An Enzyme Immunoanalytical System Based on Sequential Cross

Anal. Chem. 2005, 77, 4091-4097

An Enzyme Immunoanalytical System Based on Sequential Cross-Flow Chromatography Joung-Hwan Cho,†,‡ Eui-Hwan Paek,‡ Il-Hoon Cho,§ and Se-Hwan Paek*,†,‡,§

Graduate School of Biotechnology, BioDigit Laboratories Corp., and Program for Bio-Microsystem Technology, Korea University, 1, 5-ka, Anam-dong, Sungbuk-ku, Seoul 136-701, Korea, and Department of Biotechnology, Korea University, Jochiwon, Choongnam 339-800, Korea

A new enzyme immunoanalytical concept that can be used for point-of-care testing has been investigated. Enzyme as a tracer requires a separate reaction step for signal generation, which follows the completion of immune complex formation with analyte (e.g., Hepatitis B surface antigen) in a sample. This has been a major factor limiting its utilization within the laboratory. We carried out such sequential processes employing chromatographic analysis, using two crosswise-arranged membrane pads in vertical and horizontal directions. The vertically arranged pads were the same as those in the usual format for pregnancy testing, for instance, with the exception of the use of horseradish peroxidase (HRP) as tracer. By placing the horizontally arranged pads on each lateral side of the signal generation pad in the vertical arrangement, they were employed to supply substrate to the enzyme present in the immune complexes. The substrate flow was initiated after the antigen-antibody bindings to produce a signal, which was typically a color change in proportion to the analyte concentration. Under optimal conditions, the use of HRP labeling increased the detection capability of the assay ∼30 times compared to that of gold colloids. Potential advantages of using the concept investigated are (1) provision of a rapid and simple immunoassay, (2) satisfaction of a clinical need for highly sensitive determination of analyte, and (3) utilization of relatively inexpensive, portable quantitation means. Membrane strip-based immunochromatographic analysis was utilized initially for point-of-care-testing (POCT) of symptoms such as pregnancy and ovulation and, more recently, for the diagnosis of various diseases, including microbial infections and even acute myocardial infarction (AMI).1-3 Despite such expansion of its applications, the technology faced a major limitation in achieving a lower detection limit of a target analyte in sample. For signal * Corresponding author. Phone: +82-2-3290-3438, Fax: +82-2-927-2797. E-mail: [email protected]. † Graduate School of Biotechnology. ‡ BioDigit Laboratories Corp. § Program for Bio-Microsystem Technology. (1) Collinson, P. O.; Scand, J. Clin. Lab. Invest. Suppl. 1999, 230, 67-73. (2) Toshio, W.; Yuichi, O.; Hidekazu, M.; Hiroshi, K.; Yasunori, S.; Yasuhiko, O.; Takao, T.; Yasushi, K. Clin. Biochem. 2001, 34, 257-263. (3) Heeschen, C.; Goldmann, B. U.; Langenbrink, L.; Matschuck, G.; Hamm, C. W.Clin. Chem. 1999, 45, 1789-1796. 10.1021/ac048270d CCC: $30.25 Published on Web 05/24/2005

© 2005 American Chemical Society

generation, the assay has employed tracers, such as gold colloids or Latex beads, which produce a color signal detectable by the naked eye.4,5 The signal can then be readily converted to optical density for quantitation using photometric transducing means.6,7 The detection, however, often does not cover the clinical range of the minimum analyte concentration required for early diagnosis of a disease. The detection capability of the POCT system may be enhanced using signal generators, such as fluorescent substances8 or magnetic beads,9 perceivable in a low concentration. An analytical system has indeed been developed, where an immunochromatographic assay was performed using a detection antibody labeled with a fluorescent substance, and the assay result was measured with a fluorescence detector (e.g., photomultiplier tube).10,11 Such a system has provided a lower detection limit with no harmful effects. It has been further applied for a rapid test to detect AMI in ambulances and emergency rooms, for example.12,13 However, It is difficult to hand-carry the fluorescence detector, due to the requirement of a high power supply (up to near 1000 V10) and also its relative expense compared to other common means. Likewise, a device employing the magnetic tracer showed an enhanced performance,14 yet it was costly and bulky in its dimensions. Thus, such devices do not seem to offer other significant advantages than the rapid assay when compared to conventional laboratory-version enzyme-linked immunosorbent assay (ELISA) systems3,15 widely used in laboratories. As exemplified in ELISA, enzymes are an alternative tracer that has been used over the past two decades.15 The tracer (4) Paek, S. H.; Lee, S. H.; Cho, J. H.; Kim, Y. S. Methods 2000, 22, 53-60. (5) Centola, G. M.; Andolina, E.; Deutsch, A. Am. J. Reprod. Immunol. 1997, 37, 300-303. (6) Lo ¨nnberg, M.; Carlsson, J. Anal. Biochem. 2001, 293, 224-231. (7) Cho, J. H.; Paek, S. H. Biotechnol. Bioeng. 2001, 75, 725-732. (8) Choi, S.; Choi, E. Y.; Kim, D. J.; Kim, J. H.; Kim, T. S.; Oh, S. W. Clin. Chim. Acta 2004, 339, 147-156. (9) Mulvaney, S. P.; Mattoussi, H. M.; Whitman, L. J. Biotechniques 2004, 36, 602-606. (10) Rigolin, G. M.; Lanza, F.; Castoldi, G. Cytometry 1995, 20, 362-368. (11) Lundgren, J. S.; Watkins, A. N.; Racz, D.; Ligler, F. S. Biosens. Bioelectron. 2000, 15, 417-421. (12) Altinier, S.; Zaninotto, M.; Mion, M.; Carraro, P.; Rocco, S.; Tosato, F.; Plebani, M. Clin. Chim. Acta 2001, 311, 67-72. (13) Apple, F. S.; Christenson, R. H.; Valdes, R., Jr.; Andriak, A. J.; Berg, A.; Duh, S. H.; Feng, Y. J.; Jortani, S. A.; Johnson, N. A.; Koplen, B.; Mascotti, K.; Wu, A. H. Clin. Chem. 1999, 45, 199-205. (14) Ronald, T. L. U. S. Patent 6,607,922B2, 2003. (15) Ionescu, M. I.; Sanchez, Y.; Fields, H. A.; Dreesman, G. R. J. Virol. Methods 1983, 6, 41-52.

Analytical Chemistry, Vol. 77, No. 13, July 1, 2005 4091

produces an amplified signal resulting from the catalytic action and also offers different types of signals that are measurable by comparatively simple detectors (e.g., based on photometry,3,16 chemiluminometry,17 and electrochemistry18), depending on the substrate as well as the enzyme used. The enzyme reaction, as a unique property of this tracer, should be carried out separately for signal generation following the completion of the antigenantibody bindings. A standard protocol of the heterogeneous immunoassay, thus, requires washing steps for isolating the immune complexes formed on the solid surfaces from unreacted reagents. Despite the affirmative aspects of the enzyme tracer, such complex, multistep procedures of ELISA are clearly unsuitable for POCT. To eventually develop an enzyme immunosensor for POCT, we have investigated a novel analytical concept combining the method of immunochromatographic assay with the technology of enzyme signal generation. This process is capable of accomplishing the two reactions in a sequential mode, i.e., induction of antigen-antibody bindings in the vertical direction and production of catalytic signals in the horizontal direction after the complete formation of the immune complexes. Such immunoassay based on cross-flow chromatography significantly simplified the complex procedure required in ELISA and, thus, would be acceptable for POCT. In regard to the independent accomplishment of the two reactions, the concept is completely different and more advanced than that previously reported,19,20 enabling production of a highly sensitive signal. We present here the research results in experimental demonstration of the concept and characterization of its analytical performance. EXPERIMENTAL SECTION Materials. Hepatitis B surface antigen (HBsAg), polyclonal antibodies (produced from goat), and monoclonal antibodies against HBsAg were supplied by International Enzymes (Fallbrook, CA). Polyclonal antibody (produced from goat) against mouse immunoglobulin G (IgG), the antibody conjugated with horseradish peroxidase (HRP), streptavidin, N-succinimidyl-3(2-pyridyldithio)propionate (SPDP), succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC), and dithiotheritol (DTT) were obtained from Pierce (Rockford, IL). Casein (sodium salt type, extracted from milk), bovine serum albumin (purified by heat shock process, fraction V), human serum (frozen liquid), Tween 20, 3,3′,5,5′-tetramethylbenzidene (TMB), Triton X-100, Sephadex G-15, and Sephadex G-100 were obtained from Sigma (St. Louis, MO). Nitrocellulose (NC) membrane (12-µm pore size) and glass fiber membrane (Ahlstrom 8980) were purchased from Millipore (Bedford, MA). Cellulose membrane (17 CHR chromatography grade) was purchased from Whatman (Maidstone, England). HRP was supplied by Calbiochem (San Diego, CA) and insoluble TMB by Moss (Pasadena, MD). Other reagents used were of analytical grade. (16) Zhu, Q. Z.; Yang, H. H.; Li, D. H.; Chen, Q. Y.; Xu, J. G. Analyst 2000, 125, 2260-2263. (17) Ireland, D.; Samuel, D. J. Biolumin. Chemilumin. 1989, 4, 159-63. (18) Zeravik, J.; Ruzgas, T.; Franek, M. Biosens. Bioelectron. 2003, 18, 13211327. (19) Watanabe, H.; Satake, A.; Kido, Y.; Tsuji A. Anal. Chim. Acta 2001, 437, 31-38. (20) Ono, T.; Kawamura, M.; Arao, S.; Nariuchi, H. J. Immunol. Methods 2003, 272, 211-218.

4092

Analytical Chemistry, Vol. 77, No. 13, July 1, 2005

Labeling of Antibody. Labeling with Gold Colloids. Gold colloids with a mean diameter of 30 nm were synthesized using the sodium citrate method.21,22 The monoclonal antibody (0.1 mg/mL, 0.8 mL) specific to HBsAg and dialyzed in 10 mM phosphate buffer (pH 7.4, PB) was added to the gold solution (8 mL), and the solution was adjusted to pH 8.0. After reacting for 1 h, PB containing 5% casein (1 mL, casein-PB) was added and further reacted for 30 min. The reaction mixture was centrifuged at 15 000 rpm for 45 min, and the supernatant was then removed. The gold precipitates were suspended in caseinPB (total volume of 0.2 mL, i.e., 40-fold concentrate) and stored at 4 °C until used. Labeling with HRP. The antibody was also chemically coupled with HRP using cross-linkers. The antibody (total 1 mg, 0.5 mL) was activated with SMCC dissolved in dimethyl sulfoxide (DMSO) in a 20 molar excess at 4 °C for 4 h and then fractionated by Sephadex G-15 gel (10-mL gel volume) chromatography. HRP (total 2.8 mg, 0.5 mL) was dissolved in a reaction solution (100 mM PB, pH 7.4) containing 5 mM ethylenediaminetetraacetic acid disodium salt and reacted with SPDP in DMSO in 50 molar excess at room temperature for 2 h. The linker was activated using DTT (final 10 mM) at 37 °C for 1 h, and excess reagents were then removed on Sephadex G-15 column. The two activated proteins, antibody and enzyme, were immediately combined in a 1:10 molar ratio and reacted overnight at 4 °C. This mixture was purified by size exclusion chromatography using a Sephadex G-100 gel column (10 × 200 mm). Proteins within each fraction were monitored employing the Bradford assay,23 and the conjugates synthesized were determined by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (7% gel) without a reducing agent. Construction of Chromatographic Analytical Systems. Preparation of Vertically Arranged Pads (Immunostrip). To carry out antigen-antibody reactions, four different functional membrane pads were used (refer to Figure 1). A sample application pad was made by immersing a glass fiber membrane (7 × 20 mm) in PB containing 140 mM NaCl (PBS) containing 0.1% (v/v) Triton X-100 and then drying it in the ambient air. A conjugate release pad was fabricated by transferring 20 µL of a conjugate solution onto the glass membrane (7 × 5 mm). The conjugate solution was prepared by adding the antibody labeled with gold colloids (final 4-fold concentrate), which was diluted in casein-PB containing normal goat antibodies (IgG fraction, 4 mg/mL) and trehalose (20%, w/v). For indirect labeling with HRP, the polyclonal antibody (final 1 µg/mL) against mouse IgG, conjugated with HRP, was further added into the conjugate solution containing the goldlabeled antibody. To fabricate the signal generation pad, the polyclonal antibody against HBsAg was immobilized on a predetermined site of NC membrane (12-µm pore size, 7 × 25 mm) by physical adsorption. The antibody (1 mg/mL) diluted in PBS was either spotted (1 µL/strip) or dispensed (1.5 µL/cm) using a microdispenser (BioJet 3000, Biodot, Irvine, CA) onto a site at 10 mm from the bottom of the membrane strip. In case of line pattern, the goat antibody against mouse (0.5 mg/mL) was also (21) Albert, R. M.; Simmons, S. R.; Pawley. Immunocytochemistry; Oxford University Press: Oxford, U.K., 1993; pp 151-176. (22) Dykman, L. A.; Lyakhov, A. A.; Bogatyrev, V. A.; Chchyogolev, S. Y. Colloid 1998, 60, 700-704. (23) Loffler, B. M.; Kunze, H. Anal. Biochem. 1989, 177, 100-102.

Figure 1. Analytical components of the cross-flow chromatographic immunoassay device. The membrane pad components were arranged into two groups for antigen-antibody bindings (A) and enzyme reactions (B).

dispensed onto a site at 17 mm from the bottom. After incubating for 1 h, the strip was immersed in 100 mM tris buffer, pH 7.6, containing 0.5% (w/v) casein (casein-Tris) and 0.1% (v/v) Triton X-100 for 1 h. This was then dried at 37 °C for 2 h and kept in a desiccator at room temperature until used. The prepared membrane pads were arranged to be a 7-mm width in order, from the bottom, sample application pad, conjugate release pad, signal generation pad, and a cellulose membrane (7 × 20 mm) as the absorption pad induced a continuous wicking. Each contiguous membrane strip was partially superimposed and fixed on a plastic film using double-sided tape to construct an immunostrip. Preparation of Horizontally Arranged Pads. To produce an enzyme signal, two separate pads were prepared, as shown in Figure 1. The enzyme substrate supply pad was made by immersing the glass fiber membrane (12 × 15 mm) in casein-PB containing 0.1% (v/v) Triton X-100 and then drying in the ambient air. The cellulose membrane (12 × 15 mm) was also used as an absorption pad to provoke the horizontal flow. The two pads were spatially arranged on a plastic plate to be, at the time of signal generation, in 0.5-mm superimposition with both the lateral sides of the signal generation pad, respectively, and then fixed using double-sided tape. Immunochromatographic Analyses. Preparation of Standard Concentrations of HBsAg. Standard HBsAg solutions were prepared by serial dilutions of the stock supplied by the manufacturer. Diluents used were casein-PBS for the optimization of variables and human serum for the characterization of analytical systems. The diluent itself was regarded as a negative sample. Immunochromatographic Assay Using Gold as Tracer. Each of the standard solutions of HBsAg (total 150 µL) were placed into different microwells, and the immunostrips with the gold labels were placed into their respective microwells in an erect position for 15 min to absorb the aqueous solutions into the strips. The signal appeared at the area of the immobilized antibody, and the image was captured using a scanner (HP ScanJet 4670, HewlettPackard, Palo Alto, CA). The colored signal on the image was

quantified using an image analysis program (Multianalyst version 1.1, Bio-Rad Laboratories, Hercules, CA) as optical density. Immunochromatographic Assay Using HRP as Tracer. The analytical conditions for immune reactions were basically identical to those described above, excepting the use of the detection antibody indirectly labeled with HRP via secondary antibody, as mentioned. After a 15-min reaction, the horizontally arranged pads were positioned to the right and left sides of the signal generation pad, respectively, and the substrate solution for HRP was supplied to allow the progress of horizontal flow for 5 min. The color signal generated from HRP was converted to optical density using the same procedure as described. Determination of the Detection Limit. For characterizing the analytical systems, the chromatographic assays were repeated three times at each HbsAg concentration. The averages at the respective concentration were subtracted from the mean value at zero dose and used to plot the dose-response curve. The detection limit was determined as the analyte concentration corresponding to the signal value, which had been calculated by multiplying the standard deviation of the signal at the zero dose by three.6,20 Enzyme-Linked Immunosorbent Assay. With Directly Labeled HRP. The polyclonal antibody against HBsAg was diluted to 1 µg/mL in PBS. The solution was added to microwells (100 µL each) and incubated at 37 °C for 1 h. The microwells were washed 3 times with deionized water, and the residual surfaces of each well were then treated with 200 µL of caseinPBS at 37 °C for 1 h. After washing, the HBsAg standards diluted with casein-PBS containing 0.1% Tween (casein-PBS-TW) were transferred into each microwell (100 µL) and incubated at 37 °C for 1 h. The wells were washed again, and 1 µg/mL of the antiHBsAg monoclonal antibody labeled with HRP, diluted with casein-PBS-TW, was added (100 µL each) and reacted at 37 °C for 1 h. After washing, a HRP substrate solution was dispensed into the wells (200 µL each) and reacted at room temperature for 15 min. The substrate solution contained 10 µL of 3% (v/v) hydrogen peroxide in water, 100 µL of 10 mg/mL TMB in DMSO, and 10 mL of 0.05 M acetate buffer, pH 5.1. The reaction was terminated by adding 2 M sulfuric acid (50 µL each). The color was measured at an absorbance of 450 nm using a microplate reader (VERSAmax, Molecular Devices Corp., Sunnyvale, CA). With Indirectly Labeled HRP. The procedure for immunoassays employing the indirect labeling method was essentially the same as that of the direct labeling method. After complex formation between HBsAg and the antibody coated the inner surfaces of the microwell, the anti-HBsAg monoclonal antibody, which had been diluted to 1 µg/mL in casein-PBS-TW, was added into the wells (100 µL each) and incubated at 37 °C for 1 h. The wells were washed 3 times, and the anti-mouse goat antibody-HRP conjugates (100 µL each well) were diluted to 0.2 µg/mL in caseinPBS-TW and reacted under the same conditions. The next steps were identical to those mentioned above. ANALYTICAL CONCEPT Major Components of the Assay System. Enzyme labels that provide a high signal yield can be employed in an immunochromatographic assay, provided a separate reaction of the enzyme from the immune reactions is accomplished. This can be achieved by supplementing the flow of an enzyme substrate following the Analytical Chemistry, Vol. 77, No. 13, July 1, 2005

4093

Figure 2. Analytical principle of cross-flow chromatographic enzyme immunoassay. Upon sample application, a flow is induced in the vertical direction by capillary action, driving antigen-antibody bindings with the analyte (A). For signal generation, the horizontally arranged pads are attached on each lateral side of the signal generation pad (B) and then supply the enzyme substrate to the enzyme present on the pad (C). Consequently, enzyme reaction takes place for signal generation (D).

conventional lateral flow along the immunostrip.4,7 Such a sequential flow system has been constructed by utilizing two groups of membrane pads, vertically arranged pads and horizontally arranged pads (refer to Figure 1). Similarly, in a conventional system, the vertically arranged pads for antigen-antibody reactions of an aqueous medium consisted of four different membranes connected to each other toward the direction of capillary action (Figure 1A). For sample application, a glass fiber membrane pad is located at the bottom of the vertical arrangement. The next upper pad is a glass fiber membrane containing the detection antibody against analyte (e.g., HBsAg) conjugated with an enzyme (e.g., HRP) in a dry state. The signal generation pad, made of nitrocellulose with an immobilized capture antibody also specific to the analyte, is consecutively located at the top of the conjugate release pad. Finally, a cellulose membrane is placed at the top position in the arrangement to induce a continuous flow of medium by absorption. As new components that have not yet been introduced into any conventional systems, two membrane pads supplying the enzyme substrate to the immune complexes, which formed on the signal generation pad after the vertical flow, are used in a horizontal arrangement (Figure 1B). A glass fiber membrane holds the enzyme substrate and provides it to the enzyme upon connection to the signal generation pad. Cellulose membrane is employed to keep the substrate flow horizontally by absorption. At an initial stage, the two membrane pads are spatially separate from those in the vertical arrangement. Principle of Analysis. The analytical procedure using the components consists of four steps, as is shown in Figure 2. For antigen-antibody reactions, the bottom end of the vertically arranged pad is immersed in an analyte-containing specimen (e.g., serum, whole blood). The specimen is then absorbed inside the sample application pad and transferred along the strip by lateral flow (Figure 2A). When the medium reaches the pad with 4094

Analytical Chemistry, Vol. 77, No. 13, July 1, 2005

the detection antibody coupled with the enzyme accumulated initially in a dry state, the conjugate is instantly dissolved and a primary immune complex is formed by antigen-antibody reactions in the liquid phase. The immune complex is then carried to the signal generation pad in the upper position and captured on the solid surface to form a sandwich-type immune complex. Unbound substances are separated by the medium flow. For enzyme reaction, the horizontally arranged pads are partially superimposed and fixed at both of the lateral sides of the signal generation pad, respectively, after the completion of vertical flow (Figure 2B). The enzyme substrate solution is manually (or by semiautomatic action) added to the substrate supply pad. As a consequence, horizontal flow is initiated. The flow passes through the signal generation pad, which triggers the catalytic reaction of the enzyme included in the sandwich immune complexes formed on the solid surfaces (Figure 2C). A signal is then produced in proportion to the analyte concentration (Figure 2D) and can be quantified using adequate means (e.g., colorimetric detector), dependent upon the signal type. This novel analytical concept, based on sequential cross-flow chromatography, would potentially enable us to perform a highly sensitive enzyme immunoassay outside of the laboratory. Furthermore, the utilization of an enzyme label opens up opportunities to use a variety of the detection methods mentioned above. RESULTS AND DISCUSSION To demonstrate the analytical concept, we have selected HBsAg, which has been routinely measured for the test of infection by hepatitis B virus, as a model analyte.24 For early diagnosis of the infection, a minimum 2 ng/mL antigen needs to be detected.24,25 Since the primary factor controlling the detection limit is the binding affinity between antigen and antibody, the (24) Torlesse, H.; Wurie, I. M.; Hodges, M. Br. J. Biomed. Sci. 1997, 54, 256259.

Figure 3. Comparison of enzyme labeling methods toward their analytical performances in the ELISA format. An enzyme, HRP, was coupled to the detection antibody (direct labeling) or a secondary antibody binding to the detection (indirect labeling).

selection of an antibody pair with high binding affinities has been a usual requirement of the test. In this study, we have verified that the application of the novel enzyme immunochromatographic analysis significantly improves the detection capability, which, thus, facilitated the practical utilization of available antibodies even with enhanced performances. Selection of Enzyme Labeling Method. Among enzymes widely used as labels, HRP is the most sensitive signal generator, if appropriately combined with substrates such as TMB.26,27 The major drawback in utilizing the enzyme tracer in the immunochromatographic assay has been a lack of technical solutions for carrying out the consecutive enzyme reactions after the formation of immune complexes. Hence, the novel analytical concept based on cross-flow as a solution has been applied to the assay for HBsAg using HRP as signal generator. Although variable methods of chemically labeling detection antibodies with enzymes have been reported,28 a purification protocol of the conjugate has not yet been properly established. The use of the conjugate in a mixture with free antibody initiates competition between the different binding species for the analyte molecules, causing a decrease in signal generation. A potential solution to this problem is to fractionate the species to the ligand conjugated by means of column affinity chromatography,29,30 which is known as a demanding, low-yield separation process. Another approach is to employ a method of indirect labeling where the enzyme tracer is chemically coupled to a secondary antibody specific to the detection antibody. Such indirect labeling still involves the same problem with the presence of free antibody, but increases the stoichiometric binding ratio of the secondary antibody to the detection antibody, as compared to that of the detection antibody to the analyte in the direct method. To test this idea, we used a simple assay format employing a microtiter plate as solid matrix, i.e., ELISA. The two labeling methods were compared in their analytical performances using the same pair of antibodies to HBsAg (Figure 3). As expected, the assay system adopting indirect labeling showed an enhanced detection capability, by 1 order of magnitude, compared to that (25) Sato, K.; Ichiyama, S.; Iinuma, Y.; Nada, T.; Shimokata, K.; Nakashima, N. J. Clin. Microbiol. 1996, 34, 1420-1422. (26) Shrivastav, T. G. J. Immunoassay Immunochem. 2003, 24, 301-309. (27) Rye, D. B.; Saper, C. B.; Wainer, B. H. J. Histochem. Cytochem. 1984, 32, 1145-1153. (28) Kambegawa, A. Nippon Rinsho 1995, 53, 2160-2167. (29) Donald, R. H.; Mavanur, R. S. J. Immunol. Methods 2001, 249, 33-41. (30) Roger, G. M. Adv. Drug Delivery Rev. 1996, 22, 289-301.

Figure 4. Control of the background color supplementing additives into the enzyme substrate. Components added were an inert protein (e.g., casein) and a detergent that removed residual enzyme derivatives.

of the directly labeled tracer. The degree of such improvement, however, would depend on the stoichiometric increment of reaction between the two binding partners, as well as the yield of the antibody-enzyme conjugation. Optimization for the Catalytic Signal Generation. In the cross-flow chromatographic assay, after a completion of the antigen-antibody bindings by the vertical flow, the enzyme reaction is induced by the addition of the substrate solution in a horizontal flow across the signal generation pad. Such a reaction mode, in contrast to those of ELISA, may vary the kinetic equilibrium between enzyme and substrate. This might make the readjustment of the chemical composition necessary for optimization. Since the horizontal flow also washes unreacted components from the pad, supplementation of the solution with some additives may be effective for comprehensive removal. Additionally, the flow is initiated from one edge of the lateral sides of the pad to the other. Thus, the substrate can be depleted along the path, the severity of which would depend on the direction of the motion, as well as on the density of enzyme present in the signal band. All of these features unique to this novel system have been characterized in order to eventually yield optimal conditions toward its analytical performances. Chemical Composition of the Substrate Solution. Flow in the horizontal direction took two simultaneous roles, i.e., supplying the enzyme substrate for signal generation and washing the unbound remnant causing a noise. Using the manufacturerprepared substrate solutions with increased concentrations of hydrogen peroxide, we first tested whether the reaction equilibrium was shifted under the flow condition. Degraded outcomes Analytical Chemistry, Vol. 77, No. 13, July 1, 2005

4095

Figure 5. Various flow paths applicable for supplying the enzyme substrate solution. The substrate can be flown in vertical (A), diagonal (B), or horizontal (C) directions.

Figure 6. Effect of edge-to-edge substrate flow on the colorimetric signal distribution within the color band. The distribution was measured at varied concentrations of the analyte.

appeared (results not presented), most likely due to the effect of substrate inhibition,31 which suggested that the flow condition had little effect on the catalytic yield of the enzyme. To control the noise, we added an inert protein (e.g., casein) and detergent into the substrate solution, which effectively eliminated background staining (Figure 4). Since the residual enzyme conjugate present even in a trace concentration significantly contributed to the background, it was crucial to use the additives to remove the residual prior to signal generation. Flow Paths for Supplying the Enzyme Substrate. An improper guide of the substrate flow can cause an inhibition of the signal generation due to depletion of the reagent or due to its uneven distribution (Figure 5). For instance, the substrate flow in the vertical direction, consecutive to the first of the immune reactions, led to total inhibition of the signals (Figure 5A). If the flow was directed to follow a diagonal path of the pad, an inconsistent, partial inhibition was created, depending on the position (Figure 5B). To the contrary, the straight horizontal path shown in Figure 2C produced a uniform pattern of multiplex signals without time interruption (Figure 5C). Thus, this mode has been adopted. Hence, the enzyme-based signal generation can extend the utility of the novel immunosensor to the simultaneous detection of multiple analytes. (31) Metelitza, D. I.; Karasyova, E. I.; Grintsevich, E. E.; Thorneley, R. N. J. Inorg. Biochem. 2004, 98, 1-9.

4096

Analytical Chemistry, Vol. 77, No. 13, July 1, 2005

Surface Density of Immune Complexes. In the horizontal flow mode, another type of substrate depletion may arise within the site of the signal band, due to edge-to-edge flow between the lateral sides of the pad. This was caused by the enzyme reaction in the course of the flow, resulting in a nonuniform color along the cross section of the pad. The degree of nonuniformity would be a function of the density of sandwich immune complexes formed, with the capture antibody immobilized. To verify this effect, the color density distribution across the pad was measured after the entire immunoanalyses at varied concentrations of the analyte (Figure 6). While the color across the pad was uniform at a relatively low dose, it profoundly changed as the analyte dose increased. Since such consequences may limit the upper concentration range for an accurate measurement, the use of the gold label in parallel (see below) could be a potential solution for widening the dynamic range. Characterization of Analytical Performances. As described above, we have characterized the aspects unique to the novel immunosensor and optimized them prior to the assessment of its analytical performances. Furthermore, in preparing the enzyme signal generator, the detection antibody was first coupled to colloidal gold and indirectly labeled with the secondary antibodyHRP conjugates, as mentioned. This formulation offered two advantages: tracing of the immune reactions by the naked eye and enhancement of the enzyme signal arising from the indirect

Figure 7. Dose responses of two analytical systems, the conventional system with gold label and the cross-flow employing HRP tracer. Color images of each assay result were captured (A), and the respective signals were then converted to optical densities for plotting the respective dose-response curve (B). Standard deviations of replicative measurements were indicated. Coefficient of variation for replicate measurements did not exceed 13%.

labeling effect. It is noteworthy that, regardless of the presence of the enzyme conjugate bound, the detection antibody conjugated with the gold maintained the identical ability of binding with no detectable changes. This seemed to result from the antigenic nature of immunoglobulin regarding the secondary antibody normally recognizing the constant regions32 of the detection antibody remote from the antigen binding sites. Under optimal conditions, the detection capacity of this novel immunoassay system was assessed by comparing it to that of a conventional test kit employing the gold label. To this end, we obtained dose-responses to HBsAg prepared in standard concentrations with human serum, to closely imitate clinical samples (Figure 7A). If the lower detection limits obtained from the two systems were compared using only the naked eye, the enzyme label would reveal at least 10 times higher enhancement than did the gold tracer. For an accurate comparison, each dose-response curve was prepared by plotting the optical densities of the (32) Christoper, P. P.; David, J. N. Principle and Practice of Immunoassay; Stockton Press: New York, 1997; pp 13-34.

respective color after subtracting them from the backgrounds (Figure 7B). We determined the detection limit as the analyte concentration corresponding to the signal value that was calculated by tripling the standard deviation of the signal at the zero dose.6,20 These were 3.6 and 0.12 ng/mL for the gold and enzyme labels, respectively, which indicated 30 times enhancement in the detection limit using the enzyme label. The use of an enzyme as tracer, as shown, may alleviate drawbacks associated with a low sensitivity of the conventional test kit and, on the other hand, can cause a potential adverse effect. Enzymes are the catalytic protein having, in general, an intrinsic property of inhibition by some substances that may be present variably in different serum samples, which might result in an inconsistent assay result compared to the case of using the gold colloids. Such effect, however, does not seem to arise as a significant analytical problem because a washing step is involved before signal generation from the enzyme label. Indeed, enzymes have long been utilized as a valuable signal generator in immunoassays, e.g., ELISA, even for a variety of clinical examination of diseases. Most immunodiagnostic rapid assays so far have been ideally suited for disease screening in population-based epidemiological studies rather than for diagnosis of individual patients. To extend the utility of the rapid assay, several technical solutions for, particularly, highly reproducible, sensitive, and quantitative analysis have to be investigated. This study has also been performed with objectives in such directions as mentioned. In conclusion, the novel analytical method based on cross-flow chromatography enabled us to produce an amplified signal from an enzyme used as tracer. This enhanced the sensitivity by ∼30-fold compared to that of the gold label. Dynamic range of the analytical system may be limited to a low level for an accurate measurement, which, however, can be overcome through the simultaneous employment of the two signal generators, an enzyme and gold, for dual-color determination. We will continue this research to achieve three specific aims: application of the method to an analyte spectrum requiring higher sensitivities, such as cardiac troponin I as a specific marker of acute myocardial infarction; development of a plastic housing to semiautomatically switch the flow direction for a complete analysis; and investigation of a miniaturized format of immunosensor, applying the same concept. ACKNOWLEDGMENT This study was supported by a grant (01-PJ1-PG4-01PT02-0009) of the Korea Health 21 R&D Project, Ministry of Health & Welfare, Republic of Korea.

Received for review November 23, 2004. Accepted March 3, 2005. AC048270D

Analytical Chemistry, Vol. 77, No. 13, July 1, 2005

4097