To avoid these inconveniences, we developed and have validated a double-antibody. RIA for antimicrosomal antibody in the serum. The assay is shown to be ...
CLIN. CHEM. 27/1, 39-42 (1981)
Double-Antibody Radioimmunoassay of Thyroid Microsomal Antibody in Serum Viola 1. Kung,1 Paul M. Weber, and Leo V. dos Remedios We describe antimicrosomal
a double-antibody antibodies
radioimmunoassay
in serum.
Antigen
for
in centrifuged
pellets of microsome is solubilized by treatment with sodium deoxycholate solution and radiolabeled (1251) by the Chloramine T method. A 10-L aliquot of human serum or antimicrosomal antibody standard isincubatedovernight at room temperature with 1251-labeled microsome. Bound radiolabeled microsome is separated by precipitationwith goat antihuman lgG antiserum. The percentage of 1251.. labeled microsome bound by each serum is calculated, and its concentration
of antimicrosomal
antibody
interpolated
from a standard curve. The prevalence of abnormal antimicrosomal antibody was 16% in a self-referred asymptomatic or not acutely ill group of patients having a periodic multiphasic health examination. Values for patients with certain thyroid diseases generally agreed with other published sitive, precise,
data. We conclude that our assay is senand accurate, and is more efficient for
routine measurement of antimicrosomal antibody than the currently used solid-phase radioimmunoassay.
Additional Keyphrases: autoimmune prevalence
thyroid
disease
of abnormal antimicrosomal antibody
Materialsand Methods Apparatus We used of Beckman refrigerated MA 02194); CT 06470); ments, Inc.,
an L5-65 preparative ultracentrifuge (Spinco Div. Instruments, Inc., Palo Alto, CA 94304); PR-6000 centrifuge (Damon/IEC Div., Needham Heights, a Sorvall Omni-Mixer (DuPont Co., Newtown, and a gamma scintillation counter (LKB InstruRockville, MD 20852).
Reagents All the phosphate solutions we used contained 10 mmol of sodium phosphate per liter and had a pH of 7.4 unless otherwise specified. “Phosphate-buffered isotonic saline” was this phosphate solution plus 0.15 mol of NaC1 per liter. We used bovine serum albumin, sodium deoxycholate, Chloramine T, and sodium metabisulfite (Sigma Chemical Co., St. Louis, MO 63178); Sephadex G-50, medium (Pharmacia Fine Chemicals, Piscataway, NJ 08854); carrier-free 12Sf (ICN Chemical & Radioisotope Div., Irvine, CA 92715); Fujizoki hemagglutination test kit for antithyroglobulin and antimicrosomal antibodies (Ames Co., Division of Miles Laboratories, Inc., Elkhart, IN 46514); and goat antihuman IgG antiserum (Antibodies Inc., Davis, CA 95616).
Procedures Serum from patients with autoimmune thyroid disease frequently contains antibodies to thyroid microsomal antigen or thyroglobulin, or both (1). In clinical practice, antimicrosomal and antithyroglobulin antibodies are measured as an aid in diagnosing primary myxedema, Hashimoto’s disease, and Graves’ disease (2). A double-antibody radioimmunoassay (RIA) for antithyroglobulin in serum has been developed (3, 4), but the complexity of microsome antigen has prevented the application of this simple double-antibody technique to measurement of antimicrosomal antibody. Roitt et al. (5) indicated that microsomes in the smooth endoplasmic reticulum contain an antigen that is composed of a lipoprotein deriving from the membrane of the small vesicles of the endoplasmic reticulum, which contain newly synthesized thyroglobulin. Two current methods for detecting antimicrosomal antibody are the tanned erythrocyte hemagglutination technique (6) and the solid-phase competitive binding RIA (7). The latter is more sensitive and specific (8), but has some disadvantages. First, a constant, relatively large supply of human thyroid tissue is required to obtain enough microsomes to coat the plastic cups in the microtiter plate. Second, every plastic cup must be cut from the plate and placed in a tube for radioactivity counting. To avoid these inconveniences, we developed and have validated a double-antibody RIA for antimicrosomal antibody in the serum. The assay is shown to be accurate, sensitive, and efficient for routine use. Department
of Nuclear
Medicine,
Kaiser-Permanente
Center, 280W. MacArthur Blvd., Oakland, CA 9611. Received June 30, 1980; accepted Sept. 2, 1980.
Medical
All the purification procedures were performed at 4 #{176}C. Purification of thyroglobu un. Thyroglobulin was purified according to Bodlaender et al. (4). After thyroglobulin had been salted out with ammonium sulfate, it was additionally purified by gel filtration on Sephadex G-200 (9). The purified thyroglobulin was stored frozen at -20 #{176}C until use as an inhibitor in the antimicrosomal antibody radioassay. Isolation and solubilization of microsomal antigen. Thyroid glands, taken at autopsy, were immediately frozen and stored
until
use.
With
scissors,
we cut the
small pieces, which were suspended containing 0.25 mol of sucrose per a Waring Blendor. Crude microsomes form by differential centrifugation
frozen
glands
into
in phosphate solution liter and homogenized in were obtained in pellet
(10). To measure its microsomal and thyroglobulin antigenicities we resuspended the microsome pellet in phosphate-buffered isotonic saline solution with use of the Omni-Mixer at maximum speed for 15 s. The titer of antimicrosomal or antithyroglobulin antibody in patient’s serum was assayed with the Fujizoki hemagglutination test kit, in which microsomes or thyroglobulin-coated erythrocytes would agglutinate with serum containing antimicrosomal or antithyroglobulin antibody. The microsomal or thyroglobulin antigen added to the hemagglutination test would competitively bind to its antibody, thus inhibit the agglutination of erythrocytes. By comparing the antimicrosomal or antithyroglobulin antibody titers before and after adding microsomal fractions, the inhibition of hemagglutination by microsomal or thyroglobulin antigenicity in the fraction could be calculated. In the crude microsomal fraction, the ratio thyroglobulin antigenicity/ microsomal antigenicity exceeded 100 by the inhibition test.
CLINICAL CHEMISTRY. Vol. 27. No. 1, 1981
39
The contaminating thyroglobulin was removed from the microsomal fraction, to eliminate its interference with RIA of antimicrosomal antibody. The microsome pellets was resuspended in phosphate solution containing 20 mg of sodium deoxycholate per liter, centrifuged at 100 000 X g for 1 h, add the supernate discarded. Microsomal and thyroglobulin antigenicities in the pellet were measured by the inhibition of hemagglutination after each washing. After 10 washes, there was no substantial thyroglobulin in the pellet, but thost of the microsomal antigen remained. Then the microsomal antigen was solubilized from the pellet by resuspending it overnight in the phosphate solution containing 2 g of sodium deoxycholate per liter; the resuspension volume was 0.2 mL/g of thyroid gland. The homogenate was centrifuged at 10 000 x g for 30 mm and the supernate was stored frozen until iodination. The protein concentration in the 2 g/L deoxycholate microsomal fraction was 3 g/L as assayed by Bradford’s method (11). lodination of microsomal antigen. Microsome was labeled by a modification of the method of Hunter and Greenwood (12), by mixing for 30 s at room temperature 1 mCi of Na’251, 20 AL of solubilized microsome (60 Ag of protein), 10 AL of 0.5 mmol/L sodium phosphate (pH 7.4), 30 AL of water, and 10 1zL of 2.5 g/L Chloramine T. Then 10 AL of 6.4 g/L sodium metabisulfite was introduced and the solution was again mixed for 30 s to stop the iodination reaction, and 20 AL of a 10 g/L solution of bovine serum albumin was added to minimize adsorption of sample protein to the tube. The reaction mixture was added to a 1 X 15-cm column of Sephadex G-50 pre-equilibrated with the phosphate solution containing, per liter, 0.5 g of sodium deoxycholate, 50 mmol of NaCI, and 0.5 g of bovine serum albumin. The column was eluted with the same solution. The firt peak of radioactivity represented labeled microsome with a specific activity of 2 to 5 Ci/g of protein. Each peak fraction was precipitated in the collection tube to minimize nonspecific binding. The labeled fraction was diluted 10-fold in phosphate-buffered isotonic saline solution; then 10 AL of normal human serum and 200 AL of goat antiserum
were
added,
and
the
mixture
was
incubated
over-
night at 4 #{176}C. After centrifugation (200 X g, 20 mm, 4 #{176}C), the supernate was stored at 4 #{176}C and radioimmunoassayed within two weeks. Standard antimicrosonzal antibody. A patient’s serum with high antimicrosomal antibody but undetectable antithyroglobulin antibody was selected. The IgG fraction of this serum was partly purified according to published procedures (13), diluted in normal serum, calibrated by solid-phase RIA (7), and used as the standard in our RIA. Antibody content was expressed in arb. units X 1031L, 1 arbitrary unit being defined as the amount equivalent to 0.9 g of the antimicrosomal antibody standard. RIA for antimicrosomal antibody. Before assay, the labeled microsome was diluted to a radioactivity of 30000 counts/mm per 200 AL with phosphate-buffered isotonic saline solution containing 0.5 g of bovine serum albumin per liter; 200 AL was pipetted into a 12 X 75-mm glass tube, and 10 AL of serum or standard antimicrosomal antibody IgG added. Finally, to prevent any contaminating radiolabeled thyroglobulin from binding to the test serum, 10 Ag (50 AL) of pure thyroglobulin was added to each tube. After vortex-mixing, the mixture was left at room temperature overnight, then 50 AL of goat antihuman IgG serum was added as the second antibody, and the mixture was incubated at 37 #{176}C for an additional 2 h; 3 mL of isotonic saline was added and the resulting precipitate centrifuged down (200 X g, 20 mm, 4 #{176}C). The supernate was decanted and the inverted tube allowed to drain for 10 mm. Any remaining droplets were removed with cotton swabs. The radioactivity in the pellet was counted in a gamma counter for 1 mm and the percentage bound was calculated as (counts in 40
CLINICAL CHEMISTRY, Vol. 27, No. 1, 1981
10
8 C
2
6
‘a
C
4
0
0 2
0 10
30
100
300
antibody (arb. units x 103/L)
Antimicrosomal
Fig.1. Radioimmunoassay dose-responsecurve for antimicrosomal antibody in serum pellets/total count) X 100. Sera with less than 2% binding were pooled and used as normal serum to measure nonspecific binding, the value for which was subtracted to give the specific binding percentage for each test serum.
Results and Discussion Dose-Response
Curve
On semilogarithmic paper, we constructed a calibration curve by plotting percentage of specific binding vs concentration of antimicrosomal antibody standard (Figure 1). The concentrations of antimicrosomal antibody in unknown sera were interpolated from the standard curve. Concentrations exceeding 100 arb. units X 103/L were diluted in normal serum and retested. Maximum binding of ‘251-labeled microsome by antimicrosomal antibody was 10%, 10-fold the 1% maximum binding of the solid-phase RIA (7). With this advantage, we needed much less ‘251-labeled material for each test (3 X 10 cpm) than that needed for the solid-phase RIA (3 X 10 cpm). Requiring as little as 10 iiL of serum, this new test is sensitive enough to detect 5 arb. units X 10 of antimicrosomal antibody per liter of serum.
Analytical
Variables
Precision was evaluated (Table 1) by determining the coefficient of variation (CV) within-assay, from duplicate measurements of 60 sera at each of the two concentrations, and between-assay, from four observations of each of the five sera at three concentrations. The within-assay CV ranged from 7.3 to 8.6%, the between-assay CV from 9.5 to 14.0%. Recovery study. We assessed analytical recovery by use of 19 sera with values ranging from 500 Between assay 30-100
101-1000 >1000
7356 53.7
432 7157
647 6.2
38.8 1024
8.6 12.8
9.5 14.0
Table 2. Distribution of Antimicrosomal Antibody Values by Clinical Diagnosis in 157 Thyroid-Scan Patients No. patlenis
Clinical diagnosis
with normal antlmicrosomal antibody 7000 343->7000
Normal Diffuse nontoxic
20 13
1 (5) 7 (35)
Solitary
37
4 (10)
32 8
8 (20) 2 (20) 19 (86) 2 (50) 43 (57)
nodule
Multinodular goiter Thyroiditis
Graves’ disease Toxic nodular goiter Total
2 2 114
Abnormal range arb. units
40->7000 660->7000
antibody were added to the sera, and recovery was calculated. Among the 19 sera, seven had a recovery percentage within ±10% of 100%, ten were within ±10% to ±20% of 100%, and the remaining two were 76% and 122%. The overall average recovery percentage was 100%. Split sample comparisons. Eighty-eight specimens were analyzed by both the Nichols Institute, San Pedro, CA 90731, and by us. These data covered a wide range from 7000 arb. units X 103/L. Because the solid-phase RIA (7) used by the Nichols Institute has a sensitivity and detection limit of 20 arb. units X 103/L, the 50 samples they reported as 10000 and the other three values of 5200, 5600, and 8250 arb. units X 1Ol/L by our technique. The values for the remaining 31 samples, ranging from 20 to 7000 arb. units X 103/L, were compared by the statistical methods of Westgard and Hunt (14), by which random, constant, and proportional errors were shown to be correlated (along with other parameters) with the standard deviation of residuals, y-intercept, and slope of a standard linear regression analysis. For the entire set, the primary indicators gave random errors of 457 arb. units X loYL, constant errors of 65 arb. units X 103/L, and proportional errors of 1.05 arb. units X 10VL. Linear regression analysis between our method (y) and the reference method (x) gave the equation: y = -65.12 + 1.05x, with a correlation coefficient of r = 0.97 (p 5000 Total
No. without thyroid disease
8 13 6 3 30
No. with thyroid disease
7
6 4 3 20
Stability and efficiency. Five batches of microsomal antigen were prepared to test our new double antibody method. The solubilized microsomal fraction can be stored in aliquots at -20 #{176}C until iodination. For each gram of thyroid gland, enough microsomal activity is obtained for 10 iodinations, and each iodination gives enough material for 1000 assays. Unlabeled microsomal material is stable at -20 #{176}C for at least three months, but ‘251-labeled microsome is stable for only two weeks at 4 #{176}C. The solid-phase RIA of antimicrosomal antibody currently used by other laboratories (7) requires a large supply of human thyroid, to obtain enough microsomes to coat the microtiter plate. In contrast, because we label the microsomal antigen, relatively little thyroid is required for our test. Coating time for the solid-phase RIA is 24 h, and incubation of the coated plate with serum and tracer requires another 24 h. Our test requires only overnight incubation of the tracer with serum, plus precipitation for 2 h with goat antihuman IgG antiserum. In the solid-phase RIA, in order to measure the radioactivity, each plastic cup must be cut from the microtiter plate, placed in a test tube, and put into a counter. In our test, glass test tubes are used for incubation, thus saving labor. The sensitivity of the solid-phase RIA is 20 arb. units X 103/L with 100 AL of serum; the average CV was 17.7% (15). The sensitivity of our method is 5 arb. units X 103/L with 10 AL of serum; the average within-assay CV was 7.0%, and the between-assay CV was 12.1%. The maximum binding of ‘25I-labeled antimicrosomal antibody to the microtiter plate in the solid-phase RIA is only 1% of the total counts, whereas maximum binding of ‘251-labeled microsome by antimicrosomal antibody is 10% by our method. Evidently, our double-antibody RIA is more efficient than the current solid-phase test.
ClinicalApplications Sera from two groups of patients were assayed for antimicrosomal antibody. Of 157 patients referred to the Nuclear Medicine Laboratory for thyroid imaging, antimicrosomal antibody values were normal (
