Pemphigoid vegetans represents a bullouspemphigoid variant

Patient's IgG autoantibodies identify the major bullous pemphigoidantigen

Lawrence S. Chan, MD,a, b Michael A. Dorman, MD,a Amr Agha, MD,aTadashi Suzuki, MD,a Kevin D. Cooper, MD,b and Ken Hashimoto, MDaDetroit and Ann Arbor, Michigan

Pemphigoid vegetans is a rare disease; only four cases have been reported. The relationshipof pemphigoid vegetans to bullouspemphigoid is not completely defined.We describe a casein a 77-year-old woman who had multiple well-circumscribed, erythematous, erosive, andvegetating plaques in the axillae and inframammary areas and in the neck. Microscopic ex­amination revealed epidermal hyperplasia, dermoepidermal junctional separation, andprominent dermal eosinophilia. In the perilesional skin IgG was deposited linearly at thebasement membrane zone. In addition, our patient's IgG autoantibodies, primarily the IgG4subclass, reacted with the epidermal side of salt-split normal human skin and with the 230kd major bullous pemphigoid antigen. We conclude that pemphigoid vegetans is best clas­sified as a bullous pemphigoid variant. (1 AM ACAD DERMATOL 1993;28:331-5.)

Bullous pemphigoid (BP) is an immune-mediatedsubepidermal bullous dermatosis that is immuno­pathologically characterized by IgG deposition atthe basement membrane zone (BMZ). Most pa­tients have circulating anti-BMZ antibodies thatreact with the epidermal side of salt-split skin sub­strate.' These antibodies detect the 230kd major orthe 160 to 180 kd minor BP antigens synthesized bykeratinocytes.v" Winkelmann and Sus first reporteda distinct clinical entity they named pemphigoidvegetans that was characterized by pemphigus veg­etans-like lesions clinically, subepidermal patho­logic separation histologically, and in vivo-boundand circulating anti-BMZ antibodies immunologi­cally.i' Subsequently, three similar cases have beenreported.6-8 Indirect immunoelectron microscopicexamination performed in one case localized theimmune reaction products in the upper lamina luci­da. 8 We report an additional patient with pemphi-

Fromthe DepartmentofDermatology, WayneSlate UniversitySchoolof Medicine," and the Department of Dermatology, University ofMichiganSchool of Medicine."

Reprint requests: LawrenceS. Chan, MD, S-EUniversity Health Cen­ter, 4201 81. Antoine,Detroit, MI 48201.

Copyright© 1993 by the American Academyof Dermatology.

0190.9622/93 $1.00+.1016/4/39261

goid vegetans and identify the specific BMZ anti­gens detected by the patient's autoantibodies.

CASE REPORT

A 77-year-old woman had skin lesions of unknownduration. Examination revealed multiple confluent, well­demarcated erythematous plaques that were symmetri­cally located in the patient's axillae (Fig. 1), inframam­mary areas, and on the neck. The plaques had erosiveandvegetating serpiginous borders with numerous peripheralvesicles and pustules. Nikolsky's sign was absent, and themucosae were spared. Findings of two lesional skin biopsyspecimens revealed epidermal hyperplasia, dermoepider­mal junctional separation, and diffuse eosinophilic infil­tration in the papillary dermis. Acantholysis was absent.Direct immunofluorescence microscopic examination of aperilesionalbiopsy specimen detected linear IgG deposi­tion along the skin BMZ. IgM, IgA, C3, and fibrin werenot present. No immune deposit was observed on kerati­nocyte cell surfaces.

Laboratory studies revealed the following abnormalresults: leukocyte count, 13.7 X 109/L; hemoglobin, 104gm/L; and albumin, 22 gm/L. Results of an auramine­rhodamine fluorescent stain of a sputum sample revealeda few (I +) acid-fast bacilli. A chest x-ray film demon­strated peripheral adenopathy and extensive cavitary in­filtrates in the left upper lobe. Serologic studies revealeda l:4 titer of rapid plasma reagin and a reactive fluores-

331

332 Chan et al.Journal of the American Academy of Dermatology

February 1993

Fig. 1. Vegetating plaques and peripheral vesicles and pustules in patient's right axilla.

Fig. 2. Indirect immunofluorescence microscopic examination detects the patient's IgGanti-BMZ antibodies that react with the epidermal side of salt-split skin at a 1:10 dilution(solid arrows) but not with the dermal side(hollow arrow).

cent treponemal antibody, absorbed. Results of a Vene­real Disease Research Laboratory test of spinal fluid wasnonreactive. Results of an auramine-rhodamine fluores­cent stain of spinal fluid and a skin biopsy specimen didnot reveal any acid-fast bacilli. Mycobacterial culturesfrom sputum, spinal fluid, and skin showed no growth af­ter 8 weeks.

A diagnosis of pemphigoid vegetans was made. Dap­sone, 50 mg.'day, was initiated and increased to 100 mg!day. The patient's lesions gradually resolved during thenext 2 weeks.

Immunofluorescence studies. Indirect immunofluores­cence microscopic examination was performed with a 1:3

and a I:10 dilution of the patient's serum on I mol/L so­dium chloride salt-split normal human skin substrate,1

followed by visualization with fluorescein-conjugatedgoat anti-human IgG. To determine the subclasses of thepatient's autoantibodies, indirect immunofluorescencewas performed with the patient's serum, followed by flu­orescein-conjugated monoclonalanti-human IgG1, IgG2,IgG3, and IgG4 (Sigma Diagnostics, St. Louis, Mo.). Tocompare the tissue reactivity of our patient's autoanti­bodies with that of BP control, indirect immunofluores­cence was performed on normal human oral mucosa,conjunctiva, and rat tongue substrates.

Immunoblot studies. Normal human cultured kerati-

Journal of the American Academy of DermatologyVolume 28, Number 2, Part 2 Autoantigens in pemphigoid vegetans 333

PT N

80­

MW BP1 PT BP2 N

A B1 2 3 4 5 6 7

kd kd

205- .- 230-+

160-117- 130--

95~

Pemphigoidvegetans

Tissue antigen BP antigen

Rat tongue + +Human

Conjunctiva + +Oral mucosa + +Salt-split skin + (Com- + (Epider-

bined) mal side)

Table I. Expression of pemphigoid vegetansantigen and bullous pemphigoid antigen in BMZof various epithelia

nocytes were prepared as previously described." Normalhuman fibroblasts were cultured after the epidermis hadbeen removed with trypsin in Dulbecco's modified eaglemedium (GIBCO-BRL, Grand Island, N.Y.) supple­mented with 10% fetal bovine serum. At confluency thecells were harvested with a high ionic detergent-contain­ing extraction buffer that consisted of 65 mrnol/L Tris­HCl (pH 6.8) (Sigma), 10 mmol/L sodium ethylenedi­aminetetraacetic acid (EM Science, Gibbstown, N.J.),2% sodium dodecyl sulfate (SDS) (Bio-Rad Laborato­ries, Richmond, Calif.), and 2 mmol/L phenylmethyl­sulfonyl fluoride (Sigma). Keratinocytes were also har­vested with a low ionic detergent-containing extractionbuffer that consisted of 0.2 mol/L Tris-HCl (pH 7.5),0.17% SDS, and 1.25% Nonidet P40 (Sigma). The de­tails of antigen preparation, SDS-polyacrylamide gelelectrophoresis, and electrophoretic transfer to nitro­cellulose paper were identical to that previously de­scribed.? After transfer the nitrocellulose paper was incu­bated with a blocking buffer.? reacted with a 1:50dilutionof serum, followed by reaction with a 1:2000 dilution ofperoxidase-conjugated goat anti-human IgG (Fc-spe­cific, Caltag Laboratories, Inc., South San Francisco,Calif.). The reaction was visualized with a standardizedsubstrate."

Results

To determine whether our patient had circulating IgGantibodies against the BMZ, indirect immunofluores­cence studies were performed on salt-split normal humanskin. With a 1:10 dilution of the patient's serum, our studydetected IgG anti-BMZ antibodies that reacted exclu­sively with the epidermal side of salt-split skin (Fig. 2),identical to the reaction pattern of a BP control serum. Ata 1:3 dilution the patient's serum reacted to the dermalside of salt-split skin. The patient's IgG4 subclass anti­bodies reacted strongly with skin BMZ, and IgG1 anti­bodies reacted weakly; IgG2 and IgG3 antibodies showedno reaction. Thus our patient's anti-BMZ antibodies were

Fig. 3. Western blots of keratinocyte antigens were re­acted with our patient's serum (PT), and the serum frompatients with bullous pemphigoid (BP) and from healthyindividuals (N).On high ionic detergent extracts (panelA),our patient's serum (lane 3) and the serum from two pa­tients with BP (lanes 2 and 4) react to a 230 kd major BPantigen (arrow). No keratinocyte antigen is detected bynormal serum (lane 5). Lane 1 represents molecularweight markers (MW). On low ionic detergent extracts(panel B\ our patient's serum (lane 6) reacts weakly tothree antigens of 230,160, and 130 kd (solid arrows), butreacts strongly to a 95 kd antigen (hollow arrow). A nor­mal serum shows negative reaction (lane 7).

predominantly IgG4 subclass. Our patient's IgG au­toantibodies and a BP control serum also reacted withBMZ of normal human oral mucosa, conjunctiva, and rattongue (Table I).

Immunoblot studies on high ionic detergent-extractedkeratinocyte antigens revealed that our patient's serumdetected a 230 kd major BP antigen, which was identicalto that detected by positive control serum from twopatients who had immunopathologically documented BP(Fig. 3, panel A). Serum from a healthy person showeda negative reaction.

Immunoblot studies on low ionic detergent-extractedkeratinocyte antigens showed that our patient's serumweakly reacted to three antigens of 230, 160, and 130 kd,but strongly reacted to a 95 kd keratinocyte antigen (Fig.3,B). This 95 kd antigen was not detected by serum fromfour patients whohadimmunopathologicallydocumentedHP, although serum from a fifth patient who had BPdemonstrated faint reactivity in this area of the blot. This95 kd protein was not detected by serum from patientswhohad epidermolysisbullosa acquisita (onepatient) andpemphigus vegetans (one patient), or by serum from twohealthy persons.

Immunoblot studies on fibroblast antigens revealedthat serum from our patient, from a patient who had BP,and from a healthy person did not react with any fibro­blast antigens. A special serum sample reacted with a 105kd fibroblast antigen as previously reported. 10

Journal of the American Academy of Dermatology334 Chan et al. February 1993

Table II. Summary of features in all reported cases of pemphigoid vegetans

Case 1 Case 2 Case 3 Case 4 Case 5

Reference 5 6 7 8 Present caseSex F F M M FOnset age (yr) 23 28 82 75 77Clinical findings

Intertriginous + + + + +vegetation

Vesicle/pustule +/+ +/+ -/- +/- +/+Mucosae

Histologic findingsEpidermal + + + + +

hyperplasiaSubepidermal + + + + +

blisterEosinophilic + + + + +

infiltrateAcantholysis

DIF* (BMZ IgG, M, C3, IgG, C3 IgG ± C3 IgG, C3 IgGdeposits) fibrin

IIFt (anti-BMZ + + + +Combined epidermalantibody) and dermal

IBM (antigen ND ND ND Upper LL NDlocalization)

Biochemical (antigen) ND ND ND ND 230 kd BP antigen;characterization) also 160, 130,95 kd

Treatments Sulfapyridine TCand TC TC Dapsone(effective) antibiotic

+,Positive; -, negative; DIF. direct immunofluorescence; IIF, indirect immunofluorescence; [EM, immunoelectron microscopy; LL, lamina lucida;T'C, topical corticosteroids; ND, not done.*DIF on patient's skin biopsy.tnF with patient's serum.

DISCUSSION

We describe an additional case of pemphigoidvegetans. Our patient had lesions similar to thosereported by othersS-8 (Table II). Her histologicfindings were consistent with those previously re­portedS-8 (Table II), as were the results of direct im­munofluorescence-f (Table II). Indirect immuno­fluorescence microscopic examination revealed thatanti-BMZantibodies were present in three of fourpreviously reported patients who had pemphigoidvegetans-" (Table II). The anti-BMZ antibodies inour patient specifically reacted with the epidermalside of salt-split skin, a finding consistent with thediagnosis of BP. 1 In addition, our patient's anti­BMZ antibodies reacted with the BMZ of normalhuman oral mucosa, conjunctiva, and rat tongueepithelium, a reaction identical to that of BP controlserum (Table I). The patient's anti-BMZ anti­bodies belonged predominantly to the IgG4 sub-

class, a relatively ineffective complement activator.This may account for the absence of C3 depositionin the patient's perilesional skin.

Immunoblot and immunoprecipitation studiesare the most definitivemethods in classifyingsubepi­dermal bullous dermatoses that are characterized bycirculating anti-BMZ antibodies. With the highionic detergent-extracted keratinocyte antigens as asource, our patient's serum detected a 230 kd majorBP antigen identical to that of serum from BP pa­tients. This finding provides definitive evidence thatpemphigoid vegetans is a BP variant.

With a fibroblast extract as a source, our patient'sserum did not detect any fibroblast antigen. Thisnegative finding suggests that the BMZ antigenrecognized by our patient's serum at the dermal sideof salt-split skin is not synthesized by fibroblasts.Our finding is comparable to that of a previousstudy, which reported that serum reacting with both

Journal of the American Academy of DermatologyVolume 28, Number 2, Part 2

the epidermal and dermal sides of salt-split skin de­tects BP antigens in epidermal extracts and detectsno antigens in dermal extracts.l!

With the low ionic detergent-extracted kerati­nocyte antigens as a source, our patient's serumstrongly reacted with a 95 kd keratinocyte antigen.This antigen was not detected by the serum fromfour patients who had BP, one patient who had epi­dermolysis bullosa acquisita, 1 patient who hadpemphigus vegetans, and 2 healthy persons and wasonly minimally detected by another BP serum. It isnot known whether this 95 kd protein is a uniqueantigen or a breakdown product of BP antigen. Thedetectability ofthis 95 kd antigen in a low ionic de­tergent-extracted source and not in a high ionic de­tergent-extracted source suggests that its antigenic­ity is better preserved by nonionic detergents.Because pemphigoid vegetans is rare and BP serumhas been reported to react with a 97 kd antigen inepidermal extract.I it remains to be determinedwhether this 95 kd antigen is specific for pemphig­oid vegetans.

REFERENCES

1. Woodley D, Sauder D, Talley MJ, et al. Localization ofbasement membrane components after dermal-epidermaljunction separation. J Invest DermatolI983;81:149-53.

Autoantigens in pemphigoid vegetans 335

2. Stanley JR, Hawley-Nelson P, Yuspa SH, et al. Charac­terization of bullous pemphigoid antigen: a unique base­ment membrane protein of stratified squamous epithelia.Cell 1981;24:897-903.

3. Labib RS, Anhalt GJ, Patel HP, et al. Molecular hetero­geneity of the bullous pemphigoid antigen as detected byimmunoblotting. J Immunol1986;136:1231-5.

4. Mueller S, Klaus-Kovtun V, Stanley JR. A 230-kd basicprotein is the major bullous pemphigoid antigen. J InvestDermatol 1989;92:33-8.

5. Winkelmann RK, Su WPD. Pemphigoid vegetans. ArchDermatol 1979;115:446-8.

6. Kuokkanen K, Helin H. Pemphigoid vegetans: report of acase. Arch Dermatol 1981;117:56-7.

7. AI-Najjar A,ReillyGD,BleehenSS.Pemphigoidvegetans:a casereport. ActaDermVenereal (Stockh) 1984;64:450-2.

8. Veda Y,NashiroK, SekiY, etal. Pemphigoidvegetans. BrJ Dermatol1989;120:449-53.

9. Chan LS, Hammerberg C, Cooper KD. Cicatricial pem­phigoid: identification of two distinct sets of epidermal an­tigens by IgA and IgG class circulating autoantibodies.Arch DermatolI990;126:1466-8.

10. Chan LS, Fine J-D, Briggaman RA, et al. Identification ofa novelautoimmune bullous dermatosis and a novel 105-kdbasement mem brane protein that is defective in dystrophicepidermolysis bullosa. J Invest DermatoI1991;96:556A.

11. Niimi Y, Zhu XJZ, Bystryn J-C. Identification of base­ment membrane zone antigen defined by antibodies thatreact to both the epidermal and dermal side of 1 M sodiumchloride split skin. J Invest DermatoI199l;97:312-7.