Development of a Rapid Microparticle-enhanced Turbidimetric Immunoassay for Plasma Fatty Acid-binding Protein, an Early Marker of Acute Myocardial Infarction

¹ Department of Physiology, Cardiovascular Research Institute, Maastricht (CARIM), Maastricht University, The Netherlands;
² Roche Diagnostics, a Division of F. Hoffmann-La Roche Ltd., Basel, Switzerland;
^a author for correspondence: fax 31-43-3671028, e-mail glatz{at}fys.unimaas.nl

Heart-type fatty acid-binding protein (FABP; M_r 14 500) is an early plasma marker of acute myocardial infarction (AMI) (1)(2)(3). FABP shows increased plasma concentrations within 3 h after AMI and returns to a normal reference interval within 12–24 h, thereby resembling myoglobin (4)(5). However, the diagnostic specificity and sensitivity of FABP for AMI detection are better than those of myoglobin (6)(7). FABP is also found in skeletal muscle, but the determination of the plasma ratio of myoglobin over FABP allows the discrimination between myocardial and skeletal muscle injury (4)(5).

The application of FABP as an early plasma marker in routine clinical practice requires the availability of a rapid assay system. Several immunochemical assays for FABP have been described, taking an assay time of 2–16 h (2)(3)(8)(9). Recently, a one-step ELISA for plasma FABP with a total performance time of 45 min (10) and an amperometric enzyme immunosensor assay taking 27 min per plasma sample were described (11). However, these assays are of limited use for routine clinical practice. We describe here a microparticle-enhanced turbidimetric immunoassay for FABP that offers the advantages of being precise, easy to perform, rapid, and fully automated.

Latex reagent was prepared by physical adsorption onto carboxylated latex particles of three monoclonal anti-human FABP antibodies that recognize distinct epitopes (12), which were then stored in 10 mmol/L Tris-HCL, pH 8.0, containing 5 g/L bovine serum albumin, 0.01 g/L Tween 20, 1 g/L NaN₃, and 50 g/L sucrose. The assay was performed by using a COBAS^® MIRA Plus analyzer (F. Hoffmann-La Roche Ltd.). Briefly, 75 µL of latex reagent was mixed with 155 µL of reaction buffer (22 mmol/L phosphate buffer, pH 7.0, 350 mmol/L NaCl, 2 g/L bovine serum albumin, 1 g/L NaN₃, 20 mL/L normal rabbit serum (heat inactivated), and 2.57 g/L polyvinylpyrrolidone K90). After incubation at 37 °C for 75 s, 25 µL of sample was added, and the absorbance change of the reaction mixture was measured at a wavelength of 550 nm from about 1 min to ~8 min after the addition of sample. FABP concentration of sample was interpolated automatically from the calibration curve (calibration mode of COBAS^® MIRA Plus: LOGIT/LOG 5).

Calibration curves were recorded with tissue-derived human heart-type FABP (13) or recombinant FABP (donated by Dr. T. Börchers, Münster, Germany), diluted with storage buffer to a 1.5 mg/L FABP solution, and stored for up to 2 months at 4 °C.

FABP concentrations in blood samples were also determined by using an earlier established ELISA of the antigen-capture type (10) and recombinant human heart-type FABP as standard. Taking a 25-fold sample dilution into account, the measuring range of this ELISA is between 0 and 300 µg/L FABP with a detection limit of 5 µg/L FABP (10). For the determination of the reference range of FABP in plasma, this assay was used in a more sensitive mode. By a fivefold predilution of the sample, prolonging the sample incubation time to 2 h and the enzyme reaction (coloring time) to 30 min, the analytical range is 30 times lower (0–10 µg/L FABP), and the detection limit (10) is 0.25 µg/L FABP. For method comparison, 163 EDTA plasma samples of patients with a confirmed clinical diagnosis of AMI (14) were analyzed. For the determination of the reference range of FABP sera from 102 blood donors (54 males) were used. Control sera or plasma specimens were prepared by adding recombinant human heart-type FABP to a pool of blood from 10 healthy donors. All samples were stored at -80 °C.

A typical five point calibration curve, obtained by automated dilution of the standard solution, is shown in Fig. 1 . The analytical detection limit, calculated as the FABP concentration corresponding to the absorbance difference 2 SD found for the zero calibrator and assessed by analyzing the calibrators 11 times (which was repeated four times), was 1.1 ± 0.3 µg/L FABP. With postdilution, the test range could be extended from 150 µg/L up to 2400 µg/L FABP. With tissue-derived heart-type FABP and the recombinant protein, identical results were obtained (data not shown), confirming the same antigenicity found earlier of both proteins (10).

View larger version (27K): [in this window] [in a new window]   Figure 1. Characteristics of the latex FABP assay. Top, shape of the calibration curve for FABP by measuring five calibrators 11 times in one assay. Inset, analysis of the prozone effect (Heidelberger curve) applying samples with an up to 16-fold higher FABP concentration than the highest calibrator (150 µg/L FABP). Bottom, difference plot of 163 plasma samples in which FABP concentration was measured with the ELISA and the latex immunoassay. Mean difference (solid line), 95% confidence interval for the mean difference (mean ± 2 SE; · – · ), and the limits of agreement (mean ± 2 SD; ····) are shown.

To analyze whether an antigen excess phenomenon (prozone effect) may affect the test, a dilution series of FABP in storage buffer was prepared and analyzed. As shown in the inset to Fig. 1 , FABP concentrations between 150 and 2400 µg/L all yielded a signal that was above the limit defined by the measuring range (more than the test range).

Two pools of serum (controls) with different concentrations of FABP were used to assess intraassay precision by running 11 replicates of each sample pool in a single analytical run. An EDTA plasma sample with a high content of added recombinant human heart-type FABP was also measured. Interassay precision was assessed by measurement of the controls 24 times over a period of 2 weeks. The intraassay CV was 2–6%, and the between-day CV was 3–10% (Table 1 ).

Analysis of three plasma samples with different concentrations of FABP and diluted (up to 5 times) with 9 g/L NaCl indicated linearity of the assay (data not shown). Analytical recovery tests were performed by adding a low (7 µg/L) and a high (97 µg/L) concentration of recombinant human heart-type FABP to pools of unaffected plasma (heparinized, citrated, and EDTA plasma) and of serum. The mean recovery was 95%, the range was 90–102%. Additionally, plasma and serum samples with low and high FABP concentrations were mixed in ratios of 0%, 25%, 50%, 75%, and 100%. The mean "mixing recovery" was 101%, with a range of 98–106%, of the calculated concentrations of the mixed samples (data not shown).

For comparison of the latex assay with an established immunoassay (sandwich ELISA), 163 samples from patients with confirmed AMI were analyzed on the same day by both methods. By linear regression analysis, the slope and intercept (±SE) were 1.11 ± 0.02 µg/L and -0.45 ± 1.76 µg/L, respectively. The standard error of estimate (S_yx) was 17.3 µg/L, and the correlation coefficient was 0.98. Agreement between latex assay and ELISA was analyzed in a difference plot according to Bland and Altman (15). Despite the fact that the ELISA values were slightly but significantly lower than the latex values (mean difference = -5.7 µg/L, P <0.05), a good agreement between both methods can be concluded from the small limits (mean ± 2 SD, SD = 19.0). For only those samples that were showing a FABP value between 0 and 20 µg/L in the ELISA, the correlation between the latex assay and the sandwich ELISA was y_{(latex assay)} = 1.16 x _(ELISA) - 0.23 (r = 0.80, n = 82), and the mean difference value ± SD was -1.5 ± 4.3 µg/L FABP.

Serum samples obtained from 102 healthy subjects (44.5 ± 12.9 years) were analyzed for their FABP contents by using the latex assay and, for comparison, also with the ELISA method. In 81 (80%) of the tested sera, the FABP concentration was between <1.1 (detection limit, 53 cases below this value) and 5 µg/L and in 3 cases (3%) above 10 µg/L (highest value, 14.1 µg/L). Of all sera assayed with the ELISA, none had a FABP value above 5 µg/L FABP, whereas the mean concentration was 1.8 ± 0.9 µg/L FABP, ranging from 0.5 to 4.6 µg/L (median value, 1.3 µg/L; 95% confidence interval, 1.4–1.8 µg/L).

The turbidimetric latex immunoassay described here represents a substantial improvement over most widely used heterogeneous immunoassays for quantifying human heart-type FABP in serum or plasma. Three distinct monoclonal antibodies directed against FABP are attached to the microparticles. These antibodies show no cross-reactivity with other human FABPs, such as intestinal-type and liver-type FABP (12). With a performance time of 10 min, the assay is much faster than other FABP assays published to date (2)(3)(8)(9)(10)(11). The detection limit of 1.1 µg/L FABP is similar to that of reported sandwich ELISAs (9)(10), considering that in the latter, a predilution of the sample is indispensable to avoid matrix effects. The validity of the present latex assay for FABP was verified by experimental results in precision (interassay and intraassay), recovery, and dilution tests. No prozone phenomenon was observed for FABP concentrations up to 2400 µg/L, implying a wide analytical range. The comparison of our latex assay with an earlier published ELISA (10) by measuring FABP in 163 plasma samples of AMI patients showed a good correlation.

The average plasma FABP concentration measured in 102 sera of healthy blood donors (about 1.1 µg/L with the latex assay, 1.8 µg/L with the ELISA) is of similar magnitude as the mean values of 3.65 µg/L (n = 100) reported by Ohkaru et al. (9) and of 1.6 µg/L (n = 79) reported by Wodzig et al. (10). However, applying the ELISA and a fivefold predilution of the samples, we found a reference range between 0.5 and 4.6 µg/L FABP, whereas with the latex assay, only ~80% of samples were in this range, the remainder showing FABP concentrations up to 14 µg/L. This discrepancy could relate to interference in the latex agglutination test by sample components in this homogeneous assay system (no separation steps) or to underestimation by the ELISA from suppressed antibody-antigen interaction as a result of insufficient predilution, which was 5-fold instead of the usual 25-fold, of the samples. It should be noted, however, that for samples from AMI patients and FABP concentrations in the range 0–20 µg/L, we found a good correlation between the latex assay and ELISA. Finally, Ohkaru et al. (9), applying an ELISA method, also reported a FABP concentration >5 µg/L in 15% of cases of healthy volunteers. Unfortunately, the predilution factor used in this study and the reference range were not given (9).

For the detection of AMI, upper reference values of 10 and 12 µg/L FABP have been described (7)(14). With regard to this discriminator value, the latex assay will be a reliable tool for the early diagnosis of AMI, because in only 1 case of 102 sera tested was a slightly higher FABP value (14.1 µg/L) found. In addition, in plasma from AMI patients, FABP concentrations generally show a rapid increase after onset of symptoms (2)(3)(4), so that analyses of FABP in two samples taken at a short time interval and expression of the change in concentration may further increase the diagnostic power of this marker.

In conclusion, we found that the simplicity, reproducibility, and full automation in a widely used clinical chemistry analyzer like the COBAS MIRA seem to be factors of choice for the FABP latex immunoassay for routine clinical diagnosis of AMI.

Acknowledgments

We thank Dr. J. Górski, Bialystok, Poland, for providing EDTA plasma samples from patients with confirmed AMI, and Dr. J.-P. Chapelle, Liège, Belgium, for providing sera from healthy blood donors. We would also like to thank Maurice Pelsers for expert technical assistance and Drs. S. Eda and H. Hager for stimulating discussions.

References

1. Adams JE, Abendschein DR, Jaffe AS. Biochemical markers of myocardial injury: is MB creatine kinase the choice for the 1990's?. Circulation 1993;88:750-763. [Free Full Text]
2. Tanaka T, Hirota Y, Sohmiya K, Nishimura S, Kawamura K. Serum and urinary human heart fatty acid-binding protein in acute myocardial infarction. Clin Biochem 1991;24:195-201. [Web of Science][Medline] [Order article via Infotrieve]
3. Kleine AH, Glatz JFC, Van Nieuwenhoven FA, Van der Vusse GJ. Release of heart fatty acid-binding protein into plasma after acute myocardial infarction in man. Mol Cell Biochem 1992;116:155-162. [Web of Science][Medline] [Order article via Infotrieve]
4. Van Nieuwenhoven FA, Kleine AH, Wodzig KWH, Hermens WTh, Kragten HA, Maessen JG, et al. Discrimination between myocardial and skeletal muscle injury by assessment of the plasma ratio of myoglobin over fatty acid-binding protein. Circulation 1995;92:2848-2854. [Abstract/Free Full Text]
5. Yoshimoto K, Tanaka T, Sohmiya K, Tsuji R, Okamoto F, Kawamura K, et al. Human heart-type cytoplasmic fatty acid-binding protein as an indicator of acute myocardial infarction. Heart Vessels 1995;10:304-309. [Web of Science][Medline] [Order article via Infotrieve]
6. Glatz JFC, Haastrup B, Hermens WT, de Zwaan C, Barker J, McNeil CJ, et al. Fatty acid-binding protein and the early detection of myocardial infarction: the EUROCARDI Multicenter Trial. [Abstract]. Circulation 1997;96(Suppl I):215.
7. Ishii J, Wang JH, Naruse H, Taga S, Kinoshita M, Kurokawa H, et al. Serum concentrations of myoglobin vs heart-type cytoplasmic fatty acid-binding protein in early detection of acute myocardial infarction. Clin Chem 1997;43:1372-1378. [Abstract/Free Full Text]
8. Knowlton AA, Burrier RE, Brecher P. Rabbit heart fatty acid-binding protein: isolation, characterization, and application of a monoclonal antibody. Circ Res 1989;65:981-988. [Abstract/Free Full Text]
9. Ohkaru Y, Asayama K, Ishii H, Nishimura S, Sunahara N, Tanaka T, Kawamura K. Development of a sandwich enzyme-linked immunosorbent assay for the determination of human heart type fatty acid-binding protein in plasma and urine by using two different monoclonal antibodies specific for human heart fatty acid-binding protein. J Immunol Methods 1995;178:99-111. [Web of Science][Medline] [Order article via Infotrieve]
10. Wodzig KWH, Pelsers MMAL, Van der Vusse GJ, Roos W, Glatz JFC. One-step enzyme-linked immunosorbent assay (ELISA) for plasma fatty acid-binding protein. Ann Clin Biochem 1997;34:263-268.
11. Siegmann-Thoss C, Renneberg R, Glatz JFC, Spener F. Enzyme immunosensor for diagnosis of myocardial infarction. Sensors Actuators 1996;B 30:71-76.
12. Roos W, Eymann E, Symannek M, Duppenthaler J, Wodzig KWH, Pelsers MMAL, Glatz JFC. Monoclonal antibodies to human heart fatty acid-binding protein. J Immunol Methods 1995;183:149-153. [Web of Science][Medline] [Order article via Infotrieve]
13. Van Nieuwenhoven FA, Vork MM, Surtel DAM, Kleine AH, Van der Vusse GJ, Glatz JFC. High-yield two-step chromatographic procedure for purification of fatty acid-binding protein from human heart. J Chromatogr 1991;570:173-179. [Web of Science][Medline] [Order article via Infotrieve]
14. Wodzig KWH, Kragten JA, Modrzejewski W, Górski J, Van Dieijen-Visser MP, Glatz JFC, Hermens WT. Thrombolitic therapy does not change the release ratios of enzymatic and non-enzymatic myocardial marker proteins. Clin Chim Acta (in press)..
15. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307-310.