Clinical Chemistry 43: 824-831, 1997;
(Clinical Chemistry. 1997;43:824-831.)
© 1997 American Association for Clinical Chemistry, Inc.
Modified and improved anti-acetylcholine receptor (AChR) antibody assay: comparison of analytical and clinical performance with conventional anti-AChR antibody assay
Bruno Ferrero1,a,
Giuseppe Aimo2,
Roberto Pagni2,
Bruno Bergamasco1,
Maria R. Bongioanni1,
Lodovico Bergamini1 and
Luca Durelli1
1
Department of Neurosciences, and
2
Laboratory "Baldi e Riberi", University Hospital "Molinette", Via Cherasco 15, I-10126, Turin, Italy.
a Author for correspondence. Fax Int +31 11 6963487; e-mail chinigo{at}golgi.molinette.unito.it
 |
Abstract
|
|---|
We developed a modified anti-acetylcholine receptor (AChR) antibody (Ab)
assay based on a radioreceptor assay and a calibration curve. We
compared the analytical and clinical performances of this modified
assay with those of the conventional anti-AChR Ab radioreceptor assay.
Serum specimens were from patients with myasthenia gravis (MG) (n
= 156) and from control subjects (n = 106). The modified assay
demonstrated lower within-assay (4.0–6.6%) and between-assay
(5.3–7.8%) CVs, greater linearity, lower cost, and shorter assay time
than the conventional method. ROC curve analysis indicated almost
identical specificity and sensitivity (>0.92) for these two anti-AChR
Ab assays. The modified and conventional assays were also equivalent
for blocking anti-AChR Ab assay. Moreover, the modified anti-AChR Ab
assay, differently from the conventional assay, allowed us to reveal
anti-AChR Ab concentration differences among different clinical grades
of MG.
Key Words: indexing terms: myasthenia gravis • radioreceptor assay • intermethod comparison
|
Introduction
|
|---|
Myasthenia gravis (MG) is an autoimmune disease characterized by
fluctuating muscular weakness that usually becomes more evident upon
repetitive motor activity
(1).1
Anti-acetylcholine receptor (AChR) antibody (Ab) is involved in the
pathogenesis of the disease. It causes the loss of functional AChRs at
the postsynaptic membrane that underlies the defect of neuromuscular
transmission (2)(3)(4). Circulating anti-AChR Ab can be
detected in 75–90% of myasthenic patients
(3)(5)(6), and detection of these
Abs is useful for diagnosis and follow-up
(7)(8). The exact relation between antibody
concentration and clinical state is not clearly established as yet,
although concentrations are higher in severe generalized MG
(5)(9), and clinical improvement is usually
associated with a decrease of Ab concentration (7).
Conventional radioreceptor assay methods for determining
anti-AChR Ab concentration are lengthy and expensive (8).
For each assay, at least four different radiolabeled human (or fetal
calf) AChR amounts and four different rabbit anti-human IgG amounts, in
duplicate, are usually required to immunoprecipitate the anti-AChR Ab
and radiolabeled AChR complexes. For this reason, this method can be
rarely used for frequently repeated determination of longitudinal Ab
concentration variation. The great between-assay variation further
reduces its use for serial Ab studies. We developed a faster, easier,
and more reproducible technique for anti-AChR Ab detection, by
interpolating unknown anti-AChR Ab values from a radioreceptor assay
calibration curve (10)(11)(12)(13). Here we describe this
technique, its analytical performance, and the results, obtained over 5
years from 156 myasthenic patients and 106 nonmyasthenic subjects.
Comparison of results obtained with our technique and with the
conventional method demonstrates that our technique is a reliable and
cheap alternative to the conventional anti-AChR Ab assay, and suggests
its use for serial anti-AChR Ab determination.
|
Materials and Methods
|
|---|
subjects
We studied 106 nonmyasthenic subjects and 156 myasthenic patients.
Nonmyasthenic subjects (n = 106, 49 females, 57 males, median age
45 years, range 15–73) included healthy controls (n = 19),
patients with nonneurologic autoimmune diseases (n = 11), and
patients with nonmyasthenic neurologic diseases (n = 76). In
myasthenic patients (n = 156, 87 females, 69 males, median age 38
years, range 9–74), diagnosis was established according to appropriate
clinical history and positive response to anticholinesterase drugs
(edrophonium in 148 patients or piridostigmine in 8). In 89 myasthenic
patients, diagnosis was electromyographically confirmed by the
decremental response to repetitive nerve stimulation at 3 Hz
(1). Patients were graded according to Osserman's
classification (14): remission (n = 21 patients),
grade I (pure ocular disease, n = 31), grade II-A (mild
generalized disease, n = 32), grade II-B (moderatly severe
generalized disease, with bulbar symptoms, n = 45), grade III
(acute disease, with severe bulbar symptoms, and often rapid
progression to respiratory insufficiency, n = 16), and grade IV
(chronic severe disease, n = 11). Among the 156 myasthenic
patients, 73 had been thymectomized 1–8 years before the study (11 had
a thymoma: four women, seven men, median age 57 years, range 49–74),
and 85 were treated with immunosuppressive drugs. Serum aliquots
obtained from myasthenic patients and controls were stored at
-70 °C until used.
The procedures followed were in accordance with the Helsinki
Declaration of 1975, as revised in 1983.
achr preparation
Fetal calf muscle was immediately dissected after sacrifice and
frozen at -70 °C until used. The muscle was homogenized in a
blender apparatus for 1 min in four volumes of a cold buffer containing
50 mmol/L NaCl, 2 mmol/L sodium EDTA, 1 mmol/L sodium ethylene
glycol-bis (ß-aminoethyl ether)-tetraacetic acid (EGTA), 50 mmol/L
Tris-HCl, pH 7.4 (buffer A), and 2 mmol/L phenylmethylsulfonyl fluoride
(Sigma Chemical Co., St. Louis, MO). Homogenate was centrifuged at
40 000g for 30 min at 4 °C (J2–21 M/E Centrifuge;
Beckman Instruments, Fullerton, CA). Pellets were recovered, diluted
1:2 with buffer A, rehomogenized, and then Triton X-100 was added to a
final concentration of 20 mL/L. The mixture was gently stirred for
3 h at 4 °C. After the extraction, the material was centrifuged
at 40 000g for 1 h at 4 °C. The supernatant was
recovered and filtered through a glass wool column to remove lipids.
Final AChR content from crude extract was determined by the
diethylaminoethyl cellulose (DE-81; Whatman, Maidstone, UK) disc assay
and expressed as nanomoles of AChR per liter of extract
(8)(15). We used fetal calf AChR as antigen
for the anti-AChR Ab assay because (a) it is easier to
obtain than human AChR; (b) antigen–Ab reactivity shows
little variability among different AChR preparations lower than that of
human AChR (9); and (c) fetal calf AChR has
been demonstrated to react with human anti-AChR Abs to the same extent
or better than the human antigen (8).
anti-achr ab assay
Anti-AChR Abs detected in most myasthenic patients are directed
against epitopes located outside the ACh binding site (binding
anti-AChR Ab) (3). Abs directed against the ACh binding
site can be detected in ~40% of myasthenic patients: These Abs block
iodine 
-bungarotoxin ([125I]BGT) binding to AChR
(blocking anti-AChR Ab) (2). Therefore, a complete
serologic diagnostic test for MG requires determination of both
blocking and binding anti-AChR Ab.
Conventional radioreceptor assay for binding anti-AChR Ab.
All sample sera were diluted 1:10 in PBS + 1 mL/L Triton X-100.
Fetal calf AChR, diluted at 1 nmol/L in buffer A, was radiolabeled with
a 2 molar excess of [125I]BGT (specific activity 555.0
kBq/µg; DuPont, Wilmington, DE) for 4 h at room
temperature and dispensed in 1-mL aliquots. Each aliquot was incubated
in duplicate with increasing amounts (5, 10, 25, 50 µL) of sample
serum diluted with normal serum (respectively 45, 40, 25, 0 µL) to a
final volume of 50 µL to reach the same immunoglobulin concentration
in all samples and to reduce the analytical imprecision due to
different biological matrices (16). Incubation was
performed overnight at 4 °C. Therefore, 100 µL of a 0.75% rabbit
anti-human IgG solution (Dako, Glostrup, Denmark) was added to each
sample to give a complete Ab precipitation
(3)(8). After 2 h of incubation at room
temperature, samples were centrifuged at 4000g for 30 min at
4 °C and then washed twice in PBS + 5 mL/L Triton X-100 to remove
unbound materials. After each wash, centrifugation was repeated at
4000g for 20 min at 4 °C. Final pellet radioactivity,
proportional to precipitated receptor–toxin immune complexes, was
revealed by a gamma counter (LKB Multigamma 1261; LKB, Uppsala,
Sweden) and expressed as counts per minute (cpm). The slope of the
straight line, obtained by plotting pellet radioactivity against the
respective sample serum amount, was calculated by linear regression
analysis. When the slope showed a positive value but r was
<0.9, anti-AChR Ab was always retested. Each slope, expressed as
cpm/µL of serum, was multiplied by the sample serum
dilution used and then divided by the cpm of 1 pmol of
[125I]BGT. The result, expressed as pmol of anti-AChR
Ab/µL of sample serum, was multiplied by
101
, obtaining the concentration of anti-AChR Ab, usually
expressed as nmol/L. To validate the analytical procedure, in each
group of determinations four different nonmyasthenic and two myasthenic
sera of well-known concentrations were tested as negative and positive
controls, respectively.
Modified radioreceptor assay for binding anti-AChR Ab.
A
sample of pooled sera from 15 healthy subjects was reassayed >20 times
by the conventional anti-AChR Ab method. Being always negative for
anti-AChR Ab, it was chosen as the zero calibrator and used for
diluting patient sera. A sample of pooled sera from six myasthenics
with high-concentration anti-AChR Ab (median value 71.7 nmol/L, range
52.2–112.1) was serially diluted ninefold from 1:10 to 1:2560 with the
zero calibrator. Each dilution was stored in 50-µL aliquots at
-70 °C, and anti-AChR Ab concentration (ranging from 76.8 nmol/L
for the 1:10 dilution to 0.3 nmol/L for the 1:2560 dilution) was
calculated by taking the mean of 10 conventional Ab determinations.
Mean and SD of anti-AChR Ab concentrations (nmol/L) of these nine
calibration dilutions were, respectively, 76.8 ± 6.51, 38.4
± 2.95, 19.2 ± 1.33, 9.6 ± 0.89, 4.8 ± 0.41,
2.4 ± 0.19, 1.2 ± 0.11, 0.6 ± 0.05, and 0.3 ±
0.03. These nine myasthenic serum dilutions, in duplicate, were used to
obtain a calibration curve, by plotting, for each dilution, the mean of
the duplicate values of revealed bound radioactivity against the
corresponding anti-AChR Ab concentration. Three positive controls at
3.5 ± 0.31, 27.5 ± 2.21, and 61.0 ± 4.93 nmol/L
anti-AChR Ab (means ± SD of 10 conventional assays) and two
negative controls (tested by 10 conventional assays) were stored in
50-µL aliquots at -70 °C. Fifty microliters of the zero
calibrator, nine calibration serum dilutions, and three positive and
two negative controls and unknown samples were incubated overnight at
4 °C with 1 mL of [125]IBGT-AChR complex (at 1 nmol/L
in buffer A), obtained as in the conventional method. Finally,
toxin–receptor–Ab complexes were precipitated by rabbit anti-human
IgG and the bound radioactivity counted, as described above. All tests
were performed in duplicate and the anti-AChR antibody concentration
was calculated by the average of duplicate antibody concentrations: If
duplicate antibody concentrations differed from their own mean by
>10% (mean ± 3 SD of the within-assay precision of the modified
anti-AChR Ab assay; this value was chosen to exclude incorrect Ab
values with the probability of 99%), the sample was always retested.
The anti-AChR Ab assay was repeated if only one of the three positive
control serum Ab concentrations, measured by the modified assay,
differed by more than the mean ± SD of Ab concentrations
determined by 10 conventional assays (i.e., if the Ab concentration of
the 3.5 ± 0.31 nmol/L control serum is <3.29 nmol/L or >3.81
nmol/L, the modified assay must be repeated). Patient anti-AChR Ab
concentrations were determined by using Wiacalc (Pharmacia-LKB,
Uppsala, Sweden) software, by comparing the unknown sample
radioactivity (average of duplicate values) with the same radioactivity
on the calibration curve and interpolating the corresponding Ab value
(expressed in nmol/L). This modified anti-AChR Ab assay method includes
a four-parameter polynomial logistic regression statistical model
(17) to construct the calibration curve. The equation
utilized by this fitting algorithm has the general form: response
= D + (D - A)/[1 + (concentration/C)B], where A, B,
C, D are, respectively, the reference estimate (cpm), the slope factor,
the turning point (nmol/L), and the estimated blank (cpm); the response
is expressed as measured cpm and the concentration as nmol/L. This
curve is fitted to the means of the measured calibration points by
means of a weighted least-squares method; such an analysis procedure
minimizes errors related to an excessively theoretic or experimental
calibration curve (11)(13). Fig. 1
shows an example of the calibration curve obtained by a zero
calibrator and nine calibration points (from 0.3 to 76.8 nmol/L).
View larger version (25K):
[in this window]
[in a new window]
|
Figure 1. Example of anti-AChR Ab calibration curve obtained by nine
dilutions of a 76.8 nmol/L calibrator and a four-parameter polynomial
logistic regression analysis.
|
|
Conventional radioreceptor assay for blocking anti-AChR
Ab.
The procedure to assay serum Ab directed against the ACh
binding site (able to block BGT binding to AChR)
(3)(5) was very similar to the binding
anti-AChR Ab assay described above. The only difference was that the
blocking anti-AChR Ab concentration was determined by incubating AChR
overnight at 4 °C with increasing amounts (5, 10, 25, 50 µL) of
myasthenic serum, and then with a 2-molar excess of
[125I]BGT for 4 h at room temperature (in opposite
order to the binding anti-AChR Ab assay). Toxin–receptor–Ab complexes
were then precipitated by rabbit anti-human IgG and counted as
described above. Ab concentration (nmol/L) was calculated by linear
regression analysis by using the four different dilutions in duplicate.
This value is the total anti-AChR Ab concentration diminished by the
amount of blocking anti-AChR Ab; therefore, the blocking anti-AChR Ab
concentration can be calculated as the difference between the binding
anti-AChR Ab concentration and this value.
Modified radioreceptor assay for blocking anti-AChR Ab.
AChR was incubated overnight at 4 °C with 50 µL of myasthenic
serum, and then with a 2-molar excess of [125I[BGT for
4 h at room temperature (in opposite order to the anti-AChR Ab
binding assay procedure). Toxin–receptor–Ab complexes were then
precipitated by rabbit anti-human IgG, counted as described above, and
the Ab concentration (nmol/L) was interpolated from an anti-AChR Ab
calibration curve, obtained as described above. This value is the total
anti-AChR Ab concentration diminished by the amount of the blocking
anti-AChR Ab concentration; therefore, the blocking anti-AChR Ab
concentration can be calculated as the difference between the binding
anti-AChR Ab concentration and this value.
Analytical evaluation.
The performance of the
conventional and modified methods for the anti-AChR Ab assay was
assessed by evaluating: (a) the detection limit of the
assays (defined as the lowest concentration of anti-AChR Ab that is >4
SD than the mean signal of 70 control sera) (3);
(b) within-assay (three serum samples at anti-AChR Ab
concentrations of ~1.8, 11, and 58 nmol/L in 10 replicates assayed
three times) and between-assay (the same serum samples run in duplicate
in 15 consecutive assays over >120 days) precision, as well as
construction of a precision profile (derived from the duplicate
measurements of unknown anti-AChR Ab concentration samples processed
during the present study); (c) reproducibility of the
calibration curve of the modified method: When the anti-AChR Ab
concentration of any one of the calibrators or positive controls,
determined with Wiacalc software, differed by more than ± 2 SD
from the average concentration of the 10 conventional assays on the
same sample, the assay run was always repeated; and (d)
accuracy testing, comprising checks of linearity (determined with one
serially diluted serum sample in triplicate).
statistics
If data distribution was gaussian and variances equal, values were
compared by Student's t-test and analysis of variance. When
data distribution was not gaussian, Fisher's exact test and
Mann–Whitney U-test were used to compare two or more
groups. All tests were two-tailed and the level of significance was set
at P <0.05. For all sample groups we calculated the mean,
SD, median, minimum, and maximum serum concentrations and the
percentages of anti-AChR Ab values that exceeded the cutoff value of
0.4 nmol/L. Diagnostic test performance was further evaluated by
documenting the diagnostic accuracy of each binding anti-AChR Ab test
with ROC curves and calculation of the areas under these curves (AUC)
(18). Antibody concentrations, calculated in the same
myasthenic sera by using the two anti-AChR Ab assay methods, were
compared by linear regression analysis and used to evaluate the
agreement between the two methods, estimating the "limits of
agreement" (19). Statistical analyses were done with the
Astute statistics computer program (DDU software; The University of
Leeds, Leeds, UK). The calibration curve was obtained with the weighted
least-squares method by using the Wiacalc computer program.
|
Results
|
|---|
muscle achr content
AChR content in the crude Triton X-100 extract varied, with
different muscle samples, between 2.2 and 8.5 nmol/L. To determine the
optimal AChR concentration to be used as antigen for the anti-AChR Ab
assay, three different myasthenic sera (whose anti-AChR Ab
concentrations were around the mean of the majority of myasthenic sera
concentrations) were tested in quadruplicate with different dilutions
(from 0.25 to 4 nmol/L) of three different fetal calf AChR samples. The
best AChR concentration needed to precipitate all the anti-AChR Ab in
myasthenic serum was between 0.75 and 1.5 nmol/L, in keeping with other
reports (5)(8). For this reason we always used
the AChR concentration of 1 nmol/L.
analytical evaluation of the two different binding anti-achr ab
assay methods
Precision.
Anti-AChR Ab concentrations of three
Ab-positive myasthenic sera (at low, middle, and high Ab concentration)
were assayed by both the conventional and modified methods in 10
replicates to assess within-assay precision. In addition, the same
serum samples were run in 15 consecutive assays over >120 days to
assess between-assay precision. These variabilities are evaluated as
mean CVs (Table 1
). Within-assay CVs ranged from 8.4% to 11.6% (median 9.9%)
with the conventional method, and from 4.0% to 6.6% (median 5.5%)
with the modified method. Between-assay CVs ranged from 12.2% to
14.7% (median 12.3%) with the conventional method, and from 5.3% to
7.2% (median 6.6%) with the modified method. Fig. 2
shows the precision profiles of the two assay methods. The
first three ranges of anti-AChR Ab concentrations 
2.2 nmol/L (0–0.4,
0.4–0.9, and 0.9–2.2 nmol/L) showed CVs that decreased from 19.7% to
9.8% with the conventional method and from 14.2% to 6.1% with the
modified method. The reproducibility of the modified procedure for the
remaining part of the range of the test (2.2–245.6 nmol/L) was always
much lower (range 3.8–5.9%) than that obtained with the conventional
procedure (range 8.5–10.5%).
View this table:
[in this window]
[in a new window]
|
Table 1. Within- and between-assay precision of the conventional
and modified anti-AChR Ab assay methods, with three different
myasthenic sera (at low, medium, and high Ab
concentration).
|
|
Reproducibility of the modified assay.
With the modified
method, a high calibration curve reproducibility was shown in >98% of
assay runs. In fact, only 2 of 148 calibration curves, over 5 years of
assays, were rejected (and therefore repeated) because only one of the
positive control Ab values differed by more than ± SD from the
average value of 10 conventional assays on the same sample.
Test sensitivity and specificity.
Before developing the
modified anti-AChR Ab method described here, we established the
anti-AChR Ab cutoff value for our laboratory with the conventional
assay and 70 control sera (19 healthy and 51 nonmyasthenic patients).
The Ab values distribution was approximately normal, with a range of
0–0.54, mean 0.072, and SD 0.079 nmol/L. Therefore we chose 0.4 nmol/L
(mean + 4 SD) (3)(5) as the cutoff value of
the assay, the Ab range below such a cutoff value including >99.9% of
the healthy subjects. We recalculated the cutoff value both for the
conventional and modified assays by using the data of the 106
nonmyasthenic subjects presented here: The cutoff value of 0.4 nmol/L
was always effective for both methods. On the basis of this cutoff
value, all sera were divided in anti-AChR Ab "positive" (>0.4
nmol/L) or "negative" (0.4 nmol/L). Mean, SD, and range of
anti-AChR Ab concentrations of 156 myasthenic patients and 106
nonmyasthenic subjects determined with the two assay methods are
presented in Table 2
. Test percent sensitivity, i.e., the percentage of myasthenic
sera positive for anti-AChR Ab, was 125 of 156 patients (80.1%), being
higher in patients with active disease (114 of 135 patients: 84.4%)
than in patients in remission (11 of 21 patients: 52.4%) (P
<0.001). It was also higher in patients with generalized MG (96 of 104
patients: 92.3%) than in patients in remission (11 of 21 patients:
52.4%) (P <0.001) or in pure ocular myasthenic patients
(18 of 31 patients: 58.1%) (P <0.005). The difference
between the number of positive sera in patients with or without MG was
highly significant (P <0.001). There was no significant
difference in test percent sensitivity between the conventional and
modified assays. Although Ab concentration range of different MG
clinical grades stretched over two to three orders of magnitude,
anti-AChR Ab concentration of patients in remission or with pure ocular
MG was lower (P <0.01) than that of patients with
generalized MG, with both methods. The anti-AChR Ab concentration of
mild generalized myasthenic patients (grade II-A) was lower
(P <0.05) than that of moderately severe generalized
myasthenic patients (grade II-B) with only the modified method. Between
the other grades of generalized MG, Ab concentrations were not
significantly different. Test specificity can be expressed as the ratio
between the number of nonmyasthenic sera without detectable anti-AChR
Ab and the total number of nonmyasthenic sera tested. Among 106
subjects without MG, a positive ("false positive") anti-AChR Ab
serum was found in one rheumatoid arthritis patient treated with
penicillamine (5)(8), with both methods. As
shown in Table 2
, specificity of both methods was 99.1% (105 of 106
control subjects).
Linearity.
The myasthenic calibration serum was serially
diluted ninefold from 1:10 to 1:2560 in the zero calibrator to reach
the same immunoglobulin concentration in all samples. Anti-AChR Ab
assay was performed for all dilutions in triplicate with both anti-AChR
Ab assay methods. The recoveries of calibrator myasthenic serum
dilutions, i.e., (anti-AChR Ab measured/expected) x 100, were better
for the modified method (ranging between 96% and 107%) than for the
conventional method (ranging between 87% and 123%). Linear regression
analysis of observed vs expected Ab concentrations revealed highly
significant correlation (r >0.99) and a slope near unity
for both methods.
ROC curve analysis.
Fig. 3
represents the ROC curves and the corresponding AUCs for the
conventional and modified anti-AChR Ab assay methods. The AUCs of both
methods are >0.92, without significant difference between each other.
View larger version (19K):
[in this window]
[in a new window]
|
Figure 3. ROC curves of the conventional and modified
anti-AChR Ab assays for cases of MG (n = 156) vs
nonmyasthenic controls (n = 106).
|
|
Agreement between the conventional and modified anti-AChR Ab
assays.
The linear correlations between the results obtained with
the conventional and modified methods in the 156 myasthenic patients
showed a slope of 0.94 and a correlation coefficient of 0.992. To
estimate the agreement between the two anti-AChR Ab assays
(19), we plotted the differences between Ab concentrations
obtained with both methods vs their means. Because Ab differences were
proportional to the mean, a logarithmic transformation of data was done
(Fig. 4
). Distribution of log-transformed data was approximately normal
with a mean difference of -0.025 and a SD of 0.098 on the log scale.
Therefore, the lower and upper limits of agreement at 95% are
respectively 0.167 and -0.217. These limits can be calculated as
mean ± 1.96 SD. The antilogs of these limits are 1.182 and 0.805,
respectively. In ~95% of cases the anti-AChR Ab concentration
measured with the modified assay will therefore be 18.2% higher to
19.9% lower than the Ab concentration measured with the conventional
assay.
View larger version (18K):
[in this window]
[in a new window]
|
Figure 4. Scatter plot of the differences between anti-AChR Ab
concentrations obtained with the conventional and modified methods vs
their means.
|
|
blocking anti-achr ab
We determined, by both the conventional and modified methods, the
blocking Ab concentration of 106 nonmyasthenic and 156 myasthenic sera
(already assayed for binding anti-AChR Ab). The number of myasthenic
sera positive for blocking anti-AChR Ab, with both methods, was 97 of
156 patients (62.2%), being higher in patients with active disease (90
of 135 patients: 66.7%) than in patients in remission (7 of 21
patients: 33.3%) (P <0.001), and in patients with
generalized MG (77 of 104 patients: 74.0%) than in patients in
remission (7 of 21 patients: 33.3%) (P <0.001) or in
pure ocular myasthenic patients (13 of 31 patients: 41.9%)
(P <0.01). Among 106 nonmyasthenic subjects, no
blocking Ab-positive ("false positive") serum was found
(specificity 100%). The difference between the number of positive sera
in patients with or without MG was highly significant (P
<0.001). The Ab concentration of patients in remission or with pure
ocular MG was lower (P <0.05) than that of patients with
generalized MG, regardless of the method used. As for the other grades
of generalized MG, Ab concentrations were not significantly different.
|
Discussion
|
|---|
The conventional anti-AChR Ab assay method
(3)(5)(8) is a radioreceptor assay
that requires a reaction between an excess of radiolabeled antigen
([125I]BGT-AChR) and various dilutions (usually four or
five in duplicate) of an antibody (anti-AChR Ab) of unknown
concentration. After washing unbound radiolabeled antigen, the antibody
concentration can be calculated from the slope of a regression line
obtained by plotting the precipitated radioactivity vs microliters of
sera used. We describe a modified radioreceptor anti-AChR Ab assay
method based on a calibration curve: The concentration of an unknown
radiolabeled analyte is interpolated by referring the precipitated
radioactivity to a calibration curve obtained by plotting radioactivity
vs concentration of samples of known concentration.
We have demonstrated that this modified method possesses higher
reproducibility and lower within-assay and between-assay variabilities
than the conventional method (Table 1
). In addition, this modified
method displayed an outstanding precision profile across the entire
dynamic range, particularly in the very low concentration range (Fig. 2
), the range where a low precision increases the probability of a
wrong diagnosis of MG. Comparison of the two procedures revealed good
correlation and agreement: The Ab concentration measured with the
modified method may differ from the Ab concentration measured with the
conventional method by 19.5% below to 18.2% above in 95% of cases.
Other important advantages of the modified method compared with the
conventional method are better linearity, lower cost, and greater
rapidity (about one-fourth of reagents and assay time for a single
myasthenic serum assay), with very similar sensitivity (80.1%) and
specificity (99.1%) (Table 2
), as confirmed by ROC curve analysis
(Fig. 3
).
The anti-AChR Ab modified assay method was, in addition, more useful
than the conventional method in the study of the correlation of
clinical state and anti-AChR Ab concentration. The relation between
antibody concentration and disease severity in myasthenic patients is
controversial (3)(20)(21)(22)(23). Although anti-AChR
Ab concentrations are usually higher in severe generalized MG
(9) and clinical improvement is usually associated with
decreased Ab concentration (7), it is not possible to
predict disease severity on the basis of Ab concentration. In
individual patients, serial anti-AChR Ab determinations appear to be
correlated with changes of the clinical state induced by treatments,
such as thymectomy, plasma exchange, corticosteroids, cytostatic drugs,
or total body irradiation (7)(24), but changes
of MG severity may sometimes occur without detectable change of
anti-AChR Ab concentration and vice versa (25). The only
repeatedly confirmed finding was that patients with ocular MG or
patients in remission show either lower anti-AChR Ab concentration or
rarer Ab positivity than patients with active generalized disease
(6)(26). Using either the conventional or
modified anti-AChR Ab assay method, we confirmed a lower percentage of
Ab-positive myasthenic patients and a lower anti-AChR Ab concentration
in patients in remission or with ocular MG than in patients with active
generalized disease. However, with the modified method, but not with
the conventional method, we found a significant difference of anti-AChR
Ab concentration among myasthenic patients of different clinical
severity (namely, grade II-A and grade II-B of Osserman's
classification). Many other studies involving the conventional
anti-AChR Ab method were unable to detect any anti-AChR Ab
concentration difference among different clinical grades of generalized
MG (3)(8)(20). The lower
within-assay and between-assay variabilities of the modified vs
conventional method, reducing an important source of assay imprecision,
make the modified method more useful to describe the correlation
between clinical state and anti-AChR Ab concentration in several
studies of individual myasthenic patients.
In addition, the same analytical and clinical performances of the
modified anti-AChR Ab assay described here were maintained by using, as
immunoprecipitating agent, a 100 g/L solution of Staphylococcus
aureus Protein A (Sigma Chemical Co.) (Staph A) instead of rabbit
anti-human IgG (data not shown) (6)(27).
Advantages of using Staph A include low cost and a short incubation
time (15 min). Only such a short incubation time allows the kinetic
studies needed to calculate binding avidity of anti-AChR Ab to AChR to
be performed (27).
In the blocking Ab assay, we found that some myasthenic patients
positive for binding Ab were also positive for blocking Ab. Vincent and
Newsom-Davis (5) suggested that blocking Abs could
displace BGT binding to AChR, reducing the precipitated radioactivity
at the anti-AChR binding Ab assay and increasing the number of false
"seronegative" myasthenic patients. Differently from Vincent and
Newsom-Davis and other authors (28), we found no patient
positive for blocking and negative for binding Ab. In our experience,
therefore, the use of both binding and blocking Ab assays did not
result in a reduced number of anti-AChR Ab "seronegative"
myasthenic patients. The modified method must be, however, preferred to
the conventional one also for blocking Ab assay because of its lower
cost and greater rapidity, without difference in specificity and
sensitivity.
In summary, the modified anti-AChR Ab assay proposed in this study
displays higher precision and linearity than the conventional method,
maintaining the same sensitivity and specificity. In addition, it has a
lower cost, shorter assay time, and greater clinical usefulness than
the conventional method. Therefore we suggest that this method be used
when numerous repeated determinations are needed. For example, this
method can be used for evaluating serial changes of anti-AChR Ab
concentrations in the individual myasthenic patient and the possible
correlation between anti-AChR Ab concentration, clinical state, and
treatments.
|
Acknowledgments
|
|---|
We thank Cecilia Gotti (CNR Center of Cytopharmacology, Department
of Pharmacology, University of Milano) for her advice. This work has
been supported in part by grant no. 292 from Telethon Italia.
|
Footnotes
|
|---|
1 Nonstandard abbreviations: MG, myasthenia gravis; AChR,
acetylcholine receptor; Ab, antibody; BGT, -bungarotoxin; AUC, area
under the curve; and Staph A, Staphylococcus aureus Protein
A. 
|
References
|
|---|
-
Desmedt JE, Borenstein S. Diagnosis of myasthenia gravis by nerve stimulation. Ann N Y Acad Sci 1976;274:174-188.
-
Almon RR, Andrew CG, Appel SN. Serum globulin in myasthenia gravis: inhibition of -bungarotoxin binding to acetylcholine receptors. Science 1974;186:55-57.
[Abstract/Free Full Text]
-
Lindstrom JM, Seybold ME, Lennon VA, Whittingham S, Duane DD. Antibody to acetylcholine receptor in myasthenia gravis. Prevalence, clinical correlates and diagnostic value. Neurology 1976;26:1054-1059.
[Abstract/Free Full Text]
-
Protti MP, Manfredi A, Horton RM, Bellone M, Conti-Tronconi BM. Myasthenia gravis: recognition of a human autoantigen at a molecular level. Immunol Today 1993;14:363-368.
[ISI][Medline]
[Order article via Infotrieve]
-
Vincent A, Newsom-Davis J. Acetylcholine receptor antibody as a diagnostic test for myasthenia gravis: results in 153 validated cases and 2967 diagnostic assays. J Neurol Neurosurg Psychiatry 1985;48:1246-1252.
[Abstract/Free Full Text]
-
Oger J, Kaufman R, Berry K. Acetylcholine receptor antibodies in myasthenia gravis: use of a qualitative assay for diagnostic purposes. Can J Neurol Sci 1987;14:297-302.
[ISI][Medline]
[Order article via Infotrieve]
-
Dau PC, Lindstrom JM, Cassel JK, Denis EH, Shev EE, Spitter LE. Plasmapheresis and immunosuppressive drug therapy in myasthenia gravis. N Engl J Med 1977;297:1134-1140.
[Abstract]
-
Gotti C, Mantegazza R, Clementi F. New antigen for antibody detection in myasthenia gravis. Neurology 1984;34:374-377.
[Abstract/Free Full Text]
-
McAdams MR, Roses AD. Comparison of antigenic sources for acetylcholine receptor antibody assays in myasthenia gravis. Ann Neurol 1980;8:61-66.
[ISI][Medline]
[Order article via Infotrieve]
-
Rodbard D, Lewald JE. Computer analysis of radioligand assay data. Methods Enzymol 1975;37:3-22.
-
Naus AJ, Kuppens PS, Borst A. Calculation of radioimmunoassay standard curves. Clin Chem 1977;23:1624-1627.
[Abstract/Free Full Text]
-
McDonald M. Calculation program for weighted logit-log RIA data reduction. Clin Chem 1981;23:1624-1629.
-
Raab G. Comparison of a logistic and a mass-action curve for RIA. Clin Chem 1983;29:1757-1760.
[Abstract/Free Full Text]
-
Perlo VP, Poskanzer DC, Schwab RS, Viets HR, Osserman KE, Genkins G. Myasthenia gravis: evaluation of treatment in 1355 patients. Neurology 1966;16:431-439.
[Free Full Text]
-
Schmidt J, Raftery MA. A simple assay for the study of solubilized acetylcholine receptor. Anal Biochem 1973;52:349-354.
[ISI][Medline]
[Order article via Infotrieve]
-
Skoog DA. Principles of instrumental analysis 1985:195-198 Saunders College Publishing Philadelphia. .
-
Ekins RP, Newman GB, O'Riordan JLH. Theoretical aspects of "saturation" and radioimmunoassay. Hayes R Goswitz F Murphy B eds. Radioisotopes in medicine: in vitro studies 1968:59-100 US Atomic Energy Commission Oak Ridge, TN. .
-
Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine [Review]. Clin Chem 1993;39:561-577.
[Abstract/Free Full Text]
-
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307-310.
-
Vincent A, Newsom-Davis J. Acetylcholine receptor antibody characteristics in myasthenia gravis. I. Patients with generalised myasthenia gravis or disease restricted to ocular muscles. Clin Exp Immunol 1982;49:257-265.
[ISI][Medline]
[Order article via Infotrieve]
-
Sano M, Lambert EH, McCormick DJ, Lennon VA. Muscle acetylcholine receptors complexed with autologous IgG reflect seropositivity but not necessarily in vivo binding. Neurology 1992;42:218-222.
[Abstract/Free Full Text]
-
Martino G, Twaddle G, Brambilla E, Grimaldi LME. Detection of anti-acetylcholine receptor antibody by an ELISA using human receptor from a rhabdomyosarcoma cell line. Acta Neurol Scand 1994;89:18-22.
[ISI][Medline]
[Order article via Infotrieve]
-
Oda K. Differences in acetylcholine receptor–antibody interactions between extraocular and extremity muscle fibers. Ann N Y Acad Sci 1993;681:238-255.
[ISI][Medline]
[Order article via Infotrieve]
-
Durelli L, Ferrio MF, Urgesi A, Poccardi G, Ferrero B, Bergamini L. Total body irradiation for myasthenia gravis: a long-term follow-up. Neurology 1993;43:2215-2221.
[Abstract/Free Full Text]
-
Vincent A, Newsom-Davis J, Newton P, Beck N. Acetylcholine receptor antibody and clinical response to thymectomy in myasthenia gravis. Neurology 1983;33:1276-1282.
[Abstract/Free Full Text]
-
Oosterhuis HJGH, Limburg PC, Hummel-Tappel E, The TH. Anti-acetylcholine receptor antibodies in myasthenia gravis. : Part 2: clinical and serological follow-up of individual patients. J Neurol Sci 1983;58:371-385.
[ISI][Medline]
[Order article via Infotrieve]
-
Ferrero B, Durelli L, Cavallo R, Dutto A, Aimo G, Pecchio F, Bergamasco B. Therapies for exacerbation of myasthenia gravis: the mechanism of action of intravenous high-dose immunoglobulin G. Ann N Y Acad Sci 1993;681:563-566.
[ISI][Medline]
[Order article via Infotrieve]
-
Clarke CE, Shephered DI, Yuill GM, Smaje JC, Wilson PB. Deficiencies in anti-acetylcholine receptor antibody measurement in myasthenia gravis. J Neurol Neurosurg Psychiatry 1991;54:454-456.
[Abstract/Free Full Text]
Top
Abstract
Introduction
