Time-Resolved Fluorescence Immunoassay of Thyroxine in Serum: Immobilized Antigen Approach

Analytical Biochemistry 276, 171–176 (1999) Article ID abio.1999.4342, available online at http://www.idealibrary.com on

Time-Resolved Fluorescence Immunoassay of Thyroxine in Serum: Immobilized Antigen Approach Fengbo Wu, 1 Yongyuan Xu, Tao Xu, Yanzheng Wang, and Shiquan Han Laboratory of Immunoassay, Department of Isotope, China Institute of Atomic Energy, P.O. Box 275-39, Beijing, People’s Republic of China 102413

Received June 4, 1999

With T 4– bovine IgG as a solid-phase antigen, we have developed a direct competitive-type immunoassay for serum total thyroxine (TT 4), which depends on the competitive distribution of europium-labeled anti-T 4 monoclonal antibody between solid-phasebound T 4 and the T 4 in the sample or standard. The captured fraction of the tracer was measured after a dissociation-enhancement step. Four different T 4 protein conjugates were synthesized, of which T 4– bovine IgG was selected as the most favorable for the preparation of solid-phase antigen. The sensitivity was 3.5 ng/ml with a sample volume of 20 ml. T 4 values obtained by this procedure agreed well with those obtained by RIA (r 5 0.967, n 5 38) and EG&G Wallac TRFIA (r 5 0.926, n 5 64). All other quality criteria was also fulfilled with respect to precision, accuracy, and dynamic range. © 1999 Academic Press

dioisotopic immunoassays have shown many advantages in such aspects as sensitivity, safety, convenience, and easy for automation. With nonradioisotopic reagents as labels, the measurement of serum T 4 has been reported in several ways (1– 4). Here we describe a new competitive immunoassay for serum T 4, which is characterized by the use of solid-phase antigen and the DELFIA system. In this assay, T 4 released from serumbinding proteins by a coblocking reagent competes with the immobilized T 4 for binding to the europium-labeled ant-T 4 monoclonal antibodies; the bound fraction of the tracer is then quantified by measuring the fluorescence after a dissociation-enhancement step. The developmental details and performance characteristics of the assay are discussed. MATERIALS AND METHODS

Materials Thyroxine (T 4) 2 is an iodine-containing hormone produced and secreted by the thyroid gland. As a catalyst of oxidation reactions in the body, it has an important effect on the regulation of metabolic rate and has been proven of high value in the diagnosis of thyroid diseases or in the evaluation of therapeutic effects. For a long time, RIA has been the common method for the measurement of serum T 4; however, because of the rapid development of nonradioisotopic immunoassays, especially time-resolved fluorescence immunoassay (TRFIA) and chemiluminescence immunoassay, nonra1 To whom correspondence should be addressed. Fax: 186-1069357195. 2 Abbreviations used: T 4, thyroxine; TRFIA, time-resolved fluorescence immunoassay; DTTA-Eu, N9-[p-Isothiocyanatobenzyl]-diethylene-triamine-N 1 ,N 2 ,N 3 ,N 3 -tetraacetate-Eu 31; BSA, bovine serum albumin; EDC, 1-ethyl-3-(-3-dimethylaminopropyl) carbodiimide; MCC, maximal capturing capacity; TCA, trichloroacetic acid.

0003-2697/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

Anti-T 4 monoclonal antibody (Product No. 6901) was obtained from Mediz Biochemica Co., Finland. Microtitration strips (Product No. 1244-550, 12 wells per strip) are products of NUNC Co., Denmark. N9-[p-Isothiocyanatobenzyl]-diethylene-triamine-N 1 ,N 2 ,N 3 ,N 3 -tetraacetate-Eu 31(DTTA-Eu) was from Tianjing Radio-Medical Institute, China. BSA was from Shenzheng JinMei Bioengineer Co., China. Other reagents were purchased from Sigma Co. (U.S.A.). The chromatographic separation system was product of Bio-Rad Co., mainly including a Model EP-1 Econo Pump and a Model EM-1 Econo UV monitor. The fluorescence was measured by Arcus 1230 fluorometer (LKB-Wallac). Buffers The coating buffer was 100 mM sodium carbonate buffer (pH 9.0), containing 0.9% NaCl, 0.04% NaN 3. The blocking buffer is 50 mM Tris–HCl, pH 8.0, containing 0.9% NaCl, 0.04% NaN 3, 0.5% BSA, 0.4% sta171

172

WU ET AL.

bilizing reagent. The assay buffer was 100 mM Tris– HCl, pH 8.4, containing 0.8 mg/ml ANS, 2.0 mg/ml sodium salicylate, 0.1% BSA, 0.04% NaN 3, 0.9% NaCl, 0.04% Tween 20. The wash solution was double-distilled water containing 0.04% Tween 20, with the pH adjusted to 8.0 by Tris. The extraction solution for fluorescence enhancement was prepared according to Hemmila et al. (5), with an identical fluorescence enhancement and fluorescence background (generally less than 400 cps) compared with the commercial DELFIA enhancement solution. T 4 Standards L-T 4 was weighed out and dissolved in 50 mM NaOH solution, containing 20% glycerol. The resulting concentrated stock solution was stored at 220°C. Whenever it was used, the stock solution was diluted to the desired concentrations in T 4-free human serum. Comparison Methods For comparison, we used the commercial DELFIA T 4 kit (Product No. 1244-030, EG&G Wallac Co.) and the DP-R radioimmunoassay T 4 kit (Product No. IMK 419, produced in our institute); these assays were performed exactly according to the instructions obtained by the manufacturers. Specimens Human serum samples were provided by 301 Hospital, Beijing, China. Procedures Preparation of T 4–protein conjugates. Four kinds of T 4–protein derivatives (bovine IgG, BSA, Tg, and ovalbumin as the carrier protein, respectively) were prepared with 1-ethyl-3-(-3-dimethylaminopropyl) carbodiimide (EDC) as a coupling reagent. Briefly, 0.6 g of L-T 4 was dissolved with 25 ml methanol in a two-neck reactor and the solution was heated to 60°C. HCl gas, generated by the reaction of NaCl and concentrated H 2SO 4, was then led into the T 4 solution. After a 2-h incubation, the solution was cooled to room temperature and the reaction was allowed to proceed for another 2 h. The solvent was then distilled at a low pressure and the resulting solid thyroxine methyl was washed twice separately with 5 ml water-free alcohol and 8 ml water-free ether. Thyroxine methyl (0.39 g) was obtained after drying. One gram of protein (one of the four proteins) was dissolved in 120 ml distilled water, and the pH was adjusted to 5.0 –5.5 with 0.1 N HCl. T 4 methyl (0.3 g) in 40 ml DMF, together with 300 mg EDC in 30 ml water, was added dropwise successively with constant stir-

ring. After a 1-h incubation at room temperature, 400 mg EDC in 10 ml water was added. After fully mixed, the reaction mixture was put in static at 4°C for 18 h and then dialyzed against distilled water to completely remove EDC and thyroxine methyl. The contents of the dialysis bag were filtered through a 0.2-mm (pore size) filter (Minipore Co.) to remove the insoluble sediment and then lyophilized and stored at 4°C. The molecular ratio of T 4 to protein was determined by measuring the absorbance of the conjugate solution (0.01% in distilled water) at 242 nm. The contribution to the absorbance by the protein was subtracted by the total absorbance to yield the absorbance of the coupled T 4; the T 4 concentration can then be calculated and the molecular ratio of the T 4 to protein can be determined by dividing the T 4 concentration by the protein concentration. They were 11.2, 4.1, 13.0, and 4.3 with respect to T 4– bovine IgG, T 4–BSA, T 4–Tg, and T 4– ovalbumin, respectively. Preparation of tracer. One milligram of anti-T 4 mAb in 50 mM sodium carbonate buffer, pH 9.5, was added to a small amber bottle containing 1 mg DTTAEu 31. After being fully mixed, the mixture was allowed to stand for 24 h at room temperature. The unreacted DTTA-Eu 31 was separated from the tracer by sizeexclusion chromatography on a 50 3 1.5-cm column of Sephadex G-50 (Pharmacia, Uppsala, Sweden), eluting with 50 mM Tris–HCl buffer, pH 7.75, containing 0.9% NaCl, 0.05% NaN 3. Eight milliliters of tracer solution was gathered. The concentration of Eu 31 was determined by fluorescence measurement. The concentration of the eluted protein was determined spectrophotometrically at 280 nm and calculated using following experimental equation: Concentration of europium-labeled protein (mg/ml) 5 $A 280 2 ~0.008 3 @Eu 31#!%/1.43 The molar ratio of Eu 31 to the mAb can be determined to be 4.1, obtained with the concentration of Eu 21 divided by the concentration of mAb. To stabilize the tracer solution, we added BSA to give a final concentration of l mg/ml and stored the preparation at 4°C. Coating of the microtitration wells. The T 4– bovine IgG conjugate was immobilized by absorption to the well surface with a concentration of 15 mg/ml. We added 200-ml aliquots of T 4– bovine IgG buffer to each well of the microtitration strip and left the conjugate to adsorb at 32°C overnight. We then washed the strips twice with the wash solution, added 250 ml of blocking buffer per well, and incubated for 4 h at 32°C. After removal of the blocking buffer, the strips were stored at 4°C in sealed and moist bay until use.

173

TIME-RESOLVED FLUORESCENCE IMMUNOASSAY OF THYROXINE

FIG. 1. Effect of the coating concentration of the four T 4 conjugates on the fluorescence signal (triplicate measurement) with a constant tracer amount. 1, T 4-– bovine IgG; 2, T 4-–Tg; 3, T 4–BSA; 4, T 4– ovabumin.

Protocol of DELFIA. We added 20 ml of T 4 standards or samples to the coated wells, followed by adding 200 ml of the europium-labeled antibody. After a 1.5-h incubation with gentle shaking at room temperature, we aspirated the solution of the wells and washed the wells six times with the wash solution and added 200 ml of enhancement solution into each well; after agitating the strips for 5min, we measured the fluorescence in Arcus 1230 fluorometer. The data were processed automatically using a smoothed spline curve fit function; the standard curve and the assay values were given out directly. RESULTS

Assay Optimization The selection of T 4–protein conjugates. Under identical conditions we prepared above four different T 4– protein conjugates; also under identical conditions we coated them separately in the coating buffer with a increasing concentration of each conjugates. After washing and blocking, we tested the T 4 activity of the four immobilized T 4 conjugates for binding to a constant amount of tracer (Fig. 1). The maximal capturing capacity (MCC) of the coated T 4 for the tracer was used to evaluate the T 4 activity of the coated conjugates, MCC, together with other characteristics of these conjugates is listed in Table 1. Figure 1 shows that, with the increase of the coating concentration of the conjugates, the binding signal, corresponding to the coated T 4 activity, initiated with a rapid increase and then reached a plateau. Of the four tested conjugates, only the coated T 4– bovine IgG showed the highest capacity for binding the labeled antibody. In our assay T 4– bovine IgG was selected as the best for coating because it gave the highest signal

and required the least amount of conjugate for effective coating. The coating concentration of the T 4– bovine IgG was determined to be 15 mg/ml, for this coating concentration gave the greatest signal difference between the fluorescence reading of the zero standard and that of the highest standard (240 ng/ml). The concentration of the tracer. With a coating concentration of 15 mg/ml, the most suitable concentration of the labeled antibody was determined to be about 80 ng/ml, obtained with 1500-fold dilution of the originally gathered tracer elution; higher or lower dilution unfavorably affected the assay characteristics with respect to the detection limit and dynamic range (data not shown). With this tracer concentration, the nonspecific binding of the tracer was less than 0.37% of the Bo value when without the coated antigen. Kinetics of the immunoreaction. The effect of incubation time on the immunoreaction was also investigated. The fluorescence signal of each standard point increased with response to the increase of the incubation time from the beginning to 180 min. Within the incubation duration of 90 –150 min, the shape of the standard curve and the measured T 4 values of the samples remained relatively unchanged. One hundred twenty minutes was shown to be favorable for incubation. Analytical Characteristics Standard curve and detection limit. A typical standard curve in the present assay with T 4– bovine IgG as the solid-phase antigen is shown in Fig. 2. The detection limit of the assay was 3.5 ng/ml, defined as the T 4 concentration corresponding to the mean fluorescence reading of 24 determinations of zero standard minus two standard deviations. Precision. The within-run precision were calculated by replicate analysis of three clinical serum samples in a single assay and the between-run precision by duplicate measurement of these samples in 18 different runs. The mean T 4 values of the three tested serum

TABLE 1

Characterization of the T 4–Protein Conjugates Carrier protein

Molecular mass, kDa

Molecular ratio (T 4/protein)

MCC a

BSA Ovalbumin Bovine Tg Bovine IgG

60 46 660 160

4.1 4.3 13.0 11.2

100 91 122 212

a

MCC of T 4–BSA was prescribed to be 100, MCC of the other three conjugates were relative to this criterion.

174

WU ET AL.

FIG. 2. A typical standard curve of the present T 4 assay and its precision profile.

were 48.8, 109.1, and 223.0 ng/ml, the within-run CVs were 4.1, 3.2, and 7.8%, and the between-run CVs were 5.5, 4.3, and 7.2%, respectively. Dilution test. The dilution linearity was assessed by assaying samples serially diluted with T 4-free human serum. There was a good agreement between the expected and the observed values, as is shown in Table 2. Recovery. We added a known amount of exogenous T 4 (47–181.2 ng/ml) to serum pools. The recovery was assessed by analyzing the samples before and after the addition and subtracting the endogenous T 4, ranging from 92 to 103%, with a mean value of 98.3%. Correlation with other methods. For comparison, we analyzed 64 clinical samples by the present method and by the commercially DELFIA T 4 kit from Wallac. The regression equation is as follows: y 5 0.89x 1 15.28, r 5 0.929. We also compared the results of 38 clinical samples by the present method and by the RIA T 4 kit produced in our institute. The regression equation is as follows: y 5 1.05x 1 10.6, R 5 0.967. Within the whole T 4 concentration range of the tested samples (6.5– 401 ng/ml), the values obtained by the present

FIG. 3. Correlation of the present method with Wallac DELFIA: 41 samples from healthy subjects, 6 samples from hypothyroid patients, 11 samples from hyperthyroid patients, and 6 samples from pregnant women.

method were in good agreement with those obtained by the commercial DELFIA and RIA, as shown in Figs. 3 and 4. DISCUSSION

The solid-phase competitive immunoassay for haptens has been well established with different configurations (6), of which, the following two designs are used most commonly: A, antibody immobilized and labeled antigen as detection probe; B, antigen immobilized and labeled antibody as detection probe. For design A, the direct antibody or the second antibody against the direct antibody may be used for coating, the latter of which has been adopted more often for the development of commercial kits because it makes it possible to use the second antibody-coated solid phase as a com-

TABLE 2

Dilution of the Samples with T 4-Free Human Serum Dilution factor

Sample

Original sample

2

4

8

16

Expected Observed Expected Observed Expected Observed

— 277.2 — 108.0 — 48.8

138.60 148.20 54.00 56.10 24.4 28.3

69.30 70.3 27.00 25.50 12.2 11.5

34.65 30.50 13.5 10.8 6.1 8.3

17.33 12.90 6.75 6.70 3.05 —

FIG. 4. Correlation of the present method with RIA: 21 samples from healthy subjects, 6 samples from hypothyroid patients, 11 samples from hyperthyroid patients, and 6 samples from pregnant women.

TIME-RESOLVED FLUORESCENCE IMMUNOASSAY OF THYROXINE

mon binder for a variety of hapten assays, and endows the immunoassay with improved precision and reproducibility. In addition, it allows the amount of direct antibody, detection probe, and the standard or sample to be changed more freely according to the assay demands for different haptens when optimizing the assays. However, compared with the antigen-coating approach, the coated antibody is more likely to lose its immunoreactivity, and there is a technologic requirement for the coating and blocking, or have to choose a robust antibody for long storage of the kits. Conversely, by the present antigen-immobilized approach, the hapten activity on the solid phase is not as sensitive to the environmental factors as that of the coated antibody because the variation of the configuration of the hapten conjugate caused by the environmental factors, such as temperature and humidity, does not decrease the activity of the coupled hapten obviously. Meanwhile, the configuration change of the antibody may be vital for its activity because the disappearing of the antibody binding sites is not likely to be compensated by the newly appeared ones (if it is possible), as is the case for the coated hapten–protein conjugate. The T 4– bovine IgG-coated wells in our experiments are stable for at least 6 months stored at room temperature. Until the present there has seldom been an article for assaying serum total thyroxine with T 4–protein conjugate coating on the surface of strip wells or tubes as solid-phase antigen. This is partly because of the high T 4 concentrations in human serum, some of which in a hyperthyroid patient can be more than 400 ng/ml. This implies that the solid phase must immobilize a large amount of T 4 conjugate to compete with the T 4 in samples or standards, otherwise the assay cannot be established or cannot correctly measure samples with a high T 4 level. In our experiments, of the four T 4 conjugates used, the T 4– bovine IgG showed the highest capture capacity for the labeled antibody. We postulate that this is ascribed to the different structure of the carrier proteins which is relative to the different T 4 coupling and coating characteristics. To obtain enough antibody capture capacity of the solid-phase thyroxine, the following factors were considered in our present study: (1) The extent of the interprotein crosslinking in the preparation of the T 4 conjugate was well controlled by adjusting the amounts and adding times of EDC. A mild crosslinking among the proteins made the conjugate more beneficial to coating solid phase, but overcrosslinking generated nonsoluble aggregates and decreased the recovery of the desired T 4 conjugate. (2) The molecular ratio of T 4 to protein in the conjugate should be as high as possible. This was mainly affected by the preparation procedure and the material composition. Before being coupled to the proteins, the methylation of T 4 is important for increasing the T 4 molecular ratio.

175

In the DELFIA system, the immunoreaction should be carried out at neutral pH or in a weak base solution to avoid the unwanted leakage of Eu 31 from tracer. Under these pH conditions, trichloroacetic acid (TCA) has been reported to inactivate the serum-binding proteins of several haptens (7, 8) and, thus, has been used as a blocker to release haptens from their relative binding proteins in direct DELFIA. In the present study, 0.2– 0.3 M TCA–Na, which works satisfactorily in the DELFLA of cortisol and progesterone (7, 8), could not effectively dissociate T 4 from TBG whenever the serum samples from pregnant woman were tested because of high TBG concentrations. Only when the concentration of TCA–Na reached 0.6 M or more could the blocking be considered to be complete; however, at this TCA–Na concentration the immunoreaction between the coated thyroxine and the labeled antibody was seriously inhibited. Therefore, TCA–Na was abandoned in our assay design. A coblocker consisting of 2 mg/ml sodium salicylate and 0.8 mg/ml ANS proved to be effective without a obvious restraint to the immunoreaction. The use of DELFIA in China has been thought to be questionable (9) because China is a rare-earth-element-rich country and the enhancement step of the DELFIA is susceptible to the contamination of these elements. Countering this viewpoint, we first prepared enhancement solution (background fluorescence: 348 cps) in an ordinary laboratory without any air-cleaning apparatus, and then let this bottle of enhancement solution stand in the same laboratory for various periods of time with the mouth of the bottle (diameter: 1.5 cm) open. The volume of the enhancement solution was 25 ml. After a 1-, 7-, and 24-h exposure to the air, the background of the enhancement solution was counted as 400, 640, and 980 cps, respectively. This means the enhancement solution can still work well even if it is open to air for as long as 24 h because the enhancement solution with a background of 1000 cps is not thought to be too high to assay analytes such as TSH and anti-HIV antibodies with a requirement of high analytical sensitivity. Practically, the enhancement solution contacts with air only in the fluorescence enhancement and measurement step, which is generally less than 20 min. According to these results, it can be generally concluded that the contamination problem does not pose a barrier for the utilization of DELFIA in China. The serum TT 4 TRFIA described here is rapid and simple; all the quality criteria required for clinical measurement can be fulfilled with respect to detection limit, precision, accuracy, and dynamic range. The determinations by the present method correlated well with these by the commonly used commercial DELFIA or RIA. Moreover, it is considered that the present

176

WU ET AL.

immobilized antigen approach may be more suitable to assay haptens with a relatively low concentration in samples or to assay the free fraction of the haptens, because in these situations the required amount of the coated antigen can be easily realized with an appropriate hapten conjugate, or, more directly, by using the solid phase which can covalently bind the hapten (10, 11).

5.

6.

7.

REFERENCES 1. Papanastasiou-Diamandi, A., Bhayana, V., and Diamandis, E. P. (1989) A time-resolved fluoroimmunoassay of total thyroxine in serum. Ann. Clin. Biochem. 26, 238 –243. 2. Sturgess, M. I., Weeks, I., Mpoko, C. N., Laing, I., and Woodhead, J. S. (1986) Chemiluminescent labeled-antibody assay for thyroxine in serum, with magnetic separation of the solid-phase. Clin. Chem. 32, 532–535. 3. Izquierdo, J. M., Sotorrio, P., and Quiros, A. (1984) Enzyme immunoassay of thyroxine with a centrifugal analyzer. Clin. Chem. 28, 123–125. 4. Fetrua, B., Genetef, F., Savaron, M. L., Moulin, C., Salard, L., and Masseyeff, R. (1986) A novel enzyme immunoassay for total thyroxine using immobilized antibody and hydrophobic chroma-

8.

9.

10.

11.

tography purified thyroxine–peroxidase conjugate. J. Immunol. Methods 87, 137–143. Hemmila, I., Dakubu, S., Mukkala, V., Siitari, H., and Lovgren, T. (1984) Europium as a label in time-resolved immunofluorometric assays. Anal. Biochem. 137, 335–343. Khosravi, M. J., and Papanastasiou-Diamandi, A. (1993) Hapten-heterologous conjugates evaluated for application to free thyroxine immunoassays. Clin. Chem. 39, 256 –262. Dechaud, H., Bador, R., Claustrat, F., Desuzinges, C., and Mallein, R. (1988) New approach to competitive lanthanide immunoassay: Time-resolved fluoroimmunoassay of progesterone with labeled analyte. Clin. Chem. 34, 501–504. Eskola, J. U., Nanto, V., Meurling, L., and Lovgren, T. N. E. (1985) Direct solid-phase time-resolved immunofluorometric assay of cortisol in serum. Clin. Chem. 31, 1731–1734. Yiang, X., Chang, W,, and Ci, Y.. (1995) Immunoassay: A comprehensive review and prospect. Prog. Chem. 7, 83–97. [in Chinese] Niveleau, A., Sage, D., Reynaud, C., Bruno, C., Legastelois, S., Thomas, V., and Dante, R. (1993) Covalent linking of haptens, proteins and nucleic acids to a modified polystyrene support. J. lmmunol. Methods 159, 177–187. Hofstetter, O., Hofstetter, H., Then, D., et al. (1997) Direct binding of low molecular weight haptens to ELISA plates. J. Immunol. Methods 210, 89 –92.