skin (Fig 1, C), however, showed an absence of staining in type VII collagen– deficient skin (Fig 1, D).
Fluorescence overlay antigen mapping on salt-split skin All AECP and EBA sera stained the dermal side of 1-mol/L salt-split skin (Fig 2, A and C). Using the multicolor FOAM technique with antitype VII collagen monoclonal antibody (LH7:2) as the topographic reference marker, the 4 AECP sera showed a nonoverlap fluorescence pattern at the epidermal BMZ (Fig 2, B). This pattern was characterized by isolated spots of red fluorescent type VII collagen pointing downward from beneath the yellow-green fluorescent laminin 5 in the BMZ. The 6 EBA sera, however, showed an overlap of the green fluores-
546 Vodegel et al
cent antitype VII collagen antibody and the red fluorescent topographic reference marker, yielding a more or less yellow fluorescent linear pattern (Fig 2, D). Concomitant double staining of the geometric verification marker HSP in the suprabasal epidermal cells showed almost complete overlap between red and green fluorescent HSP antigen, indicating the geometric reliability of the overlay images in all instances.
DISCUSSION In this study we describe two IIF methods that differentiate AECP and EBA, subepidermal bullous dermatoses characterized by dermal-binding IgG anti-BMZ autoantibodies directed against the dermal side of 1 mol/L salt-split skin. The first IIF method using knockout skin differentiated between AECP and EBA by determining loss of IgG binding to laminin 5– or type VII collagen– deficient skin substrates. Laminin 5–negative skin was obtained from a patient with JEB-H. The loss of laminin 5 in such patients may be the consequence of a mutation in genes encoding either the ␣-, -, or ␥-laminin 5 subunits. Accordingly, skin from a patient with JEB-H carrying a mutation in either LAMB3 (as in this case) or LAMC2 may retain low levels of antilaminin 5 reactivity as a consequence of binding with the laminin subunit ␣3 in laminin 6 (␣31␥1). This nuance is of particular relevance because most patients with AECP have autoantibodies directed against the ␣ subunit of laminin 5.20,21 However, severe reduction of IgG binding is still indicative for AECP, because binding with normal intensity is found in EBA. Another pitfall might be the presence of heterogeneous autoantibodies in the serum binding to more than one BMZ autoantigen.14 In these cases knockout substrate fails because positive binding is found in any BMZ antigen– deficient substrate. The FOAM technique using salt-split skin as test substrate provides another option for IIF microscopy to discriminate between sera binding to laminin 5 or type VII collagen. The computer-aided FOAM technique enhances the resolution limit of the light microscope from nearly 250 nm to around 60 to 65 nm.17 The estimated distance between the type VII collagen and laminin 5 is about 100 nm.17 Actually, this distance may be increased by swelling of collagen in salt-split skin and thus seems well within the resolution limit of the technique. It was proposed that the type VII collagen would spike beneath the level of the laminin 5.22 Using the computer-aided FOAM technique, we found the expected pattern in which isolated spots of type VII collagen localized just beneath the laminin 5 in the BMZ. Currently, it is not known if it is possible to differentiate between
J AM ACAD DERMATOL APRIL 2003
anti-p200 pemphigoid23 and AECP using the FOAM technique. The diagnosis of AECP is on the basis of criteria described previously (1) chronic subepithelial blistering lesions of mucous membranes and skin, (2) in situ and circulating IgG anti-BMZ autoantibodies against the lower lamina lucida at its interface with the lamina densa, and (3) circulating IgG autoantibodies that immunoprecipitate epiligrin/laminin 5 from human keratinocyte extracts, culture media, or both.24 The evaluation of anti-BMZ autoantibody binding to the dermal side of salt-split skin can now be extended by IIF on knockout substrates, by FOAM, or both to meet criterion 2. For confirmation of the diagnosis of AECP, immunoprecipitation remains the gold standard (as mentioned in criterion 3). Reliable data of negative IIF on salt-split skin in patients with AECP or EBA are difficult to find.25,26 These types of data would require immunoelectron microsocopy of the skin in each patient showing IgG, IgA, or both at their EBMZ. Our findings indicate that the majority of patients showing dermal staining by IIF on salt-split skin have EBA (32/36), not AECP (2/36). Therefore, it appears that the majority of patients with dermal staining in salt-split skin have EBA, not AECP. We conclude that AECP can be differentiated from EBA by IIF using knockout substrate, the FOAM technique, or both. The former technique is easier to perform; the latter is more powerful. REFERENCES 1. Domloge-Hultsch N, Gammon WR, Briggaman RA, Gil SG, Carter WG, Yancey KB. Epiligrin, the major human keratinocyte integrin ligand, is a target in both an acquired autoimmune and an inherited subepidermal blistering skin disease. J Clin Invest 1992;90:1628-33. 2. Domloge-Hultsch N, Anhalt GJ, Gammon WR, Lazarova Z, Briggaman RA, Welch M, et al. Antiepiligrin cicatricial pemphigoid: a subepithelial bullous disorder. Arch Dermatol 1994;130: 1521-9. 3. Leverkus M, Schmidt E, Lazarova Z, Brocker EB, Yancey KB, Zillikens D. Antiepiligrin cicatricial pemphigoid: an underdiagnosed entity within the spectrum of scarring autoimmune subepidermal bullous diseases? Arch Dermatol 1999;135:1091-8. 4. Bernard P, Vaillant L, Labeille B, Bedane C, Arbeille B, Denoeux JP, et al. Incidence and distribution of subepidermal autoimmune bullous skin diseases in three French regions: bullous diseases French study group. Arch Dermatol 1995;131:48-52. 5. Zillikens D, Wever S, Roth A, Weidenthaler-Barth B, Hashimoto T, Brocker EB. Incidence of autoimmune subepidermal blistering dermatoses in a region of central Germany. Arch Dermatol 1995;131:957-8. 6. Lazarova Z, Yee C, Darling T, Briggaman RA, Yancey KB. Passive transfer of anti-laminin 5 antibodies induces subepidermal blisters in neonatal mice. J Clin Invest 1996;98:1509-18. 7. Ghohestani RF, Rousselle P, Nicolas JF, Claudy A. A and b subunits of laminin-5 are target antigens in a subset of patients with cicatricial pemphigoid. J Invest Dermatol 1996;106:846. 8. Lazarova Z, Yancey KB. Reactivity of autoantibodies from pa-
Vodegel et al 547
J AM ACAD DERMATOL VOLUME 48, NUMBER 4
9.
10.
11.
12.
13.
14.
15.
16.
17.
tients with defined subepidermal bullous diseases against 1 mol/L salt-split skin. Specificity, sensitivity, and practical considerations. J Am Acad Dermatol 1996;35:398-403. Egan CA, Lazarova Z, Darling TN, Yee C, Cote T, Yancey KB. Anti-epiligrin cicatricial pemphigoid and relative risk for cancer. Lancet 2001;357:1850-1. Nieboer C, Boorsma DM, Woerdeman MJ, Kalsbeek GL. Epidermolysis bullosa acquisita; immunofluorescence, electron microscopic and immunoelectron microscopic studies in four patients. Br J Dermatol 1980;102:383-92. Vodegel RM, Kiss M, De Jong MCJM, Altmeyer A, Molnar K, Husz S, et al. The use of skin substrates deficient in basement membrane molecules for the diagnosis of autoimmune bullous disease. Eur J Dermatol 1998;8:83-5. De Jong M, Bruins S, Heeres K, Jonkman MF, Nieboer C, Boorsma DM, et al. Bullous pemphigoid and epidermolysis bullosa acquisita; differentiation by fluorescence overlay antigen mapping. Arch Dermatol 1996;132:151-7. Gammon WR, Fine JD, Forbes M, Briggaman RA. Immunofluorescence on split skin for the detection and differentiation of basement membrane zone autoantibodies. J Am Acad Dermatol 1992;27:79-87. Jonkman MF, Schuur J, Dijk F, Heeres K, De Jong MCJM, van der Meer JB, et al. Inflammatory variant of epidermolysis bullosa acquisita with IgG autoantibodies against type VII collagen and laminin alpha3. Arch Dermatol 2000;136:227-31. Verrando P, Hsi BL, Yeh CJ, Pisani A, Serieys N, Ortonne JP. Monoclonal antibody GB3, a new probe for the study of human basement membranes and hemidesmosomes. Exp Cell Res 1987;170:116-28. Shimizu H, McDonald JN, Gunner DB, Black MM, Bhogal BS, Leigh IM, et al. Epidermolysis bullosa acquisita antigen and the carboxy terminus of type VII collagen have a common immunolocalization to anchoring fibrils and lamina densa of basement membrane. Br J Dermatol 1990;122:577-85. Bruins S, de Jong MC, Heeres K, Wilkinson W, Jonkman MF, van
18.
19.
20.
21.
22.
23.
24.
25.
26.
der Meer JB. Fluorescence overlay antigen mapping of the epidermal basement membrane zone: III, topographic staining and effective resolution. J Histochem Cytochem 1995;43:64956. Bruins S, de Jong MC, Heeres K, Wilkinson W, Jonkman MF, van der Meer JB. Fluorescence overlay antigen mapping of the epidermal basement membrane zone: II, color fidelity. J Histochem Cytochem 1995;43:715-22. Nishizawa Y, Uematsu J, Owaribe K. HD4, a 180 kDa bullous pemphigoid antigen, is a major transmembrane glycoprotein of the hemidesmosome. J Biochem (Tokyo) 1993;113:493-501. Lazarova Z, Hsu R, Yee C, Yancey KB. Antiepiligrin cicatricial pemphigoid represents an autoimmune response to subunits present in laminin 5 (alpha3beta3gamma2). Br J Dermatol 1998;139:791-7. Chan LS, Majmudar AA, Tran HH, Meier F, Schaumberg-Lever G, Chen M, et al. Laminin-6 and laminin-5 are recognized by autoantibodies in a subset of cicatricial pemphigoid. J Invest Dermatol 1997;108:848-53. Rousselle P, Keene DR, Ruggiero F Champliaud MF, Rest M, Burgeson RE. Laminin 5 binds the NC-1 domain of type VII collagen. J Cell Biol 1997;138:719-28. Zillikens D, Ishiko A, Jonkman MF, Chimanovitch J, Shimizu H, Hashimoto T, et al. Autoantibodies in anti-p200 pemphigoid stain skin lacking laminin 5 and type VII collagen. Br J Dermatol 2000;143:1043-9. Yancey KB, Kirtschig G, Yee C, Lazarova Z. Studies of patients with anti-epiligrin cicatricial pemphigoid. J Dermatol 1995;22: 829-35. Ghohestani RF, Nicolas JF, Rousselle P, Claudy AL Diagnostic value of indirect immunofluorescence on sodium chloridesplit skin in differential diagnosis of subepidermal autoimmune bullous dermatoses. Arch Dermatol 1997;133:1102-7. Kirtschig G, Wojnarowska F. Autoimmune blistering diseases: an update of diagnostic methods and investigations. Clin Exp Dermatol 1994;19:97-112.
A
ntiepiligrin cicatricial pemphigoid (AECP) is a recently described1,2 subepidermal blistering disease characterized by IgG anti– basement membrane zone (BMZ) autoantibodies against laminin 5 (␣33␥2). Probably the disease is underreported3 as indicated by the incidence of CP (0.871.16/million4,5) and the finding that approximately 5% to 15% of patients with CP have IgG autoantibodies against laminin 5.6,7 A potential explanation for the presumed rarity of AECP may be grounded in the difficulty of distinguishing these patients from those with other forms of CP or mucosal predominant forms of epidermolysis bullosa acquisita (EBA) on the basis of routine clinical, histopathologic, and immunopathologic evaluations. It is particularly challenging to distinguish AECP from EBA because anti-BMZ autoantibodies in these patients both bind the dermal side of 1-mol/L salt-split skin.8 This distinction is significant given the recent demonstration From the Departments of Dermatology, Groningen University Hospitala and Medical College of Wisconsin.b Funding sources: Zon/MW. Conflict of interest: None identified. Accepted for publication September 22, 2002. Reprints not available from authors. Correspondence: Prof Dr Marcel F. Jonkman, MD, Department of Dermatology, Groningen University Hospital, PO Box 30.001, 9700 RB Groningen, the Netherlands. E-mail:m.f.jonkman@ derm.azg.nl. Copyright © 2003 by the American Academy of Dermatology, Inc. 0190-9622/2003/$30.00 ⫹ 0 doi:10.1067/mjd.2003.99
542
that in 10 of 35 patients with AECP, a solid cancer was found. The incidence is increased compared with the normal population.9 Although AECP and EBA can be differentiated by immunoelectron microscopy (immunoreactants are found above and below the lamina densa in patients with AECP and EBA, respectively1,2,10) this specialized investigative technique is expensive, time-consuming, and increasingly less available in many institutions. Immunoblotting cannot reliably be used to distinguish patients with AECP from those with EBA because approximately half of the latter lack circulating antiBMZ autoantibodies. Moreover, immunoblotting can yield false-negative findings as a result of loss of conformational epitopes in denatured preparations of such autoantigens. The diagnosis of AECP would be easier to make if it could be achieved with immunofluorescence microscopy techniques. We report the successful use of two indirect immunofluorescence microscopy (IIF) methods—IIF on antigen-deficient skin (“knockout” skin)11 and fluorescence overlay antigen mapping (FOAM)12 on 1-mol/L salt-split skin— for differentiation of AECP from EBA.
MATERIAL AND METHODS Patients Serum characterization. Serum samples were collected from 4 patients with AECP and 6 patients with mechanobullous IgG-mediated EBA. The sera had been characterized by immunoblotting, immunoprecipitation, or both as described previously.8,12
J AM ACAD DERMATOL VOLUME 48, NUMBER 4
The AECP sera reacted with the ␣3 subunit of laminin 5 in immunoblotting. The EBA sera reacted in immunoblot using dermal extract with a 290-kd antigen corresponding with type VII collagen. By IIF, using 1-mol/L salt-split skin,13 all 10 sera showed binding of IgG4 antibodies exclusively to the dermal side of the substrate at a titer of at least 40. Indirect immunofluorescence microscopy on knockout skin substrate. To assess the diagnostic value of skin substrates lacking defined antigens, we used frozen skin specimens lacking either laminin 5 or type VII collagen. These knockout substrates were obtained from (1) a patient with Herlitz junctional epidermolysis bullosa (JEB-H), who was homozygous for the LAMB3:R635X mutation; and (2) a patient with recessive dystrophic epidermolysis bullosa EB (Hallopeau-Siemens type), who was compound heterozygous for 2 null mutations in COL7A1 (M1I/353delG and 352-354insCCCCCTTGCAA). Characterization of these substrates by IIF using an extensive panel of monoclonal antibodies against epidermal BMZ antigens has been described elsewhere.14 The laminin 5– deficient skin did not bind the antilaminin 5 monoclonal antibody GB3,15 whereas the type VII collagen– deficient skin did not bind the antitype VII collagen monoclonal antibodies II-32 and LH7:2.16 For IIF, air-dried cryostat-cut sections of 4-m thickness were mounted on Polysine glass slides and encircled with a hydrophobic emulsion (PAP pen; Dako, Glostrup, Denmark). The sections were then overlaid with AECP and EBA serum samples diluted 1:4 to 1:8 in 0.01 mol/L phosphate-buffered saline (PBS) (pH 7.3). After 30 to 60 minutes incubation in a moist chamber at ambient temperature, followed by thorough washing with PBS, the sections were stained for 30 minutes with either Fc␥specific fluorescein isothiocyanate (FITC)-conjugated goat F(ab⬘)2 antihuman IgG (Protos Immunoresearch, Burlingame, Calif) or Alexa 488conjugated goat antihuman IgG (Molecular Probes Europe, Leiden, The Netherlands). Proper antibody dilutions were made in PBS supplemented with 5⫻ crystallized ovalbumin 2% wt/vol (Serva, Heidelberg, Germany). After a final wash in PBS, the sections were coverslipped under fresh PBS/glycerol (50% vol/vol) and examined with a fluorescence microscope (see below). Fluorescence overlay antigen mapping on salt-split skin. Originally, the computer-aided FOAM procedure was developed to distinguish BMZ immunoreactants in the skin of patients with bullous pemphigoid from those with EBA.12,17,18 For the purpose of this study, the FOAM technique was modified to enable topographic mapping of serum anti-
Vodegel et al 543
BMZ autoantibodies instead of in vivo bound antiBMZ antibodies.12 An essential modification involved the use of salt-split human skin as the test substrate.13 Fresh skin, obtained from breast reduction, was trimmed free of fat tissue and cut into samples of 4 to 6 mm using a biopsy punch. The skin samples were then incubated for 72 to 96 hours at 4°C in an excess of 1 mol/L NaCl solution supplemented with 1 mmol/L phenylmethylsulfonyl fluoride. When small splits appeared at the epidermal BMZ (as verified at regular time intervals by microscopic examination of cryostat-cut sections from parallel blocks) the skin samples were blotted on filter paper to remove excess moisture and stored at ⫺80°C. The level of cleavage in this substrate was always checked by immunofluorescence microscopy using monoclonal antibodies against type XVII collagen (1D119) and laminin 5 (GB3). Topographic mapping of AECP and EBA autoantibodies was performed on cryostat sections of 1 mol/L salt-split skin (see above). The mapping procedure included 3 consecutive immunostaining steps of 30 minutes each (see below) at ambient temperature, alternated by washing steps with 0.01 mol/L PBS for 30 minutes. The first staining step involved binding of rabbit antihuman type VII collagen (a gift of Dr L. Bruckner-Tuderman, Mu¨ nster, Germany). The second staining step was applied as a cocktail including (1) patient serum (1:4-1:20) and (2) rat IgG1a anti– heat shock protein73 (Stressgen Biotechnology, Victoria, British Columbia). The third staining step involved fluorescent detection of the antibodies bound in steps 1 and 2. For this purpose, we used a cocktail of (1) FITC-conjugated donkey anti-rat IgG (Jackson Immunoresearch, West Grove, Pa); (2) lissamine rhodamine sulfonyl chloride (LRSC)– conjugated donkey antirat IgG (Jackson Immunoresearch); (3) LRSC-conjugated donkey antirabbit IgG (Jackson Immunoresearch); and (4) FITC-conjugated mouse antihuman IgG4 (Zymed Laboratories, San Francisco, Calif). This staining procedure yielded: (1) red fluorescent type VII collagen, used as topographic EBMZ reference marker; (2) dual green and red fluorescent HSP73, used as epidermal geometric verification marker; and (3) green fluorescent human IgG4 of the respective AECP or EBA anti-BMZ autoantibodies. Detection of the IgG4 subclass instead of total human IgG had the advantage of producing much lower background staining in the dermal compartment. The specificity of the staining steps was carefully checked according to well-established principles (eg, replacement of the relevant antibodies by irrelevant ones and/or by staining reagents yielding purposely undesired specific fluorescence).
544 Vodegel et al
J AM ACAD DERMATOL APRIL 2003
Fig 1. IIF staining for serum IgG on knockout skin substrates lacking either laminin 5 (A and C) or type VII collagen (B and D). Antiepiligrin cicatricial pemphigoid serum shows faint (or absent) IgG binding in laminin 5– deficient skin (A), but clearly positive linear IgG binding to epidermal BMZ in type VII collagen– deficient skin (B). EBA serum shows opposite binding patterns: positive in laminin 5– deficient skin (C) and negative in type VII collagen– deficient skin (D). Note that antinuclear antibody in EBA serum binds to both knockout substrates.
The sections were examined with a microscope (Leica DMRA) equipped with two switchable light sources (XBO75W and HBO50W) and 8 switchable (and motor-driven) filter cubes for selective incident light fluorescence. For observation of FITC fluorescence (blue narrow-band excitation) we used the XBO75W lamp and a Chroma high Q filter set containing (1) exciter filter HQ 480/40; (2) dichroic beam splitter Q 505 LP; and (3) emitter filter HQ 535/50. For LRSC fluorescence (green excitation) we used the HBO50W lamp and a Chroma Texas red high Q filter set with (1) exciter HQ 560/55, (2) beam splitter Q 595 LP, and (3) emitter HQ 645/75. The microscopic detection of nonoverlapping red and green fluorescent spots at the BMZ in split skin was accomplished by blue-green excitation (HBO50W lamp) using a Chroma dual band filter set with (1) exciter F/TXRD X 495/20 and 570/20, (2)
beam splitter F/TXRD BS 510 and 590, and (3) emitter F/TXRD M 535/40 and 630/50. A PL Apo 40⫻/0.80 dry objective was used throughout this study. For digital analysis of FOAM patterns, pairs of red-and-green fluorescent images were acquired as separate 16-bit, gray-scale FTS images, using a cooled Pictor 416XT CCD camera (Meade, Irvine Calif). The acquisition of gray-scale fluorescent images at selective wavelength bands (see filter sets) has the advantage of being able to be converted into images with primary colors of red and green, without lack of color fidelity.12,18 Further digital processing of images involved (1) conversion of gray-scale image pairs into either pseudo-red or pseudo-green images, (2) overlay of the pseudo-red and green image pairs, and (3) correction of image shift (if present) by pixel shifting of the geometric verification marker.
J AM ACAD DERMATOL VOLUME 48, NUMBER 4
Vodegel et al 545
Fig 2. Fluorescence overlay antigen mapping on 1 mol/L salt-split skin illustrating staining patterns for AECP serum (A and B) and EBA serum (C and D). Both AECP (A) and EBA sera (C) contain IgG4 autoantibody that binds to dermal side of salt-split skin (green fluorescence). Overlay images (B and D) show distribution patterns of IgG4 anti-BMZ antibodies relative to red fluorescent type VII collagen (used as topographic reference marker in this study). AECP serum yields nonoverlap BMZ pattern (B) distinct from yellow overlap pattern of EBA serum sample (D). Note dual red and green fluorescent epidermal geometric verification marker, HSP73, shows overlap in both cases, indicating geometric fidelity of overlay images.
RESULTS Indirect immunofluorescence microscopy on knockout skin IIF with AECP sera showed negative or faint staining on JEB-H skin (the faint staining being a result of cross-reactivity of the autoantibodies to laminin 6 of the epidermal BMZ in laminin 5– deficient skin) (Fig 1, A). In contrast, AECP sera showed positive linear staining in type VII collagen– deficient skin (Fig 1, B). IIF with EBA sera showed linear staining along the epidermal BMZ in laminin 5– deficient skin (Fig 1, C), however, showed an absence of staining in type VII collagen– deficient skin (Fig 1, D).
Fluorescence overlay antigen mapping on salt-split skin All AECP and EBA sera stained the dermal side of 1-mol/L salt-split skin (Fig 2, A and C). Using the multicolor FOAM technique with antitype VII collagen monoclonal antibody (LH7:2) as the topographic reference marker, the 4 AECP sera showed a nonoverlap fluorescence pattern at the epidermal BMZ (Fig 2, B). This pattern was characterized by isolated spots of red fluorescent type VII collagen pointing downward from beneath the yellow-green fluorescent laminin 5 in the BMZ. The 6 EBA sera, however, showed an overlap of the green fluores-
546 Vodegel et al
cent antitype VII collagen antibody and the red fluorescent topographic reference marker, yielding a more or less yellow fluorescent linear pattern (Fig 2, D). Concomitant double staining of the geometric verification marker HSP in the suprabasal epidermal cells showed almost complete overlap between red and green fluorescent HSP antigen, indicating the geometric reliability of the overlay images in all instances.
DISCUSSION In this study we describe two IIF methods that differentiate AECP and EBA, subepidermal bullous dermatoses characterized by dermal-binding IgG anti-BMZ autoantibodies directed against the dermal side of 1 mol/L salt-split skin. The first IIF method using knockout skin differentiated between AECP and EBA by determining loss of IgG binding to laminin 5– or type VII collagen– deficient skin substrates. Laminin 5–negative skin was obtained from a patient with JEB-H. The loss of laminin 5 in such patients may be the consequence of a mutation in genes encoding either the ␣-, -, or ␥-laminin 5 subunits. Accordingly, skin from a patient with JEB-H carrying a mutation in either LAMB3 (as in this case) or LAMC2 may retain low levels of antilaminin 5 reactivity as a consequence of binding with the laminin subunit ␣3 in laminin 6 (␣31␥1). This nuance is of particular relevance because most patients with AECP have autoantibodies directed against the ␣ subunit of laminin 5.20,21 However, severe reduction of IgG binding is still indicative for AECP, because binding with normal intensity is found in EBA. Another pitfall might be the presence of heterogeneous autoantibodies in the serum binding to more than one BMZ autoantigen.14 In these cases knockout substrate fails because positive binding is found in any BMZ antigen– deficient substrate. The FOAM technique using salt-split skin as test substrate provides another option for IIF microscopy to discriminate between sera binding to laminin 5 or type VII collagen. The computer-aided FOAM technique enhances the resolution limit of the light microscope from nearly 250 nm to around 60 to 65 nm.17 The estimated distance between the type VII collagen and laminin 5 is about 100 nm.17 Actually, this distance may be increased by swelling of collagen in salt-split skin and thus seems well within the resolution limit of the technique. It was proposed that the type VII collagen would spike beneath the level of the laminin 5.22 Using the computer-aided FOAM technique, we found the expected pattern in which isolated spots of type VII collagen localized just beneath the laminin 5 in the BMZ. Currently, it is not known if it is possible to differentiate between
J AM ACAD DERMATOL APRIL 2003
anti-p200 pemphigoid23 and AECP using the FOAM technique. The diagnosis of AECP is on the basis of criteria described previously (1) chronic subepithelial blistering lesions of mucous membranes and skin, (2) in situ and circulating IgG anti-BMZ autoantibodies against the lower lamina lucida at its interface with the lamina densa, and (3) circulating IgG autoantibodies that immunoprecipitate epiligrin/laminin 5 from human keratinocyte extracts, culture media, or both.24 The evaluation of anti-BMZ autoantibody binding to the dermal side of salt-split skin can now be extended by IIF on knockout substrates, by FOAM, or both to meet criterion 2. For confirmation of the diagnosis of AECP, immunoprecipitation remains the gold standard (as mentioned in criterion 3). Reliable data of negative IIF on salt-split skin in patients with AECP or EBA are difficult to find.25,26 These types of data would require immunoelectron microsocopy of the skin in each patient showing IgG, IgA, or both at their EBMZ. Our findings indicate that the majority of patients showing dermal staining by IIF on salt-split skin have EBA (32/36), not AECP (2/36). Therefore, it appears that the majority of patients with dermal staining in salt-split skin have EBA, not AECP. We conclude that AECP can be differentiated from EBA by IIF using knockout substrate, the FOAM technique, or both. The former technique is easier to perform; the latter is more powerful. REFERENCES 1. Domloge-Hultsch N, Gammon WR, Briggaman RA, Gil SG, Carter WG, Yancey KB. Epiligrin, the major human keratinocyte integrin ligand, is a target in both an acquired autoimmune and an inherited subepidermal blistering skin disease. J Clin Invest 1992;90:1628-33. 2. Domloge-Hultsch N, Anhalt GJ, Gammon WR, Lazarova Z, Briggaman RA, Welch M, et al. Antiepiligrin cicatricial pemphigoid: a subepithelial bullous disorder. Arch Dermatol 1994;130: 1521-9. 3. Leverkus M, Schmidt E, Lazarova Z, Brocker EB, Yancey KB, Zillikens D. Antiepiligrin cicatricial pemphigoid: an underdiagnosed entity within the spectrum of scarring autoimmune subepidermal bullous diseases? Arch Dermatol 1999;135:1091-8. 4. Bernard P, Vaillant L, Labeille B, Bedane C, Arbeille B, Denoeux JP, et al. Incidence and distribution of subepidermal autoimmune bullous skin diseases in three French regions: bullous diseases French study group. Arch Dermatol 1995;131:48-52. 5. Zillikens D, Wever S, Roth A, Weidenthaler-Barth B, Hashimoto T, Brocker EB. Incidence of autoimmune subepidermal blistering dermatoses in a region of central Germany. Arch Dermatol 1995;131:957-8. 6. Lazarova Z, Yee C, Darling T, Briggaman RA, Yancey KB. Passive transfer of anti-laminin 5 antibodies induces subepidermal blisters in neonatal mice. J Clin Invest 1996;98:1509-18. 7. Ghohestani RF, Rousselle P, Nicolas JF, Claudy A. A and b subunits of laminin-5 are target antigens in a subset of patients with cicatricial pemphigoid. J Invest Dermatol 1996;106:846. 8. Lazarova Z, Yancey KB. Reactivity of autoantibodies from pa-
Vodegel et al 547
J AM ACAD DERMATOL VOLUME 48, NUMBER 4
9.
10.
11.
12.
13.
14.
15.
16.
17.
tients with defined subepidermal bullous diseases against 1 mol/L salt-split skin. Specificity, sensitivity, and practical considerations. J Am Acad Dermatol 1996;35:398-403. Egan CA, Lazarova Z, Darling TN, Yee C, Cote T, Yancey KB. Anti-epiligrin cicatricial pemphigoid and relative risk for cancer. Lancet 2001;357:1850-1. Nieboer C, Boorsma DM, Woerdeman MJ, Kalsbeek GL. Epidermolysis bullosa acquisita; immunofluorescence, electron microscopic and immunoelectron microscopic studies in four patients. Br J Dermatol 1980;102:383-92. Vodegel RM, Kiss M, De Jong MCJM, Altmeyer A, Molnar K, Husz S, et al. The use of skin substrates deficient in basement membrane molecules for the diagnosis of autoimmune bullous disease. Eur J Dermatol 1998;8:83-5. De Jong M, Bruins S, Heeres K, Jonkman MF, Nieboer C, Boorsma DM, et al. Bullous pemphigoid and epidermolysis bullosa acquisita; differentiation by fluorescence overlay antigen mapping. Arch Dermatol 1996;132:151-7. Gammon WR, Fine JD, Forbes M, Briggaman RA. Immunofluorescence on split skin for the detection and differentiation of basement membrane zone autoantibodies. J Am Acad Dermatol 1992;27:79-87. Jonkman MF, Schuur J, Dijk F, Heeres K, De Jong MCJM, van der Meer JB, et al. Inflammatory variant of epidermolysis bullosa acquisita with IgG autoantibodies against type VII collagen and laminin alpha3. Arch Dermatol 2000;136:227-31. Verrando P, Hsi BL, Yeh CJ, Pisani A, Serieys N, Ortonne JP. Monoclonal antibody GB3, a new probe for the study of human basement membranes and hemidesmosomes. Exp Cell Res 1987;170:116-28. Shimizu H, McDonald JN, Gunner DB, Black MM, Bhogal BS, Leigh IM, et al. Epidermolysis bullosa acquisita antigen and the carboxy terminus of type VII collagen have a common immunolocalization to anchoring fibrils and lamina densa of basement membrane. Br J Dermatol 1990;122:577-85. Bruins S, de Jong MC, Heeres K, Wilkinson W, Jonkman MF, van
18.
19.
20.
21.
22.
23.
24.
25.
26.
der Meer JB. Fluorescence overlay antigen mapping of the epidermal basement membrane zone: III, topographic staining and effective resolution. J Histochem Cytochem 1995;43:64956. Bruins S, de Jong MC, Heeres K, Wilkinson W, Jonkman MF, van der Meer JB. Fluorescence overlay antigen mapping of the epidermal basement membrane zone: II, color fidelity. J Histochem Cytochem 1995;43:715-22. Nishizawa Y, Uematsu J, Owaribe K. HD4, a 180 kDa bullous pemphigoid antigen, is a major transmembrane glycoprotein of the hemidesmosome. J Biochem (Tokyo) 1993;113:493-501. Lazarova Z, Hsu R, Yee C, Yancey KB. Antiepiligrin cicatricial pemphigoid represents an autoimmune response to subunits present in laminin 5 (alpha3beta3gamma2). Br J Dermatol 1998;139:791-7. Chan LS, Majmudar AA, Tran HH, Meier F, Schaumberg-Lever G, Chen M, et al. Laminin-6 and laminin-5 are recognized by autoantibodies in a subset of cicatricial pemphigoid. J Invest Dermatol 1997;108:848-53. Rousselle P, Keene DR, Ruggiero F Champliaud MF, Rest M, Burgeson RE. Laminin 5 binds the NC-1 domain of type VII collagen. J Cell Biol 1997;138:719-28. Zillikens D, Ishiko A, Jonkman MF, Chimanovitch J, Shimizu H, Hashimoto T, et al. Autoantibodies in anti-p200 pemphigoid stain skin lacking laminin 5 and type VII collagen. Br J Dermatol 2000;143:1043-9. Yancey KB, Kirtschig G, Yee C, Lazarova Z. Studies of patients with anti-epiligrin cicatricial pemphigoid. J Dermatol 1995;22: 829-35. Ghohestani RF, Nicolas JF, Rousselle P, Claudy AL Diagnostic value of indirect immunofluorescence on sodium chloridesplit skin in differential diagnosis of subepidermal autoimmune bullous dermatoses. Arch Dermatol 1997;133:1102-7. Kirtschig G, Wojnarowska F. Autoimmune blistering diseases: an update of diagnostic methods and investigations. Clin Exp Dermatol 1994;19:97-112.