Anti-desmoplakin antibodies in erythema multiforme and Stevens-Johnson syndrome sera: pathogenic or epiphenomenon?

ARTICLE

Auteur(s) : Emanuele COZZANI¹ emanuele.cozzani@unige.it, Giovanni DI ZENZO², Valentina CALABRESI², Marzia CAPRONI³, Donatella SCHENA⁴, Pietro QUAGLINO⁵, Angelo V. MARZANO⁶, Paolo FABBRI³, Alfredo REBORA¹, Aurora PARODI¹

¹ Di.S.E.M. Section of Dermatology, University of Genoa, viale Benedetto XV, 7, 16132, Genoa, Italy

² Molecular and Cell Biology Laboratory, Istituto Dermopatico dell’Immacolata, IDI-IRCCS, Rome 00167, Italy

³ II Department of Dermatological Sciences, University of Florence, via degli Alfani, 37, 50121, Florence

⁴ Clinica Dermatologica, Università di Verona

⁵ Department of Biomedical Sciences and Human Oncology, Section of Clinics and Oncological Dermatology, University of Turin, via Cherasco 23, 10126 Turin

⁶ Institute of Dermatological Sciences, University of Milan–IRCCS Ospedale Maggiore of Milan, Italy

Reprints: E. COZZANI

Erythema multiforme (EM), Stevens-Johnson syndrome (SJS), and toxic epidermal necrolysis (TEN) are skin disorders with acute inflammatory eruptions with a broad spectrum of clinical manifestations [1-4]. The lesions are probably produced by a delayed-type hypersensitivity reaction to various antigens such as medications and/or infections, and, in some cases, the epidermal damage is mediated by cytotoxic T cells [5, 6].

Foedinger and co-workers found, in patients with severe EM, autoantibodies against desmoplakins (Dp) I and II [7-10]. In addition, the same autoantibodies have been recently detected in a patient with oral EM [11] and have also been identified as components of the antigenic complex characteristic of paraneoplastic pemphigus [12] and as a target of pemphigus autoantibodies [13]. Lastly, periplakin, an additional component of the plakin family, has been recognized as the target of circulating autoantibodies in the sera of patients with TEN [14]. On the other hand, whether autoantibodies, in particular against plakins, components of desmosomal keratinocytes, play a pathogenic role or are the result of an epitope spreading phenomenon is still not fully understood.

The aim of the present study was to characterize the keratinocyte antigens recognized by autoantibodies from EM, TEN and SJS patients, by analyzing the reactivity of sera on normal human skin and rat bladder epithelium by indirect immunofluorescence (IIF), immunoblotting (IB) and to assess if patient sera reacted with target proteins in native forms as well, by immunoprecipitation (IP) with keratinocyte extracts.

Materials and methods

Thirty-three patients were recruited by the Italian Group of Immunopathology (GIP). According to Roujeau's criteria [3], they were clinically classified into 4 groups: 1) EM minor, when the patients presented characteristically shaped skin lesions (target lesions) symmetrically distributed, with or without blisters, without mucosal involvement; 2) EM major, when the patients presented lesions distributed acrally, namely typical target or raised atypical target lesions, mucosal erosions and skin detachment on less than 10% body surface area (BSA); 3) Stevens-Johnson syndrome (SJS) characterized by widespread lesions, macules with blisters or flat atypical target lesions, mucosal erosions and skin detachment on less than 10% BSA; 4) Toxic epidermal necrolysis (TEN) characterized by widespread lesions, macules with blisters or flat atypical target mucosal erosions and skin detachment larger than 30% BSA. All patients developed EM, SJS or TEN as a first event. None had have recurrences.

Two patients had TEN, 1 SJS, 9 EM major and 21 EM minor. Twenty-one patients had taken drugs, 9 had a history of viral infection (HHV 1) and 4 had both taken drugs and had a viral infection (HHV 1).

All sera were collected the same day or the day after the occurrence of the skin or mucosal lesions and were studied in IIF, IB and IP. As for IIF, normal human skin (obtained from aesthetic mammoplastic surgery with prior patient's consent) and rat bladder epithelium were used as substrates, according to standard procedure [15]. As for IB, as previously described [16, 9], normal human keratinocytes were used as a source of antigens. Briefly, epidermal proteins were extracted from normal human keratinocytes, fractionated under reducing conditions by 6% SDS-PAGE and blotted onto polyvinylidene fluoride membrane Immobilon-P (Millipore, Bedford, MA). Filters were incubated with blocking solution (5% milk, 0.1% Tween 20 in Tris-buffered saline) for 2 hours at room temperature. Serum samples were diluted 1:50 and incubated with blocking solution on filters overnight at 4 ̊C. Anti-Dp I and II monoclonal antibodies diluted 1: 100 were used as reference (Progen, Heidelberg, D). After washing (0.1% Tween 20 in Tris-buffered saline), the filters were incubated with an alkaline phosphatase-conjugated rabbit antihuman IgG (H + L) (Southern Biotechnology Associates Inc., Birmingham, AL) for 1 hour at RT, washed, and stained with 330 mg/ml of nitro-blue tetrazolium and 165 mg/ml of 5-bromo-4- chloro-3-indolyl phosphate (Roche Diagnostics, Basel, CH).

IP was performed on extracts of human keratinocytes, grown to near confluence and incubated overnight with S-35-labeled amino acids (PerkinElmer Life, Shelton, CT). Cells were extracted in 1% NP-40 in TBS with 1 mM Pefabloc (Roche Diagnostics GmbH, Roche Applied Science, Mannheim, D), Antipain and Leupeptin (both 1 μg/mL; from Sigma-Aldrich, St. Louis. MO) and Complete (Proteinase Inhibitors Cocktail, Roche Applied Science, Indianapolis, IN). A particle-free supernatant was prepared by centrifugation of the cell extracts at 100,000 g for 1 hour at 4 ̊C. Labeled extracts were sequentially preabsorbed with normal human serum and protein A/G PLUS-Agarose (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and then incubated with the sera of all patients and human controls as well as monoclonal antibodies against Dp I and II (Progen, Heidelberg, D) for 1 h at 4 ̊C. Thereafter, antigen-antibody complexes were precipitated with protein A/G PLUS-Agarose (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and separated by SDS-PAGE, as described above. The precipitated antigens were visualized by autoradiography. Twenty sera from normal healthy individuals were used as controls.

Results

No sera carried antibodies to the intercellular and/or dermo-epidermal junctions (table 1). Antinuclear antibodies were positive in 5 sera when normal human skin was used as a substrate and in 7 sera on rat bladder epithelium (table 1).

Table 1 Clinical and immunopathological findings of the 33 patients.

^A Immunoblotting (IB) results are presented as molecular weight (in kDa) of protein recognized by patient's serum.

^B Immunoprecipitation (IP) from protein extracts of radiolabeled cultured human keratinocytes of 250- and 215-kDa molecular mass proteins with 10 patient sera.

^C Immunoprecipitation of Dp I and II with a specific monoclonal antibody from keratinocytes extracts and following immunoblotting with 10 patient sera; IIF, indirect immunofluorescence; NHS, normal human skin; RBE, rat bladder epithelium; ND, not done; ANA, antinuclear antibodies; TEN, toxic epidermal necrolysis; SJS, Stevens Johnson syndrome; EM, erythema multiforme.

In IB, 10 sera reacted with polypeptides of 215 and/or 250-kDa that co-migrate with Dp I and II, as proven by an anti-Dp I and II monoclonal antibody used as reference (figure 1). In particular, a polypeptide of 250 kDa was recognized by the SJS serum (# 3 in table 1), by 4 of 9 EM major sera (44%)(#4.5.7.9 in table 1), and by 5 of 21 EM minor sera (24%) (# 18,23,27,30,32 in table 1), while a polypeptide of 215 kDa was bound by the SJS serum (#3 in table 1), by 3 of 9 EM major sera (33%) (# 4,5,7 in table 1), and by 1 of 21 EM minor sera (#18 in table 1).

As for IP, although weak signals corresponding to 250 and 215 polypeptides were obtained with both patient and control sera, none of the 10 patients’ sera, mentioned above as reacting in IB, immunoprecipitated polypeptides of 215 and/or 250-kDa from radiolabeled extracts (table 1 and figure 2).

To assess the real nature of the proteins recognized by IB and co-migrating with Dp I and II (previously described as targets of a subset of EM patients [7]), keratinocyte extracts were immunoprecipitated with monoclonal antibodies against Dp I and II and analyzed by IB using the 10 positive patients’ sera. Two of 10 positive sera, 1 SJS (# 3 in table 1) and 1 EM minor (# 18 in table 1) reacted with Dp I and II when denaturated by IB procedure (figure 3). Interestingly, we obtained weak signals corresponding to 250 and 215 polypeptides with both some patients’ and control sera. These results suggest that human sera IgG in a normal healthy individual also have a weak ability to bind 250 and 215 polypeptides and this ability results in the background reactivity present in both in figures 2 and 3. These weak signals produced by both patient and normal sera are not disease-related and could be due to “sticky” IgGs reacting with polypeptides without any specificity or cross-reacting IgGs specific for a different target.

Altogether, 9 EM and the SJS sera reacted against denaturated polypeptides of 250 kDa and more rarely of 215 kDa. Of note, 1 EM and 1 SJS patient's sera recognized denaturated epitopes of Dp I and II that were not recognized as native ones, while the remaining 8 sera bound to different unknown denaturated antigens that co-migrate with but are not Dp I and DpII.

Discussion

EM, SJS and TEN probably result from a hyperacute apoptogenic insult on epidermal keratinocytes. However, the problem whether the humoral immune response plays any pathogenetic role remains unsolved. Foedinger et al. [7] found antibodies against Dp I and II in a subset of EM patients, who in addition to target lesions, exhibited widespread tense blisters and extensive mucosal erosions, and demonstrated, by passive transfer of serum into newborn mice, their in vivo-binding to the keratinocyte surface. In addition, they [8] suggested that EM patients with anti-Dp I and II antibodies belonged to a subset of EM histologically characterized by suprabasal acantholysis in the lesional skin and mucous membranes. They found also that anti-Dp antibodies were already present in the early phase of the disease and bound to an epitope within a Dp domain (YSYSYS motif representing amino acids 1739-1744 at the extreme end of the carboxy terminus) that is crucial for the interaction of keratin filaments with desmosomes, confirming that anti-DP antibodies can contribute to the tissue damage [9, 17]. These data suggested a potential pathogenetic role of anti-Dp I and II, at least in this peculiar subset of EM patients.

Such a point of view is not, however, convincing, being based on a peculiar subset of EM, and, in fact, is not shared by other Authors for whom “it seems more prudent to conclude that the identified autoantibodies represent an epiphenomenon due to exposure of desmosomal epitopes as a part of epidermal damage characteristic of EM, rather than a pathogenetic factor” [18].

The involvement of intracellular antigens in the onset of blistering lesions is, however, suggested by some experimental data in sub-epithelial autoimmune bullous disorders. In particular, mucous membrane pemphigoid sera targeting an intracellular domain of β4 integrin have been reported to cause the dermal-epidermal separation in an organ culture model based on oral mucosa [19] and the passive transfer in neonatal mice of antibodies against BP230 peptides to induce sub-epidermal blister formation [20]. On the other hand, Di Zenzo et al. described the dynamics of the humoral response to BP180 (a hemidesmosomal component that is the target of pathogenic autoantibodies in bullous pemphigoid) in mice grafted with skin obtained from transgenic mice expressing human BP180. They proved that antibodies develop first against extracellular epitopes and are followed by the emergence of IgG against intracellular epitopes [21]. Interestingly, the latter were associated with the graft loss, suggesting that their development correlates with the onset of tissue damage.

In the present study, we have shown that 9 patients with EM and 1 SJS had circulating autoantibodies against a protein of 250 kDa and, rarely, to a polypeptide of 215 kDa. These proteins were bound by patient sera only when denaturated by an IB procedure. In particular, one EM minor and the SJS sera reacted with denaturated epitopes of DpI and II, suggesting that they resulted from the epidermal damage produced by aggressive autoreactive T cells previously demonstrated to be preponderantly present within the lesional epidermis [5, 6]. Altogether these findings suggested that the infiltration of such cells into the epidermis may cause the damage that induces denaturation of desmoplakin I and II and renders binding intracellular targets such as DpI and II accessible to autoantibodies.

The difference between our findings and the data of Foedinger and co-workers [6, 7] could depend on the different immunopathological features of their patients with respect to ours. In fact, Foedinger et al., showed that 5 of 7 EM major sera, possessing autoantibodies against DpI and II, stained keratinocyte cell membranes of normal human skin by IIF [7], while none of our patient sera was able to give the same staining pattern. These findings, together with the observation that only patients with EM major and autoantibodies against DpI and II showed suprabasal acantholysis in lesional skin and mucous membranes, suggest that Foedinger's subset of patients could be re-classified as a variant of pemphigus with an unusual phenotype of EM (a phenotype occasionally also observed in PNP).

In conclusion, our findings suggest that the rare reactivity against DpI and II detected in EM patient sera just represents an epiphenomenon that plays only a secondary role in the pathogenesis of the disease.

Disclosure

Financial support: none. Conflict of interest: none.

References

1 JC Huff, WL Weston, M.G. Tonnesen Erythema multiforme: a critical review of characteristics, diagnosis criteria, and causes J Am Acad Dermatol 1983; 8: 763-775.

2 DA Wetter, M.D. Davis Recurrent erythema multiforme: clinical characteristics, etiologic associations, and treatment in a series of 48 patients at Mayo Clinic, 2000 to 2007 J Am Acad Dermatol 2010; 62: 45-53.

3 J.C. Roujeau Stevens-Johnson syndrome and toxic epidermal necrolysis are severity variants of the same disease which differs from erythema multiforme J Dermatol 1997; 24: 726-729.

4 P. Fritsch European Dermatology Forum: skin diseases in Europe. Skin diseases with a high public health impact: toxic epidermal necrolysis and Stevens-Johnson syndrome Eur J Dermatol 2008; 18: 216-217.

5 RJ Margolis, MG Tonnesen, TJ Harrist et al. Lymphocyte subsets and Langerhans cells/indeterminate cells in erythema multiforme J Invest Dermatol 1983; 81: 403-406.

6 MG Tonnesen, TJ Harrist, BU Wintroub et al. Erythema multiforme: microvascular damage and infiltration of lymphocytes and basophils J Invest Dermatol 1983; 80: 282-286.

7 D Foedinger, GJ Anhalt, B Boecskoer et al. Autoantibodies to desmoplakin I and II in patients with erythema multiforme J Exp Med 1995; 181: 169-179.

8 D Foedinger, B Sterniczky, A Elbe et al. Autoantibodies against desmoplakin I and II define a subset of patients with erythema multiforme major J Invest Dermatol 1996; 106: 1012-1016.

9 D Foedinger, A Elbe-Burger, B Sterniczky et al. Erythema multiforme associated human autoantibodies against desmoplakin I and II: biochemical characterization and passive transfer studies into newborn mice J Invest Dermatol 1998; 111: 503-510.

10 G Hinterhuber, M Binder, Y Marquardt et al. Enzyme-linked immunosorbent assay for detection of peptide-specific human antidesmoplakin autoantibodies Br J Dermatol 2005; 153: 413-416.

11 N Fukiwake, Y Moroi, K Urabe et al. Detection of autoantibodies to desmoplakin in a patient with oral erythema multiforme Eur J Dermatol 2007; 17: 238-241.

12 GJ Anhalt, SC Kim, JR Stanley et al. Paraneoplastic pemphigus. An autoimmune mucocutaneous disease associated with neoplasia N Engl J Med 1990; 323: 1729-1735.

13 E Cozzani, MG Dal Bello, A Mastrogiacomo et al. Anti-desmoplakin antibodies in pemphigus vulgaris Br J Dermatol 2006; 154: 624-628.

14 GT Park, G Quan, J.B. Lee Sera from patients with toxic epidermal necrolysis contain autoantibodies to periplakin Br J Dermatol 2006; 155: 337-343.

15 L Hodge, MM Black, N Ramnarain, B. Bhogal Indirect complement immunofluorescence in the immunopathological assessment of bullous pemphigoid, cicatricial pemphigoid, and herpes gestationis Clin Exp Dermatol 1978; 3: 61-67.

16 E Cozzani, J Kanitakis, JF Nicolas et al. Comparative study of indirect immunofluorescence and immunoblotting for the diagnosis of autoimmune pemphigus Arch Dermatol Res 1994; 286: 295-299.

17 K Cauza, G Hinterhuber, U Mann et al. Internalization via plasmalemmal vesicles: a route for antidesmoplakin autoantibodies into cultured human keratinocytes Exp Dermatol 2003; 12: 546-554.

18 SM Johnson, BR Smoller, T.D. Horn Erythema multiforme associated human autoantibodies against desmoplakin I and II J Invest Dermatol 1999; 112: 395-396.

19 KC Bhol, JE Colon, A.R. Ahmed Autoantibody in mucous membrane pemphigoid binds to an intracellular epitope on human beta4 integrin and causes basement membrane zone separation in oral mucosa in an organ culture model J Invest Dermatol 2003; 120: 701-702.

20 M Kiss, S Husz, T Jánossy et al. Experimental bullous pemphigoid generated in mice with an antigenic epitope of the human hemidesmosomal protein BP230 J Autoimmun 2005; 24: 1-10.

21 G Di Zenzo, V Calabresi, EB Olasz et al. Sequential intramolecular epitope spreading of humoral responses to human BPAG2 in a transgenic model J Invest Dermatol 2010; 130: 1040-1047.