| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Central Research Institute of BodiTech Inc., Chuncheon, South Korea;2 Department of Genetic Engineering, Hallym University, Chuncheon, South Korea;3 Department of Chemistry, Seoul National University, Seoul, South Korea
aaddress correspondence to this author at: Central Research Institute of BodiTech Inc., Chuncheon 200-160, South Korea; fax 82-33-258-6889, e-mail sangwoh{at}empal.com or sangwoh{at}boditech.co.kr
Microalbuminuria (MAU), defined as a urinary albumin excretion of 30–300 mg/day, indicates a high probability of renal damage and is an accepted predictor for the early diagnosis of nephropathy in diabetic patients (1)(2). In addition, MAU has diagnostic implications in pregnancy as a predictive marker of preeclampsia (3)(4) and may play a role in identifying high risk of developing complications from cardiovascular diseases even in nondiabetic patients (5)(6)(7).
Dye-binding assays can measure serum albumin but are too insensitive for MAU testing, making immunochemical assays the most widely used MAU methods (8). These immunoassays include immunoturbidimetry, immunofluorescence, ELISA, RIA, and zone immunoelectrophoresis. Recently, Kessler and coworkers (9)(10) introduced a laser-induced fluorescence system coupled to an automated centrifugal analyzer as a nonimmunologic assay for urinary albumin. Their system was based on the probe Albumin Blue 670/580, which becomes highly fluorescent on binding to albumin.
We report a fluorescence immunochromatography assay (ICA) for quantitative determination of albumin in urine. The assay system consists of an ICA test strip in a disposable cartridge, a fluorescently labeled detector, and a laser fluorescence reader. Basically, the assay system adopts the inherent simplicity of a lateral-flow ICA and uses a competitive immunoassay mode with a simple, one-step operation (11). Briefly, fluorescently labeled albumin in the detector buffer competes with albumin in the sample for binding to an anti-albumin antibody immobilized on the test strip matrix. The more albumin is in the sample, the less the fluorescently labeled albumin reacts with the anti-albumin antibody and, thus, the lower the accumulation of fluorescence in the test line of a test strip.
We generated a monoclonal antibody against human albumin (Sigma A8763) as an immunogen and conducted immunizations, cell fusion, and screening of hybridoma cells according to a standard method (12). Monoclonal antibody 22C5 was selected among the positive clones and used as the capture antibody (3 g/L) on the test line of a test strip for this study (13). A control line was coated with anti-rabbit IgG (1 g/L; Sigma R4880) on nitrocellulose membrane by a BioJet dispenser (BioDot). We labeled the albumin competitor and rabbit IgG control with activated Alexa Fluor 647 (Molecular Probes) in sodium bicarbonate buffer (pH 8.3) and made detector buffer by mixing two fluorescent conjugates in 0.1 mmol/L phosphate buffer (pH 6.0). The system components and the principle for the one-dimensional fluorescence reader for scanning of fluorescence intensity have been described in detail elsewhere (14).
We tested the fluorescence ICA system at albumin concentrations of 0–600 mg/L. We mixed 10 µL of detector (2 µg of fluorescent competitor and 80 ng of fluorescent rabbit IgG control), 10 µL of 1 mol/L potassium phosphate buffer (pH 6.0), and 80 µL of urine sample in the test aliquot and loaded the mixture in the cartridge well. After the sample was allowed to react for 10 min, the cartridge was inserted in the laser fluorescence reader for scanning of fluorescence intensity.
Intensity profiles, in relative fluorescence units (RFUs), for the different albumin concentrations in the samples are shown in Fig. 1B
. The first and second peaks represent the RFUs for the control line and the test line, respectively. Whereas the RFUs on the control lines were constant, the signals on the test lines changed dramatically depending on the albumin concentrations in the samples.
|
A commercially available multicalibrator set (Kamiya Biomedical) and a pure human albumin were used to construct a calibration curve. The albumin stock solution (1000 mg/L) was diluted with calibrator diluent to final concentrations of 5, 10, 20, 50, 100, 200, 300, 500, and 600 mg/L. The calibrator diluent contained, per liter, 2.7 g of KH2PO4, 0.9 g of K2HPO4, 4.5 g of NaCl, and 0.5 g of EDTA (pH 6.0) and was used as control material for all assays (9). The RFUs recorded on the test and control lines at a given albumin concentration were converted into an area value (test, AT, control, AC) by a fitting algorithm. The calibration curve was obtained from the area ratios (AT/AC) and the albumin concentrations in the samples (r2 = 0.994; Fig. 1A
).
We next compared the fluorescence ICA system with a RIA (EURO Diagnostic Products Co.). The RIA was performed on a Cobra 5010 II analyzer (Quantum). Urine samples were collected from 81 patients who visited Hallym University Medical Center (Chuncheon, Korea). The results were compared by use of Medcalc, Ver. 6.12, software (Medcalc Inc.). A Bland–Altman difference plot analysis showed a mean (SD) difference of –1.2 (46.5) mg/L and little disagreement between the two assays at the mean urinary albumin concentration (Fig. 1C
). However, the agreement between the two assays was lower in samples with concentrations >300 mg/L. The Passing–Bablok regression analysis yielded a slope of 1.06 (95% confidence interval, 0.988–1.153) and a y-intercept of –9.63 mg/L (95% confidence interval, –16.7 to –6.4 mg/L), indicating statistically good agreement between the two methods (P <0.05).
We conducted precision studies to evaluate the analytical performance of the fluorescence ICA system with urine reference materials. We prepared four diluted control samples covering the albumin concentration range usually encountered in clinical practice (Table 1
). The intra- and interassay CV for the new immunoassay system were, respectively, 7.2% and 7.7% at 15 mg/L, 4.5% and 7.1% at 92.3 mg/L, 5.2% and 3.6% at 180 mg/L, and 5.4% and 3.2% at 342 mg/L. The analytical precision was also calculated with 10 replicates in one analytical run from the same controls. The measured albumin values of 15, 92.3, 180, 342 mg/L were 100.1%, 108.5%, 112.8%, and 97.6%, respectively, of the expected values, with a mean measured value that was 105% of the expected. To confirm the reliability of the immunoassay, we also analyzed the parallelism in a series of twofold serial dilutions of the 600 mg/L albumin calibrator. The results obtained were compared with the expected results by linear regression and showed a slope (SD) of 1.026 (0.03) with correlation coefficient of 0.997 (see Fig. 1
in the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol50/issue6/).
|
We tested assay interference by adding to urine pools (20 mg/L albumin) urea (1 mol/L), creatinine (10 g/L), and human hemoglobin (10 mg/L). The concentrations added to the urine pools were well above the maximum concentrations encountered in clinical samples. The overall mean albumin concentration of 50 samples tested was 21.4 mg/L (CV = 7.2%). Because the albumin concentrations were all within 2 SD of the reference value, we considered that none of tested materials affected the assay. The limit of detection (3 SD above the value for the zero calibrator; n = 10) and limit of quantification (lowest concentration measured with a CV <10%) of the fluorescence ICA system were 1.6 and 4.65 mg/L, respectively. The developed assay system was sensitive to the pH of urine specimens and operated best at pH 6.0. Because the pH range of the collected urine specimens fell within the range typical of urine (pH 4.5–8.0), 1 mol/L potassium phosphate buffer (pH 6.0) was integrated into the assay component.
Fully automated immunoassay formats are available for quantification of urinary albumin in large numbers of samples. However, most of these methods are impractical or expensive. The criteria for point-of-care testing include affordable cost, a disposable device, and minimum maintenance/technical expertise required to perform tests (15). The sample should be applied directly to the device, which should require only a small sample volume, and the assays should have a rapid turnaround time with good accuracy. There are some point-of-care devices for determination of MAU in urine, such as the ImmunoDip (Diagnostic Chemicals Limited) and Micral Urine Test Strip (Roche Diagnostics). Despite their many advantages, one drawback of these commercial test devices is that they give only negative, threshold, or positive results without displaying quantitative values for urinary albumin. Given the different principles of the assays compared, the results obtained with the fluorescence ICA agree well with the results obtained with the independent RIA. Considering the detection limit, imprecision, linearity, and working range, the fluorescent ICA is comparable to other, well-known immunoassays and appears to be suitable for determination of urinary albumin.
Acknowledgments
This study was supported by the Nano-Core Technology Development Program (M1-0213-05-0003) and the National Research Laboratory Program (M1-0104-00-0164) of the Korean Ministry of Science and Technology.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  Y. Luo, M. Chen, Q. Wen, M. Zhao, B. Zhang, X. Li, F. Wang, Q. Huang, C. Yao, T. Jiang, et al. Rapid and Simultaneous Quantification of 4 Urinary Proteins by Piezoelectric Quartz Crystal Microbalance Immunosensor Array Clin. Chem., December 1, 2006; 52(12): 2273 - 2280. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
