BP230- and BP180-Specific Auto-Antibodies

in Bullous Pemphigoid
Sybille Thoma-Uszynski, Wolfgang Uter,w Susanne Schwietzke, Silke C. Hofmann, Thomas Hunziker,z
Philippe Bernard,y Regina Treudler,z Christos C. Zouboulis,z Gerold Schuler, Luca Borradori,# and
Michael Hertl
Departments of Dermatology and wMedical Statistics, Biometry and Epidemiology, Friedrich-Alexander-University Erlangen-Nürnberg, Germany; zDepartment of
Dermatology, University Hospital, Bern, Switzerland; yDepartment of Dermatology, University of Reims, France; zDepartment of Dermatology, Campus Benjamin
Franklin, Charité University Medicine Berlin, Germany; #Department of Dermatology, University Hospital, Geneva, Switzerland

Bullous pemphigoid is a subepidermal blistering disease associated with auto-antibodies (auto-ab) to BP180 and
BP230. We developed ELISAs utilizing baculovirus-encoded recombinant proteins of BP230 and BP180 and studied
their diagnostic and prognostic values by assessing the profile of the auto-ab response in 127 patients with BP. 39
patients had focal involvement, whereas 88 had generalized disease; 51 individuals served as controls. The results
indicate: (1) BP180 IgG reactivity was associated with an overall sensitivity of 0.953 and specificity of 0.940; (2) 105
of 127 BP patients also displayed BP230 auto-reactivity, the global diagnostic performance of which, however, was
moderate compared to BP180-auto-reactivity (sensitivity 0.815 vs 0.953, specificity 0.648 vs 0.940); (3) 101 patients
(79.5%) had concordant BP180 and BP230 reactivity; (4) the association between the presence of BP230 auto-
reactivity and focal involvement was stronger than in generalized disease (odds ratio (OR) 17.7 vs 10.2),
independently from BP180 auto-ab profile; (5) correlation of total IgG with IgG1 and IgG4 was variable for both
BP180 and BP230. Collectively, the global diagnostic properties of the BP180-ELISA outperform those of the
BP230-ELISA. Presence of BP230 auto-reactivity, however, supports the diagnosis of BP and might be indicative
for the extent of the disease.
Key words: adhesion molecules/auto-immunity/BP180/BP230 auto-antibodies/bullous pemphigoid/hemidesmosome
J Invest Dermatol 122:1413 –1422, 2004

Bullous pemphigoid (BP) is an auto-immune blistering
disorder of the skin that usually affects the elderly. The
disease is associated with tissue-bound and circulating
auto-antibodies (auto-ab) directed against components of
the epidermal basement membrane zone (Jordon et al,
1967). Clinically, BP is characterized by generalized tense
subepidermal blisters arising on apparently normal or
erythematous skin and occasionally involvement of mucous
membranes (Lever, 1953). Previous studies have demon-
strated that BP auto-ab predominantly recognize BP180
(BPAG2 or type XVII collagen) and/or BP230 (BPAG1) (Labib
et al, 1986; Stanley et al, 1988; Giudice et al, 1992), two
structural components of hemidesmosomes, junctional
complexes promoting the adhesion of basal keratinocytes
Abbreviations: auto-ab, autoantibodies; BP, bullous pemphigoid;
BP180ex, baculoprotein containing the ECD of BP180 (residue
485–1497); BP230-N, baculoprotein containing the NH2-terminus
of BP230 (residue 1–1307); BP230-C1, baculoprotein containing
the COOH-terminus (residue 1881–2649) of BP230; BP230-C2,
baculoprotein containing the COOH-terminus (residue 2077–2649)
of BP230; ECD, extracellular domain; GST, glutathione-S-transfer-
ase; IIF, indirect immunofluorescence; OD, optical density; OR,
odds ratio; ROC, receiver operating characteristics; RT, room
temperature
to the underlying basement membrane zone (reviewed in
Borradori et al (1999)).
The cytoplasmic protein BP230, that was originally
identified as a target antigen of BP (Stanley et al, 1988) is
a cytoplasmic component of hemidesmosomes that be-
longs to the plakin family of cytolinkers (reviewed in Guo
et al (1995), Borradori et al (1999), Ruhrberg et al (1997)).
This protein is implicated in the linkage of the keratin
intermediate filament network to the basal cell side (Guo
et al, 1995; Ruhrberg et al, 1997; Borradori et al, 1999;
Koster et al, 2003).
The transmembrane protein BP180 has a large extra-
cellular domain (ECD) which consists of 15 interrupted
collagenous subdomains (Giudice et al, 1992; Hopkinson
et al, 1992). It serves as a cell-surface receptor, contributing
to the maintenance of dermo-epidermal cohesion by
binding most likely to the extracellular matrix component
laminin 5 (Giudice et al, 1992; Hopkinson et al, 1992;
Borradori et al, 1999; Nievers et al, 2000; Nishiyama et al,
2000). Furthermore, BP180 is probably involved in epider-
mal differentiation by shedding and facilitating detachment
of keratinocytes from the basal cell layer (Giudice et al,
1992; Hopkinson et al, 1992; Borradori et al, 1999; Nievers
et al, 2000; Nishiyama et al, 2000). Mutations of the BP180
gene (COL17A1) may cause a clinical variant of non-Herlitz

Copyright r 2004 by The Society for Investigative Dermatology, Inc.

1413
1414 THOMA-USZYNSKI ET AL
junctional epidermolysis bullosa, a heritable disorder char-
acterized by skin fragility and blistering (McGrath et al,
1995; Pulkkinen et al, 1998).
The commonly accepted idea that auto-ab to BP180 are
critical in subepidermal blister formation in BP patients has
been supported by a passive murine transfer model, in
which rabbit antibodies raised against the murine homo-
logue of the human immunodominant region of BP180, the
NC16A domain, were able to reproduce all key features of
BP (Liu et al, 1993). Furthermore, serum levels of IgG auto-
ab directed against the ECD of BP180 were directly related
to the severity of BP (Haase et al, 1998; Schmidt et al, 2000)
or distinct clinical phenotypes of BP (Hofmann et al, 2002).
Auto-ab from patients with BP and gestational pemphi-
goid, a clinical variant occurring during pregnancy, have
been found to be predominantly directed against distinct
antigenic regions of the ECD of BP180: Specifically, the
NC16A subdomain, a region adjacent to the membrane
spanning domain of BP180, contains major epitopes
recognized by most sera from patients with BP and ges-
tational pemphigoid (Giudice et al, 1993; Matsumura et al,
1996). Preadsorption of patients’ IgG with recombinant
NC16A protein leads to almost complete abolishment of
BP180-specific serum reactivity (Zillikens et al, 1997b; Hata
et al, 2000; Lin et al, 2000). The COOH-terminal region of the
ECD of BP180 has been shown to be recognized by
approximately one-third of BP sera (Murakami et al, 1998;
Nakatani et al, 1998; Hofmann et al, 2002) as well as by sera
from a subset of patients with cicatricial pemphigoid (CP), a
related auto-immune bullous disorder characterized by a
scarring phenotype of the mucous membranes (Balding
et al, 1996; Bedane et al, 1997; Murakami et al, 1998; Nie
et al, 1999).
In contrast, the cytoplasmic BP230 as target for auto-ab
has been the focus of only few studies (Ishiko et al, 1993;
Savage et al, 1999; Seishima et al, 1999a, b; Skaria et al,
2000). Preliminary studies by Husz et al suggest that anti-
BP230 antibodies directly contribute to tissue damage
(Husz et al, 2002).
The purpose of this study was to investigate the
diagnostic properties of auto-ab against BP180 and
BP230 in BP serum samples utilizing newly established
ELISA assays with baculovirus-encoded recombinant pro-
teins and the relationship between auto-ab profiles and
distinct clinical variants of BP.
Results
Production of baculovirus-encoded BP180- and BP230
recombinant proteins Using baculovirus expression, rela-
tively large BP180- and BP230 recombinant proteins were
successfully produced and purified. After purification,
BP180- and BP230-like recombinant proteins displayed
the expected size and were specifically immunoreactive
using appropriate monoclonal antibodies directed at the
tags (data not shown) and specific antisera (see Figs 1
and 2).
IgG reactivity of BP sera against the ECD of BP180,
BP180ex, is highly diagnostic for BP in all clinical
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
variants Sera from BP patients and controls were tested
against recombinant BP180ex which was immobilized onto
ELISA microtiter plates. The BP180ex recombinant protein
used for ELISA preparation presumably possessed a
conformation which allowed specific detection of BP180-
specific auto-ab and thus perhaps similar to the native ECD
of BP180 (Fig 3). This notion is supported by the fact that
BP180ex reactive sera were also reactive to BP180ex-virus-
infected insect cells, but not the appropriate controls, e.g.,
wild-type-virus/vector-infected insect cells did not react
with BP180-reactive serum and irrelevant control serum did
not react with BP180ex-infected insect cells (Fig 2C). Based
on the maximization of the Youden index, the cut-off point
for the ELISA with BP180ex protein was set at 0.207 OD
units which corresponds to a Youden index of 0.893 (Table
I). The diagnostic accuracy of this assay was high regarding
both sensitivity and specificity (Tables I and II). The
distribution of IgG levels (OD’s) in the two clinical variants
and the controls, respectively, is shown in Fig 3. The
difference between ‘‘generalized’’ versus ‘‘focal’’ involve-
ment was significant, indicating that higher levels of BP180-
specific IgG are correlated to the severity of the clinical
variant.
IgG reactivity of BP sera against BP230 Further char-
acterization of the BP sera was performed utilizing BP230
recombinant proteins. Similar to BP180ex, recombinant
BP230 proteins were recognized both, as native proteins
expressed by baculovirus-infected insect cells and after
purification under denaturing conditions in immobilized form
(Fig 2B, C). Again, the cutpoints for the ELISA with BP230-
C1, BP230-C2, and BP230-N proteins were determined by
maximization of the Youden indices and corresponding
ROC curves were generated (Fig 4, Table I).
Among clinical variants, BP230 reactivity is strongly
associated with focal BP The results of the polytomous
logistic regression analysis confirmed the highly significant
association between the diagnosis of BP (both clinical
variants) and positive BP180 reactivity. Independently from
this, IgG reactivity to BP230 was significantly associated
with the outcome, in particular with ‘‘focal BP’’ (Table III,
Fig 5). In the multifactorial analysis (polytomous logistic
regression) adjusted for BP 180 reactivity and age, IgG
reactivity to BP230 was significantly associated with the
outcome ‘‘focal BP’’ (OR 17.7, 95% CI: 2.7–116.0) and, to a
lesser extent, with the outcome ‘‘generalized BP’’ (OR 10.2,
95% CI: 1.7–60.7).
Diagnostic properties of BP180 and BP230 IgG reactivi-
ty A comparison between ‘‘positive’’ seroreactivity towards
BP180 and BP230, as defined above, is shown in Table II.
Discordance is fairly symmetric (p ¼ 0.64, exact McNemar
test). Concordance between the two diagnostic classifica-
tions as quantified with Cohen’s kappa coefficient is
moderate: k ¼ 0.45 (95% CI: 0.31–0.59). Diagnostic proper-
ties of ‘‘BP230 seropositivity’’; regarding both sensitivity and
specificity were less favorable than BP180 IgG reactivity
(Table II). Table III describes the joint distribution of BP180
and BP230 reactivity within the clinical variants.
122 : 6 JUNE 2004 AUTO-ANTIBODY PROFILES AND BULLOUS PEMPHIGOID 1415

Figure 1
Bullous pemphigoid (BP)230 recombinant proteins utilized in this ELISA study. (A) BP230 is an intracellular hemidesmosomal component which
consists of a central rod flanked by globular end domains. (B) The recombinant BP230 fusion proteins are 6xHis- and glutathione-S-transferase
(GST) tagged at their NH2 terminus. Two COOH-terminal constructs (BP230-C1, containing residues 1881–2649 and BP230-C2, residues 2077–
2649) and one NH2-terminal construct (BP230-N, residues 1–1307) were expressed. The purified BP230 recombinant proteins were transferred to
nitrocellulose membranes and Coomassie stained. (C) SF21 insect cells infected with recombinant BP230 baculovirus for 3 d were recognized by a
representative BP serum.

Figure 2
Bullous pemphigoid (BP)180 recombinant protein utilized in this ELISA study. (A) BP180 is a type II transmembrane protein containing an
intracellular, a transmembrane, and an extracellular domain with 15 interrupted collagenous regions. (B) The recombinant fusion protein BP180ex is
6xHis- and glutathione-S-transferase (GST)-tagged at its NH2 end (residues 485–1497). The purified BP180ex protein was transferred to
nitrocellulose membranes and Coomassie stained. (C) SF21 insect cells infected with BP180ex baculovirus for 3 d were recognized by a
representative BP serum.

BP230 domains and clinical variants BP230 reactive BP
sera recognize epitopes within the COOH- and NH2-
terminal regions (BP230-C1, BP230-C2, and BP230-N)
without significant differences between the two clinical
variants (after adjustment for multiple testing) (Fig 5).
Relationship of IgG subtypes within clinical variants
against BP230 and BP180 In addition to total BP230 and
BP180 IgG, the major auto-ab subtypes (IgG1 and IgG4)
were analyzed. Thus, it appeared that in active BP, IgG4 was
the predominant IgG subtype reactive with the COOH-
1416 THOMA-USZYNSKI ET AL
terminus, but not the NH2-terminus of BP230 (Fig 6A–C).
Regarding the IgG subtype reactivity against BP180
(BP180ex), IgG1 and IgG4 were nearly equally distributed
(Fig 6D).
Correlation between COOH- and NH2-terminal con-
structs of BP IgG reactivity to the overlapping COOH-
terminal BP230 constructs, BP230-C1 and BP230-C2,
revealed a high correlation (0.94, po0.0001 unadjusted,
BP230-C1 vs BP230-C2). When IgG reactivity against NH2-
terminal versus COOH-terminal BP230-like recombinant
proteins was plotted, correlation was only moderate
(r ¼ 0.43, po0.0001, unadjusted). This indicates that re-
levant epitopes are distributed over both globular domains
of BP230; however, relevant BP230 epitopes of the COOH-
terminus are harbored in BP230-C2 (residues 2077–2649).
Discussion
The aim of this study was to analyze the relationship
between IgG auto-ab directed against the intracellular
BP230 and the transmembranous BP180 proteins, the
major auto-antigens of BP, as diagnostic markers in BP.
For this purpose, we developed novel ELISA assays utilizing
four recombinant eukaryotic proteins encompassing im-
munodominant regions of BP230 (BP230-C1, BP230-C2,
BP230-N) and BP180 (BP180ex), respectively.
The results demonstrate a highly diagnostic BP180
ELISA utilizing the entire ECD of BP180 which detected
BP180 as a specific auto-ab in the majority of the patients’
sera. Throughout both clinical variants (focal vs generalized
BP), BP180 reactivity was highly associated with the
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
Figure 3
IgG reactivity against the extracellular
domain of bullous pemphigoid (BP)180ex
is highly indicative for BP. BP patient
and control serum samples were tested
by ELISA using recombinant BP180ex.
Sera of patients with BP (n ¼ 127) and of
51 controls were incubated with 0.5 mg of
immobilized BP180ex protein. IgG binding
was detected with an alkaline-phospha-
tase-labeled anti-human IgG ab. Samples
were run in duplicate and mean outer
diameter reading at 405 nm of a serum
sample with BP230 recombinants. Data
are displayed as box plots: The box
shows first and third quartiles, the median
(as line), and the arithmetic mean (as dot),
the whiskers represent values below the
first quartile and above the third quartile
within the 1.5-fold inter-quartile range,
respectively, and outliers beyond the
whiskers are shown as squares. The
dotted lines indicate the cut-off values
as evaluated by maximization of Youden
index. The inset visualizes the corre-
sponding receiver operating characteristic
curve. There was a highly significant
difference between clinical variants and
controls as well as between ‘‘generalized’’
and ‘‘focal’’ BP.
diagnosis of BP. In contrast, BP230-specific IgG was also
detected in the majority of the BP sera, although less
frequently than BP180-specific IgG and was associated
with a less severe clinical variant of BP.
At present, it is unclear whether BP230 auto-ab reactivity
is part of the disease initiation and directly contributes to
blister formation or whether these auto-ab represent an
epiphenomenon. A serological approach towards elucida-
tion of the potential functional role of BP230 auto-ab was
the thorough analysis of auto-ab profiles in BP patients.
Although in terms of diagnostic power, BP230-specific IgG
was outperformed by anti-BP180 IgG, the majority of our
patients also showed BP230 reactivity, particularly those
with focal BP. Thus, valuable diagnostic information was
obtained by the present novel BP230 ELISA in addition to
the already established BP180 ELISA.
ELISA systems detecting auto-ab specific for defined
regions of the ECD of BP180 utilized mainly prokaryotic
BP180-like recombinant proteins (Giudice et al, 1994;
Zillikens et al, 1997a; Nakatani et al, 1998). Zillikens et al
detected IgG reactivity against the NC16A domain of BP180
in 94% of 50 BP sera (Zillikens et al, 1997a). Another ELISA
utilizing bacterial recombinant proteins demonstrated the
presence of IgG auto-ab against the NC16A subdomain and
the COOH-extremity of BP180 in 96% and 38% of the BP
sera, respectively (Nakatani et al, 1998). A recent study from
our laboratory with eukaryotic BP180-like recombinant
proteins demonstrated that auto-ab against the NH2-
terminus of the BP180 ectodomain were associated with a
more extensive phenotype of BP while auto-ab against the
COOH-terminus of BP180 were not (Hofmann et al, 2002).
Consistent with previous passive transfer studies with
rabbit sera reactive with the murine homologue of the
122 : 6 JUNE 2004 AUTO-ANTIBODY PROFILES AND BULLOUS PEMPHIGOID 1417

Table I. Cutpoints (outer diameter (OD) values with maximum corresponding Youden Index (YI)a with sensitivity and specificity,
including exact 95% confidence intervals (CIs))

Cutpoint Yia Sensitivity Specificity

BP230-C1 Total IgG 40.229 0.575 0.654 0.922
(0.564–0.736) (0.811–0.978)
IgG1 40.131 0.769 0.984 0.784
(0.944–0.998) (0.647–0.887)
IgG4 40.104 0.765 1.00 0.765
(0.977–1.00) (0.625–0.872)
BP230-C2 Total IgG 40.197 0.469 0.685 0. 784
(0.597–0.765) (0.647–0.887)
IgG1 40.126 0.796 0.992 0.804
(0.957–0.999) (0.669–0.902)
IgG4 40.173 0.831 0.890 0.941
(0.822–0.938) (0.838–0.988)
BP230-N Total IgG 40.170 0.568 0.724 0.843
(0.638–0.800) (0.714–0.930)
IgG1 40.106 0.757 0.992 0.765
(0.957–0.999) (0.625–0.872)
IgG4 40.089 0.737 0.992 0.745
(0.957–0.999) (0.604–0.857)
BP180ex Total IgG 40.207 0.893 0.953 0.940
(0.900–0.983) (0.835–0.988)
IgG1 40.122 0.838 0.898 0.940
(0.831–0.944) (0.835–0.988)
IgG4 40.081 0.760 1.00 0.760
(0.977–1.00) (0.618–0.869)
a
BP, bullous pemphigoid; Youden index (sensitivity þ specificity
1).

Table II. Concordance between IgG reactivity against bullous Hofmann et al, 2002). In line with the latter, here we also
pemphigoid (BP)180 and BP230, and global diagnostic found a significant difference of BP180 auto-ab levels
properties of BP230 IgG reactivity between generalized and focal involvement (Hofmann et al,
2002).
BP230a
A recent serological study showed that auto-ab against
Pos. Neg. Total BP230 were mainly directed against the B -the linker, and C
b subdomains of the COOH-terminus of BP230 (Skaria et al,
BP180 Pos. 101 23 124
2000). This analysis, which is an extensive serological study
Neg. 19 35 54 utilizing an ELISA system with baculovirus-derived eukar-
Pos. Neg. yotic BP230 recombinant proteins, confirms these findings
showing that both, NH2- and COOH-terminus of BP230 are
BP Yes 105 22 127 major immunodominant regions targeted by IgG auto-ab in
No 15 36 51 BP. Several observations support the hypothesis that
a
BP230-specific auto-ab are relevant for the pathogenesis
‘‘Pos.’’: either BP230-C1-, BP230-C2- or BP230-N-reactive IgG above
cutpoint. of BP and not just an epiphenomenon: (1) Immunization of
b
Using BP180 as an alternative gold standard’’, i.e. case definition rabbits with BP230-derived peptides appeared to increase
criterion:
Sensitivity of BP230: 0.815 (95% CI: 0.735–0.879);
the local inflammatory response following epidermal da-
Specificity of BP230: 0.648 (95% CI: 0.506–0.773). mage (Hall et al, 1993); (2) BP230-specific auto-ab were
readily detectable in the majority of patients even even after
a relatively short disease duration (1.5 mo); (3) preliminary
NC16A domain (Liu et al, 1993), our findings further support results by Husz et al, indicate that BP230 is relevant for
the notion that auto-ab against the ECD of BP180 are pathogenesis of BP: Immunization of rabbits with human
pathogenic. Recent studies from our group and others with BP230 fragments yielded high-titer BP230-specific poly-
prokaryotic and eukaryotic BP180 recombinant proteins clonal rabbit serum which, after subcutaneous injection into
also suggested a possible relationship between disease neonatal mice induced an inflammatory response and skin
activity and auto-ab levels against the BP180 ectodomain fragility. Immunohistological analysis revealed IgG deposi-
(Haase et al, 1998; Dopp et al, 2000; Schmidt et al, 2000; tion along the basement membrane zone, few animals even
1418 THOMA-USZYNSKI ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 4
Cutpoint determination by receiver operating characteristic (ROC) curves for bullous pemphigoid (BP)230-reactive total IgG. Detection of
serum IgG against the NH2-terminus (BP230-N) and COOH-terminus (BP230-C1 and BP230-C2) of BP230 is shown as ROC curves. Sera of patients
with BP (n ¼ 127) and 51 controls were incubated with 0.5 mg of immobilized BP230-N, BP230-C-1, BP230-C-2 proteins. IgG binding was detected
with an alkaline phosphatase (AP)-labeled anti-human IgG ab. Samples were run in duplicate and mean OD reading at 405 nm of a serum sam-
ple with BP230-like recombinant proteins. The solid lines indicate the cut-off values as evaluated by maximization of Youden index.

Table III. Comparison of IgG reactivity against BP180 and/or
BP230 and clinical variant of BP
Clinical variant of BP
Auto-ab reactivity ‘‘Generalized’’ ‘‘Focal’’ Controls
BP 180 only 16 5 2
BP 230 only 2 3 14
Both 69 31 1 specificity for distinct BP230 regions, i.e., NH2-terminus
Neither 1 0 34
Auto-ab, auto-antibodies; BP, bullous pemphigoid.
developed subcutaneous blisters (Husz et al, 2002). In
addition to participating to the inflammatory reaction, auto-
ab to BP230 may interfere with the function of BP230 and
its molecular interactions with the other components of the
hemidesmosomal adhesion complex. Further evidence that
functional BP230 is essential for skin stability comes from
BP230-null mutant mice which exhibit an increased;
however, intraepidermal skin fragility due to the lack of the
inner hemidesmosomal plate (Guo et al, 1995).
Since the various IgG subclasses have distinct functional
properties and their secretion is differentially regulated, we
have also assessed the relative isotype distribution among
the major IgG subclasses specific for both BP180 and
BP230, most studies reported a preferential detection of
IgG1 and IgG4 auto-ab against BP180, while BP180-specific
IgG2 and IgG3 auto-ab were detected in few BP serum
samples only (Bernard et al, 1990; Dopp et al, 2000; Laffitte
et al, 2001). In this study, BP180-reactive IgG1 and IgG4
auto-ab were almost equally distributed in the studied
patient cohort. A recent study from our laboratory, however,
indicated that IgG1 auto-ab against the immunodominant
NH2-terminal ECD of BP180 were predominant in acute
onset BP, while IgG4 auto-ab predominated in remittent BP
(Hofmann et al, 2002). Noteworthy, a dual IgG1 and IgG4
response against the NH2-terminus of the ECD of BP180
was associated with a more severe skin involvement in BP
(Hofmann et al, 2002). Two additional previous studies
reported a predominance of BP180-specific IgG1 (Bernard
et al, 1990; Laffitte et al, 2001), while another study
detected more BP180-specific IgG4 than IgG1 (Dopp et al,
2000).
In this study, IgG reactivity against BP230 was mediated
both by IgG1 and IgG4 auto-ab. There was no significant
difference of the IgG1 and IgG4 titers with regard to their
(BP230-N) or COOH-terminus (BP230-C1/C2). Noteworthy,
IgG4 but not IgG1 auto-ab against the C-terminus of BP230
were particularly elevated in the sera of BP patients with
focal BP. In a previous study, auto-ab binding to the
immunodominant COOH-terminal region of BP230 predo-
minantly belonged to the IgG4 and IgG1 subclasses (Skaria
et al, 2000) while a serological study by Bernard et al
showed also a predominant IgG4 restriction of anti-BP230
auto-ab (Bernard et al, 1990).
The concept that, in addition to IgG4, complement-fixing
IgG1 plays a key role in the initiation of BP is supported by
the clinical finding that, in addition to IgG, perilesional C3 is
in most cases detectable at the basement membrane in BP
patients (Jordon et al, 1967). In summary, the present novel
ELISA assays utilizing eukaryotic recombinant proteins of
the NH2- and COOH-terminal regions of BP230 and the
entire BP180 ectodomain are highly useful for rapid and
sensitive analysis of the auto-ab profile, which reflects
disease activity of BP. In this investigation, auto-ab against
the BP180 ectodomain were highly diagnostic for BP and
their titers were correlated with the severity of BP. Auto-ab
against BP230 were also found in the majority of the BP
serum samples, and thus are of confirmatory diagnostic
value. Noteworthy, the presence of anti-BP230 IgG auto-ab
was particularly associated with a less severe clinical
manifestation of BP and may thus be of potential prognostic
value.
Materials and Methods
Patients and control sera Serum samples were obtained from
patients with clinically active BP (n ¼ 127), and control donors
122 : 6 JUNE 2004
Figure 5
Bullous pemphigoid (BP)230 IgG levels in clinical
variants. Total IgG responses against the BP230 recom-
binant proteins were compared between serum samples
from patients with ‘‘generalized’’ versus ‘‘focal’’ BP. Box
plot presentation depicting median, arithmetic mean, and
range from first to third quartile as in Fig 3. BP230-
specific IgG reactivity against all three BP230 recombi-
nants in both generalized and focal BP. Overall, higher,
significantly different IgG levels were observed in
localized BP. Thus, BP230 IgG-reactivity was significantly
associated with the outcome ‘‘focal’’ BP.
which included healthy volunteers and patients with unrelated
conditions (n ¼ 51). Patients and volunteers gave written consent
to participate in this study which was adherent to the Declaration
of Helsinki Guidelines and which has been approved by the
institutional review board. Serum samples were stored at
201C
prior to analysis.
Diagnosis of BP was based on the following criteria: (i) typical
clinical presentation with tense cutaneous blisters, (ii) histopatho-
logical evidence of subepidermal blister formation, (iii) linear IgG
and/or C3 deposits at the dermo-epidermal junction of perilesional
skin by direct immunofluorescence.
Severity of BP was classified according to the extent and
number of blisters as either focal disease (only few bullae at limited
areas of the body involving maximum 20% of total body surface) or
generalized disease with total body surface involvement greater
than 20% and wide spread innumerable bullae.
Patient characteristics This analysis includes 127 patients with
BP (39 with focal involvement and 88 with generalized disease) and
51 age- and sex-matched controls with unrelated skin conditions
such as herpes zoster, erysipelas or atopic dermatitis or allergic
conditions. Among the patients, there were 51.1% females, with no
significant difference between the two clinical groups (p ¼ 0.85,
Fisher’s exact test). Median age was 80 y (range: 30–101 y), with no
significant differences between the clinical variants (p ¼ 0.45,
Wilcoxon–Mann–Whitney test) (Table IV). Furthermore, the duration
of disease (overall median: 1.5 mo) did not differ significantly
between ‘‘generalized’’ and ‘‘focal’’ cases (p ¼ 0.99, Wilcoxon–
Mann–Whitney test) (Table IV). Oral mucosa involvement was
observed in three patients with focal and ten patients with
generalized disease, none of which displayed a scaring phenotype
or ocular involvement.
The majority of patients was untreated when the serum samples
were obtained. The few patients under therapy were receiving
various regimens of immunosuppressive drugs and/or local
treatment. They did not preferentially belong to either one of the
defined clinical categories.
Expression and purification of BP180 and BP230 recombinant
proteins The recombinant BP proteins were produced as glu-
tathione-S-transferase (GST) and 6xHis tag tagged fusion proteins
AUTO-ANTIBODY PROFILES AND BULLOUS PEMPHIGOID 1419
in a baculovirus expression system as described below (Fig 1).
cDNA expression constructs for human BP180 and BP230 were
generated as previously reported (Borradori et al, 1997; Skaria
et al, 2000). In brief, cDNAs for BP180 or BP230 were used as
a template to amplify by Pfu DNA polymerase with forward and
reverse primers containing either an EcoR I or a Not I site
the desired target sequence (Stratagene, La Jolla, California). The
amplified cDNA fragments were subsequently subcloned into the
pAcGHLT-A baculovirus transfer vector (Pharmingen, San Diego,
California) by ligation of double-digested PCR fragments into the
correspondingly cut vectors. Correctness of both constructs was
verified by sequencing. pAcGHLT-A vectors were cotransfected
with Baculo-DNA (SixPack, Clontech, Palo Alto, California) into
Sf21 insect cells and recombinant baculoviruses were amplified
using a previously described protocol (Amagai et al, 1994). The
recombinant proteins and the GST/6xHis control protein (gener-
ated by cotransfection of an empty pAcGHLT-A vector) were
expressed in Sf21 cells as previously described (Haase et al, 1998).
BP230- and BP180 recombinant proteins were purified as follows:
3 d after infection, 10  106 baculovirus-infected insect cells were
suspended in 1 mL of 6 M guanidine-hydrochloride, incubated for
24 h at
201C and then centrifuged (2000  g, 15 min). Solubilized
proteins were purified by affinity-chromatography over Nickel–NTA
agarose (Quiagen, Hilden, Germany) using the batch method
according to the manufacturer’s protocol. The BP230 and BP180
recombinant proteins and the GST/6xHis control protein were
purified under denaturing conditions by adding 6 M guanidine-
hydrochloride to the lysis buffer and 8 M urea to washing and
elution buffers. Finally, purified proteins were gradually dialyzed
against PBS supplemented with 8–3 M urea at 41C for 48 h
(BP180ex, GST/6xHis). Affinity-purified proteins were stored in
aliquots at –801C. Protein concentrations were determined by a
commercial kit (DC Protein Assay, Bio-Rad, Munich, Germany)
according to a modified protocol by Lowry. BP180ex, BP230-N,
and BP230-C1/C2 were fractionated by 10% SDS-Page and
visualized by Coomassie staining and western blot with a
monoclonal mouse anti-GST antibody (1:2000, clone GST3-4C,
Zymed, San Francisco, California) to verify molecular weight and
immunoreactivity (Figs 1B and 2B).
ELISA analysis utilizing BP230 and BP180 recombinant
proteins Optimal conditions for the ELISA were established by
1420 THOMA-USZYNSKI ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 6
(A—D) IgG subclass analysis (IgG1 and IgG4) of bullous pemphigoid (BP)230 and BP180 reactive auto-antibodies (auto-ab). IgG1 and IgG4
auto-ab against BP230 (A–C) and BP180 (D) were compared between the sera from patients with ‘‘generalized’’ versus ‘‘focal BP’’. Box plot
presentation depicting median, arithmetic mean, and range from first to third quartiles as in Fig 3. There was neither a significant correlation for IgG1
or IgG4 reactivity against BP230 (BP230-N; BP230-C1/C2) nor to BP180 (BP180ex) with the outcome ‘‘generalized’’ versus ‘‘focal’’ BP.

various chessboard titrations. The affinity-purified recombinant
proteins of BP230 and BP180 were immobilized on 96-well
polystyrene plates (Maxisorb Immunoplate, Nunc, Wiesbaden,
Germany) by coating each well with 0.5 mg of BP230-N, BP230-C1,
BP230-C2, and BP180ex, or a molar equivalent amount of the
control protein GST/6xHis in 100 ml of 0.1 M bicarbonate buffer, pH
9.8, at 41C for 5 h. ELISA plates were then washed 5  with TTBS
(TBS pH 7.5 with 0.05% Tween 20; Merck, Nürnberg, Germany)
and blocked for 2 h with 100 ml of 5% skimmed milk powder (MP) in
TTBS at room temperature (RT). BP and control sera were diluted
in blocking-buffer 50-fold and added to the wells in duplicate. After
an incubation period of 12 h at 41C and consecutive washes,
reactions with alkaline phosphatase (AP)-labeled goat anti-human
IgG (g) (1:6000, Kirkegaard and Perry, Gaithersburg, Maryland),
monoclonal biotinylated mouse anti-human IgG1 (1:800, clone
HP6069) or monoclonal biotinylated mouse anti-human IgG4
(1:500, clone HP6025) (both Zymed, San Francisco, California),
were performed in the wells 1–2 h. After incubation of biotinylated
antibodies for 1 h, wells were washed as above and incubated with
streptavidin-AP (ZyMAX, Zymed) diluted at 1:2000 in TBS for 1 h at
RT. After another series of washes, 100 ml of p-nitrophenylpho-
sphate (Kirkegaard and Perry) or the peroxidase substrate 2,20 -
azino-bis[3-ethylbenzthiazoline-6-sulfonic acid] (Sigma, Deisenho-
fen, Germany) were added to the wells and optical density (OD)
was measured at 405 nm when a defined patient serum (positive
reference control) reached an OD of 1.0 as determined by a Wallac
1420 microplate reader (Wallac, Freiburg, Germany).
All extinction data were standardized based on the positive and
negative internal standards, which were set to 1 and 0 OD U,
respectively. The background reactivity of IgG to the control protein
GST/6xHis was subtracted from the mean OD values obtained with
BP180-N and BP180-C proteins. For detection of IgG1 and IgG4
reactivity to BP180 and BP230, the GST/6xHis control protein was
not routinely used, because a relevant total IgG reactivity against
GST/6xHis had not been detected in any of the serum samples by
ELISA. To evaluate plate-to-plate variability, each plate included
identical internal controls (two positive serum samples, one control
serum).
122 : 6 JUNE 2004
Table IV. Clinical and immunopathological characteristics of
the studied BP patients
Patient description Generalized Focal Total
N 88 39 127
Age (y, median) 80 78 80
Disease duration 1.5 1.25 1.5
(mo, median and range) (0.25–60) (0.25–123) (0.25–123)
IIF (SSS)a
Epidermal 86 38 124
Dermal þ epidermal 1 — 1 Hartmannstrae 14, D-91052 Erlangen, Germany. Email: sybille.
Negative 1 1 2 thoma-uszynski@derma.imed.uni-erlangen.de
a
IIF, indirect immunofluorescence; SSS, saline split skin.
2 of 127 false-negative against case definition; sensitivity: 0.984 (95%
CI: 0.944–0.998).
Statistical analysis In view of the skewed distribution of OD
values, which was scarcely amenable to log-transformation, OD
values were analyzed with non-parametric statistical methods, e.g.,
Spearman rank correlation coefficient and Wilcoxon–Mann–Whitney
or Kruskal–Wallis test for overall equality of mean scores. p-values
were adjusted for multiple testing with the Bonferroni method,
values less than 0.05 were considered significant. The distribution of
data is shown as ‘‘box plot’’: The box shows first and third quartiles,
the median (as line) and the arithmetic mean (as dot), the whiskers
represent values below the first quartile and above the third quartile
within the 1.5-fold inter-quartile range, respectively, and outliers
beyond the whiskers are shown as squares. For data analysis, the
statistical software package SAS, version 8.2, was used (SAS
Institute, Cary, North Carolina). Receiver operating characteristic
(ROC) curves were compiled by additionally utilizing a freeware SAS
macro package (Witte, 2002) to evaluate the diagnostic properties of
BP230 and BP180 IgG antibodies with respect to correctly
classifying cases of BP. As there are no ‘‘natural’’ cutpoints for
these antibody levels, and as there are no previous studies which
had provided evidence concerning an appropriate diagnostic
threshold, an ‘‘optimum’’ cutpoint was determined data adaptively
by choosing the OD values with the largest Youden-Index
(YI ¼ sensitivity þ specificity
1).
For those analyses requiring dichotomization of ELISA results
as positive and negative, BP180 seroreactivity was considered
positive, if the OD value for total against BP180ex was above the
cutpoint (0.207 U) and negative otherwise. BP230 seroreactivity
was regarded as positive, if OD values for either total IgG against
BP230-C1, BP230-C2, or BP230-N were above the cutpoint (Table
I). To assess the independent association between positive BP230
seroreactivity (i.e., any single subclass above cutpoint) and the
diagnosis BP, adjusting for BP180 reactivity as well-established
factor, and age as potential confounder, a polytomous logistic
regression analysis, with the two clinical variants of the disease
versus control as outcomes, was performed. The degree of
association between BP230 seroreactivity on one hand, and
localized and generalized involvement, respectively, on the other
hand, was quantified with the odds ratio (OR), accompanied by the
corresponding 95% confidence interval (CI). Unique non-proprie-
tary reagents described will be made freely available to qualified
scientists.
We are grateful to Dr Giovanna Zambruno, Dr Peter von den Driesch,
and Dr Nicolas Hunzelmann for providing sera of BP patients. This
work was supported by grants from the Deutsche Forschungsge-
meinschaft (He1602/6-1 to M.H.), European Community (QLG1-CT-
AUTO-ANTIBODY PROFILES AND BULLOUS PEMPHIGOID 1421
2001-02007 to L.B. and M.H.), the Interdisciplinary Center for Clinical
Research (IZKF; C9 to M.H.), the ELAN Fonds (0203291 to S.T.U.) at
the University of Erlangen, Germany and the Swiss National Founda-
tion of Scientific Research (32-56727.99 to L.B.). The authors are also
indebted to Dr J.R. Stanley (University of Pennsylvania, Phildelphia)
and Dr J. Uitto, (Jefferson Medical College, Philadelphia), who
generously provided cDNAs for human BP230.
DOI: 10.1111/j.0022-202X.2004.22603.x
Manuscript received August 27, 2003; revised January 12, 2004;
accepted for publication January 26, 2004
Address correspondence to: Sybille Thoma-Uszynski, Department
of Dermatology, Friedrich-Alexander-University of Erlangen-Nürnberg,
References
Amagai M, Hashimoto T, Shimizu N, et al: Absorption of pathogenic auto-
antibodies by the extracellular domain of pemphigus vulgaris antigen
(Dsg3) produced by baculovirus. J Clin Invest 94:59–67, 1994
Balding SD, Prost C, Diaz LA, et al: Cicatricial pemphigoid autoantibodies react
with multiple sites on the BP180 extracellular domain. J Invest Dermatol
106:141–146, 1996
Bedane C, McMillan JR, Balding SD, et al: Bullous pemphigoid and cicatricial
pemphigoid autoantibodies react with ultrastructurally separable epi-
topes on the BP180 ectodomain: Evidence that BP180 spans the lamina
lucida. J Invest Dermatol 108:901–907, 1997
Bernard P, Aucouturier P, Denis F, et al: Immunoblot analysis of IgG subclasses of
circulating antibodies in bullous pemphigoid. Clin Immunol Immunopathol
54:484–494, 1990
Borradori L, Koch PJ, Niessen CM, et al: The localization of bullous pemphigoid
antigen 180 (BP180) in hemidesmosomes is mediated by its cytoplasmic
domain and seems to be regulated by the beta4 integrin subunit. J Cell
Biol 136:1333–1347, 1997
Borradori L, Sonnenberg A: Structure and function of hemidesmosomes:
More than simple adhesion complexes. J Invest Dermatol 112:411–418,
1999
Dopp R, Schmidt E, Chimanovitch I, et al: IgG4 and IgE are the major
immunoglobulins targeting the NC16A domain of BP180 in Bullous
pemphigoid: Serum levels of these immunoglobulins reflect disease
activity. J Am Acad Dermatol 42:577–583, 2000
Giudice GJ, Emery DJ, Diaz LA: Cloning and primary structural analysis of the
bullous pemphigoid autoantigen BP180. J Invest Dermatol 99:243–250,
1992
Giudice GJ, Emery DJ, Zelickson BD, et al: Bullous pemphigoid and herpes
gestationis autoantibodies recognize a common non-collagenous site on
the BP180 ectodomain. J Immunol 151:5742–5750, 1993
Giudice GJ, Wilske KC, Anhalt GJ, et al: Development of an ELISA to detect anti-
BP180 autoantibodies in bullous pemphigoid and herpes gestationis. J
Invest Dermatol 102:878–881, 1994
Guo L, Degenstein L, Dowling J, et al: Gene targeting of BPAG1: Abnormalities
in mechanical strength and cell migration in stratified epithelia and
neurologic degeneration. Cell 81:233–243, 1995
Haase C, Budinger L, Borradori L, et al: Detection of IgG autoantibodies in the
sera of patients with bullous and gestational pemphigoid: ELISA studies
utilizing a baculovirus- encoded form of bullous pemphigoid antigen 2.
J Invest Dermatol 110:282–286, 1998
Hall RP III, Murray JC, McCord MM, et al: Rabbits immunized with a peptide
encoded for by the 230-kD bullous pemphigoid antigen cDNA develop an
enhanced inflammatory response to UVB irradiation: A potential animal
model for bullous pemphigoid. J Invest Dermatol 101:9–14, 1993
Hata Y, Fujii Y, Tsunoda K, et al: Production of the entire extracellular domain of
BP180 (type XVII collagen) by baculovirus expression. J Dermatol Sci
23:183–190, 2000
Hofmann S, Thoma-Uszynski S, Hunziker T, et al: Severity and phenotype of
bullous pemphigoid relate to autoantibody profile against the NH2- and
COOH-terminal regions of the BP180 ectodomain. J Invest Dermatol
119:1065–1073, 2002
Hopkinson SB, Riddelle KS, Jones JC: Cytoplasmic domain of the 180-kD
bullous pemphigoid antigen, a hemidesmosomal component: Molecular
and cell biologic characterization. J Invest Dermatol 99:264–270, 1992
Husz S, Kiss M, Janossy T, et al: Experimental mouse model of bullous
pemphigoid using antibodies generated against an antigenic epitope of
1422 THOMA-USZYNSKI ET AL
hemidesmosomal plaque protein BP230. J Invest Dermatol 119:724,
2002 (abstract)
Ishiko A, Shimizu H, Kikuchi A, et al: Human autoantibodies against the 230-kD
bullous pemphigoid antigen (BPAG1) bind only to the intracellular domain
of the hemidesmosome, whereas those against the 180-kD bullous
pemphigoid antigen (BPAG2) bind along the plasma membrane of the
hemidesmosome in normal human and swine skin. J Clin Invest 91:
1608–1615, 1993
Jordon RE, Beutner EH, Witebsky E, et al: Basement zone antibodies in bullous
pemphigoid. JAMA 200:751–756, 1967
Koster J, Geerts D, Favre B, et al: Analysis of the interactions between BP180,
BP230, plectin and the integrin alpha6beta4 important for hemidesmo-
some assembly. J Cell Sci 116:387–399, 2003
Labib RS, Anhalt GJ, Patel HP, et al: Molecular heterogeneity of the bullous
pemphigoid antigens as detected by immunoblotting. J Immunol 136:
1231–1235, 1986
Laffitte E, Skaria M, Jaunin F, et al: Autoantibodies to the extracellular and
intracellular domain of bullous pemphigoid 180, the putative key
autoantigen in bullous pemphigoid, belong predominantly to the IgG1
and IgG4 subclasses. Br J Dermatol 144:760–768, 2001
Lever WF: Pemphigus. Medicine 32:1–123, 1953
Lin MS, Fu CL, Giudice GJ, et al: Epitopes targeted by bullous pemphigoid T
lymphocytes and autoantibodies map to the same sites on the bullous
pemphigoid 180 ectodomain. J Invest Dermatol 115:955–961, 2000
Liu Z, Diaz LA, Troy JL, et al: A passive transfer model of the organ-specific
autoimmune disease, bullous pemphigoid, using antibodies generated
against the hemidesmosomal antigen, BP180. J Clin Invest 92:2480–2488,
1993
Matsumura K, Amagai M, Nishikawa T, et al: The majority of bullous pemphigoid
and herpes gestationis serum samples react with the NC16a domain of the
180-kDa bullous pemphigoid antigen. Arch Dermatol Res 288:507–509,
1996
McGrath JA, Gatalica B, Christiano AM, et al: Mutations in the 180-kD bullous
pemphigoid antigen (BPAG2), a hemidesmosomal transmembrane col-
lagen (COL17A1), in generalized atrophic benign epidermolysis bullosa.
Nat Genet 11:83–86, 1995
Murakami H, Nishioka S, Setterfield J, et al: Analysis of antigens targeted by
circulating IgG and IgA autoantibodies in 50 patients with cicatricial
pemphigoid. J Dermatol Sci 17:39–44, 1998
Nakatani C, Muramatsu T, Shirai T: Immunoreactivity of bullous pemphigoid
(BP) autoantibodies against the NC16A and C-terminal domains of the
180 kDa BP antigen (BP180): Immunoblot analysis and enzyme-linked
immunosorbent assay using BP180 recombinant proteins. Br J Dermatol
139:365–370, 1998
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
Nie Z, Hashimoto T: IgA antibodies of cicatricial pemphigoid sera specifically
react with C-terminus of BP180. J Invest Dermatol 112:254–255, 1999
Nievers MG, Kuikman I, Geerts D, et al: Formation of hemidesmosome-like
structures in the absence of ligand binding by the (alpha)6(beta)4 integrin
requires binding of HD1/plectin to the cytoplasmic domain of the (beta)4
integrin subunit. J Cell Sci 113 (Pt 6):963–973, 2000
Nishiyama T, Amano S, Tsunenaga M, et al: The importance of laminin 5 in the
dermal-epidermal basement membrane. J Dermatol Sci 24 (Suppl. 1):
S51–S59, 2000
Pulkkinen L, Uitto J: Hemidesmosomal variants of epidermolysis bullosa.
Mutations in the alpha6beta4 integrin and the 180-kD bullous pemphigoid
antigen/type XVII collagen genes. Exp Dermatol 7:46–64, 1998
Ruhrberg C, Watt FM: The plakin family: Versatile organizers of cytoskeletal
architecture. Curr Opin Genet Dev 7:392–397, 1997
Savage PA, Boniface JJ, Davis MM: A kinetic basis for T cell receptor
repertoire selection during an immune response. Immunity 10:485–492,
1999
Schmidt E, Obe K, Brocker EB, et al: Serum levels of autoantibodies to BP180
correlate with disease activity in patients with bullous pemphigoid. Arch
Dermatol 136:174–178, 2000
Seishima M, Izumi T, Kanoh H, et al: Demonstration of antibody to 230 kDa
bullous pemphigoid antigen in lichen planus-like keratosis. Eur J
Dermatol 9:393–396, 1999a
Seishima M, Izumi T, Kitajima Y: Antibody to bullous pemphigoid antigen 1 binds
to the antigen at perilesional but not uninvolved skin, in localized bullous
pemphigoid. Eur J Dermatol 9:39–42, 1999b
Skaria M, Jaunin F, Hunziker T, et al: IgG autoantibodies from bullous pemphigoid
patients recognize multiple antigenic reactive sites located predominantly
within the B and C subdomains of the COOH-terminus of BP230. J Invest
Dermatol 114:998–1004, 2000
Stanley JR, Tanaka T, Mueller S, et al: Isolation of complementary DNA for bullous
pemphigoid antigen by use of patients’ autoantibodies. J Clin Invest
82:1864–1870, 1988
Witte S: Macro for generation of ROC-curves. Proceedings of the Sixth
Conference of SAS Application in Research and Development, Dortmund,
2002
Zillikens D, Mascaro JM, Rose PA, et al: A highly sensitive enzyme-linked
immunosorbent assay for the detection of circulating anti-BP180
autoantibodies in patients with bullous pemphigoid. J Invest Dermatol
109:679–683, 1997a
Zillikens D, Rose PA, Balding SD, et al: Tight clustering of extracellular BP180
epitopes recognized by bullous pemphigoid autoantibodies. J Invest
Dermatol 109:573–579, 1997b