Anal. Chem. 1998, 70, 2110-2114
Automatic and Simultaneous Analysis of Lens culinaris Agglutinin-Reactive r-Fetoprotein Ratio and Total r-Fetoprotein Concentration Hideo Katoh,* Kenji Nakamura, Takumi Tanaka, Shinji Satomura, and Shuji Matsuura
Osaka Research Laboratories, Wako Pure Chemical Industries Ltd., 6-1 Takada-cho Amagasaki, Hyogo, Japan 661
An automated analytical method for analyzing r-fetoprotein (AFP) carbohydrate chain microheterogeneity based on competitive reaction between lectin and anti-AFP monoclonal antibody in liquid phase is described. The antibody used binds to all species of AFP molecule without Lens culinaris agglutinin (LCA); however, its binding reaction to LCA-reactive AFP was inhibited by LCA. Sulfated tyrosine octamer was conjugated to the antibody, and sulfated tyrosine pentamer and peroxidase were conjugated to other monoclonal antibodies, respectively. Serum reacted with three anti-AFP monoclonal antibodies and LCA in liquid phases, and two types of immune complex were observed. The two types were separated directly by the liquid-phase binding assay system equipped with an anion-exchange column. Peroxidase activity of immune complex was determined fluorophotometrically. Total AFP concentration and the ratio of LCA-reactive AFP in samples were calculated simultaneously, using the sum of the two peaks and the ratio of peaks obtained by LCA inhibition to sum of two peaks. The results correlated well with conventional methods. The method is simple and convenient for routine clinical assays.
Among these techniques, a combination of Lens culinaris agglutinin (LCA) affinity electrophoresis and antibody-affinity blotting12 shows higher diagnostic and prognostic potential and has been used for earlier recognition of HCC compared to imaging modalities,13-16 for discriminating benign liver diseases from HCC with highest specificity,9,13 and for monitoring in treatment responses and disease recurrence.17,18 However, this method is complicated and time-consuming. Accordingly, an automated method for determination of LCA-reactive AFP is needed for measuring large quantities of samples. During a search of known monoclonal antibodies against human AFP, a unique antibody was found. This antibody recognizes AFP obtained from benign and malignant liver disease similarly; however, the antibody is inhibited in its binding to LCAreactive AFP by LCA. During this search, we developed a new automatic immunoassay method19 for human chorionic gonadotropin,20 thyroxine,21 and R-fetoprotein22,23 based on a liquid-phase binding assay (LBA) system. In LBA, an antibody reacts with
* Corresponding author: (telephone) 6-499-9106; (facsimile) 6-499-1524; (email)
[email protected]. (1) Abelev, G. I. Cancer Res. 1968, 28, 1344-1355. (2) Taketa, K. Hepatology 1990, 12, 1420-1432. (3) Yoshima, H.; Mizuochi, T.; Ishii, M.; Kobata, A. Cancer Res. 1980, 40, 42764281. (4) Yamashita, K.; Hitoi, A.; Tsuchida, Y.; Nishi, S.; Kobata, A. Cancer Res. 1983, 43, 4691-4695. (5) Smith, C. J.; Kelleher, P. C. Biochim. Biophys. Acta 1973, 317, 231-235. (6) Aoyagi, Y.; Suzuki, Y.; Isemura, M.; Nomoto, M.; Sekine, C.; Igarashi, K.; Ichida, F. Cancer 1988, 61, 769-774. (7) Kerckaert, J.-P.; Bayard, B.; Biserte, G. Biochim. Biophys. Acta 1979, 576, 99-108.
(8) Breborowicz, J.; Mackiewicz, A.; Breborowicz, D. Scand. J. Immunol. 1981, 14, 15-20. (9) Taketa, K.; Sekiya, C.; Namiki, M.; Akamatsu, K.; Ohta, Y.; Endo, Y.; Kosaka, K. Gastroenterology 1990, 99, 508-518. (10) Shimizu, K.; Katoh, H.; Yamashita, F.; Tanaka, M.; Tanikawa, K.; Taketa, K.; Satomura, S.; Matsuura, S. Clin. Chim. Acta 1996, 254, 23-40. (11) Kinoshita, N.; Suzuki, S.; Matsuda, Y.; Taniguchi, N. Clin. Chim. Acta 1989, 179, 143-152. (12) Shimizu, K.; Taniichi, T.; Satomura, S.; Matsuura, S.; Taga, H.; Taketa, K. Clin. Chim. Acta 1993, 214, 3-12. (13) Taketa, K.; Endo, Y.; Sekiya, C.; Tanikawa, K.; Koji, T.; Taga, H.; Satomura, S.; Matsuura, S.; Kawai, T.; Hirai, H. Cancer Res. 1993, 53, 5419-5423. (14) Sato, Y.; Nakata, K.; Kato, Y.; Shima, M.; Ishii, N.; Koji, T.; Taketa, K.; Endo, Y.; Nagataki, S. N. Engl. J. Med. 1993, 328, 1802-1806. (15) Aoyagi, Y.; Saitoh, A.; Suzuki, Y.; Igarashi, K.; Ogro, M.; Yokota, T.; Mori, S.; Suda, T.; Isemura, M.; Asakura, H. Hepatology 1993, 17, 50-52. (16) Shiraki, K.; Takase, K.; Tameda, Y.; Hamada, M.; Kosaka, Y.; Nakano, T. Hepatology 1995, 22, 802-807. (17) Yamashita, F.; Tanaka, M.; Satomura, S.; Tanikawa, K. Eur. J. Gastroenterol. Hepatol. 1995, 7, 627-633. (18) Yamashita, F.; Tanaka, M.; Satomura, S.; Tanikawa, K. Gastroenterology 1996, 111, 996-1001. (19) Nakamura, K.; Satomura, S.; Tanaka, T.; Matsuura, S. Anal. Sci. 1992, 8, 157-160. (20) Nakamura, K.; Satomura, S.; Matsuura, S. Anal. Chem. 1993, 65, 613616. (21) Hara, T.; Nakamura, K.; Satomura, S.; Matsuura, S. Anal. Chem. 1994, 66, 351-354. (22) Nakamura, K.; Imajo, N.; Yamagata, Y.; Katoh, H.; Fujio, K.; Tanaka, T.; Satomura, S.; Matsuura, S. Anal. Chem. 1998, 70, 954-957. (23) Yamagata, Y.; Katoh, H.; Nakamura, K.; Tanaka, T.; Satomura, S.; Matsuura, S. J. Immunol. Methods, in press.
2110 Analytical Chemistry, Vol. 70, No. 10, May 15, 1998
S0003-2700(97)01280-8 CCC: $15.00
R-Fetoprotein (AFP) has been used as a tumor marker of hepatocellular carcinoma (HCC).1 However, serum AFP concentrations of patients with benign liver diseases and HCC frequently overlapped.2 AFP has a single asparagine-linked sugar chain as revealed by structural analysis of human AFP obtained from HCC and yolk sac tumor.3,4 AFP structural heterogeneity has been investigated by lectin affinity column chromatography,5,6 gel electrophoresis,7-10 and lectin-antibody sandwich immunoassay.11
© 1998 American Chemical Society Published on Web 04/16/1998
Figure 1. Presumed immune complexes 1 and 2.
antigen without a solid phase. Consequently, an immune reaction is completed under controlled and defined liquid-phase conditions. The reaction mixture of LBA is subjected to high-performance liquid chromatography (HPLC) followed by detection of the immune complex by a postcolumn enzyme reaction. The combination of the unique antibody against AFP and the LBA system solved many problems for authentic analysis of LCA-reactive AFP. EXPERIMENTAL SECTION Materials. Tyrosine pentamer and octamer were synthesized using a solid-phase method by peptide synthesizer (System 990E Synthesizer, Beckman) and then all hydroxy residues on tyrosine were substituted by sulfate ester, and the amine terminal was coupled with maleimide residue. These maleimidyl sulfated tyrosine oligomers are termed YS5 (pentamer) and YS8 (octamer). LCA-nonreactive and -reactive AFP were purified from human placenta and HuH-7 cell culture24 by LCA immobilized agarose gel column chromatography, respectively. An anion-exchange column 4.6-mm i.d. × 10 mm was packed with POROS DEAE resin (PerSeptive Biosystems). AFP concentration and the ratio of LCA-reactive AFP were measured by IMx AFP Dynapack (Abbott Laboratories) and AFP differentiation kit-L (Wako Pure Chemical Industries Ltd.), respectively. Anti-AFP Monoclonal Antibodies. Three anti-AFP monoclonal antibodies (clones A4-4, WA-1, and WA-2) were selected from our panel of antibodies and purified as immunoglobulin G. These antibodies recognize different AFP epitopes. In particular clone A4-4’s epitope is assumed to include a nearby sugar chain as shown in Figure 1. Three antibodies were digested with pepsin followed by reduction of F(ab′)2 to form Fab′. Clone WA-1’s Fab′ was conjugated to peroxidase (POD, Toyobo Co., Ltd.) using sulfosuccinimidyl 4-(p-maleimidophenyl) butyrate (Pierce), and Fab′-POD conjugate was purified. Clones WA-2 and A4-4 were conjugated with YS5 and YS8, Fab′-YS5 and Fab′-YS8, respectively. (24) Nakabayashi, K.; Taketa, K.; Miyano, K.; Yamane, T.; Sato, J. Cancer Res. 1982, 42, 3858-3863.
Apparatus. A Shimadzu model LC-9A HPLC system (Shimadzu Co.) equipped with a postcolumn flow-through coil (0.25mm i.d. × 20 m) was used. Figure 2 shows a scheme of the experimental apparatus. An anion-exchange column was used with an elution buffer of 50 mM Tris-HCl buffer, pH 8.0, containing 0.3 M sodium chloride and 50 mM Tris-HCl buffer, pH 8.0, containing 3 M sodium chloride. The total flow rate was 2 mL/ min. Controlled mixing of two buffers was performed with a stepwise gradient. The gradient profile is summarized in Figure 3. A substrate buffer for POD reaction was 317 mM 4-acetoamidophenol (fluorogenic substrate for POD)25 and 38 mM H2O2 in 15 mM citrate buffer, pH 5.5. After column separation, column and substrate buffer were mixed in a 10:1 ratio and incubated in the flow-through coil at 60 °C for 30 s. The fluorescent product by POD was detected by an excitation wavelength of 328 nm and an emission wavelength of 432 nm. Immune Reaction. The first reagent consisted of 1 mg/mL LCA and 175 nM Fab′-YS5 with 50 mM N-(2-acetamide)-2aminoethanesulfonate buffer, pH 6.5. The second reagent consisted of 400 nM Fab′-POD and 70 nM Fab′-YS8 with 50 mM N-(2-acetamide)-2-aminoethanesulfonate buffer, pH 6.5. Reaction time course is as follows: 100 µL of the first reagent and 10 µL of sample are mixed and incubated at 8 °C for 20 min. Then, 10 µL of the second reagent is combined with the reaction mixture and incubated at 8 °C for 20 min. Then 80 µL of the final reaction mixture is injected to HPLC system. “Complex 1”, consisting of Fab′-POD, AFP, and Fab′-YS5 eluted at the first step, and “complex 2”, consisting of Fab′-POD, AFP, Fab′-YS5, and Fab′YS8 eluted later (Figure 3). Calculation of AFP Concentration and Ratio of LCAReactive AFP. The concentration of AFP was calculated from the sum of two peak areas using a standard curve. The ratio of LCA-reactive AFP (AFP-L3%) was calculated from the peak area of complex 1 and the sum of two peak areas by a comparison of standard curves. The calculation formulas were as follows:
total AFP:
complex 1 + complex 2
AFP-L3%:
complex 1 × 100 complex 1 + complex 2
RESULTS Analysis of Immune Reaction. In this method, a presumed immune complex is summarized in Figure 1. LCA-nonreactive AFP (AFP-L1) was bound by three antibodies (complex 2). On the other hand, LCA-reactive AFP (AFP-L3) was inhibited in its binding to A4-4 antibody (Fab′-YS8) by LCA, and AFP-L3 bound only two antibodies (Fab′-POD and Fab′-YS5, complex 1). Complexes 1 and 2 were separated by anion-exchange column chromatography on the basis of their anionic charge difference because complex 1 contains 5 sulfate residues in its immune complex, while complex 2 contains 13 sulfate residues. Each complex eluted in a stepwise gradient as two separate peaks (Figure 3). Effect of LCA Concentration. The concentration of LCA in the first reagent was determined from 0 to 2 mg/mL (Figure 4). The percentage of complex 1 obtained by purified AFP-L1 was (25) Shiga, M.; Yakata, K.; Aoyama, M.; Sasamoto, K.; Takagi, M.; Ueno, K. Anal. Sci. 1995, 11, 195-201.
Analytical Chemistry, Vol. 70, No. 10, May 15, 1998
2111
Figure 2. Schematic diagram of the HPLC system.
Figure 3. Chromatographic pattern of the reaction products and gradient profile.
Figure 4. Effect of LCA concentration in reagent: (b) AFP-L1 100 ng/mL, (O) AFP-L3, 100 ng/mL, (×) percent difference of complex 1 between AFP-L1 and AFP-L3.
almost the same; however, the percentage of complex 1 obtained by purified AFP-L3 was elevated with relation to LCA concentration. These results suggest that LCA inhibited binding of A4-4YS8 to AFP-L3. The percentage difference of complex 1 between purified AFP-L1 and AFP-L3 was constant above 0.5 mg/mL LCA; 1 mg/mL LCA was used. 2112 Analytical Chemistry, Vol. 70, No. 10, May 15, 1998
Figure 5. Effect of Fab′-YS8 concentration in reagent: (b) AFPL1 100 ng/mL, (O) AFP-L3 100 ng/mL, (×) percent difference of complex 1 between AFP-L1 and AFP-L3.
Effect of Antibody Concentration. The concentration of antibodies used in LBA is important to the immune reaction. Fab′POD and Fab′-YS5 recognize an independent epitope of AFP. Therefore, they can bind to all AFP molecules when in excess (data not shown). Optimum concentrations of Fab′-POD and Fab′-YS5 were found to be 400 and 175 nM, respectively. Fab′-YS8 competed with LCA for binding, so the differences in percent complex formed using AFP-L3 and AFP-L1 were affected by Fab′-YS8 concentration. In the case of Fab′-YS8 concentration in a range between 50 and 100 nM, a maximum difference of percent AFP-L3 and AFP-L1 complex formed is obtained, when Fab′-YS8’s concentration in the second reagent is 70 nM as shown in Figure 5. Effect of Reaction Temperature. Generally, lectin worked effectively at low temperature. In the present assay, lectin activity is a very important factor. Therefore, the reaction temperature was optimized for the best lectin activity. Reaction temperature was tested from 4 to 22 °C (Figure 6). The percent recovery of complex 1 decreased when the temperature exceeded 12 °C; 8 °C was chosen as an optimum temperature. Effect of Reaction Time. The effects of the first and second reaction times were investigated (Figure 7). The first incubation time was changed from 1 to 40 min and second incubation time was 20 min; the percentage of complex 1 formed was stable for
Figure 6. Effect of temperature on percent recovery of complex 1. AFP-L3 (100 ng/mL) was assayed.
Figure 8. Calibration curve of AFP concentration (A) and ratio of AFP-L3 (B). The calibration curve of AFP concentration was obtained by AFP-L1 purified from human placenta. The calibration curve of ratio of AFP-L3 was obtained by 200 ng/mL AFP samples that mixed up purified AFP-L1 and AFP-L3. Table 1. Within-Run Precision of the Ratio of AFP-L3 and AFP Concentration
Figure 7. Effect of incubation time on percentage of complex 1. (A) First incubation time was changed from 1 to 40 min and second incubation time was 20 min; (B) second incubation time was changed from 1 to 40 min and first incubation time was 20 min: (b) AFP-L1 100 ng/mL, (O) AFP-L3, 100 ng/mL, (×) percent difference of complex 1 between AFP-L1 and AFP-L3.
AFP-L1 and AFP-L3 (Figure 7A). In changing second incubation time from 1 to 40 min when the first incubation time was 20 min, the percentage of complex 1 formed by AFP-L1 decreased depending on the second incubation time, while AFP-L3 maintained a high percentage of complex 1 (Figure 7B). The difference between percentages of complex 1 of AFP-L1 and that of AFP-L3 remained constant between 10 and 30 min; 20 min was chosen as both first and second incubation times. Calibration Curves of AFP Concentration and Ratio of LCA-Reactive AFP. The calibration curve for AFP concentration
serum
no.
A B
17 17
serum
no.
C D
17 17
Ratio of AFP-L3 L3 (%) av (%) 50 20
51.8 20.1
AFP Concentration concn av (ng/mL) (ng/mL) 20 600
20.3 593.4
SD (%)
CV (%)
2.16 0.78
4.2 3.9
SD (ng/mL)
CV (%)
0.80 19.97
4.0 3.4
is linear up to 1000 ng/mL (Figure 8A). Furthermore, the ratio of AFP-L3 formed was linear from 0% to 100% (Figure 8B). The y-intercept was not 0%, because competitive conditions were not perfect. The ratio of AFP-L3 in samples differed between AFPL1 and AFP-L3 data. These calibration curves were used to convert raw data into AFP concentration and ratio of AFP-L3. Statistical Analysis. Table 1 shows within-run precision data from four sera. The sera were measured 17 times. All CVs (%) were below 5% for AFP concentration and ratio of AFP-L3. Correlation with EIA and Lectin Affinity Electrophoresis. Fifty-five sera were measured by this method and compared with both AFP concentrations and ratios of AFP-L3 determined by EIA and by LCA affinity electrophoresis, respectively (Figure 9). The relationship of AFP concentration with EIA was r ) 0.996 and the regression formula obtained is y ) 0.998x + 1.5. The results, expressed as ratio of AFP-L3 showed a correlation with lectin Analytical Chemistry, Vol. 70, No. 10, May 15, 1998
2113
Figure 9. Correlations of AFP concentration and ratio of AFP-L3. (A) AFP concentration of LBA method was compared with the conventional EIA method; (B) ratio of AFP-L3 was compared with the LCA affinity electrophoresis method.
electrophoresis (r ) 0.983). The corresponding regression formula is y ) 1.02x + 1.0. DISCUSSION The measurement of serum AFP-L3 ratios is very useful for the diagnosis and prognosis of HCC.12-18 In fact, these measurements recently have become most important for these purposes. Unfortunately, however, the LCA affinity electrophoresis method for these measurements is very complicated, requires an initial
2114 Analytical Chemistry, Vol. 70, No. 10, May 15, 1998
AFP measurement as a baseline, and has a narrow assay range. In contrast, the LBA method allows the simultaneous measurement of AFP concentration and AFP-L3 ratio. The LBA procedure is simpler than the conventional immunoassay methods that utilize solid-phase reactions. In particular, the LBA method more properly handles the delicate reaction between LCA and monoclonal antibody. By carrying out this important reaction in a liquid phase instead of a solid phase, the LBA method allows greater control of binding conditions, which ideally should take place over a narrow range. Furthermore, the LBA method separates two analytes in a single analysis using two types of sulfated peptides, YS5 and YS8, to label monoclonal antibodies. These antibodies are separated by the differential ionic binding of their conjugate forms to an anion-exchange column. A simple stepwise salt gradient can thus be used to separate the two conjugated antibodies in a short procedure. Moreover, this procedure provides simultaneous quantitative and qualitative analyses. The LBA method can be used to assay other glycoprotein tumor markers both quantitatively and qualitatively. This ability to obtain qualitative information is particularly important for these measurements because conventional tumor markers typically exist in normal body fluids. Slightly elevated levels of these markers often indicate benign diseases. Thus, a rapid qualitative measurement is required to make a clinical determination for these markers. The LBA method can employ various lectin and monoclonal antibodies to detect new tumor markers that will be discovered in the future. However, as yet, a generalization cannot be made concerning the conformation and composition of sugar chains found in these markers. Abbreviations: AFP, R-fetoprotein; LCA, Lens culinaris agglutinin; AFP-L1, Lens culinaris agglutinin-nonreactive R-fetoprotein; AFP-L3, Lens culinaris agglutinin-reactive R-fetoprotein; YS5, sulfated tyrosine pentamer; YS8, sulfated tyrosine octamer.
Received for review November 24, 1997. February 19, 1998. AC971280C
Accepted
