- Email: [email protected]
Simultaneous Determination of α-Fetoprotein and Carcinoembryonic Antigen in Human Serum by Time-Resolved Fluoroimmunoassay
Analytical Biochemistry 276, 81– 87 (1999) Article ID abio.1999.4336, available online at http://www.idealibrary.com on
Simultaneous Determination of a-Fetoprotein and Carcinoembryonic Antigen in Human Serum by Time-Resolved Fluoroimmunoassay Kazuko Matsumoto,* ,1 Jingli Yuan,* Guilan Wang,* and Hiroko Kimura† *Department of Chemistry, Waseda University, Shinjuku-ku, Japan Science and Technology Corporation, Tokyo 169-8555, Japan; and †Department of Forensic Medicine, Juntendo University School of Medicine, Tokyo 113-8421, Japan
Received June 16, 1999
A novel simultaneous measurement method for a-fetoprotein (AFP) and carcinoembryonic antigen (CEA) in human sera by time-resolved fluoroimmunoassay (TR-FIA) is described. The proposed approach combines the use of europium-labeled anti-AFP antibody for AFP TR-FIA and biotinylated anti-CEA antibody complexed to samarium-labeled streptavidin for CEA TR-FIA. A 96-well microtiter plate coated with a mixture of anti-AFP and antiCEA monoclonal antibodies was used for the assay. After it was reacted with a solution containing AFP and CEA, a mixture of anti-AFP antibody labeled with BHHCT–Eu 31 and biotinylated anti-CEA antibody was added. The AFP concentration was determined by measuring the solid-phase fluorescence of the europium-labeled anti-AFP antibody at 615 nm. Then a BHHCT–Sm 31 -labeled streptavidin– bovine serum albumin conjugate (SA–BSA) was added to react with the biotinylated anti-CEA antibody. After the reaction, the unreacted SA–BSA was washed out, and a 0.1 M NaOH solution containing 1.0 3 10 25 M TOPO and 0.05% SDS was added to dissociate the samarium-labeled SA–BSA in the immune complex on the surface of the well into the solution. The CEA concentration was determined by measuring the solution fluorescence of 643 nm from the samariumlabeled SA–BSA. The present method gives detection limits of 0.07 ng/ml for AFP and 0.3 ng/ml for CEA. The coefficient variations of the method are less than 7%, and the recoveries are in the range of 90 – 110% for serum samples. The AFP and CEA concentrations in 27 human serum samples were determined by the present method as well as by single assay for comparison. A good correlation was ob-
1 To whom correspondence should be addressed. Fax: 181-3-52733489. E-mail: [email protected].
0003-2697/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
tained with the correlation coefficients of 0.990 for AFP and 0.973 for CEA. © 1999 Academic Press
Human AFP 2 and CEA are typical carcinoembryonic proteins and are known to be related to hepatoma and nonseminomatous testicular tumors, colon cancer, pulmonary cancer, and mammary cancer (1– 6). A number of immunoassay methods for human AFP and CEA have been reported (7–17), and it is possible to use various commercial AFP and CEA assay kits with appropriate sensitivity. Time-resolved fluoroimmunoassay (TR-FIA) using europium chelates as the labels has been noticed as a highly sensitive method (18 –25). However, the europium chelates used in the current Delfia and FIAgen systems have some limitations. To improve the sensitivity of TR-FIA, we recently developed a new tetradentate b-diketone label, BHHCT, for chelating Eu 31, and by using SA or a SA–BSA conjugate labeled with BHHCT–Eu 31, human serum AFP, IgE, and TSH were determined with over two orders of magnitude higher sensitivity than those of other immunometric methods (26 –28). It has been reported that some samarium and terbium chelates have strong fluorescence and long lifetime as well as europium chelates and can be used as labels in TR-FIA (25, 29). Since the chelates of the three lanthanide ions have very sharp fluorescence profiles (a half-width of ca. 10 nm) each at their own 2 Abbreviations used: AFP, a-fetoprotein; CEA, carcinoembryonic antigen; TR-FIA, time-resolved fluoroimmunoassay; BHHCT, 4,49bis(10,10,10,20,20,30,30-heptafluoro-40,60-hexanedion-60-yl)-chlorosulfo-o-terphenyl; BSA, bovine serum albumin; SA, streptavidin; IgE, immunoglobulin E; TSH, thyroid-stimulating hormone; TOPO, tri-n-octylphosphine oxide; CV, coefficient variation; arb. units, arbitrary units.
81
82
MATSUMOTO ET AL.
different wavelength, two or three components in serum can be simultaneously determined from a single sample by using the different chelates as labels in multiple-label TR-FIA. Such multiple-label assays have been reported by using Eu 31–Sm 31, Eu 31–Tb 31, or Eu 31–Tb 31–Sm 31–Dy 31 labeling (29, 30). The dual-label TR-FIA of human serum lutropin and follitropin (31), myoglobin and carbonic anhydrase III (32), AFP and free b-subunit of human chorionic gonadotropin (33), pregnancy-associated plasma protein A and free b-subunit of human chorionic gonadotropin (34), and free and total prostate-specific antigen (35) have been reported. However, the fluorescence of Sm 31 chelate is rather weak compared to that of Eu 31 chelate. The quantum yield of Sm 31 chelate is usually several hundred times lower than that of Eu 31 chelate (36), and the fluorescence lifetime of Sm 31 chelate is shorter (usually below 100 ms) (37) than that of Eu 31 chelate (usually over several hundreds microseconds). These cause the TR-FIA using Sm 31 chelate as the label to be less sensitive than that of using Eu 31 chelate (37). Because of this, TR-FIA using Sm 31 label is suitable only for the detection of a high-concentration component in the sample. The Eu 31–Sm 31 pair dual-label TR-FIA was thus used for the simultaneous measurement of lowconcentration and high-concentration components in a sample, respectively. However, the concentrations of human AFP and CEA in sera are usually low, at the ng/ml level. The current immunoassay methods for AFP and CEA give detection limits of 0.1–1 ng/ml for AFP (26) and 0.1– 0.5 ng/ml for CEA (15, 16, 38). Therefore, in the simultaneous method for human AFP and CEA, the sensitivity must be high enough not only for AFP but also for CEA. Using the reported Eu 31–Sm 31 dual-label TR-FIA (31– 34), simultaneous detection of AFP and CEA is difficult because of the lower sensitivity of the Sm 31 chelate. Although many excellent methods have been developed for the immunoassay of AFP and CEA in human serum, these are usually time and sample consuming. In addition, simultaneous measurement of AFP and CEA by enzyme immunoassay is not possible, since coloration reaction is necessary before measurement. In the present work, we developed a novel TR-FIA method for simultaneous measurement of AFP and CEA in human serum by using BHHCT–Eu 31-labeled anti-AFP antibody, biotinylated anti-CEA antibody, and BHHCT–Sm 31-labeled SA–BSA conjugate. In the present method, the emission of the Sm 31 label was amplified by introducing a biotinylated antibody– Sm 31-labeled streptavidin system. The method has the advantage of simple assay format, rapidity, high sensitivity, and small sample amount. The detection limits of the present method are 0.07 ng/ml for AFP and 0.3 ng/ml for CEA, which are low enough to measure AFP and CEA in human serum. The AFP and CEA concen-
trations in 27 human serum samples were determined by the present method as well as by the usual single assay with either of the chelates for comparison. A good correlation was obtained for the simultaneous and single assay with the correlation coefficients of 0.990 for AFP and 0.973 for CEA. MATERIALS AND METHODS
Labeling and Biotinylation of Antibodies Labeling of anti-AFP antibody and SA–BSA conjugate with BHHCT. Two milliliters of goat anti-human AFP antibody solution (Nippon Bio-Test Laboratories, Inc.; the anti-AFP antibody concentration is 0.5 mg/ml, total protein concentration is 6.3 mg/ml; immunoelectrophoresis shows that the impurity antibody does not exist) was dialyzed twice against 3 liters of saline– water at 4°C each for 24 h. To the dialyzed solution were added 8 ml of water and 84 mg of NaHCO 3, which made the solution ca. 0.1 M in NaHCO 3. The pH of the solution was adjusted to 9.1 with 0.1 M NaOH. Dry ethanol solution (180 ml) containing 3.0 mg of BHHCT was added dropwise with stirring, and the solution was further stirred for 1 h at room temperature. After a small amount of precipitate was removed by centrifuge, the unreacted free labeling reagent was separated from the labeled protein on a Sephadex G-50 gel column with 0.05 M NH 4HCO 3 of pH 8.0 as the eluent. To the labeled protein solution were added 30 mg of BSA and 20 mg of NaN 3, and the pH was adjusted to 6.4 with 1 M HCl. The solution was stored at 220°C before use. Since the labeled antibody solution contains proteins other than anti-AFP antibody, the labeling ratio was not determined. When the labeled antibody solution was used in a single assay of AFP, it was diluted 600-fold with 0.05 M Tris–HCl buffer of pH 7.8, containing 1.0 3 10 27 M EuCl 3, 0.2% BSA, 0.1% NaN 3, and 0.9% NaCl. Preparation and labeling of the SA–BSA conjugate, SA(BSA) 0.9(BHHCT) 46, were described in a previous report (27). The conjugate solution was 200-fold diluted with 0.05 M Tris–HCl of pH 7.8, containing 1.0 3 10 26 M SmCl 3, 0.2% BSA, 0.1% NaN 3, and 0.9% NaCl, and was incubated at 56°C for 2 h, before use in immunoassay. Biotinylation of anti-CEA antibody. After two dialyses of 0.5 ml of rabbit anti-human CEA antibody solution (2.4 mg/ml, IgG fraction of rabbit antiserum, Dako-immunoglobulins, Denmark) for 24 h at 4°C against 3 liters of saline–water, 0.5 ml of water, 8.4 mg of NaHCO 3, and 6 mg of sulfosuccinimidyl-6-(biotinamido) hexanoate (NHS–LC– biotin, Pierce Chemical Co.) were added with stirring. After stirring for 1 h at room temperature, the solution was incubated for 24 h at 4°C. After the solution was twice dialyzed each for 24 h at 4°C against 3 liters of 0.1 M NaHCO 3 contain-
DETERMINATION OF a-FETOPROTEIN AND CARCINOEMBRYONIC ANTIGEN
ing 0.25 g of NaN 3, 10 mg of BSA was added, and the solution was stored at 220°C before use. When the biotinylated antibody solution was used in a single assay of CEA, it was diluted 1000-fold with 0.05 M Tris–HCl buffer of pH 7.8, containing 0.2% BSA, 0.1% NaN 3, and 0.9% NaCl. When the BHHCT-labeled anti-AFP antibody and biotinylated anti-CEA antibody were used for the simultaneous measurement of AFP and CEA, a mixture of 1/600 BHHCT-labeled anti-AFP antibody solution and 1/1000 biotinylated anti-CEA antibody solution was prepared with 0.05 M Tris–HCl buffer of pH 7.8, containing 1.0 3 10 27 M EuCl 3, 0.2% BSA, 0.1% NaN 3, and 0.9% NaCl. Instrument and Measurement Conditions To measure the fluorescence properties of BHHCT– Sm 31, a BHHCT-labeled BSA solution was prepared as previously reported (26). It was reacted with SmCl 3 in buffer to give a BHHCT–Sm 31 solution. The excitation and emission spectra of BHHCT–Sm 31 were measured on an Hitachi F-4500 fluorospectrometer. The fluorescence lifetime of BHHCT–Sm 31 was measured on a Perkin–Elmer LS 50B luminescence spectrometer. TRFIA measurement of AFP and CEA was performed on an Arcus 1232 time-resolved fluorometer (Wallac Oy), to which a filter (Wallac Oy) of 643 nm for Sm 31 chelate was attached. Eu fluorescence measurement was carried out with excitation at 340 nm, emission at 615 nm, delay time of 0.2 ms, and window time of 0.4 ms (specific fluorescence was measured after a 200-ms time delay for 400 ms between each excitation pulse). For Sm, excitation at 340 nm, emission at 643 nm, delay time of 0.03 ms, and window time of 0.10 ms (specific fluorescence was measured after a 30-ms time delay for 100 ms between each excitation pulse) were used. Immunoassay of AFP and CEA Human serum samples were obtained from Juntendo University Hospital and were stored at 220°C before use. The standard AFP and CEA solutions were prepared by diluting an a-1-fetoprotein solution (Dakopatts, Denmark) and a CEA solution (Seikagaku Kogyo Co., Ltd., Japan) with 0.05 M Tris–HCl buffer of pH 7.8 containing 5% BSA, 0.1% NaN 3, and 0.9% NaCl. The immunoassay format for the simultaneous measurement of AFP and CEA is shown in Fig. 1, and is described below. One hundred microliters of 0.1 M carbonate buffer of pH 9.6 containing 5 mg/ml of antiAFP monoclonal antibody (Biostride, Inc.) and 5 mg/ml of anti-CEA monoclonal antibody (Boehringer-Mannheim Biochemica) was incubated in a 96-well microtiter plate (Fluoro Nunc module plates) for 24 h at 4°C.
83
FIG. 1. Schematic representation of the simultaneous measurement system for AFP and CEA. For the details of the procedure, see Materials and Methods.
The wells were rinsed twice with 0.05 M Tris–HCl buffer of pH 7.8 containing 0.05% Tween 20 (buffer 1) and once with 0.05 M Tris–HCl buffer of pH 7.8 (buffer 2). The coated plate was stored at 220°C before use. The antigen–antibody reaction was carried out as follows. After 50 ml of the AFP–CEA standard solution or human serum was incubated at 37°C for 1 h in the coated wells, the wells were rinsed twice with buffer 1 and once with buffer 2. After a 50-ml mixture of BHHCT–Eu 31-labeled anti-AFP and biotinylated antiCEA was incubated in the wells at 37°C for 1 h, the wells were rinsed four times with 0.05 M Tris–HCl of pH 9.1 containing 0.05% Tween 20 (buffer 3). The rinsed plate was subjected to solid-phase Eu time-resolved fluorometric measurement. After measurement, the plate was rinsed once with buffer 2, and 50 ml of the SA(BSA) 0.9(BHHCT) 46–Sm 31 solution was incubated in the wells at 37°C for 1 h. The wells were rinsed four times with buffer 3. Fifty microliters of 0.1 M NaOH containing 1.0 3 10 25 M TOPO and 0.05% SDS was added to the wells and incubated at room temperature for 10 min. The solution-phase Sm time-resolved fluorometric measurement was carried out. To compare the correlation between the simultaneous measurement and the independent single measurement for each of the two antigens, AFP and CEA were assayed also in a single-assay format, in which only one of the two antibodies was coated to the plate. The standard solution containing only either AFP or CEA and the labeled antibody containing only either BHHCT–Eu 31-labeled anti-AFP or biotinylated antiCEA were used. Other conditions for the single assay were the same as those for the simultaneous assays. The cross-reactions between the two antigens were examined with the single assays of anti-AFP antibody– CEA and anti-CEA antibody–AFP systems. RESULTS AND DISCUSSION
The Fluorescence Properties of BHHCT–Eu 31 and BHHCT–Sm 31 Because BHHCT is insoluble in water, the fluorescence properties were examined using a BHHCT-la-
84
MATSUMOTO ET AL.
FIG. 2. Excitation and emission spectra of BHHCT (bound to BSA)–Sm 31 chelate in 0.1 M NaOH, containing 0.05% SDS and 1.0 3 10 25 M TOPO. [BHHCT] 5 [Sm 31] 5 1.0 3 10 25 M. The excitation spectrum was recorded with l em 5 561 nm and the emission spectrum with l ex 5 335 nm.
beled BSA solution. The fluorescence properties of BHHCT (bound to BSA)–Eu 31 have been discussed in a previous report (26). This chelate has excitation and emission maxima at 326 nm and 611.6 nm in a Tris– HCl buffer. Its lifetime in 0.1 M Tris–HCl buffer of pH 9.1 is greater than 380 ms. Excitation and emission spectra of BHHCT (bound to BSA)–Sm 31 chelate in a 0.1 M NaOH– 0.05% SDS–1.0 3 10 25 M TOPO solution are shown in Fig. 2. The maximum excitation at 328 nm is the same as that for BHHCT–Eu 31, but the Sm 31 chelate gives three emission maxima at 561, 597, and 642 nm, which correspond to 4G 5/2 3 6H 5/2 (561 nm), 4 G 5/2 3 6H 7/2 (597 nm), and 4G 5/2 3 6H 9/2 (642 nm), respectively (39). The lifetime (642-nm emission) of the BHHCT–Sm 31 chelate was measured in 0.1 M Tris– HCl buffer of pH 9.1 and in a 0.1 M NaOH– 0.05% SDS–1.0 3 10 25 M TOPO solution. The lifetimes in the two solutions are 31.1 and 70.7 ms, respectively.
of a biotin–streptavidin system is effective for increasing the labeling number per antibody molecule and obtaining a high sensitivity (40). A biotin–streptavidin system was therefore employed to raise the sensitivity of the Sm fluorescent label. As described above, the lifetime of the BHHCT–Sm 31 chelate is only 31.1 ms in Tris–HCl buffer, and therefore the solid-phase measurement after the labeled SA–BSA reacts with biotinylated anti-CEA antibody on the solid in Tris–HCl buffer is not very sensitive (see Fig. 3). On the other hand, a solution-phase time-resolved fluoromeasurement method is usually more sensitive, if the fluorescence probe complex is transferred from the solid phase into the solution (41, 42). The mechanism of the dissociation of the solid-phase immune complex for fluorescence measurement has been discussed in a previous work (42). The dissociation solutions such as SDS– urea or SDS–TOPO are usually employed (41, 42). In the present work, the BHHCT–Sm 31-labeled SA–BSA was dissolved from the solid-phase into the solution by NaOH–SDS–TOPO, and the solution was measured for Sm fluorescence. The dissociation of BHHCT–Sm 31 reached equilibrium in ca. 10 min at room temperature, and the signal intensity did not change for a further ca. 30 min. As shown in Fig. 3, the Sm fluorescence intensity in the solution phase is higher than that in the solid phase, while the background intensity increases only slightly, and therefore a lower detection limit can be obtained in solution-phase measurement. The cross-reactions between anti-AFP antibody and CEA and vice versa were examined. The results, shown in Fig. 4, show that the cross-reaction in the present system is negligible. Furthermore, the results show
Evaluation of the Assay Format The immunoassay format for the simultaneous measurement of AFP and CEA is shown in Fig. 1. Owing to the BHHCT–Eu 31 chelate having strong fluorescence with a very long lifetime (the detection limits of BHHCT–Eu 31 chelate are 8.4 3 10 213 M in Tris–HCl buffer and 2.3 3 10 213 M in SDS–TOPO–NaHCO 3 buffer, respectively) (26), TR-FIA using BHHCT–Eu 31 has very high sensitivity even in the solid-phase measurement (26 –28). In contrast to this, the Sm 31 chelate does not have such a strong and long-lifetime fluorescence, which makes the immunoassay using a Sm label less sensitive (37). We also tested the CEA TR-FIA by direct labeling the anti-CEA antibody and measuring the fluorescence for comparison with the assay format in Fig. 1. However, a clear significant signal was not obtained below 1000 ng/ml. In such a case, introduction
FIG. 3. The relative fluorescence intensities in the solid-phase and solution-phase measurements of CEA by using BHHCT–Sm 31-labeled SA–BSA in single-label assay. The solid-phase measurement means that the fluorescence measurement was carried out on the solid-phase immune complex. The solution-phase measurement means that the fluorescence measurement was carried out after the solid-phase immune complex was dissolved into the solution by a NaOH–SDS–TOPO solution.
DETERMINATION OF a-FETOPROTEIN AND CARCINOEMBRYONIC ANTIGEN
85
ground intensities in the simultaneous assay and the single assay for CEA did not change in the presence and absence of the labeled anti-AFP antibody. The above results indicate that AFP and CEA in serum can be simultaneously assayed by using the present method. Assay of Human Serum Samples
FIG. 4. The cross-reactivity of anti-AFP antibody with CEA (A) and anti-CEA antibody with AFP (B). (A) After anti-AFP antibody-coated wells were reacted with CEA standard solutions at 37°C for 1 h, BHHCT–Eu 31-labeled anti-AFP antibody was added, and the solution was incubated at 37°C for 1 h. After washing, the Eu 31 fluorescence from the solid-phase was measured. (B) After anti-CEA antibody-coated wells were reacted with AFP standard solutions at 37°C for 1 h, biotinylated anti-CEA antibody was added, and the solution was incubated at 37°C for 1 h. After washing, BHHCT–Sm 31-labeled SA–BSA was added, and the solution was incubated at 37°C for 1 h. After washing, a 0.1 M NaOH solution containing 1.0 3 10 25 M TOPO and 0.05% SDS was added, and the solution was incubated at room temperature for 10 min. The Sm 31 fluorescence from the solution was measured.
that the existence of the BHHCT–Eu 31-labeled antiAFP antibody did not affect the CEA assay, which was confirmed by the fact that both the signal and back-
The calibration curves of simultaneous measurement of AFP and CEA are shown in Fig. 5. The curves of background-subtracted signals vs antigen concentrations are shown in Fig. 5. Since the delay time for Sm 31 fluorescence is ca. an order of magnitude smaller than that of Eu 31, the background intensity for the CEA assay is ca. 12 times larger than that of AFP. This background, caused by the coexisting proteins and plastic solid phase (a part of the background may be caused by the nonspecific adsorption of the labeled protein), can be greatly reduced by a prolonged delay time such as in the assay of AFP. However, since the lifetime of the Sm label is not so long, a longer delay time is not possible. The detection limits of the present method are calculated (41) to be 0.07 and 0.3 ng/ml for AFP and CEA, respectively, which are low enough for the assay of human serum. The average CV (n 5 4, within-run) of AFP and CEA in the ranges in Fig. 5 are 4.34 and 2.87%, respectively. The upper limits of the calibration curves are ca. 100 and ca. 1000 ng/ml for AFP and CEA, respectively. In the range below 100 ng/ml, two calibration curves show nearly straight lines when background-subtracted fluorescence signals vs antigen concentrations were plotted. However, the slopes of the two calibration curves are below 1.0,
FIG. 5. Calibration curves of the simultaneous measurement of AFP (A) and CEA (B). bg, background. The inset curves are plotted for background-subtracted signals vs the concentrations of AFP and CEA.
86
MATSUMOTO ET AL. TABLE 1
Precision in the Simultaneous Measurement of AFP and CEA in Human Sera Interassay (n 5 4) a AFP (ng/ml) Sample Sample Sample Sample Sample a
1 2 3 4
CEA (ng/ml)
Mean
CV (%)
Mean
CV (%)
1.05 3.92 5.46 10.02
5.56 5.02 5.52 6.10
6.57 4.02 7.90 13.65
5.32 3.22 4.79 6.05
Four assays using four different plates (one assay per plate).
which might be due to the different efficiency of the antigen–antibody reaction on the solid-phase surface in a different antigen concentration range or caused by the different efficiency of nonspecific bindings of the Eu 31-labeled antibody and Sm 31-labeled SA–BSA in different antigen concentration ranges. If the antigen– antibody reaction efficiency is higher at lower antigen concentration, or if the nonspecific binding efficiency is higher at lower antigen concentration, then the slope of the calibration curve would be less than 1.0. To evaluate the assay precision, four serum samples containing different AFP and CEA concentrations were measured four times using four different plates. The results are summarized in Table 1. The CVs (between-
run) of four assay results were in the range of 5.0 – 6.5% for AFP and 3.0 – 6.5% for CEA. These results show that the present method has good precision not only for the standard solution assay but also for the serum sample assay. To evaluate the accuracy of the assay, standard solutions of AFP and CEA mixtures (the concentrations of AFP and CEA are below 20 ng/ml) were added to four serum samples containing different AFP and CEA concentrations. After the concentrations of AFP and CEA in four serum samples and in the four solutions to which standard AFP and CEA were added to sera were measured using the simultaneous measurement method, the analytical recoveries of AFP and CEA were calculated, respectively. The analytical recoveries of AFP and CEA added to four serum samples are 89.4, 106.2, 96.2, and 103.4% (average recovery 5 98.8%) for AFP and 100.6, 97.2, 105.3, and 91.2% (average recovery 5 98.6%) for CEA, respectively. These results show that the analytical accuracy of the present method is good enough for the application of AFP and CEA assays in human sera. The AFP and CEA concentrations in 27 human sera were determined with both the simultaneous measurement method and the single-label assay method. Correlations of the two methods are shown in Fig. 6. The correlation coefficients of AFP and CEA are 0.990 and 0.973, respectively. These show that the two proteins can be determined by the present method with high precision.
FIG. 6. Correlation between the single-label measurement and the simultaneous measurement of AFP (A) and CEA (B) in human sera (n 5 27). The single-label measurement means that AFP and CEA were assayed independently. AFP single-label measurement: after anti-AFP antibody-coated wells were reacted with AFP standard solutions or serum samples at 37°C for 1 h, BHHCT–Eu 31-labeled anti-AFP antibody was added, and the solution was incubated at 37°C for 1 h. After washing, the Eu 31 fluorescence from the solid phase was measured, and the AFP concentrations in the samples were calculated. CEA single-label measurement: after anti-CEA antibody-coated wells were reacted with CEA standard solutions or serum samples at 37°C for 1 h, biotinylated anti-CEA antibody was added, and the solution was incubated at 37°C for 1 h. After washing, BHHCT–Sm 31-labeled SA–BSA was added, and the solution was incubated at 37°C for 1 h. After washing, a 0.1 M NaOH solution containing 1.0 3 10 25 M TOPO and 0.05% SDS was added, and the solution was incubated at room temperature for 10 min. The Sm 31 fluorescence from the solution was measured, and the CEA concentrations in the samples were calculated.
DETERMINATION OF a-FETOPROTEIN AND CARCINOEMBRYONIC ANTIGEN
Based on the above results, we can conclude that the present AFP and CEA simultaneous detection method can be practically employed to the clinical analysis of AFP and CEA. In addition, we can expect that application of the present method can be extended to other two-component simultaneous measurement systems. In the present method, because of weak fluorescence of the samarium label, the CEA assay was carried out with the use of a biotin–streptavidin amplifying system and a dissociation step for solution-phase measurement. Although this method enables high sensitivity for the CEA assay, the procedure is complicated compared with the method using the antibody labeled with fluorescence label and direct solid-phase measurement. Among lanthanide fluorescence labels, it is known that terbium labels have higher quantum yields and longer lifetimes than samarium labels. Instead of the samarium label, the simultaneous use of europium and terbium labels for two-component measurement would be a more ideal method. ACKNOWLEDGMENT The present work has been supported by CREST (Core Research for Evolutional Science and Technology) of Japan Science and Technology Corporation (JST).
REFERENCES 1. Buamah, P. K., Harris, R., James, O. F. W., and Skillen, A. W. (1986) Clin. Chem. 32, 2083–2084. 2. Gold, P., and Freedman, S. O. (1965) J. Exp. Med. 122, 467– 481. 3. Fuks, A., Banjo, C., Shuster, J., Freedman, S. O., and Gold, P. (1975) Biochem. Biophys. Acta 417, 123–152. 4. Bo¨rmer, O. P. (1982) Clin. Biochem. 15, 128 –132. 5. Primus, F. J., Kelley, E. A., Hansen, H. J., and Goldenberg, D. M. (1988) Clin. Chem. 34, 261–264. 6. Letiagin, V. P., and Vystotskaia, I. V. (1995) Vestn Ross. Akad. Med. Nauk. 4, 10 –14. 7. Christensen, R. L., Rea, M. R., Kessler, G., Crane, J. P., and Valdes, R., Jr. (1986) Clin. Chem. 32, 1812–1817. 8. John, R., Henley, R., and Shankland, D. (1986) Clin. Chem. 32, 2066 –2069. 9. Nishizono, I., Iida, S., Suzuki, N., Kawada, H., Murakami, H., Ashihara, Y., and Okada, M. (1991) Clin. Chem. 37, 1639 –1644. 10. Arakawa, H., Ikegami, T., Maeda, M., and Tsuji, A. (1993) J. Biolumin. Chemilumin. 8, 135–139. 11. Suonpa¨a¨, M., Lavi, J., Hemmila¨, I., and Lo¨vgren, T. (1985) Clin. Chim. Acta 145, 341–348. 12. Chan, M. A., Bellem, A. C., and Diamandis, E. P. (1987) Clin. Chem. 33, 2000 –2003. 13. Passelecq, B., De Bo, M., Huber, C., Gennart, J. P., Bernard, A., and Lauwerys, R. (1988) J. Immunol. Methods 109, 69 –74.
87
14. Felder, R. A., MacMillan, R. H., III, and Bruns, D. E. (1987) Clin. Chem. 33, 700 –704. 15. Mathieu, F., Mareschal, J. C., and Collet-Cassart (1989) J. Immunol. Methods 120, 65– 69. 16. Bo¨rmer, O. P., and Nustad, K. (1990) J. Immunol. Methods 127, 171–178. 17. Kimura, H., Matsuzawa, S., Tu, C., Kitamori, T., and Sawada, T. (1996) Anal. Chem. 68, 3063–3067. 18. Soini, E., and Hemmila¨, I. (1979) Clin. Chem. 25, 353–361. 19. Hemmila¨, I. (1985) Clin. Chem. 31, 359 –370. 20. Soini, E., and Lo¨vgren, T. (1987) CRC Crit. Rev. Anal. Chem. 18, 105–154. 21. Diamandis, E. P. (1988) Clin. Biochem. 21, 139 –150. 22. Diamandis, E. P., and Christopoulos, T. K. (1990) Anal. Chem. 62, 1149A–1157A. 23. Hemmila¨, I. (1988) Appl. Fluoresc. Technol. 1, 1– 8. 24. Hemmila¨, I. (1988) Scand. J. Clin. Lab. Invest. 48, 389 – 400. 25. Dickson, E. P. G., Pollak, A., and Diamandis, E. P. (1995) Pharmacol. Ther. 66, 207–235. 26. Yuan, J., Matsumoto, K., and Kimura, H. (1998) Anal. Chem. 70, 596 – 601. 27. Yuan, J., Wang, G., Kimura, H., and Matsumoto, K. (1997) Anal. Biochem. 254, 283–287. 28. Yuan, J., Wang, G., Kimura, H., and Matsumoto, K. (1998) Anal. Sci. 14, 421– 423. 29. Hemmila¨, I. (1995) J. Alloys Comp. 225, 480 – 485. 30. Xu, Y.-Y., Pettersson, K., Blomberg, K., Hemmila¨, I., Mikola, H., and Lo¨vgren, T. (1992) Clin. Chem. 38, 2038 –2043. 31. Hemmila¨, I., Holttinen, S., Pettersson, K., and Lo¨vgren, T. (1987) Clin. Chem. 32, 2281–2283. 32. Vuori, J., Rasi, S., Takala, T., and Vaananen, K. (1991) Clin. Chem. 37, 2087–2092. 33. Pettersson, K., Alfthan, H., Stenman, U., Turpeinen, U., Suonpa¨a¨, M., So¨derholm, J., Larsen, S. O., and Nørgaard-Pedersen, B. (1993) Clin. Chem. 39, 2084 –2089. 34. Qin, Q., Christiansen, M., Lo¨vgren, T., Nørgaard-Pedersen, B., and Pettersson, K. (1997) J. Immunol. Methods 205, 169 –175. 35. Merio¨, L., Pettersson, K., and Lo¨vgren, T. (1996) Clin. Chem. 42, 1513–1517. 36. Huang, H. G., Hiraki, K., and Nishikawa, Y. (1981) Nippon Kagaku Kaishi, 66 –73. 37. Bador, R., De´chaud, H., Claustrat, F., and Desuzinges, C. (1987) Clin. Chem. 33, 48 –51. 38. Liu, Y. S., Tobias, R. J., and Zurawski, V. R., Jr. (1985) Clin. Chem. 31, 191–195. 39. Morin, M., Bador, R., and De´chaud, H. (1989) Anal. Chim. Acta 219, 67–77. 40. Diamandis, E. P., Morton, R. C., Reichstein, E., and Khosravi, M. J. (1989) Anal. Chem. 61, 48 –53. 41. Kropf, J., Quitte, E., and Gressner, A. M. (1991) Anal. Biochem. 197, 258 –265. 42. Yang, X.-D., Ci, Y.-X., and Chang, W.-B. (1994) Anal. Chem. 66, 2590 –2594.
Simultaneous Determination of a-Fetoprotein and Carcinoembryonic Antigen in Human Serum by Time-Resolved Fluoroimmunoassay Kazuko Matsumoto,* ,1 Jingli Yuan,* Guilan Wang,* and Hiroko Kimura† *Department of Chemistry, Waseda University, Shinjuku-ku, Japan Science and Technology Corporation, Tokyo 169-8555, Japan; and †Department of Forensic Medicine, Juntendo University School of Medicine, Tokyo 113-8421, Japan
Received June 16, 1999
A novel simultaneous measurement method for a-fetoprotein (AFP) and carcinoembryonic antigen (CEA) in human sera by time-resolved fluoroimmunoassay (TR-FIA) is described. The proposed approach combines the use of europium-labeled anti-AFP antibody for AFP TR-FIA and biotinylated anti-CEA antibody complexed to samarium-labeled streptavidin for CEA TR-FIA. A 96-well microtiter plate coated with a mixture of anti-AFP and antiCEA monoclonal antibodies was used for the assay. After it was reacted with a solution containing AFP and CEA, a mixture of anti-AFP antibody labeled with BHHCT–Eu 31 and biotinylated anti-CEA antibody was added. The AFP concentration was determined by measuring the solid-phase fluorescence of the europium-labeled anti-AFP antibody at 615 nm. Then a BHHCT–Sm 31 -labeled streptavidin– bovine serum albumin conjugate (SA–BSA) was added to react with the biotinylated anti-CEA antibody. After the reaction, the unreacted SA–BSA was washed out, and a 0.1 M NaOH solution containing 1.0 3 10 25 M TOPO and 0.05% SDS was added to dissociate the samarium-labeled SA–BSA in the immune complex on the surface of the well into the solution. The CEA concentration was determined by measuring the solution fluorescence of 643 nm from the samariumlabeled SA–BSA. The present method gives detection limits of 0.07 ng/ml for AFP and 0.3 ng/ml for CEA. The coefficient variations of the method are less than 7%, and the recoveries are in the range of 90 – 110% for serum samples. The AFP and CEA concentrations in 27 human serum samples were determined by the present method as well as by single assay for comparison. A good correlation was ob-
1 To whom correspondence should be addressed. Fax: 181-3-52733489. E-mail: [email protected].
0003-2697/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
tained with the correlation coefficients of 0.990 for AFP and 0.973 for CEA. © 1999 Academic Press
Human AFP 2 and CEA are typical carcinoembryonic proteins and are known to be related to hepatoma and nonseminomatous testicular tumors, colon cancer, pulmonary cancer, and mammary cancer (1– 6). A number of immunoassay methods for human AFP and CEA have been reported (7–17), and it is possible to use various commercial AFP and CEA assay kits with appropriate sensitivity. Time-resolved fluoroimmunoassay (TR-FIA) using europium chelates as the labels has been noticed as a highly sensitive method (18 –25). However, the europium chelates used in the current Delfia and FIAgen systems have some limitations. To improve the sensitivity of TR-FIA, we recently developed a new tetradentate b-diketone label, BHHCT, for chelating Eu 31, and by using SA or a SA–BSA conjugate labeled with BHHCT–Eu 31, human serum AFP, IgE, and TSH were determined with over two orders of magnitude higher sensitivity than those of other immunometric methods (26 –28). It has been reported that some samarium and terbium chelates have strong fluorescence and long lifetime as well as europium chelates and can be used as labels in TR-FIA (25, 29). Since the chelates of the three lanthanide ions have very sharp fluorescence profiles (a half-width of ca. 10 nm) each at their own 2 Abbreviations used: AFP, a-fetoprotein; CEA, carcinoembryonic antigen; TR-FIA, time-resolved fluoroimmunoassay; BHHCT, 4,49bis(10,10,10,20,20,30,30-heptafluoro-40,60-hexanedion-60-yl)-chlorosulfo-o-terphenyl; BSA, bovine serum albumin; SA, streptavidin; IgE, immunoglobulin E; TSH, thyroid-stimulating hormone; TOPO, tri-n-octylphosphine oxide; CV, coefficient variation; arb. units, arbitrary units.
81
82
MATSUMOTO ET AL.
different wavelength, two or three components in serum can be simultaneously determined from a single sample by using the different chelates as labels in multiple-label TR-FIA. Such multiple-label assays have been reported by using Eu 31–Sm 31, Eu 31–Tb 31, or Eu 31–Tb 31–Sm 31–Dy 31 labeling (29, 30). The dual-label TR-FIA of human serum lutropin and follitropin (31), myoglobin and carbonic anhydrase III (32), AFP and free b-subunit of human chorionic gonadotropin (33), pregnancy-associated plasma protein A and free b-subunit of human chorionic gonadotropin (34), and free and total prostate-specific antigen (35) have been reported. However, the fluorescence of Sm 31 chelate is rather weak compared to that of Eu 31 chelate. The quantum yield of Sm 31 chelate is usually several hundred times lower than that of Eu 31 chelate (36), and the fluorescence lifetime of Sm 31 chelate is shorter (usually below 100 ms) (37) than that of Eu 31 chelate (usually over several hundreds microseconds). These cause the TR-FIA using Sm 31 chelate as the label to be less sensitive than that of using Eu 31 chelate (37). Because of this, TR-FIA using Sm 31 label is suitable only for the detection of a high-concentration component in the sample. The Eu 31–Sm 31 pair dual-label TR-FIA was thus used for the simultaneous measurement of lowconcentration and high-concentration components in a sample, respectively. However, the concentrations of human AFP and CEA in sera are usually low, at the ng/ml level. The current immunoassay methods for AFP and CEA give detection limits of 0.1–1 ng/ml for AFP (26) and 0.1– 0.5 ng/ml for CEA (15, 16, 38). Therefore, in the simultaneous method for human AFP and CEA, the sensitivity must be high enough not only for AFP but also for CEA. Using the reported Eu 31–Sm 31 dual-label TR-FIA (31– 34), simultaneous detection of AFP and CEA is difficult because of the lower sensitivity of the Sm 31 chelate. Although many excellent methods have been developed for the immunoassay of AFP and CEA in human serum, these are usually time and sample consuming. In addition, simultaneous measurement of AFP and CEA by enzyme immunoassay is not possible, since coloration reaction is necessary before measurement. In the present work, we developed a novel TR-FIA method for simultaneous measurement of AFP and CEA in human serum by using BHHCT–Eu 31-labeled anti-AFP antibody, biotinylated anti-CEA antibody, and BHHCT–Sm 31-labeled SA–BSA conjugate. In the present method, the emission of the Sm 31 label was amplified by introducing a biotinylated antibody– Sm 31-labeled streptavidin system. The method has the advantage of simple assay format, rapidity, high sensitivity, and small sample amount. The detection limits of the present method are 0.07 ng/ml for AFP and 0.3 ng/ml for CEA, which are low enough to measure AFP and CEA in human serum. The AFP and CEA concen-
trations in 27 human serum samples were determined by the present method as well as by the usual single assay with either of the chelates for comparison. A good correlation was obtained for the simultaneous and single assay with the correlation coefficients of 0.990 for AFP and 0.973 for CEA. MATERIALS AND METHODS
Labeling and Biotinylation of Antibodies Labeling of anti-AFP antibody and SA–BSA conjugate with BHHCT. Two milliliters of goat anti-human AFP antibody solution (Nippon Bio-Test Laboratories, Inc.; the anti-AFP antibody concentration is 0.5 mg/ml, total protein concentration is 6.3 mg/ml; immunoelectrophoresis shows that the impurity antibody does not exist) was dialyzed twice against 3 liters of saline– water at 4°C each for 24 h. To the dialyzed solution were added 8 ml of water and 84 mg of NaHCO 3, which made the solution ca. 0.1 M in NaHCO 3. The pH of the solution was adjusted to 9.1 with 0.1 M NaOH. Dry ethanol solution (180 ml) containing 3.0 mg of BHHCT was added dropwise with stirring, and the solution was further stirred for 1 h at room temperature. After a small amount of precipitate was removed by centrifuge, the unreacted free labeling reagent was separated from the labeled protein on a Sephadex G-50 gel column with 0.05 M NH 4HCO 3 of pH 8.0 as the eluent. To the labeled protein solution were added 30 mg of BSA and 20 mg of NaN 3, and the pH was adjusted to 6.4 with 1 M HCl. The solution was stored at 220°C before use. Since the labeled antibody solution contains proteins other than anti-AFP antibody, the labeling ratio was not determined. When the labeled antibody solution was used in a single assay of AFP, it was diluted 600-fold with 0.05 M Tris–HCl buffer of pH 7.8, containing 1.0 3 10 27 M EuCl 3, 0.2% BSA, 0.1% NaN 3, and 0.9% NaCl. Preparation and labeling of the SA–BSA conjugate, SA(BSA) 0.9(BHHCT) 46, were described in a previous report (27). The conjugate solution was 200-fold diluted with 0.05 M Tris–HCl of pH 7.8, containing 1.0 3 10 26 M SmCl 3, 0.2% BSA, 0.1% NaN 3, and 0.9% NaCl, and was incubated at 56°C for 2 h, before use in immunoassay. Biotinylation of anti-CEA antibody. After two dialyses of 0.5 ml of rabbit anti-human CEA antibody solution (2.4 mg/ml, IgG fraction of rabbit antiserum, Dako-immunoglobulins, Denmark) for 24 h at 4°C against 3 liters of saline–water, 0.5 ml of water, 8.4 mg of NaHCO 3, and 6 mg of sulfosuccinimidyl-6-(biotinamido) hexanoate (NHS–LC– biotin, Pierce Chemical Co.) were added with stirring. After stirring for 1 h at room temperature, the solution was incubated for 24 h at 4°C. After the solution was twice dialyzed each for 24 h at 4°C against 3 liters of 0.1 M NaHCO 3 contain-
DETERMINATION OF a-FETOPROTEIN AND CARCINOEMBRYONIC ANTIGEN
ing 0.25 g of NaN 3, 10 mg of BSA was added, and the solution was stored at 220°C before use. When the biotinylated antibody solution was used in a single assay of CEA, it was diluted 1000-fold with 0.05 M Tris–HCl buffer of pH 7.8, containing 0.2% BSA, 0.1% NaN 3, and 0.9% NaCl. When the BHHCT-labeled anti-AFP antibody and biotinylated anti-CEA antibody were used for the simultaneous measurement of AFP and CEA, a mixture of 1/600 BHHCT-labeled anti-AFP antibody solution and 1/1000 biotinylated anti-CEA antibody solution was prepared with 0.05 M Tris–HCl buffer of pH 7.8, containing 1.0 3 10 27 M EuCl 3, 0.2% BSA, 0.1% NaN 3, and 0.9% NaCl. Instrument and Measurement Conditions To measure the fluorescence properties of BHHCT– Sm 31, a BHHCT-labeled BSA solution was prepared as previously reported (26). It was reacted with SmCl 3 in buffer to give a BHHCT–Sm 31 solution. The excitation and emission spectra of BHHCT–Sm 31 were measured on an Hitachi F-4500 fluorospectrometer. The fluorescence lifetime of BHHCT–Sm 31 was measured on a Perkin–Elmer LS 50B luminescence spectrometer. TRFIA measurement of AFP and CEA was performed on an Arcus 1232 time-resolved fluorometer (Wallac Oy), to which a filter (Wallac Oy) of 643 nm for Sm 31 chelate was attached. Eu fluorescence measurement was carried out with excitation at 340 nm, emission at 615 nm, delay time of 0.2 ms, and window time of 0.4 ms (specific fluorescence was measured after a 200-ms time delay for 400 ms between each excitation pulse). For Sm, excitation at 340 nm, emission at 643 nm, delay time of 0.03 ms, and window time of 0.10 ms (specific fluorescence was measured after a 30-ms time delay for 100 ms between each excitation pulse) were used. Immunoassay of AFP and CEA Human serum samples were obtained from Juntendo University Hospital and were stored at 220°C before use. The standard AFP and CEA solutions were prepared by diluting an a-1-fetoprotein solution (Dakopatts, Denmark) and a CEA solution (Seikagaku Kogyo Co., Ltd., Japan) with 0.05 M Tris–HCl buffer of pH 7.8 containing 5% BSA, 0.1% NaN 3, and 0.9% NaCl. The immunoassay format for the simultaneous measurement of AFP and CEA is shown in Fig. 1, and is described below. One hundred microliters of 0.1 M carbonate buffer of pH 9.6 containing 5 mg/ml of antiAFP monoclonal antibody (Biostride, Inc.) and 5 mg/ml of anti-CEA monoclonal antibody (Boehringer-Mannheim Biochemica) was incubated in a 96-well microtiter plate (Fluoro Nunc module plates) for 24 h at 4°C.
83
FIG. 1. Schematic representation of the simultaneous measurement system for AFP and CEA. For the details of the procedure, see Materials and Methods.
The wells were rinsed twice with 0.05 M Tris–HCl buffer of pH 7.8 containing 0.05% Tween 20 (buffer 1) and once with 0.05 M Tris–HCl buffer of pH 7.8 (buffer 2). The coated plate was stored at 220°C before use. The antigen–antibody reaction was carried out as follows. After 50 ml of the AFP–CEA standard solution or human serum was incubated at 37°C for 1 h in the coated wells, the wells were rinsed twice with buffer 1 and once with buffer 2. After a 50-ml mixture of BHHCT–Eu 31-labeled anti-AFP and biotinylated antiCEA was incubated in the wells at 37°C for 1 h, the wells were rinsed four times with 0.05 M Tris–HCl of pH 9.1 containing 0.05% Tween 20 (buffer 3). The rinsed plate was subjected to solid-phase Eu time-resolved fluorometric measurement. After measurement, the plate was rinsed once with buffer 2, and 50 ml of the SA(BSA) 0.9(BHHCT) 46–Sm 31 solution was incubated in the wells at 37°C for 1 h. The wells were rinsed four times with buffer 3. Fifty microliters of 0.1 M NaOH containing 1.0 3 10 25 M TOPO and 0.05% SDS was added to the wells and incubated at room temperature for 10 min. The solution-phase Sm time-resolved fluorometric measurement was carried out. To compare the correlation between the simultaneous measurement and the independent single measurement for each of the two antigens, AFP and CEA were assayed also in a single-assay format, in which only one of the two antibodies was coated to the plate. The standard solution containing only either AFP or CEA and the labeled antibody containing only either BHHCT–Eu 31-labeled anti-AFP or biotinylated antiCEA were used. Other conditions for the single assay were the same as those for the simultaneous assays. The cross-reactions between the two antigens were examined with the single assays of anti-AFP antibody– CEA and anti-CEA antibody–AFP systems. RESULTS AND DISCUSSION
The Fluorescence Properties of BHHCT–Eu 31 and BHHCT–Sm 31 Because BHHCT is insoluble in water, the fluorescence properties were examined using a BHHCT-la-
84
MATSUMOTO ET AL.
FIG. 2. Excitation and emission spectra of BHHCT (bound to BSA)–Sm 31 chelate in 0.1 M NaOH, containing 0.05% SDS and 1.0 3 10 25 M TOPO. [BHHCT] 5 [Sm 31] 5 1.0 3 10 25 M. The excitation spectrum was recorded with l em 5 561 nm and the emission spectrum with l ex 5 335 nm.
beled BSA solution. The fluorescence properties of BHHCT (bound to BSA)–Eu 31 have been discussed in a previous report (26). This chelate has excitation and emission maxima at 326 nm and 611.6 nm in a Tris– HCl buffer. Its lifetime in 0.1 M Tris–HCl buffer of pH 9.1 is greater than 380 ms. Excitation and emission spectra of BHHCT (bound to BSA)–Sm 31 chelate in a 0.1 M NaOH– 0.05% SDS–1.0 3 10 25 M TOPO solution are shown in Fig. 2. The maximum excitation at 328 nm is the same as that for BHHCT–Eu 31, but the Sm 31 chelate gives three emission maxima at 561, 597, and 642 nm, which correspond to 4G 5/2 3 6H 5/2 (561 nm), 4 G 5/2 3 6H 7/2 (597 nm), and 4G 5/2 3 6H 9/2 (642 nm), respectively (39). The lifetime (642-nm emission) of the BHHCT–Sm 31 chelate was measured in 0.1 M Tris– HCl buffer of pH 9.1 and in a 0.1 M NaOH– 0.05% SDS–1.0 3 10 25 M TOPO solution. The lifetimes in the two solutions are 31.1 and 70.7 ms, respectively.
of a biotin–streptavidin system is effective for increasing the labeling number per antibody molecule and obtaining a high sensitivity (40). A biotin–streptavidin system was therefore employed to raise the sensitivity of the Sm fluorescent label. As described above, the lifetime of the BHHCT–Sm 31 chelate is only 31.1 ms in Tris–HCl buffer, and therefore the solid-phase measurement after the labeled SA–BSA reacts with biotinylated anti-CEA antibody on the solid in Tris–HCl buffer is not very sensitive (see Fig. 3). On the other hand, a solution-phase time-resolved fluoromeasurement method is usually more sensitive, if the fluorescence probe complex is transferred from the solid phase into the solution (41, 42). The mechanism of the dissociation of the solid-phase immune complex for fluorescence measurement has been discussed in a previous work (42). The dissociation solutions such as SDS– urea or SDS–TOPO are usually employed (41, 42). In the present work, the BHHCT–Sm 31-labeled SA–BSA was dissolved from the solid-phase into the solution by NaOH–SDS–TOPO, and the solution was measured for Sm fluorescence. The dissociation of BHHCT–Sm 31 reached equilibrium in ca. 10 min at room temperature, and the signal intensity did not change for a further ca. 30 min. As shown in Fig. 3, the Sm fluorescence intensity in the solution phase is higher than that in the solid phase, while the background intensity increases only slightly, and therefore a lower detection limit can be obtained in solution-phase measurement. The cross-reactions between anti-AFP antibody and CEA and vice versa were examined. The results, shown in Fig. 4, show that the cross-reaction in the present system is negligible. Furthermore, the results show
Evaluation of the Assay Format The immunoassay format for the simultaneous measurement of AFP and CEA is shown in Fig. 1. Owing to the BHHCT–Eu 31 chelate having strong fluorescence with a very long lifetime (the detection limits of BHHCT–Eu 31 chelate are 8.4 3 10 213 M in Tris–HCl buffer and 2.3 3 10 213 M in SDS–TOPO–NaHCO 3 buffer, respectively) (26), TR-FIA using BHHCT–Eu 31 has very high sensitivity even in the solid-phase measurement (26 –28). In contrast to this, the Sm 31 chelate does not have such a strong and long-lifetime fluorescence, which makes the immunoassay using a Sm label less sensitive (37). We also tested the CEA TR-FIA by direct labeling the anti-CEA antibody and measuring the fluorescence for comparison with the assay format in Fig. 1. However, a clear significant signal was not obtained below 1000 ng/ml. In such a case, introduction
FIG. 3. The relative fluorescence intensities in the solid-phase and solution-phase measurements of CEA by using BHHCT–Sm 31-labeled SA–BSA in single-label assay. The solid-phase measurement means that the fluorescence measurement was carried out on the solid-phase immune complex. The solution-phase measurement means that the fluorescence measurement was carried out after the solid-phase immune complex was dissolved into the solution by a NaOH–SDS–TOPO solution.
DETERMINATION OF a-FETOPROTEIN AND CARCINOEMBRYONIC ANTIGEN
85
ground intensities in the simultaneous assay and the single assay for CEA did not change in the presence and absence of the labeled anti-AFP antibody. The above results indicate that AFP and CEA in serum can be simultaneously assayed by using the present method. Assay of Human Serum Samples
FIG. 4. The cross-reactivity of anti-AFP antibody with CEA (A) and anti-CEA antibody with AFP (B). (A) After anti-AFP antibody-coated wells were reacted with CEA standard solutions at 37°C for 1 h, BHHCT–Eu 31-labeled anti-AFP antibody was added, and the solution was incubated at 37°C for 1 h. After washing, the Eu 31 fluorescence from the solid-phase was measured. (B) After anti-CEA antibody-coated wells were reacted with AFP standard solutions at 37°C for 1 h, biotinylated anti-CEA antibody was added, and the solution was incubated at 37°C for 1 h. After washing, BHHCT–Sm 31-labeled SA–BSA was added, and the solution was incubated at 37°C for 1 h. After washing, a 0.1 M NaOH solution containing 1.0 3 10 25 M TOPO and 0.05% SDS was added, and the solution was incubated at room temperature for 10 min. The Sm 31 fluorescence from the solution was measured.
that the existence of the BHHCT–Eu 31-labeled antiAFP antibody did not affect the CEA assay, which was confirmed by the fact that both the signal and back-
The calibration curves of simultaneous measurement of AFP and CEA are shown in Fig. 5. The curves of background-subtracted signals vs antigen concentrations are shown in Fig. 5. Since the delay time for Sm 31 fluorescence is ca. an order of magnitude smaller than that of Eu 31, the background intensity for the CEA assay is ca. 12 times larger than that of AFP. This background, caused by the coexisting proteins and plastic solid phase (a part of the background may be caused by the nonspecific adsorption of the labeled protein), can be greatly reduced by a prolonged delay time such as in the assay of AFP. However, since the lifetime of the Sm label is not so long, a longer delay time is not possible. The detection limits of the present method are calculated (41) to be 0.07 and 0.3 ng/ml for AFP and CEA, respectively, which are low enough for the assay of human serum. The average CV (n 5 4, within-run) of AFP and CEA in the ranges in Fig. 5 are 4.34 and 2.87%, respectively. The upper limits of the calibration curves are ca. 100 and ca. 1000 ng/ml for AFP and CEA, respectively. In the range below 100 ng/ml, two calibration curves show nearly straight lines when background-subtracted fluorescence signals vs antigen concentrations were plotted. However, the slopes of the two calibration curves are below 1.0,
FIG. 5. Calibration curves of the simultaneous measurement of AFP (A) and CEA (B). bg, background. The inset curves are plotted for background-subtracted signals vs the concentrations of AFP and CEA.
86
MATSUMOTO ET AL. TABLE 1
Precision in the Simultaneous Measurement of AFP and CEA in Human Sera Interassay (n 5 4) a AFP (ng/ml) Sample Sample Sample Sample Sample a
1 2 3 4
CEA (ng/ml)
Mean
CV (%)
Mean
CV (%)
1.05 3.92 5.46 10.02
5.56 5.02 5.52 6.10
6.57 4.02 7.90 13.65
5.32 3.22 4.79 6.05
Four assays using four different plates (one assay per plate).
which might be due to the different efficiency of the antigen–antibody reaction on the solid-phase surface in a different antigen concentration range or caused by the different efficiency of nonspecific bindings of the Eu 31-labeled antibody and Sm 31-labeled SA–BSA in different antigen concentration ranges. If the antigen– antibody reaction efficiency is higher at lower antigen concentration, or if the nonspecific binding efficiency is higher at lower antigen concentration, then the slope of the calibration curve would be less than 1.0. To evaluate the assay precision, four serum samples containing different AFP and CEA concentrations were measured four times using four different plates. The results are summarized in Table 1. The CVs (between-
run) of four assay results were in the range of 5.0 – 6.5% for AFP and 3.0 – 6.5% for CEA. These results show that the present method has good precision not only for the standard solution assay but also for the serum sample assay. To evaluate the accuracy of the assay, standard solutions of AFP and CEA mixtures (the concentrations of AFP and CEA are below 20 ng/ml) were added to four serum samples containing different AFP and CEA concentrations. After the concentrations of AFP and CEA in four serum samples and in the four solutions to which standard AFP and CEA were added to sera were measured using the simultaneous measurement method, the analytical recoveries of AFP and CEA were calculated, respectively. The analytical recoveries of AFP and CEA added to four serum samples are 89.4, 106.2, 96.2, and 103.4% (average recovery 5 98.8%) for AFP and 100.6, 97.2, 105.3, and 91.2% (average recovery 5 98.6%) for CEA, respectively. These results show that the analytical accuracy of the present method is good enough for the application of AFP and CEA assays in human sera. The AFP and CEA concentrations in 27 human sera were determined with both the simultaneous measurement method and the single-label assay method. Correlations of the two methods are shown in Fig. 6. The correlation coefficients of AFP and CEA are 0.990 and 0.973, respectively. These show that the two proteins can be determined by the present method with high precision.
FIG. 6. Correlation between the single-label measurement and the simultaneous measurement of AFP (A) and CEA (B) in human sera (n 5 27). The single-label measurement means that AFP and CEA were assayed independently. AFP single-label measurement: after anti-AFP antibody-coated wells were reacted with AFP standard solutions or serum samples at 37°C for 1 h, BHHCT–Eu 31-labeled anti-AFP antibody was added, and the solution was incubated at 37°C for 1 h. After washing, the Eu 31 fluorescence from the solid phase was measured, and the AFP concentrations in the samples were calculated. CEA single-label measurement: after anti-CEA antibody-coated wells were reacted with CEA standard solutions or serum samples at 37°C for 1 h, biotinylated anti-CEA antibody was added, and the solution was incubated at 37°C for 1 h. After washing, BHHCT–Sm 31-labeled SA–BSA was added, and the solution was incubated at 37°C for 1 h. After washing, a 0.1 M NaOH solution containing 1.0 3 10 25 M TOPO and 0.05% SDS was added, and the solution was incubated at room temperature for 10 min. The Sm 31 fluorescence from the solution was measured, and the CEA concentrations in the samples were calculated.
DETERMINATION OF a-FETOPROTEIN AND CARCINOEMBRYONIC ANTIGEN
Based on the above results, we can conclude that the present AFP and CEA simultaneous detection method can be practically employed to the clinical analysis of AFP and CEA. In addition, we can expect that application of the present method can be extended to other two-component simultaneous measurement systems. In the present method, because of weak fluorescence of the samarium label, the CEA assay was carried out with the use of a biotin–streptavidin amplifying system and a dissociation step for solution-phase measurement. Although this method enables high sensitivity for the CEA assay, the procedure is complicated compared with the method using the antibody labeled with fluorescence label and direct solid-phase measurement. Among lanthanide fluorescence labels, it is known that terbium labels have higher quantum yields and longer lifetimes than samarium labels. Instead of the samarium label, the simultaneous use of europium and terbium labels for two-component measurement would be a more ideal method. ACKNOWLEDGMENT The present work has been supported by CREST (Core Research for Evolutional Science and Technology) of Japan Science and Technology Corporation (JST).
REFERENCES 1. Buamah, P. K., Harris, R., James, O. F. W., and Skillen, A. W. (1986) Clin. Chem. 32, 2083–2084. 2. Gold, P., and Freedman, S. O. (1965) J. Exp. Med. 122, 467– 481. 3. Fuks, A., Banjo, C., Shuster, J., Freedman, S. O., and Gold, P. (1975) Biochem. Biophys. Acta 417, 123–152. 4. Bo¨rmer, O. P. (1982) Clin. Biochem. 15, 128 –132. 5. Primus, F. J., Kelley, E. A., Hansen, H. J., and Goldenberg, D. M. (1988) Clin. Chem. 34, 261–264. 6. Letiagin, V. P., and Vystotskaia, I. V. (1995) Vestn Ross. Akad. Med. Nauk. 4, 10 –14. 7. Christensen, R. L., Rea, M. R., Kessler, G., Crane, J. P., and Valdes, R., Jr. (1986) Clin. Chem. 32, 1812–1817. 8. John, R., Henley, R., and Shankland, D. (1986) Clin. Chem. 32, 2066 –2069. 9. Nishizono, I., Iida, S., Suzuki, N., Kawada, H., Murakami, H., Ashihara, Y., and Okada, M. (1991) Clin. Chem. 37, 1639 –1644. 10. Arakawa, H., Ikegami, T., Maeda, M., and Tsuji, A. (1993) J. Biolumin. Chemilumin. 8, 135–139. 11. Suonpa¨a¨, M., Lavi, J., Hemmila¨, I., and Lo¨vgren, T. (1985) Clin. Chim. Acta 145, 341–348. 12. Chan, M. A., Bellem, A. C., and Diamandis, E. P. (1987) Clin. Chem. 33, 2000 –2003. 13. Passelecq, B., De Bo, M., Huber, C., Gennart, J. P., Bernard, A., and Lauwerys, R. (1988) J. Immunol. Methods 109, 69 –74.
87
14. Felder, R. A., MacMillan, R. H., III, and Bruns, D. E. (1987) Clin. Chem. 33, 700 –704. 15. Mathieu, F., Mareschal, J. C., and Collet-Cassart (1989) J. Immunol. Methods 120, 65– 69. 16. Bo¨rmer, O. P., and Nustad, K. (1990) J. Immunol. Methods 127, 171–178. 17. Kimura, H., Matsuzawa, S., Tu, C., Kitamori, T., and Sawada, T. (1996) Anal. Chem. 68, 3063–3067. 18. Soini, E., and Hemmila¨, I. (1979) Clin. Chem. 25, 353–361. 19. Hemmila¨, I. (1985) Clin. Chem. 31, 359 –370. 20. Soini, E., and Lo¨vgren, T. (1987) CRC Crit. Rev. Anal. Chem. 18, 105–154. 21. Diamandis, E. P. (1988) Clin. Biochem. 21, 139 –150. 22. Diamandis, E. P., and Christopoulos, T. K. (1990) Anal. Chem. 62, 1149A–1157A. 23. Hemmila¨, I. (1988) Appl. Fluoresc. Technol. 1, 1– 8. 24. Hemmila¨, I. (1988) Scand. J. Clin. Lab. Invest. 48, 389 – 400. 25. Dickson, E. P. G., Pollak, A., and Diamandis, E. P. (1995) Pharmacol. Ther. 66, 207–235. 26. Yuan, J., Matsumoto, K., and Kimura, H. (1998) Anal. Chem. 70, 596 – 601. 27. Yuan, J., Wang, G., Kimura, H., and Matsumoto, K. (1997) Anal. Biochem. 254, 283–287. 28. Yuan, J., Wang, G., Kimura, H., and Matsumoto, K. (1998) Anal. Sci. 14, 421– 423. 29. Hemmila¨, I. (1995) J. Alloys Comp. 225, 480 – 485. 30. Xu, Y.-Y., Pettersson, K., Blomberg, K., Hemmila¨, I., Mikola, H., and Lo¨vgren, T. (1992) Clin. Chem. 38, 2038 –2043. 31. Hemmila¨, I., Holttinen, S., Pettersson, K., and Lo¨vgren, T. (1987) Clin. Chem. 32, 2281–2283. 32. Vuori, J., Rasi, S., Takala, T., and Vaananen, K. (1991) Clin. Chem. 37, 2087–2092. 33. Pettersson, K., Alfthan, H., Stenman, U., Turpeinen, U., Suonpa¨a¨, M., So¨derholm, J., Larsen, S. O., and Nørgaard-Pedersen, B. (1993) Clin. Chem. 39, 2084 –2089. 34. Qin, Q., Christiansen, M., Lo¨vgren, T., Nørgaard-Pedersen, B., and Pettersson, K. (1997) J. Immunol. Methods 205, 169 –175. 35. Merio¨, L., Pettersson, K., and Lo¨vgren, T. (1996) Clin. Chem. 42, 1513–1517. 36. Huang, H. G., Hiraki, K., and Nishikawa, Y. (1981) Nippon Kagaku Kaishi, 66 –73. 37. Bador, R., De´chaud, H., Claustrat, F., and Desuzinges, C. (1987) Clin. Chem. 33, 48 –51. 38. Liu, Y. S., Tobias, R. J., and Zurawski, V. R., Jr. (1985) Clin. Chem. 31, 191–195. 39. Morin, M., Bador, R., and De´chaud, H. (1989) Anal. Chim. Acta 219, 67–77. 40. Diamandis, E. P., Morton, R. C., Reichstein, E., and Khosravi, M. J. (1989) Anal. Chem. 61, 48 –53. 41. Kropf, J., Quitte, E., and Gressner, A. M. (1991) Anal. Biochem. 197, 258 –265. 42. Yang, X.-D., Ci, Y.-X., and Chang, W.-B. (1994) Anal. Chem. 66, 2590 –2594.