| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Laboratory of Neuroimmunology and
2
Division B, Istituto di Ricovero e Cura a Carattere Scientifico, Neurological Institute `Mondino', via Palestro 3, University of Pavia, 27100 Pavia, Italy.
3
Experimental Neuroimmunotherapy and
4
Neuroimmunology Unit, Department of
Biotechnology, San Raffaele Scientific Institute, via Olgettina
58, 20132 Milan, Italy.
5
Unit of Biometric, Istituto di Ricovero e Cura a
Carattere Scientifico, Policlinico S. Matteo, University of Pavia,
27100 Pavia, Italy.
6
`Casa Sollievo della Sofferenza', 71013 San Giovanni
Rotondo (FG), Italy.
a Author for correspondence. Fax 39-0382-380286; e-mail francid{at}cpbim1.unipv.it
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: After TE671 cell homogenization, the crude AChR extract was used for plate coating. Anti-AChR antibodies were determined in 207 MG patients and in 77 controls.
Results: The mean intra- and interassay CVs (for two samples with different anti-AChR antibody concentrations) were 9.7% and 15.7%, respectively. Test sensitivity and specificity, for generalized MG, were 79.5% (95% confidence interval, 72.8–85.0%) and 96.1% (89.0–99.1%). The detection limit was 2 nmol/L. Anti-AChR antibody concentrations from 53 MG patients, as tested with our ELISA, showed good agreement with an RIPA with a mean difference (SD) of 1.0 (5.6) nmol/L.
Conclusion: Our ELISA is a simple screening test for the diagnosis of MG and enables rapid and inexpensive patient follow-up. © 1999 American Association for Clinical Chemistry
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-bungarotoxin (125I--BuTx), a potent acetylcholine
antagonist. The main disadvantages of this sensitive method are the use
of radioisotopes and the limited availability of human muscles.
Furthermore, the presence of -BuTx deprives a part of the anti-AChR
antibodies of potential binding sites.
Various authors have proposed alternative solid-phase
immunoenzymatic methods to overcome some of the above-mentioned
limitations. One method involves the coating of polystyrene wells
directly with muscle extracts (2)); another method uses
muscle-derived AChR, which in turn is captured by means of precoating
with -BuTx (3)) or with anti-AChR monoclonal
antibodies (4)). The main drawbacks of these methods are
the paucity of AChR in relation to other muscle proteins
(2)) and the inadequacy of AChR capture, whether by
-BuTx or by anti-AChR monoclonal antibodies (3)(4)).
The effect of these drawbacks is a reduction in the diagnostic
sensitivity of these methods.
The human rhabdomyosarcoma cell line TE671, which expresses AChR on its surface (5)), has been used as an alternative to human muscle as a source of AChR (6)(7)(8)(9)(10)). Radioimmunoprecipitation assays (RIPAs) with TE671 cell-derived AChR as antigen have been developed (6)(7)(8)) or reevaluated (9)). Our own group previously proposed a TE671 cell-based ELISA for the detection of serum anti-AChR antibodies (10)). Here we report the development and a thorough validation of this ELISA.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cell culture
TE671 cell line (American Type Culture Collection) was
cultured at 37 °C, 5% CO2, in RPMI-1640 supplemented
with 2 mmol/L glutamine, 100 kU/L penicillin, 10 mg/L
streptomycin, 25 mmol/L HEPES (all products from Sigma Chemical Co.),
and 100 mL/L fetal calf serum. Confluent cells were harvested every
5–7 days.
preparation of crude AChR EXTRACT
Cells were washed in homogenization buffer
containing 5 mmol/L Tris (pH 7.8), 10 mmol/L EDTA, 10 mmol/L
ethyleneglycol-bis-(ß-aminoethylether)-N,N,N',N,-tetraacetic
acid, 0.1 g/L phenylmethylsulfonyl fluoride, and 0.2
g/L sodium azide, and then removed with a cell scraper. Cells were
centrifuged at 500g for 15 min and then resuspended in
homogenization buffer at 4 °C. After incubation for 30 min, cells
were homogenized with a Polytron for 30 s on ice. Homogenized
cells were centrifuged at 500g for 15 min at 4 °C.
Supernatant containing membrane debris was centrifuged at
40 000g for 45 min at 4 °C. The pellet was resuspended
in homogenization buffer at 4 °C, sonicated for 3 min on ice, and
frozen in aliquots at -80 °C.
TE671 cell-based elisa optimization
We optimized the coating concentration of the AChR preparation by
testing different dilutions of the preparation (total protein
concentration, 5 mg/L to 5 g/L) against a reference serum (anti-AChR
antibody concentration of 120 nmol/L, positive control) and against a
pool of sera from 10 blood donors (negative control). The total protein
concentration of the AChR preparation was taken as an index of its AChR
content.
We evaluated the ability of the TE671 preparation to bind specifically
to 125I--BuTx by testing with monoiodinated
125I--BuTx (0.5–550 fmol/well; Amersham) in microwells
coated with 100 µL of the AChR preparation, which contained a total
protein concentration of either 0.50 or 0.75 g/L, or with 50 fmol
125I--BuTx/well in microwells coated with 100 µL of
serially diluted AChR preparation (total protein concentration, 5 mg/L
to 5 g/L). The specific activity of 125I--BuTx was
~2000 Ci/mmol. The AChR preparation was diluted with
phosphate-buffered saline (PBS). Single microwells were coated with 100
µL AChR preparation/well and incubated overnight at 4 °C. After
wells were washed with PBS, 200 µL of 50 g/L nonfat dry milk in PBS
was added to each well to block nonspecific binding sites. After a 1-h
incubation at 4 °C, wells were washed three times. We used 5
g/L nonfat dry milk in PBS to make the appropriate dilutions of
125I--BuTx. Each dilution was tested twice. Microwells
were incubated for 2 h at room temperature. After a triple washing
cycle, each microwell was inserted into appropriate tubes, and the
radioactivity was measured.
We also evaluated the effect on assay specificity of antibodies directed against other non-AChR autoantigens that the AChR preparation could contain. We performed a study of preadsorption on four serum samples: two samples, from MG patients, had anti-AChR antibody concentrations of 40 and 12 nmol/L; the other two samples, from SLE patients, were positive for anti-AChR antibodies at a low concentration (see Results). Briefly, PC12 cells (clone 16A, courtesy of Dr. E. Clementi, Department of Pharmacology, University of Milan, Italy) were cultured at 37 °C, 5% CO2, in RPMI-1640 supplemented exactly as for TE671 cell culture, with the addition of 50 mL/L horse serum. Cells were harvested at confluence every 3–4 days, and crude extraction of the PC12 cells was carried out as for TE671 cells. Each serum sample (50 µL) was diluted twofold in 20 g/L nonfat dry milk in PBS and incubated overnight at 4 °C with 50 µL of the PC12 cell preparation. This preparation had a total protein content of 5 g/L. Membranes were spun down (10 000g for 15 min), and the supernatant was re-incubated for 1 h at 4 °C with 50 µL of the PC12 cell preparation. After centrifugation, the supernatant was tested for the presence of anti-AChR antibodies with the TE671 ELISA. In a similar study, the same four serum samples were also tested for anti-AChR antibodies after preadsorption with a TE671 cell preparation (total protein content of 5 g/L), using the same procedure as for PC12 cells.
te671 cell-based elisa
We determined the total protein concentration of the AChR
preparation from TE671 cells (BCA Protein Assay; Pierce). The
preparation was diluted with PBS at the optimal coating concentration
(0.5 g/L). Microwell plates (Polysorp; Nunc) were coated with 100 µL
diluted AChR preparation/well and incubated overnight at 4 °C. After
the wells were washed with 100 µL PBS/well (a volume that was
maintained for subsequent washing cycles), 200 µL of 50 g/L nonfat
dry milk in PBS was added to each well to block nonspecific binding
sites. After 1 h at 4 °C, wells were washed three times. Using
20 g/L nonfat dry milk in PBS, we prepared 1:100 and 1:1000 dilutions
for each sample and control. Each dilution (100 µL/well) was
then tested twice. The plates were incubated with light shaking
(100 rpm) for 2 h at room temperature. After a triple washing
cycle, an alkaline phosphatase-conjugated, rabbit anti-human IgG
antibody (Dakopatts), diluted 500-fold in 20 g/L nonfat dry milk in
PBS, was added (100 µL/well). Both serum samples and the labeled
antibody were diluted in nonfat dry milk/PBS to minimize nonspecific
binding of the antibody to protein-free sites of the AChR preparation
(e.g., through their Fc portions). After an incubation of 1 h at
room temperature, wells were washed five times.
p-Nitrophenylphosphate (Sigma) at the concentration of 1 g/L
in diethanolamine buffer (pH 9.6) was added (100 µL/well). After
1 h of incubation, or when the calibrator diluted 1:10
reached the absorbance value of 1.4–1.5, the absorbances of the
reaction products were read at 405 nm.
calibration
The assays for the detection of antibodies to AChR are based on
the immunoprecipitation of the 125I-labeled
-BuTx/anti-AChR antibody complex. Anti-AChR antibody concentrations
were expressed as picomoles of 125I-labeled -BuTx
precipitated by 1 mL of serum. The calibrator for the present ELISA was
a serum sample with high anti-AChR antibody concentration (120 nmol/L)
from an MG patient. The patient had no serum paraproteins and had
concentrations of IgG, IgA, and IgM within the health-related reference
intervals. Previously, this sample had been tested repeatedly
with a reference RIPA, with human muscle as the source of antigens, and
then aliquoted and stored at -80 °C. The serum was diluted serially
(1:10 to 1:100 000) in 20 g/L nonfat dry milk in PBS. Anti-AChR
antibody concentrations were calculated by comparison between the mean
absorbance of each sample and that generated by the serially
diluted calibrator serum (calibration curve). Another serum sample with
known anti-AChR antibody concentration (10 nmol/L) from another MG
patient was used as positive control. A pool of sera from 10 blood
donors was the negative control.
statistical analysis
The following methods were used for the statistical
analysis: log regression analysis for the evaluation of calibration
curves; the calculation of coefficients of variation (CVs) for within-
and between-run imprecision; a plot for the assessment of the
agreement between two methods in accordance with Bland and Altman
(12)); and an ROC curve (13)) for the
evaluation of the diagnostic accuracy. Ninety-five percent confidence
intervals are provided in parentheses for the indices of diagnostic
accuracy and for data on imprecision.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
When tested at the total protein concentrations of 0.50 g/L (Fig. 1
A) and 0.75 g/L (Fig. 1B
), the AChR preparation showed specific
binding with 125I--BuTx. This specific binding was also
confirmed by tests on the AChR preparation at serial dilutions (Fig. 1C
). All correlations were significant (P <0.05). The
concentration of 125I--BuTx-binding sites can be
estimated as ~35 fmol/well at the total protein concentration of the
AChR preparation used in the ELISA.
|
The effects on the anti-AChR antibody concentrations of preadsoption with PC12 or TE671 cell preparations showed that (a) in serum samples from the two seropositive MG patients, these concentrations were slightly decreased (~10%) after preadsorption with the PC12 cell preparation and markedly decreased (~60%) after preadsorption with the TE671 cell preparation; and (b) in serum samples from the two SLE patients, the concentrations were below the detection limit after preadsorption with either PC12 or TE671 cell preparations.
calibration curves and detection limit
The calibration curve obtained by serial dilutions of the
reference serum is shown in Fig. 2
. Each point corresponds to the mean absorbance, for the
relative dilution, of eight different assays. A good log-linear
regression (r2 = 0.997) was found for the
intervals corresponding to dilutions of 1:10 to 1:100 000. Measured
values of up to 32-fold dilutions of two serum samples with high
anti-AChR antibody concentrations ranged from 95% to 115% of the
expected values. The detection limit was 2 nmol/L, which is the
concentration corresponding to a signal 3 SD above the mean of four
replicates of the pool of control sera. The calibration curves
of the ELISA presented here are constructed with a reference serum that
has been tested previously for the presence of anti-AChR antibodies
with a human muscle-based immunoprecipitation RIPA. This enables a
quantitative correlation between two different methods. We stored our
calibrator in large quantities at -80 °C.
|
precision
Within- and between-run imprecision was calculated for samples at
low and high anti-AChR antibody concentrations. Each sample was tested
15 times per analytical run and in triplicate in 15 analyses performed
at a monthly interval. At the low anti-AChR antibody concentration (5.1
nmol/L), the respective intra- and interassay CVs were 12.6%
(7.1–21.2%) and 20.0% (12.7–29.2%). At the high anti-AChR antibody
concentration (64.0 nmol/L), the respective intra- and the interassay
CVs were 6.9% (2.9–13.9%) and 11.5% (6.0–19.1%). There were no
significant differences in precision when the relative data of our
assay were compared with those of the reference RIPA (data not shown).
analytical accuracy
A plot of the respective differences of the two methods vs the
means of the two methods for anti-AChR antibody determination is shown
in Fig. 3
. For this purpose, 53 of the 207 serum samples of MG patients
were tested. When two sera with low anti-AChR antibody
concentrations were each mixed with a serum with a high anti-AChR
antibody concentration, the recovery was 94% and 90% of the expected
calculated value. Polyclonal IgG, lipids, hemoglobin, and bilirubin
showed no interference with anti-AChR antibody determination: sera with
high concentrations of these substances, mixed with a serum with
anti-AChR antibody concentration of 25 nmol/L, had negligible effects
on the expected values (data not shown).
|
diagnostic accuracy
Anti-AChR antibody concentrations above the detection limit were
found in 146 of 176 (79%) patients with generalized MG, and in 17 of
31 (55%) patients with ocular MG (Fig. 4
). Patients with ocular MG had lower anti-AChR antibody
concentrations than those with generalized disease (P =
0.02). Samples from five MG patients with anti-AChR antibody
concentrations of <2 nmol/L in the reference RIPA were negative in
ELISA. Anti-AChR antibody concentrations showed no statistically
significant correlation either with the severity of symptoms or with
the presence of thymoma in patients with generalized MG. When the
control group was considered intoto, two SLE patients and the patient
with pseudotumor orbitae showed low concentrations (<4.0 nmol/L) of
serum anti-AChR antibodies.
|
The ROC curve for anti-AChR antibodies, which we plotted by matching
the positive "diseased" group (all MG patients, n = 207) and
the negative "control" group (all non-MG patients, n = 77) is
shown in Fig. 5
. The area under the curve was 0.868 (0.823–0.905). Correlated
diagnostic accuracy indexes were as follows: sensitivity, 75.8%
(69.4–81.5%); specificity, 96.1% (89.0–99.1%); and positive
likelihood ratio for the point of the greatest efficiency of the test,
19.5 (8.3–45.7). If the two groups were taken separately, sensitivity
for generalized and ocular MG was, respectively, 79.5% (72.8–85.2%),
and 54.8% (36.0–72.4%). The specificity was 96.1% (89.0–99.1%)
for both groups. The respective positive likelihood ratios were
20.4 (8.7–47.9) and 14.1 (6.0–33.0).
|
clinical follow-up
Sixteen MG patients were followed longitudinally. Their
serum anti-AChR antibody concentrations varied as follows: in 7
of 16 patients, the antibody concentrations decreased in parallel with
clinical improvement whether induced (5)) or not
(2)) by immunosuppressive drugs; in 6 of 16 patients (4
clinically stationary, 2 with worsening symptoms), the antibody
concentrations remained almost unchanged; in 3 of 16 patients who
complained of worsening symptoms, the antibody concentrations were
increased (one patient withdrew from immunosuppressive therapy).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
For the generalized form of MG, this TE671 cell-based ELISA showed a sensitivity of 79%, near the 87–88% reported in the literature (14)(15)). For the ocular form of MG, sensitivity decreased to 55%, which falls within the reported 45–70% range (14)(15)). The method showed a detection limit of 2 nmol/L, which is higher than that of conventional RIPAs based on human muscle-derived AChR but similar to that of TE671 cell-based RIPAs (6)(7)(8)(9)). The assay precision data of our ELISA compare well with those obtained with the reference RIPA.
The TE671 ELISA correlated well with the reference human muscle-based
RIPA. In contrast with previous reports on TE671 cell-based RIPAs
(6)(7)(8)(9)), we found that anti-AChR antibody concentrations
tested with our ELISA were not significantly lower than those tested
with the reference RIPA. In those methods (6)(7)(8)(9)), which
used TE671 cells as the source of antigen, AChRs were bound to
125I--BuTx and used for the immunoprecipitation of
anti-AChR antibodies. Whereas AChR sites would be occupied by
125I--BuTx in RIPAs, in our ELISA such sites are
available for the binding of serum anti-AChR antibodies. However, the
diagnostic sensitivity of this ELISA is lower than that of human
muscle-based RIPAs, although it is comparable to that of TE671
cell-based RIPAs (6)(7)(8)(9)). The number of coated AChRs
available for specific antibody binding is limited by the surface area
of microwells. This limitation is probably compensated for by the
availability of free -BuTx binding sites. Such sites are important
for the performance of this ELISA: we found a good dose–response
relationship when we tested the specific binding between the TE671
cell-derived AChR preparation and 125I--BuTx. These
sites, which are only a portion of the AChR sites available for
anti-AChR antibody binding, probably allow our ELISA to overcome the
disadvantage of microwell surface area constraints.
Regarding diagnostic specificity, two patients with SLE and one patient with pseudotumor orbitae showed low serum anti-AChR concentrations. The two SLE serum samples, after preadsorption with the PC12 or with the TE671 cell preparation, were negative for anti-AChR antibodies. Because PC12 cells are particularly rich in cytoskeletal proteins (16)) and because SLE patients may present serum antibodies against these proteins (17)), it is likely that the positivity of the two SLE serum samples was caused by the recognition by antibodies of non-AChR proteins, probably cytoskeletal proteins. The preadsorption of the two MG sera caused only a slight decrease in their anti-AChR antibody concentration. Sera from all healthy controls were negative for anti-AChR antibodies. Moreover, we previously had evaluated the ability of a pool of six monoclonal antibodies, each of which recognized a different region of AChR, to compete with the binding of the reference serum with the AChR preparation (10)). This binding was blocked by as much as 30%. The low inhibition might depend on the heterogeneity of AChR epitopes. The pool of six monoclonal anti-AChR antibodies would only partially compete with all the naturally occurring anti-AChR polyclonal antibodies.
The advantages of TE671 cells in comparison with human muscle as a source of antigen include (a) greater homogeneity, because ischemia and other variables may influence muscle preparations; (b) easy availability; and (c) no risk of infection through the manipulation of potentially infected human tissues.
The present method for the nonradioactive detection of anti-AChR antibodies provides good overall analytical and diagnostic performance; however, it fails to detect very low anti-AChR antibody concentrations. Accordingly, it might yield more false-negative results. Data from the longitudinal follow-up of parallel antibody/clinical variations in single MG patients showed that anti-AChR antibody concentrations correlated well with clinical improvement subsequent to immunosuppressive treatment. This finding confirms the reliability of the method, which may therefore be useful, even as an addition to standard RIPAs, for easy and inexpensive monitoring of immunosuppressive therapy. We believe our ELISA can be used for specific antibody screening in MG and is particularly suitable for laboratories that are not equipped for RIPAs and/or have no access to human muscle tissue.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
-BuTx, 125I-labeled -bungarotoxin; RIPA, radioimmunoprecipitation assay; SLE, systemic lupus erythematosus; and PBS, phosphate-buffered saline. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-bungarotoxin binding sites extracted from TE671 cells. J Neuroimmunol 1988;19:149-157.
[Web of Science][Medline]
[Order article via Infotrieve]
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J. Ricny, L. Simkova, and A. Vincent Determination of Anti-Acetylcholine Receptor Antibodies in Myasthenic Patients by Use of Time-resolved Fluorescence Clin. Chem., March 1, 2002; 48(3): 549 - 554. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
