Development of a Micro Total Analysis System Incorporating

Integrated Microfluidic Device for Serum Biomarker Quantitation Using Either Standard ... Smart devices ... Rapid screening and evaluation of antioxid...
0 downloads 0 Views 213KB Size
Download PDF
Anal. Chem. 2005, 77, 1684-1688

Development of a Micro Total Analysis System Incorporating Chemiluminescence Detection and Application to Detection of Cancer Markers Kazuhiko Tsukagoshi,* Naoya Jinno, and Riichiro Nakajima

Department of Chemical Engineering and Materials Science, Faculty of Engineering, Doshisha University, Kyotanabe 610-0321, Japan

We developed a micro total analysis system (µ-TAS) incorporating chemiluminescence detection, in which the chemiluminescence reaction of isoluminol isothiocyanato (ILITC) (as a chemiluminescence reagent for labeling)s microperoxidase (as a catalyst)shydrogen peroxide (as an oxidant) was adopted. The analysis system performed the following three processes on a microchip: immune reaction for high selectivity, electrophoresis for formation and transportation of the sample plug, and chemiluminescence detection for high sensitivity. The three processes were compactly integrated onto the microchip to give the µ-TAS. The microchip contained two microchannels that crossed at an intersection, while the ends of the microchannels accessed four reservoirs. As the first process, the immune reaction was performed using an antibody-immobilized glass bead. The glass bead was placed in one of the reservoirs along with antigen (analyte) and a known amount of ILITC-labeled antigen to set up a competitive immune reaction. For electrophoresis, as the second process, the reactant after the immune reaction was fed electrophoretically into the intersection resulting in a sample plug. The sample plug was then moved into another reservoir containing hydrogen peroxide solution. At this point, chemiluminescence detection was performed as the third process: the labeled antigen mixed with the hydrogen peroxide and the catalyst included in the migration buffer to produce chemiluminescence. Chemiluminescence was detected by a photomultiplier tube located under the reservoir. The µ-TAS described here was capable of determining, with high selectivity and sensitivity, human serum albumin or immunosuppressive acidic protein as a cancer marker in human serum. Since the beginning of the last century, a number of chemiluminescence (CL) systems in a variety of species have been characterized. CL detection has many advantages, as demonstrated in FIA,1 HPLC,2 and capillary electrophoresis (CE),3 such as the following: (1) high detection sensitivity, in particular, a * To whom correspondence should be addressed: E-mail: ktsukago@ mail.doshisha.ac.jp. (1) Qin, W. Anal. Lett. 2002, 35, 2207-2220. (2) Ohba, Y.; Kuroda, N.; Nakashima, K. Anal. Chim. Acta 2002, 465, 101109. (3) Liu, Y.-M.; Cheng, J.-K. J. Chromatogr., A 2002, 959, 1-13.

1684 Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

detection limit down to the single-molecule level can be accomplished in CE, which is comparable with that achieved using laser-induced fluorescence; (2) it has a wide linear range of response signal, which is beneficial for analyte quantification; (3) it uses inexpensive reagents and apparatus as well as easy and rapid measurement; and (4) no light source or spectroscopes are required, and thus instrument configuration is very simple. The past decade has witnessed rapid progress in micro total analysis system (µ-TAS) development, referred to as a lab-on-achip. µ-TAS have been widely used to perform chemical and biochemical analyses and are also recognized as powerful tools for genomics, proteomics, and metabolomics.4-6 Microchip-based analytical systems mostly involve capillary electrophoresis (microchip capillary electrophoresis; MCE). The extremely small sample size (usually at the picoliter level) makes achieving high detection sensitivity a challenge. Common detection schemes for MCE include laser-induced fluorescence and mass spectroscopy.7,8 Although the CL detection method is very promising, little attention has been paid to its application in MCE. Mangru and Harrison performed an interesting study on CL detection for MCE to monitor horseradish peroxidase (HRP) and fluorescein-conjugated HRP using the luminol CL system.9 Arora et al. demonstrated a novel wireless electrochemiluminescence detector for MCE involving a U-shaped floating electrode.10 Liu et al. developed a microchip with poly(dimethylsiloxane) and used it for MCE with CL detection.11 We also successfully developed MCE with a CL detector,12-14 based on knowledge obtained through our research using an ordinary CE with a CL detector.15,16 (4) Roddy, E. S.; Xu, H.; Ewing, A. G. Electrophoresis 2004, 25, 229-242. (5) Sheehan, A. D.; Quinn, J.; Daly, S.; Dillon, P.; O’Kennedy, R. Anal. Lett. 2003, 36, 511-537. (6) Greenwood, Paul A.; Greenway, Gillian M. Trends Anal. Chem, 2002, 21, 726-740. (7) Tu, J.; Anderson, L. N.; Dai, J.; Peters, K.; Carr, A.; Loos, P.; Buchanan, D.; Bao, J. J.; Liu, C.; Wehmeyer, K. R. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2003, 789, 323-335. (8) Tachibana, Y.; Otsuka, K.; Terabe, S.; Arai, A.; Suzuki, K.; Nakamura, S. J. Chromatogr., A 2003, 1011, 181-192. (9) Mangru, S. D.; Harrison, D. J. Electrophoresis 1998, 68, 2301-2307. (10) Arora, A.; Eijkel, J. C. T.; Morf, W. E.; Manz, A. Anal. Chem. 2001, 73, 3282-3288. (11) Liu, B.-F.; Ozaki, M.; Utsumi, Y.; Hattori, T.; Terabe, S. Anal. Chem. 2003, 75, 36-41. (12) Hashimoto, M.; Tsukagoshi, K.; Nakajima, R.; Kondo, K.; Arai, A. J. Chromatogr., A 2000, 867, 271-279. (13) Tsukagoshi, K.; Hashimoto, M.; Suzuki, T.; Nakajima, R.; Arai, A. Anal. Sci. 2000, 16, 1111-1112. 10.1021/ac040133t CCC: $30.25

© 2005 American Chemical Society Published on Web 02/11/2005

Figure 1. Illustration of the µ-TAS incorporating an immunoassay and CL detection.

On the other hand, the immunoassay is one of the most important analytical tools and is used in clinical diagnoses and biochemical studies because of its extremely high selectivity. For example, the immunoassay is an indispensable technique for determining small amounts of a given cancer marker in a given serum sample. However, the conventional heterogeneous immunoassay requires a relatively long assay time and involves troublesome liquid-handling procedures and many expensive antibody reagents. A microchip-based system would overcome these drawbacks. Integration of analytical systems into a microchip should bring about enhanced reaction efficiency, simplify the procedures, reduce the assay time, and lower sample, reagent, and energy consumption. Microchip-based immunoassays have been examined previously,17,18 mostly in relation to miniaturization of the CE-based immunoassay. Several recent studies regarding the integration of a heterogeneous immunoassay system into microscale devices have been reported.19,20 Here, we propose a novel concept for a µ-TAS incorporating an immunoassay and CL detection. The concept is illustrated in Figure 1. The analysis system performs the following processes on a microchip: immune reaction for high selectivity, electrophoresis for formation and transportation of the sample plug, and CL detection for high sensitivity, with the three processes compactly integrated onto the microchip. The present µ-TAS was able to determine specific proteins, such as human serum albumin and immunosuppressive acidic protein as a cancer marker in serum, with high sensitivity. EXPERIMENTAL SECTION Reagents. All reagents used were commercially available and of analytical grade. Ion-exchanged water was distilled for use. Human serum albumin (HSA; MW, 66 000) and rabbit anti-human (14) Tsukagoshi, K.; Hashimoto, M.; Suzuki, T.; Nakajima, R.; Arai, A. Anal. Sci. 2001, 17, 1129-1131. (15) Tsukagoshi, K.; Nakamura, T.; Nakajima, R. Anal. Chem. 2002, 74, 41094116. (16) Tsukagoshi, K. Bunseki Kagaku 2003, 52, 1-13 and references therein. (17) Koutny, L. B.; Schmalzing, D.; Taylor, T. A.; Fuchs, M. Anal. Chem. 1996, 68, 18-22. (18) Chiem, N. H.; Harrison, D. J. Electrophoresis 1998, 19, 3040-3044. (19) Arenkov, P.; Kukhtin, A.; Gemmell, A.; Voloshchuk, S.; Chupeeva, V.; Mirzabekov, A. Anal. Biochem. 2000, 278, 123-131. (20) Sato, k.; Tokeshi, M.; Kimura, H.; Kitamori, T. Anal. Chem. 2001, 73, 12131218.

serum IgG (anti-HSA; MW, 150 000) were purchased from Sigma Chemical Co. (St. Louis, MO) and Wako Pure Chemical Industries, Ltd. (Osaka, Japan), respectively. Control human serum, immunosuppressive acidic protein (IAP; MW, 50 000) and goat anti-human serum IAP (anti-IAP) were purchased from Sanko Junyaku Co., Ltd. (Tokyo, Japan). Isoluminol isothiocyanato (ILITC) and microperoxidase were purchased from Nacalai Tesque (Kyoto, Japan) and Tokyo Chemical Industry Co. (Tokyo, Japan), respectively. Hydrogen peroxide solution (30 wt %) was purchased from Wako Pure Chemical Industries, Ltd. Labeling Procedure. Labeling using ILITC was carried out as described previously.15,21 A known amount of protein (micromole order) was added together with ILITC to a microvessel and dissolved in 100 µL of a mixture of water and triethylamine (95: 5). The solution was subjected to ultrasonication for 1 min and then mixed in the dark for 20 min using a vortex mixer. The residue obtained by evaporation from the solution was redissolved in 10 mM phosphate buffer (pH 7.3) to yield ILITC-labeled protein solution and ILITC. For further purification to remove excess ILITC, the ILITClabeled protein solution obtained as described above was subjected to column separation (PD-10 desalting columns, Amersham Pharmacia Biotech, Buckinghamshire, U.K.). ILITC and ILITClabeled protein were separated on a separation column with 10 mM phosphate buffer (pH 7.0). The fractions containing ILITClabeled protein were collected to yield pure ILITC-labeled protein solution containing no free ILITC. Preparation of Antibody-Immobilized Glass Beads.22 The surface of the glass beads was activated by holding 1 g of glass beads (1 mm in diameter) at 100°C for 6 h in 50 wt % sodium hydroxide solution (5 mL). The glass beads were washed and dried well. Alkylamino-bonded glass beads were then prepared by holding them for 6 h under reflux in 2 wt % (3-aminopropyl)triethoxysilane-toluene solution (15 mL). The immobilization of anti-HSA onto alkylamino-bonded glass beads was performed by the glutaraldehyde method as follows. The glass beads were dipped in 5 wt % glutaraldehyde-ethanol solution (2.5 mL), and the mixture was allowed to react for 24 h at room temperature. After washing with 10 mM phosphate buffer (pH 7.3), the beads were dipped in 2 mL of 12 g L-1 sodium cyanoborohydride in ethanol for 5 h. After washing with phosphate buffer, the activated glass beads were dipped in 5 g L-1 antibody phosphate buffer (2 mL) and the mixture was then held at 4 °C for 48 h. The beads were dipped in 12 g L-1 sodium borohydride solution at 4 °C for 5 h. The antibody-immobilized glass beads were then stored at 4 °C in phosphate buffer. Microchip. A schematic layout of a microchip (12.5 mm × 35.0 mm) (type U, Shimadzu Corp., Kyoto, Japan) is shown in Figure 1. The microchips made of quartz consisted of sample load and separation microchannels 20 µm deep and 50 µm wide, as well as holes 2.0 mm in diameter drilled into the chip to facilitate access to the microchannels and to serve as reservoirs (R1-R4 in Figure 1). The channel length from the intersection to R4, which corresponds to the effective separation length, was 31.0 mm. The photomultiplier tube (R7400U, Hamamatsu Co., Hamamatsu, Japan) was located under R4. (21) Hashimoto, M.; Tsukagoshi, K.; Nakajima, K.; Kondo, K. J. Chromatogr., A 1999, 832, 191-202.

Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

1685

tion channel during sample loading was returned easily to R4 during separation. This injection technique made it possible to introduce a known and small amount of sample (of the order of several tens pl). CL Detection. The sample, which migrated toward R4, came into contact with the hydrogen peroxide reagent to produce visible light at the interface between the separation channel and R4. The light was detected by the photomultiplier tube located under R4.

Figure 2. Illustration of immune reaction in R1 on the microchip.

Analytical Procedure and Apparatus. The luminol and hydrogen peroxide CL system was used in conjunction with microperoxidase as a catalyst. Phosphate buffer (10 mM, pH 7.3) containing 4 µM microperoxidase was used as the migration buffer. Phosphate buffer (10 mM, pH 10.8) containing 400 mM hydrogen peroxide was added to the detection cell (R4). After the channels were filled with migration buffer using a disposable syringe, the sample was placed in R1, with migration buffer in R2 and R3, and hydrogen peroxide solution in R4. The µ-TAS developed in the present study comprised the following three processes on a microchip: immune reaction for high selectivity, electrophoresis for formation and transportation of the sample plug, and CL detection for high sensitivity. Electrophoresis on the microchip was performed using a modified microchip electrophoresis system (MCE-2010, Shimadzu Co.). CL detection was performed using a modified CL detector (CLD-10A, Shimadzu Co.). Competitive Immunoassay. The competitive immunoassay is illustrated in Figure 2b. HSA and anti-HSA were used in the immune reaction to illustrate the model shown in Figure 2b. Solutions of ILITC-labeled HSA (5.0 × 10-7 M; 10 µL) and analyte (various concentrations of HSA; 10 µL) were mixed, and an aliquot (2 µL) of the mixture then added to R1 for the competitive immune reaction against immobilized anti-HSA on the glass beads. Immediately, the reactant (sample) in R1 was delivered electrophoretically toward the intersection point to form a sample plug. Electrophoresis. Sample plug formation was achieved simply using a cross-shaped injector as follows. The sample was delivered electrokinetically from R1 toward R2 by applying 0.3 kV for 30 s to R1, with R2 held at ground. During this process, the sample stream was pinched with 0.20 and 0.28 kV applied to R3 and R4, respectively, to prevent the sample spreading out to the separation channel, as the reagent in R4 may leak slightly into the separation channel due to electroosmotic flow. After the intersection was filled completely with the sample, 0.90 kV was applied to R3, with R4 grounded, allowing the sample to inject into and migrate down the separation channel. Simultaneously, 0.70 kV was applied to R1 and R2 during separation to hinder leakage of the sample into the separation channel. Reagent that had leaked into the separa1686

Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

RESULTS AND DISCUSSION Reagents Concentrations and Sample Loading Time. Previously, we reported CE-CL detector systems using luminolmicroperoxidase-hydrogen peroxide reagent.15,21 We examined the effects of microperoxidase and hydrogen peroxide concentrations on CL intensities. In the present study, we selected the concentrations of these reagents described in the Experimental Section with reference to the information obtained in our previous work. We examined the relationship between sample loading time and CL intensity. A 5 × 10-5 M solution of ILITC-labeled HSA was placed in R1 and analyzed using µ-TAS without going through the immune reaction process. The CL intensities due to ILITClabeled HSA increased with increasing loading time up to ∼30 s, subsequently remaining almost constant, indicating that the 30-s loading time filled the intersection with ILITC-labeled HSA sample. Consequently, all of the experiments in the present study were performed using a 30-s sample loading time. Analysis of ILITC-Labeled HSA. A mixture of ILITC and ILITC-labeled HSA, which did not go through the separation column after the labeling reaction, was analyzed using µ-TAS without the immune reaction process. The mixture was applied to R1 alone, not together with the glass bead and antigen. The mixture moved electrophoretically to the intersection to form a sample plug. The sample plug migrated further through the separation channel to R4 (CL detection cell). ILITC and ILITClabeled HSA in the mixture were separated and detected by the CL detector. The resultant electropherogram is shown in Figure 3. ILITC, which has a smaller molecular weight, migrated faster than ILITC-labeled HSA. As shown here, satisfactory baseline separation was not achieved under the present conditions. Thus, ILITC-labeled HSA containing no ILITC, which was obtained through further purification by column separation, was used for the present µ-TAS. First, ILITC-labeled HSA was analyzed in a similar way using a microchip with CL detection without the immune reaction process. ILITC-labeled HSA was determined over the range of 1.0 × 10-7-5.0 × 10-6 M with a detection limit of 1.0 × 10-7 M (S/N ) 3). Preliminary Experiments for Immune Reaction Using the Glass Bead. Adsorption of protein to the surface of untreated glass beads was examined as follows: 5.0 × 10-5 M ILITC-labeled HSA was loaded into R1 together with an untreated glass bead. The experiment was performed according to the analytical procedure described in the Experimental Section. The obtained CL intensity of ILITC-labeled HSA was compared with that obtained without the glass bead. No difference was observed in CL intensity with and without the glass bead. Thus, the adsorption of protein onto the bare surface of the glass bead was negligible in the present study.

Figure 3. Electropherogram of a mixture of ILITC and ILITC-labeled HSA. Conditions: R1, a mixture of ILITC and ILITC-labeled HSA in 10 mM phosphate buffer (pH7.3); R2 and R3, 10 mM phosphate buffer (pH 7.3) containing 4 µM microperoxidase; and R4, 10 mM phosphate buffer (pH 10.8) containing 400 mM hydrogen peroxide. Sample loading time, 30 s; labeling procedure, 5.0 × 10-5 M HSA and 5.0 × 10-4 M ILITC.

Next, we examined the adsorption of ILITC-labeled HSA, bovine serum albumin (BSA), and ovalbumin (Ova) to the antiHSA immobilized on the glass bead. ILITC-labeled HSA, BSA, or Ova (5.0× 10-5 M) was loaded into R1 together with the antiHSA immobilized glass bead. The reaction behavior using HSA in R1 is illustrated in Figure 2a. The experiment was performed as described above to obtain the CL intensity due to ILITC-labeled HSA, BSA, and Ova. Data were compared with those obtained without the glass bead. The CL intensity of ILITC-labeled HSA decreased with the glass bead, while the CL intensities of ILITClabeled BSA and Ova with the glass beads were not significantly different from those without the glass beads. Thus, no nonspecific adsorption of proteins, such as BSA and Ova, to anti-HSA on the glass beads occurred. Competitive Immune Reaction. The results described above supported the suggestion that a competitive immune reaction could be performed to determine antigen concentration using a definite amount of ILITC-labeled antigen and an antibodyimmobilized glass bead. The competitive immune reaction in R1 is illustrated in Figure 2b. We examined the nonbound fraction of the labeled reagent in the competitive binding assay. We felt that detecting the nonbound fraction of labeled reagent in the competitive immunoassay would lead to successful determination of analyte under conditions with no nonspecific adsorption onto the immobilized antibody.22 The present competitive immunoassay enabled us to compactly integrate the three processes (immune reaction, electrophoresis, CL detection) onto the microchip. That is, the present µ-TAS is not an advancement of CL detection but an entirely new approach. The µ-TAS including immune reaction, electrophoresis, and CL detection was tested using HSA and an anti-HSA immobilized glass bead as a model antigen and antibody reaction, as well as ILITC-labeled HSA. (22) Tsukagoshi, K.; Okumura, Y.; Fukaya, R.; Otsuka, M.; Fujiwara, K.; Umehara, H.; Maeda, R.; Nakajima, R. Anal. Sci. 2000, 16, 121-124.

Figure 4. Calibration curve of IAP using the µ-TAS. Conditions: R1, various concentrations of IAP, 2.0 × 10-6 M ILITC-labeled IAP, and anti-IAP immobilized glass bead in 10 mM phosphate buffer (pH 7.3); R2 and R3, 10 mM phosphate buffer (pH 7.3) containing 4 µM microperoxidase; and R4, 10 mM phosphate buffer (pH 10.8) containing 400 mM hydrogen peroxide. Sample loading time, 30 s.

HSA (analyte) was determined over the range of 1.0 × 10-75.0 × 10-6 M with a detection limit of 1.0 × 10-7 M (S/N ) 3) (correlation coefficient, 0.999) through immune reaction, electrophoresis, and CL detection on a microchip. One analysis was performed within 2 min. To determine the amount of HSA in a serum sample by the present method, a control serum solution, diluted 1:2000 by volume, was added to R1 together with 5.0 × 10-7 M ILITC-labeled HSA. The serum sample was analyzed using the µ-TAS including immunoassay, electrophoresis, and CL detection. The amount of HSA in the serum was calculated using the CL intensity of labeled HSA and calibration curve of HSA. The value of HSA obtained by the present method, 3.9 g dL-1 (the average value of 6 analyses), corresponded closely to that reported by the manufacturer (3.7-4.2 g dL-1). Thus, the present method is applicable to the determination of protein in serum samples. Analysis of IAP in Serum. A number of cancer markers have been identified to date. IAP is one of the most useful of these cancer markers,23-26 and the concentration of IAP in the blood has been shown to be increased in cancer patients. We obtained a calibration curve of IAP (analyte) using the µ-TAS by use of an anti-IAP-immobilized glass bead and ILITC-labeled IAP (2.0 × 10-6 M). IAP was analyzed by immune reaction, electrophoresis, and CL detection on the microchip. As shown in Figure 4, IAP was determined over the range of 1.0 × 10-7-5.0 × 10-6 M with a detection limit of 1.0 × 10-7 M (S/N ) 3) (correlation coefficient, 0.999). The assay showed good linearity but a narrow dynamic range. Each plot in Figure 4 is the average of 7-10 measurements. One analysis was performed within 2 min. To determine the amount of IAP in a serum sample by the present method, a control serum solution, diluted 1:50 by volume, was added to R1 together with 2.0 × 10-6 M ILITC-labeled IAP. The serum in R1 was analyzed using the µ-TAS. The amount of IAP in the serum obtained by the present method, 272 mg dL-1 (average value of 6 analyses), also corresponded closely to that (23) Tamura, K.; Shibata, Y.; Matsuda, Y.; Ishida, N. Cancer Res. 1981, 41, 32443252. (24) Kikuchi, S. Nippon Geka Gakkai Zasshi 1984, 85, 283-299. (25) Sawada, M.; Matsui, Y.; Okudaira, Y. Gan to Kagaku Ryoho 1983, 10, 12511258. (26) Ishida, N.; Tamura, K.; Shibata, Y. Igaku no Ayumi 1980, 115, 423-433.

Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

1687

reported by the manufacturer (266 mg dL-1). Thus, the present method was applicable to the determination of protein in a serum sample. The incubation time and temperature for the immune reaction are issues to be resolved for application of this method to lower concentrations of analytes and will be examined in detail in future studies. CONCLUSIONS A novel concept involving a µ-TAS incorporating an immunoassay and CL detection was developed. The analysis system performed three processes on a microchip: an immune reaction for high selectivity, electrophoresis for formation and transportation of the sample plug, and CL detection for high sensitivity. The three processes were compactly integrated onto the microchip. The present µ-TAS gave calibration curves for HSA and IAP in the 1.0 × 10-7-5.0 × 10-6 M range with a detection limit of 1.0

1688

Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

× 10-7 M (S/N ) 3). HSA and IAP (a cancer marker) in a human serum were determined by the system with high selectivity and sensitivity. One analysis was performed within 2 min. ACKNOWLEDGMENT This work was supported by a Grant-in-Aid for Scientific Research (C) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan. The authors also acknowledge the financial support for this research by Doshisha University’s Research Promotion Fund. We thank Mr. Akihiro Arai for his help with the microchip.

Received for review July 15, 2004. Accepted December 7, 2004. AC040133T