Arch Dermatol Res (2004) 295 : 434–441
DOI 10.1007/s00403-003-0441-4
434
O R I G I N A L PA P E R
Ana María Abréu-Vélez · Maria Mercedes Yepes ·
Pablo Javier Patiño · Wendy B. Bollag ·
Fernando Montoya (senior)
A sensitive and restricted enzyme-linked immunosorbent assay
for detecting a heterogeneous antibody population in serum
from people suffering from a new variant of endemic pemphigus
Received: 4 June 2003 / Revised: 13 September 2003 / Accepted: 14 November 2003 / Published online: 17 January 2004
© Springer-Verlag 2004
Abstract We recently described a new variant of endemic pemphigus foliaceus (EPF) in El Bagre, Colombia,
that resembles Senear-Usher syndrome and identified autoantibodies to desmoglein 1 (Dsg1), as well as to multiple
known and unknown antigens including plectins, in the
serum of these patients. Here, we developed a cost-effective ELISA assay capable of detecting the heterogeneous
antibody population observed in these EPF patients, and
useful for serum epidemiological studies. A protein extract obtained from trypsin-digested fresh bovine skin and
further purified on a concanavalin A matrix was used as
antigen. This extract contains an important conformational epitope (a 45 kDa tryptic fragment of the Dsg1
ectodomain), which is recognized by antibodies in serum
from patients with all varieties of pemphigus foliaceus
(PF), and from half of those with pemphigus vulgaris with
active clinical disease. The cut-off and threshold values
were normalized using human serum obtained from both
endemic and non-endemic areas for PF. The efficiency of
this ELISA was tested using 600 serum samples from
controls and patients diagnosed with EPF, non-endemic
PF and other bullous diseases. The overall sensitivity and
specificity of the assay were determined to be 95% and
72%, respectively, with reproducibilities of 98% (intraassay) and 95% (interassay). Comparing the ELISA with
other tests to detect EPF autoantibodies, this ELISA was
the most sensitive, followed by direct immunofluorescence
(DIF), indirect immunofluorescence using anti-IgG4 monoclonal antibodies and immunoprecipitation (IP), respectively. The most specific assay was IP, followed by DIF.
Immunoblotting to Dsg1 exhibited both poor sensitivity
and poor specificity, although plectins were well visualized. We conclude that this ELISA is an excellent tool for
field serological studies, allowing testing of multiple serum
samples simultaneously and for detecting, with appropriate restriction and sensitivity, the heterogeneous antibody
population seen in patients with this variant of EPF. Finally, autoantibody serum levels obtained with this ELISA
correlated well with the clinical activity and extent of disease in patients with El Bagre EPF.
Keywords Autoimmunity · ELISA · Pemphigus
A. M. Abréu-Vélez (✉) · W. B. Bollag
Institute for Molecular Medicine and Genetics,
Medical College of Georgia,
CB 2803, 1120 15th Street, Augusta, GA 30912-2630, USA
Tel.: +1-721-7210689, Fax: +1-706-7217915,
e-mail: aavelez@mail.mcg.edu
Abbreviations BMZ Basement membrane zone · BP
Bullous pemphigoid · BP180 Bullous pemphigoid 180
kDa antigen · ConA Concanavalin A · CPF Cazanave’s
pemphigus foliaceus · DIF Direct immunofluorescence ·
ELISA Enzyme-linked immunosorbent assay · EPF
Endemic pemphigus foliaceus · IB Immunoblotting · IIF
Indirect immunofluorescence · IP Immunoprecipitation ·
mAb Monoclonal antibody · PBS Phosphate-buffered
saline · PBS– Phosphate-buffered saline lacking divalent
cations · PF Pemphigus foliaceus · PV Pemphigus
vulgaris · SLE Systemic lupus erythematosus
M. M. Yepes
Epidemiology and Public Health Program, School of Medicine,
University of Antioquia, Medellin, Colombia
Introduction
This paper is part of the doctoral (PhD) thesis of Ana María
Abréu Vélez, MD, while at the University of Antioquia.
F. Montoya (senior)
Basic Biomedical Science Corporation,
University of Antioquia, Medellin, Colombia
P. J. Patiño
Group of Primary Immunodeficiencies,
University of Antioquia, Medellin, Colombia
Endemic pemphigus foliaceus (EPF) is the only known
autoimmune disease that occurs in one relatively well-defined geographic area [1]. EPF has been described in foci
in the South American tropical rain forest [1], and perhaps
in Tunisia [2]. Patients with EPF and other types of pem-
435
phigus foliaceus (PF), e.g. the sporadic form (also known
as Cazanave’s pemphigus foliaceus, CPF) and pemphigus
erythematosus (PE), have autoantibodies directed against
desmoglein 1 (Dsg1) [3], a ubiquitin carrier protein, desmocollins, envoplakin, periplakin, and acetylcholine receptors,
as well as other antigens including, among others, one of
168 kDa [4, 5, 6, 7, 8, 9, 10, 11, 12]. Many studies have
focused on the autoantibodies to Dsg1, and the development of an ELISA assay for measurement of autoantibodies against both Dsg1 and Dsg3 has been an important advance in the diagnosis and monitoring of pemphigus [13,
14, 15, 16, 17, 18]. However, this ELISA utilizing recombinant Dsg1 and Dsg3 has some disadvantages including
its high cost. Since EPF is mostly localized in underdeveloped countries, the cost of the ELISA represents a significant drawback [18]. In addition, the recent discovery
of new autoantigens in autoimmune blistering diseases
(e.g. Dsg4 in PV) [19] make this available ELISA directed
against one or two antigens less accurate for serum epidemiological studies in areas of high prevalence of endemic pemphigus.
Patients suffering from a new variant of EPF in El Bagre
(Colombia) with features of Senear-Usher syndrome exhibit a heterogeneous autoantibody population recognizing
Dsg1, periplakin, envoplakin, and desmoplakins, as well as
several unknown antigens [20, 21]. This El Bagre EPF, occurring in a poor rural area in the Colombian rainforest, has
raised the need to perform cost-effective serum epidemiological studies. Thus, we aimed to develop an ELISA with
the versatility to obtain large amounts of antigen that would
maintain conformational epitopes in tropical weather, so
that serum samples would not need to be transported overseas for analysis. We also aimed to prevent cross-reactivity
in this ELISA with antibodies presumably directed against
antigens from animals, plants or microorganisms that prevail in the endemic area of pemphigus, using serum from
normal donors both within and outside the area of EPF to
normalize the assay. As antigen(s) we used a trypsin-digested extract from fresh bovine skin with superficial dermal remnants. This extract has been shown to contain major epitopes, including the ectodomain of the mature form
of Dsg1 [22], specifically recognized using immunoprecipitation (IP) by serum from patients with all types of PF
including its variants (endemic and sporadic) [20, 21, 23,
24].
Materials and methods
Patients and serum samples
All patients participated willingly in this study and signed or
agreed to a consent form. Since 98% of the patients with EPF are
illiterate, we explained to them the purposes of this research in the
presence of a witness from the community. In addition, these research protocols were approved by the Institutional Review Board
and Human Assurance Committee in accordance with the Scientific and Ethics committees of the Colombian Institute of Tropical
Diseases (ICMT), as part of the Institute of Health Sciences (CES),
the Sectional Direction of Health of Antioquia State (DSSA), and
the University of Antioquia.
A total of 600 serum samples were tested. Samples were distributed into three groups as follows:
– Group 1: serum from patients with PF disease. Samples were
obtained from 170 patients with PF comprising 100 EPF patients from rural areas around El Bagre, Colombia [19, 20, 21],
15 fogo selvagem (FS) patients from Brazil [25], 35 CPF patients from Colombia, and 20 PE (also known as Senear-Usher
syndrome) patients [26] from Colombia and Spain. All samples
from PF patients were tested by indirect IF [27], IB [24], and IP
[24]. Of the EPF patients, 80% had received systemic corticosteroids (between 10 and 40 mg/day) according to their clinical
condition. We assessed the severity of the disease based on the
percentage of skin compromised as used to measure the extent
of burns [28], low response to corticosteroids, and autoantibody
titers obtained by IIF.
– Group 2: normal serum. Samples of normal serum were tested,
100 from the Colombia Red Cross Serum Bank and 150 from
normal donors in the endemic area of EPF in El Bagre [21, 23].
Some of these last samples were obtained from individuals genetically related to the EPF patients (n=50).
– Group 3: patients with other autoimmune skin diseases. Samples
from 80 patients with bullous pemphigoid (BP) from Germany
[28], the US, and Colombia were tested. The diagnosis was confirmed by clinical histopathological and immunopathological
criteria including IIF and by ELISA [29] procedures. Samples
from 50 well-characterized PV patients, assessed by clinical, histological, and immunological criteria including DIF, IIF, IB and
IP using the same antigen preparation used in this ELISA [24,
27], and from 50 patients with SLE, assessed using the criteria of
the American Association of Rheumatologists, were also tested.
Extraction, partial purification and radiolabeling of the PF antigen(s)
Fresh bovine skin, processed within 1 h of the death of the animal,
was used as antigen source. The antigen was further purified using
a concanavalin A (ConA) matrix. This extract has previously been
clearly demonstrated to contain a major conformational epitope for
pemphigus disease [24, 30]. For this assay we also included the
basement membrane zone (BMZ) of the skin in the trypsinization
protocol, as demonstrated using hematoxylin-eosin (H&E), periodic acid-Schiff, and alcian blue staining. Extracts of both the epidermis and BMZ were used because we had previously demonstrated the presence of a lupus band in 40% of the samples from
the El Bagre EPF patients, resembling findings observed in SenearUsher syndrome, as well as histopathological features indicating
dermoepidermal junction alterations in these individuals [20, 21].
A detailed method for preparation and characterization of the antigen
is given in reference 31, as well as the validation of this method.
A schematic comparison of the proteins present in the tryptic
digest before and after the ConA affinity purification visualized
using silver stain is shown in Fig. 1. Fraction A of the ConA extract was found to contain antigens recognized by serum from EPF
patients. A portion of the ConA fraction A was also radiolabeled
by the chloramine T method using 125I as previously described [24,
30], and protein-bound radioactivity determined by precipitation
with 10% trichloroacetic acid also as previously described [24, 30].
This antigen preparation immunoadsorbs intercellular staining from
the serum of PF patients as observed by IIF [29]. However, since
ConA itself has been reported to decrease the subsequent reaction
by IIF of serum from pemphigus patients with antigen following
immunoadsorption [32], an additional two slides were also incubated with normal human serum as a negative control with and without ConA beads [32]. A slight decrease in the intercellular stain
was noticed with the ConA beads as previously described [32].
Immunoblotting, immunoprecipitation and immunoadsorption
using ConA products
IB and IP were performed as previously described [24, 30]. To test
for the presence of antigens in the extract capable of immunoab-
436
Fig. 1a–d A schematic representation of the partial purification of the antigen used in
the ELISA. In order to test the
electrophoretic profile of a soluble protein extract of trypsindigested epidermis from cow
snouts, a 7% SDS-PAGE was
run (a). b Same procedure, but
using the ConA-isolated antigen(s), stained with Coomassie
blue. Note that the protein profile is different. c Using the
ConA isolated antigen(s) all
samples from PF and EPF patients immunoprecipitated the
45 kDa antigen(s) (red arrow).
Lanes 3 and 4 serum from PF
patients, lane 5, molecular
weight standards, lanes 7 to 12
serum from EPF patients, lanes
1 and 2 and 6, normal human
serum. d Autoradiography with
the 45 kDa radiolabeled tracer
prepared by radioiodination of
the ConA-purified antigen extract
sorbing the intercellular staining observed routinely with serum
from EPF patients by IIF in human skin, antigenic fractions eluted
from the ConA column were determined by immunoadsorption, as
previously described [24, 30].
Immunofluorescence analysis
DIF and IIF using sections of normal human skin were performed
by standard methods [27]. To determine IgG subclasses, fluorescence-conjugated mouse mAb specific to human IgG1–4 (Sigma,
St Louis, Mo.) were used.
ELISA protocol
Microtiter plates (96-well, Immulon-4; Dentate Laboratory, Alexandria, Va.) were preincubated with ice-cold glutaraldehyde. Excess
crosslinker was discarded, and the PF antigen (ConA fraction A)
was diluted twofold in phosphate-buffered saline (PBS) to a final
concentration of 0.25 µg/µl per well. The antigen was then incubated in the plate, washed with PBS lacking divalent cations
(PBS–) and soaked. Nonspecific binding was diminished by blocking plates with 10% calcium lactate (Kodak) in 0.019 M Tris-HCl,
0.29 M NaCl, and 0.1% Tween-20 (Sigma). Plates were then washed
and incubated with 50 µl serum (dilution 1:100). A negative control (1:100 dilution) and positive controls (1:100, 1:200 and 1:300
dilutions) were run in each microplate. Each serum sample (primary
antibody) was incubated in triplicate for 1 h and the plates were
washed. The secondary antibody, HRP-labeled goat anti-human
IgG (Kirkegard and Perry, Gaithersburg, Md.) was then added to
each well at a 1:20,000 dilution. Finally, the plates were washed
and incubated with a solution containing 50 µl o-phenyldiamine
and 0.1% H2O2 (OPD, Sigma). The reaction was stopped by addition of 50 µl 2 N H2S04 and the optical density read at 492 nm using a microplate reader (Biorad, model 2550).
Nonrestricted reactivity was determined by two strategies: (1)
eight samples were incubated in the ELISA plate with buffer (no
serum) and the mean value of nonrestricted reactivity obtained was
subtracted from the reading from other samples; (2) the mean readings from serum samples from 150 normal individuals from the endemic area were also determined. The mean of the readings obtained
with all nonrestricted samples was used as a standard nonrestricted
reactivity value for all assays [31, 33, 34]. To determine reproducibility, four distinct batches of antigen preparations were tested.
The variation in ELISA readings was evaluated, and the final protocol was repeated 50 times [31, 33, 34]. To test the feasibility of
using this ELISA for future serum epidemiological studies, the stability of the antigen was tested under different conditions of temperature (4°C to 37°C including room temperature 22°C) and humidity. Thimerosal 0.01% (Sigma) diluted in PBS or PBS– was
added as a preservative to the antigen-coated plates.
Statistical analysis
A receiver operating characteristic (ROC) analysis was carried out
to determine a cut-off value using MedCal software (Mariakerke,
437
Table 1 Comparison between
different assays used for the
detection of PF autoantibodies.
We tested for the presence of
autoantibodies against PF antigen(s) using different techniques. Serum from patients
with EPF (Colombia), FS from
Brazil, CPF, and PE were also
tested. The results are expressed as the percentage of
positive reactions for autoantibodies against PF antigen(s).
The ELISA was the most sensitive assay, followed by DIF,
IP and IIF, respectively. The IP
was the most restricted assay.
IB exhibited both poor sensitivity and restriction (NA not
available)
Assay
EPF
(n=100)
IIF to detect intercellular staining between keratinocytes
using polyclonal IgG on human cryosections
50%
50%
53%
IIF to detect intercellular staining between keratinocytes
using IgG1 mAb subclass on human cryosections
5%
15%
10%
IIF to detect intercellular staining between keratinocytes
using IgG4 mAb subclass
84%
84%
86%
DIF to detect intercellular staining between keratinocytes
using IgG4 mAb subclass
90%
NA
94%
IB to detect a 160 kDa protein (Dsg1) using human and
bovine extract as antigens
27%
28%
27%
IP using the 45 kDa bovine tryptic antigen obtained after
Con-A purification
91%
91%
100%
ELISA for detecting PF autoantibodies using the bovine
45 kDa tryptic antigen obtained after Con-A purification
96%
100%
100%
Belgium) for Windows and GraphPad Prism (GraphPad Software,
San Diego, Calif.). The adjusted cut-off, determined after normalization with normal control serum from the endemic area, was 0.1.
Samples with antibody titers less than 0.1 were considered negative.
Results
FS
(n=15)
CPF and PE
(n=55)
reading of each sample and from the positive control
(1:100 dilution). Adjusting the relative readings to that of
the positive control permitted comparison of results from
multiple plates tested on different days. The OD492 nm
mean adjusted value was 0.4090 and the adjusted standard
deviation was 0.097. The reproducibility between assays
was 95%, and the intraassay reproducibility was 97% using this adjusted reading.
Efficiency of antigen binding to plates
Our results showed that 80% of the EPF antigen was bound
to the ELISA plates. Thus, radiolabeled antigen was measured precoating at 538,463 pCi/ml of 125I associated with
protein and postcoating at 118,063 pCi/ml remaining in
the supernatant (including all washes) after 2.5 h of incubation.
Long-term ELISA efficiency
The stability testing of the ConA antigen showed that the
coated plates retained antigenicity when stored for up to
3 months at room temperature. However, with storage, the
cut-off value was found to decrease to 0.08. When the
ConA antigen was stored at –70°C, the buffers at 4°C, the
serum (primary antibody) at –20°C, and the HRP-coupled
secondary antibody at –20°C, the ELISA was stable for at
least 1 year after production, although again, at this time
the cut-off value to be considered positive was found to
have decreased to 0.065.
Comparison with different tests
presently available for detecting PF antibodies
To determine the restriction of the ELISA relative to other
diagnostic tests, we examined serum from PF patients and
normal controls from the endemic PF area using IIF, IB,
IP, and this ELISA. Samples from half of the EPF patients
were also tested by DIF (Table 1). The ELISA was the
most sensitive assay followed by DIF and IIF using mAbs
against the IgG4 subclass. However, serum from three patients with El Bagre EPF who had been clinically cured
for more than 5 years was still positive by this ELISA. IP
was the most restricted and IB the least restricted assay.
We also calculated the cost of the ELISA assay using the
cost of the reagents required for the ELISA as well as
those necessary to produce sufficient quantities of antigen
for up to 25 96-well plates. The cost per 96-well plate of
this ELISA was approximately one-tenth that of the commercially available assay.
ELISA restriction and sensitivity
Reproducibility of the assay
In the intraassay variability experiments, which used a positive control at a 1:100 dilution, 50 plates were run and the
samples were read three times. An intraassay reproducibility of 98% was calculated. For interassay reproducibility,
an adjusted OD492 nm value was used. The mean reading of
the negative control was subtracted from the OD492 nm
The overall sensitivity and restriction of this assay were
95% and 72%, respectively. As shown in Fig. 2, in the EPF
group, 6/60 serum samples were negative (90.4% sensitivity). The six patients were in clinical remission, and
their samples only weakly immunoprecipitated the 45 kDa
PF antigen (Dsg1 ectodomain) and were negative by IIF.
However, in this new variant of EPF, some patients have a
438
Fig. 2 The presence of antibodies against the PF antigen(s) detected using the developed ELISA. A total of 600 serum samples
were tested using the ELISA as described in Methods. Autoantibody levels were determined using this ELISA in serum from CPF
patients, FS patients, normal individuals from outside the endemic
area (NHS), PE patients, BP patients, PV patients, SLE patients,
EPF patients, and unrelated normal donors from the PF endemic
area (DEA). For the numbers of patients, refer to Methods. We determined the SEM as a measure of how far the sample mean was
likely to be from the true population mean. The SEM was calculated using the expression SD/√N. With large samples, the SEM is
always small, and by itself, is difficult to interpret. For this reason,
it is more precise to use the 95% confidence interval, which was
calculated from the SEM as described in the Results
very localized form of the disease (a “frustre form”). In
the area around specific lesions of these patients, DIF was
positive for intercellular staining using anti-IgG4 mAb.
All samples from FS, CPF and PE patients were positive
by IP, recognizing conformational epitopes using the same
ConA, tryptic antigen used in this ELISA (the sensitivity
was 100%). These patients all had clinically active disease. All samples from SLE patients were negative (100%
restriction). The samples from half of the PV patients were
positive and also immunoprecipitated the 45 kDa ConA PF
antigen (50% restriction). Of the normal serum samples
from patients outside the endemic area, 2/100 were positive (2% false-positive).
Of the samples from 80 BP patients, 6 were positive
(93.7% restriction) (Fig. 2). These six samples showed high
titers of antibodies against BP 180 kDa antigen (BP180).
We pre-immunoadsorbed the serum from the BP patients
with an affinity-purified recombinant form of BP180
(NC16A) and the ELISA was repeated. The ELISA readings still remained positive after this immunoadsorption
(data not shown), indicating a lack of cross-reactivity of
anti-NC16A autoantibodies with the ConA fraction A
antigen(s) and suggesting potential PF-like autoantibodies
in BP. However, serum from BP patients may exhibit autoantibodies to regions outside the NC16A domain; indeed,
an immunoprecipitation was also performed and the samples from all the BP patients were negative. Nevertheless, it
is possible that the ConA antigenic fractions contain glycoproteins other than Dsg1 or BP180 recognized by BP serum. Therefore, it is not entirely clear whether the observed
reactivity was falsely positive or represented antibodies to
glycoproteins in the extract other than BP180 or Dsg1.
Normal serum from donors from the endemic area (particularly those genetically related to EPF patients) exhibited
Fig. 3 Reactivity against the PF antigen(s) compared in serum from
normal individuals living within and outside the endemic area
around El Bagre. Autoantibody levels were determined using the
ELISA. We tested serum from asymptomatic donors from the PF
endemic area (DEA) and from normal individuals residing outside
the PF endemic area (NHS). A marked hyperreactivity was demonstrated in serum from individuals living in the endemic area, demonstrating that those living in the same environment are also likely
exposed to the possible trigger or exacerbating factors, although
they do not develop the disease
higher antibody reactivity by this ELISA than that from outside the endemic area (Fig. 3). Of the 150 normal controls
from the endemic EPF area, the samples from 9 of 50 individuals genetically related to EPF patients living in the
endemic area were positive by the ELISA as well as by IP,
and the samples from four of these individuals were positive by IB. Control serum from 8 of 100 normal individuals
from the endemic area was also positive (Fig. 3). Thus, serum from a total of 17 asymptomatic individuals from within
the endemic area tested positive by this ELISA (Fig. 3).
ELISA values oscillate in parallel
with disease activity over time
Figure 4 illustrates the high correlation between the presence and levels of autoantibodies and the activity and extent of the disease. Clinical activity was evaluated by surface compromise according to the scale of Browder and
Lund [35], and antibody levels were determined using the
adjusted ELISA OD492 nm readings. A correlation coefficient (r2) of 0.92 was obtained. Importantly, in 40 of the
50 EPF patients who were tested three times a year, a correlation of 95%, with parallel fluctuations, was observed
between ELISA readings and the clinical condition over
time. Thus, the intake of oral steroids, the skin surface affected by the disease, and autoantibody levels measured
by this ELISA displayed a high degree of correlation.
Discussion
The ability to detect an antibody response in people living
in an endemic area of EPF and at risk of developing this
439
Fig. 4 Correlation between clinical disease activity and OD492 nm
readings using the ELISA. The x-axis represents clinical disease
activity and dosage of steroid hormone. Clinical disease activity
was evaluated and plotted based on the amount of skin surface
compromised using the method used in patients with burns [28]
(1.0 corresponds to 100% and 0.1 correspond to 10%). A group of
100 patients with El Bagre EPF were followed. The y-axis corresponds to OD492 nm readings obtained using the ELISA. A linear
correlation between the amount of skin surface affected by the disease and the ELISA readings was detected
disease is necessary for serum epidemiological surveys.
To this end, we developed an indirect ELISA assay that
detected antibodies in serum with the proper sensitivity and
restriction. This assay was able to detect antibodies directed
against PF antigens, including the ectodomain of Dsg1
[22]. The antigens used for the ELISA were derived from
trypsinization and ConA purification of epidermal and
BMZ molecules from fresh cow snouts. The tryptic ConA
fraction is recognized using IP by all serum samples from
PF and EPF patients (Brazilian and El Bagre) with clinically active disease and by serum samples from half of the
PV patients [24, 29].
A recent optimization of a commercially available
ELISA for pemphigus disease has been reported that used
serial dilutions and intraassay normalization, similar to
the procedures used in this research [36]. In fact, serum
samples from 34 patients with this new variant of EPF that
were positive in our ELISA were also tested using the commercially available ELISA, and 33 of the 34 were found
to be positive, demonstrating a high degree of correlation
between the two assays [37]. However, the set-up and
preparation of the ELISA described here is less expensive
and does not require a highly technological infrastructure,
a factor which is particularly important since most foci of
EPF are located in underdeveloped countries [2, 3].
As shown, higher antibody reactivity was detected by
this ELISA in serum from normal individuals in the endemic area (particularly from those genetically related to
EPF patients) than from those outside the endemic area
(Fig. 3), as has been also shown previously [38, 39, 40].
One factor that may account for these results is higher exposure to a cross-reactive agent(s) in the environment
within the endemic area which results in autoimmunity in
people with a genetic susceptibility. Alternatively, the high
sensitivity of this assay may allow the detection of nonpathogenic PF antibodies prior to disease onset. For these
reasons, this assay could be useful for serum epidemiological studies to detect immune conversion in people in
endemic foci of pemphigus. Another phenomenon observed
with this ELISA was the fact that the serum from three
“clinically cured” EPF patients was still positive in the
ELISA. Moreover, serum from nine disease-free relatives
of these EPF patients was positive in the assay. This may
represent an “immunological scar” or, again, non-pathogenic autoantibodies. Only larger and longer-term serum
epidemiological studies will provide an answer to this
question.
This ELISA also exhibited reasonable restriction to PF
and EPF. On the other hand, a few samples from BP patients demonstrated reactivity in this assay. Moreover, in
our studies the immunoreactivity of these samples was not
reduced after adsorption with the NC16A fusion protein
(BP180 antigen). Whether this ELISA detects reactivity of
these samples against a portion or portions of BP180 outside the NC16 A domain, with a number of other antigens,
including BP230, or other glycoproteins in the tryptic ConA
fractions is unclear at this time. Note that the presence of
minor BP autoantigens has been reported previously [41].
Additional restriction of this ELISA was presumably
also provided by the normalization of the assay using
samples from normal donors from the endemic area, since
people living in endemic areas tend to have higher exposure to the same environmental stimuli as patients, resulting in an increase in such cross-reactive autoantibodies.
Thus, several environmental agents share homology with
the ectodomain of Dsg1, including proteins from Tylorrhynchus heterochaetus, Trypanozoma cruzi, Salmonella
typhimurium, Mus musculus, Drosophila melanogaster
and Petunia hybrida microorganisms (SWISS-PROT Protein Knowledgebase TrEMBL Computer-annotated supplement to SWISS-PROT). Most of these species or their
ancestors are present in the endemic area of pemphigus
and might contribute to the detection of non-restricted immunoreactivity by ELISA in these control populations.
In conclusion, when other assays are not available, we
recommend DIF using the anti-IgG4 mAb subclass, which
exhibits greatly increased sensitivity, for the diagnosis of
PF, as has been previously demonstrated [42]. However,
here we describe the development of an ELISA that detects
a heterogeneous autoantibody population and demonstrate
its sensitivity, appropriate restriction, stability and costeffectiveness for the detection of autoantibodies to PF antigens, making it highly suitable for use in underdeveloped
countries.
Acknowledgements Initial studies were funded by the Immunodermatology Laboratory of the Medical College of Wisconsin,
Milwaukee, Wis., under the guidance of Dr. Luis A. Diaz. Later,
this project was funded by the University of Antioquia, Basic Bio-
440
medical Science Corporation, Medellin, Colombia (Abreu and
Montoya) and by the Institute for Molecular Medicine, Medical
College of Georgia, Augusta, Ga. (Abreu) and NIH # AR45212
(Bollag). We want to thank Drs. Monica Olague and Jose M. Mascaro Jr. and Ms. Argelia Lopez for their team work. We also thank
Dr. Detleff Zillikens (University of Wuerzburg, Germany) for providing serum from BP patients and Dr. George Guidice of the Department of Dermatology at the Medical College of Wisconsin for
the NC16 peptide for the immunoadsorption studies. Dr. Abreu
was the recipient of a scholarship from Colciencias, Colombia.
References
1. Castro RM, Proenca NG (1983) Semelhancas e diferencas entre o fogo selvagem e o penfigo foliaceo de Cazanave (Similarities and differences between South American pemphigus foliaceus and Cazanave’s pemphigus foliaceus). An Bras Dermatol
53:137–139
2. Morini JP, Jomaa B, Gorgi Y, Saguem MH, Nouira R, Roujeau
JC, Revuz J (1993) An endemic pemphigus foliaceus focus in
the Sousse area of Tunisia. Arch Dermatol 129:69–73
3. Koulu L, Kusumi A, Steinberg MS, Klaus-kovtun V, Stanley
JR (1984) Human autoantibodies against a desmosomal core
protein in pemphigus foliaceus. J Exp Med 160:1509–1518
4. Kazerounian S, Mahoney MG, Uitto J, Aho S (2000) Envoplakin and periplakin, the paraneoplastic pemphigus antigens,
are also recognized by pemphigus foliaceus autoantibodies.
J Invest Dermatol 115:505–507
5. Nguyen VT, Ndoye A, Shultz LD, Pittelkow MR, Grando SA
(2000) Antibodies against keratinocyte antigens other than
desmogleins 1 and 3 can induce pemphigus vulgaris-like lesions.
J Clin Invest 106:1467–1479
6. Diaz LA, Sampaio SAP, Rivitti EA, Macca LL, Roscoer JT,
Takashashi Y, Labib RS, Patel HT, Mutassim DF, Dugan AM,
Anhalt GJ (1987) An autoantibody in pemphigus serum, restriction for the 59 kDa keratin, selectively binds the surface of
the keratinocytes: evidence for an extracellular keratin domain.
J Invest Dermatol 89:287–295
7. Dugan EM, Labib RS, Anhalt GJ, Diaz LA (1989) Selective
surface radioiodination of keratinocytes in primary culture labels a 59 kD keratin and other surface proteins. Arch Dermatol
Res 281:463–469
8. Liu Z, Diaz LA, Haas AL, Giudice GJ (1992) cDNA cloning of
a novel human ubiquitin carrier protein. An antigenic domain
restrictionally recognized by endemic pemphigus foliaceus autoantibodies is encoded in a secondary reading frame of this
human epidermal transcript. J Biol Chem 267:15829–15835
9. Dmochowski M, Hashimoto T, Garrod DR, Nishikawa T
(1993) Desmocollins I and II are recognized by certain sera
from patients with various types of pemphigus, particularly
Brazilian pemphigus foliaceus. J Invest Dermatol 100:380–384
10. Kim SC, Kwon YD, Lee IJ, Chang SN, Lee TG (1997) cDNA
cloning of the 210-kDa paraneoplastic pemphigus antigen reveals that envoplakin is a component of the antigen complex.
J Invest Dermatol 109:365–369
11. Vu TN, Lee TX, Ndoye A, Shultz LD, Pittelkow MR, Dahl
MV, Lynch PJ, Grando SA (1998) The pathophysiological significance of non-desmoglein targets of pemphigus autoimmunity. Development of antibodies against keratinocyte cholinergic receptors in patients with pemphigus vulgaris and pemphigus foliaceus. Arch Dermatol 134:971–980
12. Joly P, Gilbert D, Thomine E, Zitouni M, Ghohestani R,
Delpech A, Lauret P, Tron F (1997) Identification of a new antibody population directed against a desmosomal plaque antigen in pemphigus vulgaris and pemphigus foliaceus. J Invest
Dermatol 108:469–475
13. Amagai M, Komai A, Hashimoto T, Shirakata Y, Hashimoto
K, Yamada T, Kitajima Y, Ohya K, Iwanami H, Nishikawa T
(1999) Usefulness of enzyme-linked immunosorbent assay using recombinant desmogleins 1 and 3 for serodiagnosis of pemphigus. Br J Dermatol 140:351–357
14. Ishii K, Amagai M, Hall RP, Hashimoto T, Takayanagi A,
Gamou S, Shimizu N, Nishikawa T (1997) Characterization of
autoantibodies in pemphigus using antigen-restriction enzymelinked immunoadsorbent assays with baculovirus-expressed recombinant desmogleins. J Immunol 159:2010–2017
15. Lenz P, Amagai M, Volc-Platzer B, Stingl G, Kirnbauer R
(1999) Desmoglein 3-ELISA: a pemphigus vulgaris-restriction
diagnostic tool. Arch Dermatol 135:143–148
16. Sami N, Bhol KC, Ahmed AR (2001) Diagnostic features of
pemphigus vulgaris in patients with pemphigus foliaceus: detection of both autoantibodies, long-term follow-up and treatment responses. Clin Exp Immunol 125:492–498
17. Harman KE, Gration MJ, Seed PT, Bhogal BS, Challacombe
SJ, Black MM (2000) Diagnosis of pemphigus by ELISA: a
critical evaluation of two ELISAs for the detection of antibodies to the major pemphigus antigens, desmoglein 1 and 3. Clin
Exp Dermatol 25:236–240
18. Bystryn JC, Akman A, Jiao D (2002) Limitations in enzymelinked immunosorbent assays for antibodies against desmogleins
1 and 3 in patients with pemphigus. Arch Dermatol 138:1252–
1253
19. Kljuic A, Bazzi H, Sundberg JP, Martinez-Mir A, O’Shaughnessy R, Mahoney MG, Levy M, Montagutelli X, Ahmad W,
Aita VM, Gordon D, Uitto J, Whiting D, Ott J, Fischer S,
Gilliam TC, Jahoda CA, Morris RJ, Panteleyev AA, Nguyen
VT, Christiano AM (2003) Desmoglein 4 in hair follicle differentiation and epidermal adhesion. Evidence from inherited hypotrichosis and acquired pemphigus vulgaris. Cell 113:249–260
20. Abreu-Velez AM, Maldonado JG, Jaramillo A, Patiño PJ,
Prada S, Leon W, Montoya F (1998) Immunological characterization of a unique focus of endemic pemphigus foliaceus in
the rural area of El Bagre, Colombia. J Invest Dermatol 110:
516a
21. Abreu-Velez AM, Hashimoto T, Bollag WB, Tobon Arroyave
S, Abreu-Velez CE, Londono ML, Montoya F, Beutner EH
(2003) A unique form of endemic pemphigus in northern
Colombia. J Am Acad of Dermatol 49:599–608
22. Abreu AM, Olague-Marchan M, Lopez-Swiderski A, Mascaro
JM, Giudice GJ, Diaz LA (1997) Characterization of a 45 kD
epidermal tryptic peptide recognized by pemphigus foliaceus
sera. J Invest Dermatol 108:541a
23. Abreu-Velez AM, Beutner EH, Montoya F, Bollag WB,
Hashimoto T (2003) Analyses of autoantigens in a new form of
endemic pemphigus foliaceus in Colombia. J Am Acad of Dermatol 49:609–614
24. Labib RS, Rock B, Robledo MA, Anhalt GJ (1991) The calciumsensitive epitope of pemphigus foliaceus antigen is present on
a murine tryptic fragment and constitutes a major antigenic region for human autoantibodies. J Invest Dermatol 96:144–147
25. Diaz LA, Sampaio SAP, Rivitti EA, Martins CR, Cunha PR,
Lombardi C, Almeida FA, Castro RM, Macca ML, Lavarado C
(1989) Endemic pemphigus foliaceus (fogo selvagem). Clinical
features and immunopathology. J Am Acad Dermatol 20:657–
669
26. Viera JP (1942) Penfigo foliaceo e syndrome de Senear-Usher
(Pemphigus foliaceus and Senear-Usher Syndrome). Empresa
Grafica da Revista dos Tribunas, Sao Paulo
27. Kransy SA, Beutner EH, Chorzelski TP (1978) Restrictionity
and sensitivity of indirect and direct immunofluorescent findings in the diagnosis of pemphigus. In: Beutner EH, Chorzelsi
TP, Kumar V (eds) Immunopathology of the skin, 32nd edn.
Wiley, New York, pp 207-247
28. Curtis PA, Dabney RJ (1977) Burns. Including gold chemicals
and electrical injuries. In: Sabiston DC Jr (ed) Textbook of surgery. Saunders, Philadelphia, pp 297–298
29. Zillikens D, Mascaro JM Jr, Rose PA, Liu Z, Erwing SM, Caux
F, Hoffmann RG, Diaz LA, Giudice G (1997) A highly sensitive enzyme-linked immunosorbent assay for the detection of
circulating anti-BP180 autoantibodies associated with bullous
pemphigoid. J Invest Dermatol 109:679–683
441
30. Labib RS, Rock B, Martins CR, Diaz LA (1990) Pemphigus
foliaceus antigen: characterization of an immunoreactive tryptic fragment from BALB/c mouse epidermis recognized by all
patient sera and major autoantibody subclasses. Clin Immunol
Immunopathol 57:317–329
31. McCullough KC, Bruckner L, Schaffner R, Fraefel W, Muller
HK, Kihm U (1992) Relationship between the anti-FMD virus
antibody reaction as measured by different assays, and protection
in vivo against challenge infection. Vet Microbiol 30:99–112
32. Imamura S, Takigawa M, Ofuji S (1979) Binding restrictionity
of rabbit anti-guinea pig epidermal cell sera: comparison of their
receptors with those of concanavalin A and pemphigus sera.
Acta Derm Venereol 59:113–119
33. Crowther JR (1995) ELISA theory and practice. In: Crowther JR
(ed) Methods in molecular biology. Humana Press, Totowa, NJ
34. Dow BC, Munro H, Ferguson K, Buchanan I, Jarvis L, Jordan
T, Franklin IM, McClelland M (2001) HTLV antibody screening using mini-pools. Transfus Med 11:419–422
35. Quak JJ, Balm AJ, van Dongen GA, Brakkee JG, Scheper RJ,
Snow GB, Meijer CJ (1990) A 22-kd surface antigen detected by
monoclonal antibody E 48 is exclusively expressed in stratified
squamous and transitional epithelia. Am J Pathol 136:191–197
36. Cheng SW, Amagai M, Nishikawa T (2002) Monitoring disease activity in pemphigus with enzyme-linked immunosorbent
assay using recombinant desmogleins 1 and 3. Br J Dermatol
147:261–265
37. Hisamatsu Y, Abreu-Velez AM, Amagai M, Ogawa MM, Kanzaki T, Hashimoto T (2003) Comparative study of autoantigen
profile between Colombian and Brazilian types of endemic
pemphigus by various biochemical and molecular techniques.
J Dermatol Sci 32:33–41
38. Silva dos Reis VM, Cucé LC, Rivitti EA (1991) Anatomopatologia e immunofluorescência direta e indireta das lesőes de
pênfigo foliáceo endêmico resistentes á corticoretapia (Histopathological, direct and indirect immunofluorescence studies
of skin biopsies from people affected by endemic pemphigus
foliaceus resistant to corticotherapy). Rev Inst Med Trop Sao
Paulo 33:97–103
39. Martins-Castro R, Chorzelsi TD, Jablonska S, Marquart F
(1976) Antiepithelial antibodies in healthy people living in the
endemic areas of South American pemphigus foliaceus. Castellania 4:111–112
40. Kricheli D, David M, Frusik-Zotlink M, Goldsmith D, Rabinov
M, Sulkes J, Milner Y (1999) The distribution of pemphigus
vulgaris IgG subclasses and their reactivity to desmoglein 1
and 3 in pemphigus patients and their relatives. J Invest Dermatol 112:614a
41. Zhu XJ, Bystryn JC (1983) Heterogenicity of pemphigoid antigens. J Invest Dermatol 80:16–20
42. Rock B, Martins C, Diaz LA (1989) The pathogenic effect of
IgG4 autoantibodies in endemic pemphigus foliaceus (Fogo
selvagem). New Engl J Med 320:1464–1469