ABSTRACT Antineutrophil cytoplasmic antibodies
(ANCA) have become an established
tool for the diagnosis of systemic vasculitis. The major role for ANCA
testing is in diagnosing renal insufficiency,of unknown origin, where a
positive test indicates whether the patient will benefit from
immunosuppressive treatment or not. A negative test result almost
completely rules out thepresence ofsystemic vasculitis.
In this clinical setting the major
antigens for AATCA are proteinase 3
and myeloperoxidase, and antibodies to these antigens can best be
tested by ELISA. In other clinical settings like inflammatory
bowel disease, arthritis and so on, several otherAATCA specificities
have been described and the IIF test is preferred. However, the
clinical value of these somewhat more esoteric specificities is
doubtful. New developments in assay techniques and better
knowledge ofspecific epitopes will lead to tools for the improved
diagnosis as well as follow up ofpatients during treatment, as has
already been seen with the capture assay for PR3 -ANCA.
Introduction ANCA (anti-neutrophil cytoplasmic
antibodies) is a family of
autoantibodies related to vasculitis and inflammatory disorders.
The first reports of antibodies reacting with leukocytes date from 1959
(1) and 1964 (2). Antibodies staining the cytoplasm of
neutrophils in necrotizing glomerulonephritis were seen in 1982
(3). These older studies had in large part been forgotten when
van der Woude et al. (4) in 1985 showed that ANCA was related to
Wegener's granulomatosis. Following this study, interest in ANCA
has increased each year and today these antibodies are considered major
diagnostic tools for systemic vasculitis (5). In this review we
will focus on methodological problems, new methods and discuss a
strategy for ANCA testing. Antigens The granulocyte is full of granules,
each with many different proteins,
mostly enzymes that are used in the defense against bacteria and
therefore the potential number of antigens is very large (6). It
was, however, very early shown that when the granules were separated
based on their density, all reactivities of sera from systemic
vasculitis patients were with the alpha fraction containing the
azurophil granules. The proteins from these granules could be
isolated and separated and most of the reactivities were found to be
with proteins of molecular weights around 29kD and l4OkD. Several
of the serine proteinases such as elastase, cathepsin G and proteinase
3 have molecular weights around 29kD and it was shown by sequencing
that the 29kD antigen actually was proteinase 3 (PR3) (7). The
other major antigen was shown to be myeloperoxidase (MPO) (8).
These two enzymes take part in the killing of bacteria by cleavage of
proteins (PR3) and by the generation of oxygen radicals (MPO) (6).
PR3 is a serine proteinase with a
molecular weight of 29kD consisting
of a single protein chain that is glycosylated. The sequence is
conserved and very similar to other serine proteases like elastase and
cathepsin G (9). MPO is a dimer with a molecular weight of l4OkD
consisting of one heavy (59kD) and one light chain (13.5kD). It is
glycosylated with high mannose oligosaccharide chains. The
molecule is easily cleaved by heating and has a characteristic green
colour (10). Many other antigens have been described to be
associated with ANCA, i.e. elastase, lactoferrin cathepsin G, BPI and
more recently molecules like defensins, HMG1/2, catalase etc. (11, 12)
as summarized in Figure 1.
Problems and pitfalls in the analysis
of ANCA The first method developed to detect
ANCA was indirect
immunofluorescence (IIF) on ethanol fixed granulocytes (13, 14).
By this method two patterns are seen, a cytoplasmic staining of the
granulocyte called C-ANCA and a perinuclear staining called
P-ANCA. The P-ANCA pattern arises due to a fixation artifact when
MPO is redistributed to the negative nucleus due to its positive
charge. IIF was followed by RIA and ELISA, first using different
extracts of granulocytes (15, 16) and later the purified proteins (14,
17). Thus, antibodies to PR3 are called PR3-ANCA and antibodies
to MPO are called MPO-ANCA, when the purified proteins are used in the
assay.
The two IIF patterns are presumed to
be related to antibody
reactivities to defined antigens, i.e. C-ANCA equals PR3-ANCA and
P-ANCA equals MPO-ANCA. This is, however, not always the
case. A few years ago we studied some patients with a C-ANCA
pattern by IIF and MPO-ANCA by ELISA (18). The index patient had
an atypical disease with light chain deposition. The antibodies
also had an atypical subclass distribution with predominant IgG2
autoantibodies. We could show that the antibodies were directed
to an epitope that is different from the regular epitope and probably
not redistributed during fixation. The epitope is still on the
same molecule and therefore it appears that all MPO do not redistribute
during fixation. These sera are rare and probably constitute only
around 1-2% of C-ANCA positive sera. The recently discovered
BPI-ANCA (antibodies to bactericidal permeability increasing protein),
however, can also give a C-ANCA pattern, but only in half of all cases
(19).
Fig. 1. Assay of ANCA using IIF
divides ANCA into two group,;, the
C-ANCA and the P-ANCA patterns. These patterns can arise from
antibodies to many granulocyte proteins and therefore it is
impossibileto know the specificity of ANCA from the pattern obtained on
IIF
The problem is much greater when
studying P-ANCA since many of the
antibodies to cationic proteins of the granulocyte can give rise to
this pattern and it is impossible to distinguish P-ANCA from the GS-ANA
pattern supposed to be true nuclear staining. Antibodies to MPO,
elastase, lactoferrin, BPI etc. can all give this kind of
pattern. The clinical selection of patients is important and if
the patients have systemic vasculitis the important antigen is MPO, but
if the patients have inflammatory bowel diseases, arthritis or
autoimmune hepatitis the specificity could be elastase, lactoferrin,
BPI, cathepsin G, etc. We still do not know the major dominating
antigen in these diseases.
Furthermore, not all antibodies to
granule proteins react well by
IIF. This has been demonstrated with BPI-ANCA, where about half
of sera containing this antibody do not give staining by IIF.
Another major problem is ANA that will also react with neutrophils and
give a PANCA or atypical staining. Therefore it is important to
always check for ANA; otherwise it is impossible to be sure that the
ANCA staining is not an ANA.
It is well known that the level by
ELISA does not correlate to the
titre by IIF. Samples can be found that are high by IIF and low
by ELISA or the opposite. This is probably a reflection of the
nature of the epitopes. It has been shown that the epitopes are
conformational, i.e. denaturation of the molecule destroys the epitope
(20). The molecule is presented in a different conformation when
bound to a plastic surface in the ELISA test than when fixed by ethanol
in the granules for the IIF (Fig. 2). Patients react with
different epitopes (21) on the molecule.
Fig. 2. The antigen (PR3) is
presented in different ways in the
assay. At IIF PR3 is fixed by ethanol and mixed with several
other proteins in the granula and thus epitopes can be destroyed or
hidden. In the direct ELISA the isolated protein is bound to a
solid phase, generally a plastic surface. The protein will bind
to hydrophobic areas and change conformation to a greater or lesser
degree. In the capture assay PR3 is held in place by a monoclonal
antibody in a more native conformation. This will orient the
antigen and the same surface will be accessible.
and thus they will give a low or high
reaction in the test depending on
how their epitope is presented. This question was recently
addressed (22) using a monoclonal antibody to catch the antigen instead
of direct binding to the plastic. It was shown that reactivates
that were low on antigens directly coated could be high when the
antigen was indirectly bound. This highlights how important the
characteristics of the assay are when comparing results between
laboratories and patients.
A new assay type: The capture ELISA
for detection of ANCA As mentioned
above, the three-dimensional structure of the antigen may be crucial
for the recognition and binding of antibodies, and thus for the
sensitivity of the assay. In an attempt to preserve the native
structure of the antigen and hence to increase the sensitivity of the
assays, capture ELISAs have been developed. In a capture ELISA
the plate is pre-coated with a monoclonal antibody (MoAb) capturing the
antigen (Fig. 2). A few different MoAbs have been used for the
detection of PR3-ANCA; MoAb 12.8 (23), MoAb lE8 (24) and MoAb 4A3, 4A5
and 6A6, which recognize different epitopes of PR3 (22).
Regarding the latter three, MoAb 4A3 presented the antigen in such a
way that 98% of all antibodies that most likely were against PR3 were
detected. Recently a capture-ELISA based on recombinant PR3 has
been established (25). For MPO-ANCA the MoAb 7.17 has been
developed and evaluated (23).
Sensitivity and specificity When using MoAb 4A3 a high
sensitivity could be established. The
sensitivity of the capture PR3-ANCA was 85% in C-ANCA positive sera
with a specificity of 90% obtained from disease control material
comprising patients with various forms of glomerulonephritis.
Furthermore, the capture PR3-ANCA assay showed a significantly higher
nosographic sensitivity in patients with Wegener's granulomatosis with
renal involvement (41/48, 85%) compared to CANCA by IIF (28/48,58%)
(26). A capture ELISA using MoAb 12.8 was able to detect PR3-ANCA
in 21 out of 24 (79%) patients with Wegener's granulomatosis with renal
involvement. The specificity was found to be 96% (23). MoAb
lE8 was evaluated by Merkel et al. (24), who reported a somewhat lower
sensitivity for the capture ELISA compared to the direct ELISA
method. Out of 26 known ANCA-positive sera from patients with
Wegener's granulomatosis 23 were positive, while 25 out of 26 were
positive with the direct ELISA method (24). However, in this
study the same monoclonal anti-PR3 antibody was used as a control for
background and for anti-mouse antibodies. Such a procedure may
result in a 'false low' level since it has been shown that patients
with Wegener's granulomatosis have complexes consisting of PR3 and
PR3-ANCA, which can bind to the monoclonal anti-PR3 antibody by
themselves (27).
The MoAb 7.17 capture ELISA for
detection of MPO-ANCA gave positive
results in sera from 11 out of 35 patients with crescentic
nephritis. Only serum from one out of 41 disease controls
contained MPO-ANCA by this capture ELISA, thus demonstrating the high
specificity of the test (28).
Capture PR3-ANCA detects the epitope
that is most relevant In material comprising 123 patients
with biopsy-confirmed Wegener's
granulomatosis and microscopic polyangiitis with renal involvement, we
have previously shown that the level of PR3-ANCA at diagnosis, measured
by the capture ELISA, correlates to renal outcome (29). Data
revealed that patients with a high level (> 550 U) of PR3-ANCA by
the capture ELISA had a 5-year renal survival of 60%, versus 85% for
the patients with levels below 550U, (p=0.001). There was no such
correlation to renal survival when PR3-ANCA was measured by direct
ELISA, or to MPO-ANCA levels (Fig. 3) (29). We have indications
from on-going studies that the capture assay for detection of PR3-ANCA
is more sensitive for the detection of relapses.
Potential weakness of a capture ELISA Since the capture assay uses a
monoclonal antibody to capture the
antigen, there could be a possibility that the epitope that binds to
the monoclonal antibody is the same as the one that binds to some of
the antibodies in patient sera. However, the results from the
clinical evaluations performed do not support such a hypothesis, since
the sensitivity is high by the capture method. It is also known
that most patients react to several epitopes. Another argument
against using a capture ELISA is the risk for non-specific binding and
rheumatoid factor binding to the monoclonal antibody. This
possibility of a false positive test result could easily be excluded by
using a control plate coated with an unrelated monoclonal antibody of
the same mouse Ig subclass as that used in the test assay Therefore one
can conclude that a capture ELISA for detection of PR3-ANCA has a
higher sensitivity, and an acceptable specificity. Regarding the
specificity it is important to set the cut-off level using disease
control material comprising patients with, for example, lupus, and not
healthy blood-donors.
Release of granule enzymes during
coagulation and the level of alpha I
antitrypsin A further complication when
conducting ANCA assays is the release of
granule enzymes during coagulation, which can give rise to a reduction
of the test results by binding the antibody and forming an immune
complex (22). To complicate the matter further, the proteinase
inhibitor alpha I antitrypsin (α IAT) can also bind serine proteases
and thus compete with the antibody. The level of α 1AT can
fluctuate in a patient. One can speculate that a patient with a
low level of α 1AT and a high release of enzymes can be low in the test
one day and high another day when centrifugation is done immediately
after the blood is drawn.
Low level by
capture ELISA
_____________________________ High level by
capture ELISA
---------------------------------------------------
LowlevelbydirectELISA
_.._.._.._.._.._.._.._.._.._.._.._.._.._ High level by
direct ELISA
………………………………………..
Fig. 3. The capture assay correlates
better with renal survival than
the direct assay, Patients with a high or low level of PR3-ANCA by
capture or direct ELISA at presentation were followed over time.
Correlation of ELISA to IIF A study was recently carried out to
compare IIF and ELISA for the
detection of ANCA in systemic vasculitis (30). The study
highlights the problem with IIF and the reason for the low specificity
of IIF.
Three commercial kits for IIF and 7
commercial kits for ELISA were
tested and their potential to detect ANCA in systemic vasculitis was
compared. Serum samples from 396 patients with various forms of
glomerulonephritis, including 146 patients with pauci-immune crescentic
glomerulonephritis, were analyzed. It was shown that IIF had an
overall sensitivity of 88% with a specificity of 71 %, while the ELISA
had a sensitivity of 84% with a specificity of 94%. Furthermore,
the authors could show that the problem with IIF was the presence of
samples from patients with SLE that gave a P-ANCA pattern in 68-82%,
thus demonstrating what has already been discussed - that IIF detects
many specificities while ELISA detects antibodies to the coated antigen
only.
By combining IIF and ELISA results a
higher specificity can be
obtained, as was shown in a European study (31), but in this study it
was marginally increased to 96%. The authors also discuss the
predictive value of ANCA testing and show that in a patient with strong
evidence for rapidly progressive glomerulonephritis a positive test by
ELISA provides > 90% confirmation of crescentic glomerulonephritis,
while in a patient with weak clinical evidence a positive ANCA is
inadequate to confirm a diagnosis of crescentic
glomerulonephritis. A negative test, however, rules out the
presence of crescentic glomerulonephritis.
Clinical use of ANCA ANCA in renal disease The clinical setting where ANCA
testing provides the most valuable
contribution is probably renal insufficiency of unknown origin.
The outcome of an ANCA test dramatically changes the odds as to whether
the patient will benefit from treatment with immunosuppressive drugs or
not. The likelihood that the patient has such a disease to begin
with probably ranges between 5-30%, depending on the symptoms and
signs. If antibodies are detected, the likelihood for a disease
that normally benefits from immunosuppression increases to >
95%. One reason for this high probability is that some of the
diseases that may yield 'false positive results', such as SLE or
mesangiocapillary glomerulonephritis with crescents, also benefit from
such treatment. Rapid tests are available that can be
administered at the bedside and yield results while the patient is
still in the emergency room. If the test result is positive, the
treatment can be started immediately, but patients (especially those
with low levels) should undergo a careful work-up. Thus, in the
setting of renal insufficiency of unknown origin it is important to use
a test with high sensitivity. On the other hand, a negative
result for the ANCA test will drastically decrease the probability of a
diagnosis that requires the immediate institution of immunosuppression,
especially if anti-GBM antibodies are analyzed simultaneously.
Septic endocarditis with renal involvement may yield false positive
results and is probably the most important differential diagnosis to be
kept in mind.
ANCA in systemic inflammatory disease Another clinical setting where ANCA
testing is often used is in
patients with inflammatory conditions of unknown origin. Typical
symptoms are malaise and fatigue in combination with myalgia, elevated
ESR or CRP and low grade fever. The differential diagnosis
includes many types of infections and malignancies, conditions that
deteriorate when treated with immunosuppressive drugs. When no
life-threatening organ involvement is present, there is time for
thorough diagnostic procedures. Under such circumstances the high
specificity of the test for systemic vasculitis is more important than
high sensitivity. The highest specificity is obtained when the
results from ELISA and IIF are combined.
ANCA in inflammatory bowel disease It
is sometimes difficult to
distinguish the inflammatory bowel diseases (IBD) ulcerative colitis
and Mb Crohn from each other. Both are characterized by an
unknown etiology as well as chronic remitting inflammatory processes of
the intestine. The established differentiation between Mb Crohn
and ulcerative colitis is made by clinical parameters.
Serological parameters would be helpful for the clinician in diagnosing
and categorizing patients with ulcerative colitis from Mb Crohn.
Today the use of ANCA to detect ulcerative colitis is under discussion
since ANCA, especially of the P-ANCA (or atypical ANCA) type, is found
in 40-60% in these patients. In Mb Crohn, antibodies to the
mannan of Saccharomyces cerevisiae (ASCA) can be detected in 60-70% of
patients. A patient with a clinical suspicion of IBD who has
P-ANCA is more likely to have ulcerative colitis than if the serum
contains ASCA(32,33).
ANCA in cystic fibrosis Cystic fibrosis (CF) is a genetic
disease characterized by hyperviscous
mucus and chronic suppurative lung infections. ANCA is a
relatively common finding in CF and BPI is the predominant
antigen. BPI is active against gram negative bacteria.
Recent reports have shown that there is a correlation between chronic
colonization with Pseudomonas and BPI-ANCA. Antibodies from CF
patients have been shown to react with the C-terminal portion of BPI
that is necessary for oppsonization. The autoantibodies were also
shown to inhibit the opsonization process. A negative correlation
between BPI-ANCA titre and lung function was also detected. Thus
it is possible in the future that treatment that would reduce the
antibody production may prove beneficial in CF. Today it is too
early to conclude that the ANCA test is of clinical value (34).
ANCA for the follow-up of-patients
with ANCA associated vasculitis Relapse is a common feature in
systemic vasculitis and has been
described as late as 15 years after the initial onset of disease.
Approximately 50% of the patients who are alive 5 years after a
diagnosis of small vessel vasculitis associated with PR3-ANCA have
experienced at least one relapse (35); lower figures have been reported
in MPO-ANCA associated disease (36). It is widely accepted that
relapses occur more often among patients who remain ANCA-positive
during remission. However, different views prevail concerning the
usefulness of serial ANCA measurements for the follow-up of patients
with a systemic vasculitis that was positive at the time of
diagnosis. In the original report linking ANCA to Wegener's
granulomatosis in 1985, it was stated that ANCA titres are related to
disease activity (4). Later it was claimed that treatments based
on ANCA titres are more beneficial compared to treatments based only on
clinical signs (37). This notion has been challenged, for
instance by a report from the NIH group who found changes in ANCA
titres to correlate poorly to disease activity (38, 39). Overall
there seems to be about a 50% risk of relapse within 6 months after a
significant rise in ANCA titre. A high degree of alertness is
always indicated when the level is rising. In most published
series (4, 36-39) more than 90% of the cases have positive test results
at the time of relapse. A negative ANCA result is a good
indication that the disease is under control.
ANCA during renal replacement therapy Uremia necessitating renal
replacement therapy is a common sequel of
systemic vasculitis. During dialysis treatment relapses can be
difficult to diagnose. General symptoms of vasculitis are easily
confused with symptoms of uremia, and specific symptoms such as
hematuria and raised serum creatinine concentrations cannot be
used. Furthermore, ANCA positivity by IIF have been found in 5%
of dialysis patients without any evidence of vasculitis disease
(40). Despite the immunosuppressive effect of uremia, a relapse
rate close to 10% per year have been reported for dialysis-treated
patients (41).
Relapses also occur after renal
transplantation, both extra-renal and
in the graft. Neither ANCA specificity at diagnosis (PR3 or MPO)
nor the ANCA test status at the time of transplantation seem to
influence the risk of relapse. In a recent meta-analysis we found
the average risk of relapse to be around 20% (42). Many of these
were minor extra-renal relapses and the overall graft survival is
reported to be similar to other forms of renal disease. It seems
as if modem immunosuppression protocols, used to prevent rejection, are
also active in preventing relapses.
Fig. 4. The optimal way to perform
ANCA assavs is for the clinician to
indicate whether the results are going to be used to diagnose systemic
vasculitis or other inflammatory conditions. ELISA for PR3- and
MPO-ANCA is the best choice if systemic vasculitis is suspected; if
other inflammatory conditions are suspected, IIF is to be preferred.
ANCA strategy Because of the complications in the
analysis of ANCA, an international
group of ANCA researchers has published a consensus statement on ANCA
testing (43). In this publication guidelines for ANCA testing and
reporting are presented. The group states that a minimum
requirement for a laboratory that screens with IIF should be to never
send out positive results without a confirmation by antigen-specific
ELISA, i.e. for PR3 and MPO, since ANA will react with neutrophils and
give a positive ANCA. The optimal recommendation is to use both
IIF and ELISA on all samples. Results should be reported as
C-ANCA, PANCA or atypical ANCA for assays carried out by IIF, and as
PR3-ANCA, MPO-ANCA etc. for assays done by ELISA.
Our own view is that the tests to be
used depends on the clinical
question being asked (Fig. 4). If the clinician is interested in
diagnosing systemic vasculitis with renal involvement, then ELISA using
PR3 and MPO is the method of choice. If the interest is arthritis
or inflammatory bowel disease, IIF on granulocytes is the method of
choice. When systemic vasculitis is suspected, ELISA for PR3-ANCA
and MPO-ANCA is performed and if the sample is positive the result is
sent out. If the result is negative a test for anti-GBM
antibodies is warranted since patients with systemic vasculitis and
Goodpasture syndrome can present in the same way. An IIF test for
ANCA can be performed to rule out ANCA since a few samples will be
positive by IIF and not by ELISA, but most systemic vasculitis patients
will be positive by ELISA. The clinical value of a positive
C-ANCA that is negative for antibodies to PR3 or MPO in the context of
systemic vasculitis is not known. In systemic vasculitis patients
and during follow-up, capture assays may be useful since they seem to
detect relapses very well.
ANCA of specificities other than PR3
or MPO are of limited value for
patients with systemic vasculitis and occur primarily in patients with
other inflammatory disorders such as ulcerative colitis, Mb Crohn,
autoimmune hepatitis, etc. It is of little value to know the
specificity of these esoteric ANCAs, except perhaps when studying
cystic fibrosis.
Acknowledgements The studies referred to in this
review were supported by grants from
the Swedish Medical Research Council and Riks-forbundet Njursjukas
fbrening.
