- Email: [email protected]
Electrochemical detection of HbA1c, a maker for diabetes, using a flow immunoassay system
Biosensors and Bioelectronics 22 (2007) 2051–2056
Electrochemical detection of HbA1c, a maker for diabetes, using a flow immunoassay system Tsuyoshi Tanaka a , Shoko Tsukube a , Kojiro Izawa a , Mina Okochi a , Tae-Kyu Lim b , Shugo Watanabe b , Manabu Harada b , Tadashi Matsunaga a,∗ a
Department of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-16 Naka-Cho, Koganei, Tokyo 184-8588, Japan b Malcom Co., LTD., 4-15-10 Hon-machi, sibuya, Tokyo 151-0071, Japan Received 24 April 2006; received in revised form 5 September 2006; accepted 6 September 2006 Available online 6 October 2006
Abstract An on-chip electrochemical flow immunoassay system for the detection of hemoglobin A1c (HbA1c ) was developed using anti-human hemoglobin (Hb) IgG labeled with ferrocene monocarboxylic acid (Fc–COOH) and boronate-affinity chromatography. An on-chip column packed with boronateactivated agarose beads was used for the separation of HbA1c from both non-glycated Hb and free antibody. Anti-human Hb IgG conjugated to Fc–COOH (Fc–IgG) was used for the electrochemical detection of HbA1c . The assay procedure included immunoreactions with Fc–IgG and HbA1c , separation of immunocomplexes by boronate affinity, and electrochemical detection of Fc–IgG–HbA1c immunocomplexes. The immunoreaction mixtures were injected onto a boronate-affinity column. HbA1c –antibody complexes were then trapped onto the column by the affinity of HbA1c to boronic acid. Subsequently, elution buffer containing sorbitol was applied to elute HbA1c –antibody complexes and a current was detected by applying 600 mV versus Ag/AgCl. The elution signal was an estimation of the HbA1c amount. A linear correlation between the increase of current and HbA1c concentration was obtained up to an HbA1c concentration of 500 g/ml. The HbA1c flow immunoassay was successfully achieved using hemolysates. This electrochemical flow immunoassay system enabled us to construct a novel point-of-care testing device for the monitoring of glycated proteins including HbA1c . © 2006 Elsevier B.V. All rights reserved. Keywords: On-chip electrochemical flow immunoassay; Hemoglobin A1c (HbA1c ); Ferrocene monocarboxylic acid (Fc–COOH); Boronate-affinity chromatography; Point-of-care testing device
1. Introduction Hemoglobin A1c (HbA1c ), which is irreversibly glycated on the N-terminal valine of the -chain, is well known as the main diabetes marker protein for monitoring long-term glycemic control clinically. A number of analytical techniques have been used to measure HbA1c . In the clinic, boronate-affinity chromatography (Li et al., 2002) and cation-exchange chromatography (Lafferty et al., 2002) have been widely used. In addition, the immunoturbidimetric method (Metus et al., 1999), electrophoresis (Jenkins and Ratnaike, 2003), and mass spectrometry (Jeppsson et al., 2002) have been employed clinically. Boronate-affinity chromatography is helpful in the separation
∗
Corresponding author. Tel.: +81 42 388 7020; fax: +81 42 385 7713. E-mail address: [email protected] (T. Matsunaga).
0956-5663/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2006.09.008
and purification of HbA1c from hemoglobin based on the interaction between glucose and boronic acid (Koyama and Terauchi, 1996; Tanaka and Matsunaga, 2001). Boronic acid is known to form a covalent bond between its diol group and the cis-diol group of the sugar under alkaline conditions (Rohovec et al., 2003). Ferrocene derivatives have often been used as electrochemical signaling probes in miniaturized flow immunoassay systems (Lim et al., 2002; Lim et al., 2003; Padeste et al., 2000; Wang et al., 2002). In the present study, our research group developed a miniaturized electrochemical flow immunoassay system (Lim and Matsunaga, 2001) using a cation-exchange capillary column and ferrocene-conjugated antibodies for the detection of the pregnancy marker, human chorionic gonadotropin (hCG). Antibody–antigen complexes were separated from free ferrocene-conjugated IgG antibody (Fc–IgG) on the basis of their differences in isoelectric point (pI) using an ion-exchange
2052
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
resin. Ferrocene-conjugated anti-hCG antibody was used as an electrochemical probe. Furthermore, a network of channels coated with cation-exchange resin for the separation of allergen markers has been constructed (Lim et al., 2003). The flow immunoassay system enables the generation of highly reproducible results using only minute quantities of whole blood samples within a few minutes. The integration of chromatography and electrochemical detection with the lab-on-a-chip technology enables the measurement of HbA1c with good reproducibility. In this study, an electrochemical flow immunoassay system for the detection of HbA1c was developed using ferroceneconjugated antibody and boronate-affinity chromatography. This system consisted of an on-chip column packed with boronate-activated agarose beads and flow-cell electrodes. The on-chip column packed with boronate-activated agarose beads was used for the separation of HbA1c from both non-glycated Hb and free antibody. 2. Materials and methods 2.1. Materials HbA1c was purchased from Exocell Co. (Philadelphia, PA, USA). Goat-anti human hemoglobin IgG was purchased from Nippon Biotest Laboratory Co. (Tokyo, Japan). Ferrocene monocarboxylic acid (Fc–COOH) and m-aminophenyl boronic acid-modified agarose beads were obtained from Sigma Chemical Co. (St. Louis, MO, USA). N-hydroxysuccinimide (NHS) and 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride (EDC) were purchased from Pierce (Rockford, IL, USA). PhastGel IEF 3–9 and molecular markers were purchased from GE Healthcare UK Ltd. (mersham Place, Little Chalfont, Buckinghamshire HP7 9NA, England). Blood calibrator samples containing various concentrations of HbA1c were purchased from Healthcare Technology Foundation (Tokyo, Japan). Deionized distilled water was used in all procedures. 2.2. Preparation of anti-human Hb IgG labeled with Fc–COOH Fc–COOH labeling to IgG was carried out according to the following procedure. Fc–COOH (23 mg) was dissolved in 1 ml dimethyl sulfoxide (DMSO) (solution A). EDC (0.14 mmol) and NHS (0.14 mmol) were dissolved in 1 ml DMSO (solution B). Solution A was dropped into solution B with stirring for 2 h at 25 ◦ C. The mixture was added to anti-human hemoglobin IgG solution (3 mg/ml, 300 l) in 0.13 M NaHCO3 buffer and stirred at 4 ◦ C overnight. Free Fc–COOH was removed by ultrafiltration using a Centricon YM-30 column (Millipore). The concentration of IgG was determined by the BCA protein assay method (Pierce). The average number of ferrocene moieties bound to anti-human hemoglobin IgG was determined by the concentration of iron using atomic absorption spectroscopy (model AA-6600G, Shimadzu, Kyoto, Japan). Cyclic voltammetry of IgG labeled with Fc–COOH (Fc–IgG) was performed using ALS/CH Instruments, Electrochemical Analyzer model 832A (BAS Inc. Tokyo, Japan).
2.3. Analysis of the binding affinity of Fc–IgG by ELISA The binding affinity of goat IgG labeled with Fc–COOH against both HbA1c and non-glycated Hb was investigated by ELISA in the following procedure. A polystyrene 96-well microtiter plate (Corning Inc., NY, USA) was coated with either 10 g/ml human HbA1c or Hb solution (100 l/well) and incubated for 1 h at room temperature. The plate was washed three times with 100 mM phosphate buffer containing 0.9% NaCl and 0.05% Tween 20 (PBST; pH 7.4). Then, 200 l of 100 mM phosphate buffer containing 0.9% NaCl (PBS; pH 7.4) and 0.1% (w/v) bovine serum albumin (BSA) was added to each well and incubated at room temperature for 1 h to block the active sites for non-specific adsorption. After the washing step, 100 l of 23 g/ml goat anti-human Hb IgG labeled with Fc–COOH was added and incubated for 1 h. The plate was rinsed again and 100 l alkaline phosphatase (ALP)labeled anti-goat IgG antibody (1000-fold dilution in PBS; secondary antibody) was added and incubated for 1 h. After each binding reaction, the plate was washed three times with PBST. Lumiphos 530, the substrate of ALP, was then added to the each well and luminescence was measured using a luminescent reader (Lucy2, Anthos Labtec Instruments, Salzburg, Austria). 2.4. Electrophoresis of HbA1c and anti-hemoglobin IgG complexes Each fraction collected from the boronate-affinity column was subjected to polyacrylamide gel electrophoresis with PhastGel IEF 3-9. Lenti lectin (basic, pI 8.65), lenti lectin (middle, pI 8.45), lenti lectin (acidic, pI 8.15), horse myogloblin (basic, pI 7.35), horse myogloblin (acidic, pI 6.85), human carbonic anhydrase B (pI 6.55), bovine carbonic anhydrase B (pI 5.85), -lactoglobulin A (pI 5.20), soybean trypsin inhibitor (pI 4.55), and amyloglucosidase (pI 3.50) were used as pI markers. Electrophoresis, isoelectric focusing, and silver staining of the gels (GE Healthcare UK Ltd., Phastsystem) were performed according to the manufacturer’s instructions. 2.5. Electrochemical flow immunoassay system The layout of the electrochemical flow immunoassay system is shown in Fig. 1. The on-chip column (height, 1.5 mm; width, 3 mm) was constructed using polydimethylsiloxane (PDMS). The negative molds were fabricated on a polymethylmethacrylate (PMMA) plate using an End-Mill (Toki Co., Tokyo, Japan). The PDMS elastomer was poured onto the negative molds. Flat PDMS was also prepared as a lid. The microchamber was constructed using a Machining Star 25 (Toki Co.) with a computeraided design (CAD) system. Layout files in a CAD graphics format were electronically transferred and compiled to function as motion control files for the computer-aided modeling machine (CAMM) (PNC-300, Roland DG Co., Tokyo). Both parts of PDMS were treated with oxygen plasma (18 W, Plasma Cleaner, Harrick Scientific Corporation, USA) and placed into contact
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
2053
Fig. 1. Schematic diagram of an on-chip electrochemical flow immunoassay system. The eluted species pass through the flow cell and the current associated with the HbA1c –Fc–IgG antibody complex is monitored. The applied potential is 600 mV vs. Ag/AgCl.
with each other. The column was packed with boronate-activated agarose beads. A three-electrode flow-cell system, equipped with a glassy carbon electrode, a platinum counter electrode, and an Ag/AgCl reference electrode (BAS Inc.), was used in this study (Lim et al., 2003). A constant potential was applied with a potentiostat (model LC-4C; amperometric detector, BAS Inc.). Results were recorded on a chart recorder (SP-J5C, Riken Denshi, Japan). Eluent flow was produced using a microsyringe pump (Model 260; Muromachi Kikai Co., Japan). The electrochemical oxidation current of the HbA1c –Fc–IgG complex in the eluate was measured using a three-electrode flow cell. 2.6. Separation of HbA1c –anti hemoglobin IgG complexes using boronate-affinity chromatography Separation of HbA1c –anti hemoglobin IgG complexes using on-chip boronate-affinity chromatography was carried out according to the following procedure. HbA1c solution (0–2000 g/ml, 12.5 l in PBS) was mixed with 12.5 l (3200 g) anti-human hemoglobin IgG labeled with Fc–COOH. The mixture was incubated for 1 h at 37 ◦ C and diluted 5-fold with loading buffer (50 mM taurine, 10 mM MgCl2 , and 50 mM NaCl, pH 8.6). The immunocomplexes were injected onto an on-chip column at a flow rate of 100 l/min. HbA1c –IgG complexes were trapped on the boronate-affinity column by covalent bonds between the cis-diol groups of the boronic acid and the cis-diol groups of the hemoglobin-bound glucose (HbA1c ). Subsequently, the elution buffer (the loading buffer containing 50 mM sorbitol) was used to elute the immunocomplexes. The amount of immunocomplex in each fraction was evaluated by measuring the absorbance at 415 nm for hemoglobin and at 280 nm for total proteins. Furthermore, the detection of HbA1c from hemolysates (Tominaga, 2001) diluted 20-fold with PBS were examined in the same manner.
3. Results and discussion 3.1. Characterization of IgG labeled with Fc–COOH The conjugation of Fc onto IgG was performed using NHS, which was used to modify a carboxyl group to an amine-reactive ester. The NHS ester-activated Fc was reacted with the primary amines of IgG. The hydration process was accelerated in DMSO compared with aqueous solutions. This fact suggests that the carboxyl group of Fc was successfully activated and chemically conjugated to the amine group of IgG. A maximum number of ferrocene moieties was obtained for labeling when Fc–COOH was reacted with IgG at a molar ratio of 1:10 (IgG:Fc–COOH) in the reaction mixture. The average number of ferrocene moieties bound to a single IgG molecule was estimated to be 2.6 ± 0.5 per IgG. This number was similar to that found in a previously published report (Okochi et al., 2005). The electrochemical properties of Fc–IgG in PBS were investigated by cyclic voltammetry using a glassy carbon electrode. The oxidation potential of free Fc–COOH appeared at 395 mV versus Ag/AgCl (Fig. 2A). No significant signals were observed in control anti-human hemoglobin IgG solution (Fig. 2C). On the other hand, the oxidation potential of Fc–IgG was observed at 600 mV versus Ag/AgCl (Fig. 2B). The redox potential of ferrocene derivatives depends on the functional groups present in the side chain (Padeste et al., 2000). The shift in the oxidation potential is due to the chemical conjugation of IgG to the carboxylic acid of Fc–COOH. The applied potential was set at 600 mV for the detection of immunocomplexes between HbA1c and Fc–IgG. The binding capacity of Fc–IgGs against human HbA1c and Hb was evaluated by ELISA. No differences in the binding capacities of Fc–IgG against HbA1c and Hb were confirmed. Furthermore, Fc–IgG retained 93.3% of binding capacity compared with that of native IgG. These results suggest that Fc–IgG can be used for both immunoreaction and electrochemical detection.
2054
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
Fig. 2. Cyclic voltammograms of Fc–IgG dissolved in PBS (pH 7.4) at a scan rate of 20 mV/s using a glassy carbon electrode. (A) IgG (8.7 mg/ml); (B) Fc–IgG (8.7 mg/ml); (C) Fc–COOH (10 mg/ml).
3.2. Boronate-affinity separation of HbA1c –IgG immunocomplexes The binding and elution of HbA1c –IgG immunocomplexes was evaluated using boronate-affinity column chromatography by measuring the absorbance of hemoglobin at 415 and 280 nm, respectively. A peak was observed in the eluted fractions at 415 nm (Fig. 3a). The peak increased linearly with increasing
Fig. 4. Isoelectric focusing of eluted fractions when each concentration of HbA1c was used: lane 1, HbA1c (200 g/ml); lane 2, a mixture of HbA1c (400 g/ml); anti-human Hb IgG (800 g/ml); lane 3, anti-human Hb IgG (400 g/ml) and lanes 4–7, each eluted fraction (HbA1c : 2000, 1500, 1000, and 200 g/ml).
HbA1c concentration in the 0–1000 g/ml range. Two peaks were detected in the washing and eluted fractions at 280 nm. The peak in the washing fractions corresponds to IgG since a single peak in the washing fractions was obtained in the absence of HbA1c (Fig. 3b, open square). The IgG peak decreased with increasing HbA1c concentration, while the peak in the eluted fractions increased (Fig. 3b). Therefore, the peak in the eluted fractions corresponds to the immunocomplexes of HbA1c and IgG. The identification of HbA1c –IgG immunocomplexes was carried out by isoelectric gel electrophoresis of each eluted fraction (Fig. 4). The isoelectric points were 6.85 for HbA1c (lane 1) and 6.50 for anti-human hemoglobin IgG (lane 2). The isoelectric point of the immunocomplex was similar to that of free IgG (lane 3). Each fraction eluted from the boronate-affinity column was also subjected to isoelectric gel electrophoresis when various concentrations of HbA1c were used for immunoreaction (molar ratio of HbA1c to IgG; 4:3, 4:4, 4:6, and 4:30) (lanes 4–7). The bands of HbA1c (upper band) and the immunocomplex (lower band) were detected in lane 4 when the molar ratio of HbA1c to IgG was 4:3. The upper bands disappeared at molar ratios of 4:4 to 4:30 (lanes 5–7). The density of the immunocomplex bands increased with increasing HbA1c concentration. These results suggest that HbA1c concentration is quantitatively measured by monitoring the amount of the immunocomplexes. The currents of the immunocomplexes for each molar ratio of HbA1c (10–2000 g/ml) to Fc–IgG (400–6400 g/ml IgG) were examined. Electrochemical signals from the immunocomplexes were obtained at 3:2 and 3:4 molar ratios. Based on these results, 3200 g/ml of anti-human hemoglobin IgG labeled with Fc–COOH was used in subsequent experiments. 3.3. Calibration of HbA1c by a electrochemical flow immunoassay system
Fig. 3. Elution pattern of HbA1c –IgG complexes in boronate-affinity chromatography by measuring absorbance at 415 nm (a) and 280 nm (b). The various concentrations of HbA1c were reacted with anti-human Hb IgG.
The currents in the eluted fractions were detected by a flow immunoassay system when various concentrations of HbA1c (0–2000 g/ml) were reacted with 3200 g/ml Fc–IgG (Fig. 5a). The signal from the HbA1c –Fc–IgG complex increased
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
2055
Table 1 Relationship between HbA1c concentration in hemolysates and currents hemolysate
HbA1c concentration (g/ml)
Current (nA)
HbA1c -low HbA1c -medium HbA1c -high
270 530 834
4.0 ± 1.3 8.5 ± 1.1 12.0 ± 2.1
in 6600 g/ml total Hb). HbA1c -low is considered a healthy hemolysate and HbA1c -medium and -high are hemolysates from diabetic patients. The concentrations of total hemoglobin and HbA1c were determined by HPLC. The relationship between HbA1c concentration and current was obtained using these hemolysates. The currents increased with increasing HbA1c concentration in hemolysates independent of total Hb concentration (Table 1). The errors are shown as standard deviations from three experiments on individual chips. The current values from hemolysates decreased compared with those from a pure HbA1c sample (Fig. 5) when HbA1c concentrations in both samples were the same. The decrease in current value was approximately 45%. This phenomenon may be explained as the difference in total Hb concentration per sample. 4. Conclusions
Fig. 5. Electrochemical detection of HbA1c under various concentrations: (a) the peaks after elution by sorbitol were presented, the concentration of Fc–IgG was 3200 g/ml. HbA1c concentrations were 0 g/ml (A), 200 g/ml (B), 400 g/ml (C), 1000 g/ml (D), and 2000 g/ml (E) and (b) relationship between currents and HbA1c concentration (0–2000 g/ml).
with increasing HbA1c concentration and reached a plateau at more than 1000 g/ml (Fig. 5b). The relationship between current and HbA1c was linear up to an HbA1c concentration of 500 g/ml. A required range for clinical examination of HbA1c is 200–980 g/ml, which is estimated from the HbA1c concentration in 20-fold diluted hemolysates. Therefore, the obtained detection range by proposed methods can be used clinically. On the other hand, samples containing higher HbA1c concentrations require further dilution. 3.4. Electrochemical detection of HbA1c in hemolysates In practical application, HbA1c should be expressed as a percentage of total hemoglobin. Therefore, the detection of HbA1c in the presence of total Hb was examined. Hemolysates without interference from other oxidizable substances (which are negligible in the electrochemical detection of ferrocene) were used as blood samples. Hemolysates (20-fold dilution) contained each concentration of HbA1c (HbA1c -low: 270 g/ml HbA1c in 6750 g/ml total Hb; HbA1c -medium: 530 g/ml HbA1c in 7250 g/ml total Hb; HbA1c -high: 834 g/ml HbA1c
An on-chip electrochemical flow immunoassay system for the detection of HbA1c was constructed. In this assay format, non-glycated hemoglobin can also be measured by the electrochemical detection of Hb and anti-Hb IgG complexes. The HbA1c –anti-human Hb IgG complexes were separated from Hb and free antibodies using boronate-affinity column chromatography and electrochemically detected using ferrocene-conjugated antibody as a probe. A linear correlation between elution currents and HbA1c concentration was obtained at HbA1c concentrations up to 500 g/ml, fulfilling the required range for clinical examination. These results indicate that the electrochemical flow immunoassay system enables the construction of a novel point-of-care testing device for the monitoring of HbA1c in hemolysates. In future work, point-of-care testing devices will be constructed for the simultaneous detection of total Hb and HbA1c . Acknowledgement This work was funded by a support program for technology development on the basis of academic findings from The New Energy and Industrial Technology Development Organization. References Jenkins, M., Ratnaike, S., 2003. Clin. Chem. Lab. Med. 41 (6), 747–754. Jeppsson, J.O., Kobold, U., Barr, J., Finke, A., Hoelzel, W., Hoshino, T., Miedema, K., Mosca, A., Mauri, P., Paroni, R., Thienpont, L., Umemoto, M., Weykamp, C., 2002. Clin. Chem. Lab. Med. 40 (1), 78–89. Koyama, T., Terauchi, K., 1996. J. Chromatogr. B Biomed. Appl. 679 (1–2), 31–40. Lafferty, J.D., McFarlane, A.G., Chui, D.H.K., 2002. Arch. Pathol. Lab. Med. 126 (12), 1494–1500.
2056
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
Li, Y., Jeppsson, J.O., Jornten-Karlsson, M., Linne Larsson, E., Jungvid, H., Galaev, I.Y., Mattiasson, B., 2002. J Chromatogr. B Analyt. Technol. Biomed. Life Sci. 776 (2), 149–160. Lim, T.K., Imai, S., Matsunaga, T., 2002. Biotechnol. Bioeng. 77 (7), 758– 763. Lim, T.K., Matsunaga, T., 2001. Biosens. Bioelectron. 16 (9–12), 1063–1069. Lim, T.K., Ohta, H., Matsunaga, T., 2003. Anal. Chem. 75 (14), 3316–3321. Metus, P., Ruzzante, N., Bonvicini, P., Meneghetti, M., Zaninotto, M., Plebani, M., 1999. J. Clin. Lab. Anal. 13 (1), 5–8.
Okochi, M., Ohta, H., Tanaka, T., Matsunaga, T., 2005. Biotechnol. Bioeng. 90 (1), 14–19. Padeste, C., Grubelnik, A., Tiefenauer, L., 2000. Biosens. Bioelectron. 15 (9–10), 431–438. Rohovec, J., Maschmeyer, T., Aime, S., Peters, J.A., 2003. Chemistry 9 (10), 2193–2199. Tanaka, T., Matsunaga, T., 2001. Biosens. Bioelectron. 16 (9–12), 1089–1094. Tominaga, M., 2001. Rinsho Byori 49 (12), 1199–1204. Wang, J., Ibanez, A., Chatrathi, M.P., 2002. Electrophoresis 23 (21), 3744–3749.
Electrochemical detection of HbA1c, a maker for diabetes, using a flow immunoassay system Tsuyoshi Tanaka a , Shoko Tsukube a , Kojiro Izawa a , Mina Okochi a , Tae-Kyu Lim b , Shugo Watanabe b , Manabu Harada b , Tadashi Matsunaga a,∗ a
Department of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-16 Naka-Cho, Koganei, Tokyo 184-8588, Japan b Malcom Co., LTD., 4-15-10 Hon-machi, sibuya, Tokyo 151-0071, Japan Received 24 April 2006; received in revised form 5 September 2006; accepted 6 September 2006 Available online 6 October 2006
Abstract An on-chip electrochemical flow immunoassay system for the detection of hemoglobin A1c (HbA1c ) was developed using anti-human hemoglobin (Hb) IgG labeled with ferrocene monocarboxylic acid (Fc–COOH) and boronate-affinity chromatography. An on-chip column packed with boronateactivated agarose beads was used for the separation of HbA1c from both non-glycated Hb and free antibody. Anti-human Hb IgG conjugated to Fc–COOH (Fc–IgG) was used for the electrochemical detection of HbA1c . The assay procedure included immunoreactions with Fc–IgG and HbA1c , separation of immunocomplexes by boronate affinity, and electrochemical detection of Fc–IgG–HbA1c immunocomplexes. The immunoreaction mixtures were injected onto a boronate-affinity column. HbA1c –antibody complexes were then trapped onto the column by the affinity of HbA1c to boronic acid. Subsequently, elution buffer containing sorbitol was applied to elute HbA1c –antibody complexes and a current was detected by applying 600 mV versus Ag/AgCl. The elution signal was an estimation of the HbA1c amount. A linear correlation between the increase of current and HbA1c concentration was obtained up to an HbA1c concentration of 500 g/ml. The HbA1c flow immunoassay was successfully achieved using hemolysates. This electrochemical flow immunoassay system enabled us to construct a novel point-of-care testing device for the monitoring of glycated proteins including HbA1c . © 2006 Elsevier B.V. All rights reserved. Keywords: On-chip electrochemical flow immunoassay; Hemoglobin A1c (HbA1c ); Ferrocene monocarboxylic acid (Fc–COOH); Boronate-affinity chromatography; Point-of-care testing device
1. Introduction Hemoglobin A1c (HbA1c ), which is irreversibly glycated on the N-terminal valine of the -chain, is well known as the main diabetes marker protein for monitoring long-term glycemic control clinically. A number of analytical techniques have been used to measure HbA1c . In the clinic, boronate-affinity chromatography (Li et al., 2002) and cation-exchange chromatography (Lafferty et al., 2002) have been widely used. In addition, the immunoturbidimetric method (Metus et al., 1999), electrophoresis (Jenkins and Ratnaike, 2003), and mass spectrometry (Jeppsson et al., 2002) have been employed clinically. Boronate-affinity chromatography is helpful in the separation
∗
Corresponding author. Tel.: +81 42 388 7020; fax: +81 42 385 7713. E-mail address: [email protected] (T. Matsunaga).
0956-5663/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2006.09.008
and purification of HbA1c from hemoglobin based on the interaction between glucose and boronic acid (Koyama and Terauchi, 1996; Tanaka and Matsunaga, 2001). Boronic acid is known to form a covalent bond between its diol group and the cis-diol group of the sugar under alkaline conditions (Rohovec et al., 2003). Ferrocene derivatives have often been used as electrochemical signaling probes in miniaturized flow immunoassay systems (Lim et al., 2002; Lim et al., 2003; Padeste et al., 2000; Wang et al., 2002). In the present study, our research group developed a miniaturized electrochemical flow immunoassay system (Lim and Matsunaga, 2001) using a cation-exchange capillary column and ferrocene-conjugated antibodies for the detection of the pregnancy marker, human chorionic gonadotropin (hCG). Antibody–antigen complexes were separated from free ferrocene-conjugated IgG antibody (Fc–IgG) on the basis of their differences in isoelectric point (pI) using an ion-exchange
2052
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
resin. Ferrocene-conjugated anti-hCG antibody was used as an electrochemical probe. Furthermore, a network of channels coated with cation-exchange resin for the separation of allergen markers has been constructed (Lim et al., 2003). The flow immunoassay system enables the generation of highly reproducible results using only minute quantities of whole blood samples within a few minutes. The integration of chromatography and electrochemical detection with the lab-on-a-chip technology enables the measurement of HbA1c with good reproducibility. In this study, an electrochemical flow immunoassay system for the detection of HbA1c was developed using ferroceneconjugated antibody and boronate-affinity chromatography. This system consisted of an on-chip column packed with boronate-activated agarose beads and flow-cell electrodes. The on-chip column packed with boronate-activated agarose beads was used for the separation of HbA1c from both non-glycated Hb and free antibody. 2. Materials and methods 2.1. Materials HbA1c was purchased from Exocell Co. (Philadelphia, PA, USA). Goat-anti human hemoglobin IgG was purchased from Nippon Biotest Laboratory Co. (Tokyo, Japan). Ferrocene monocarboxylic acid (Fc–COOH) and m-aminophenyl boronic acid-modified agarose beads were obtained from Sigma Chemical Co. (St. Louis, MO, USA). N-hydroxysuccinimide (NHS) and 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride (EDC) were purchased from Pierce (Rockford, IL, USA). PhastGel IEF 3–9 and molecular markers were purchased from GE Healthcare UK Ltd. (mersham Place, Little Chalfont, Buckinghamshire HP7 9NA, England). Blood calibrator samples containing various concentrations of HbA1c were purchased from Healthcare Technology Foundation (Tokyo, Japan). Deionized distilled water was used in all procedures. 2.2. Preparation of anti-human Hb IgG labeled with Fc–COOH Fc–COOH labeling to IgG was carried out according to the following procedure. Fc–COOH (23 mg) was dissolved in 1 ml dimethyl sulfoxide (DMSO) (solution A). EDC (0.14 mmol) and NHS (0.14 mmol) were dissolved in 1 ml DMSO (solution B). Solution A was dropped into solution B with stirring for 2 h at 25 ◦ C. The mixture was added to anti-human hemoglobin IgG solution (3 mg/ml, 300 l) in 0.13 M NaHCO3 buffer and stirred at 4 ◦ C overnight. Free Fc–COOH was removed by ultrafiltration using a Centricon YM-30 column (Millipore). The concentration of IgG was determined by the BCA protein assay method (Pierce). The average number of ferrocene moieties bound to anti-human hemoglobin IgG was determined by the concentration of iron using atomic absorption spectroscopy (model AA-6600G, Shimadzu, Kyoto, Japan). Cyclic voltammetry of IgG labeled with Fc–COOH (Fc–IgG) was performed using ALS/CH Instruments, Electrochemical Analyzer model 832A (BAS Inc. Tokyo, Japan).
2.3. Analysis of the binding affinity of Fc–IgG by ELISA The binding affinity of goat IgG labeled with Fc–COOH against both HbA1c and non-glycated Hb was investigated by ELISA in the following procedure. A polystyrene 96-well microtiter plate (Corning Inc., NY, USA) was coated with either 10 g/ml human HbA1c or Hb solution (100 l/well) and incubated for 1 h at room temperature. The plate was washed three times with 100 mM phosphate buffer containing 0.9% NaCl and 0.05% Tween 20 (PBST; pH 7.4). Then, 200 l of 100 mM phosphate buffer containing 0.9% NaCl (PBS; pH 7.4) and 0.1% (w/v) bovine serum albumin (BSA) was added to each well and incubated at room temperature for 1 h to block the active sites for non-specific adsorption. After the washing step, 100 l of 23 g/ml goat anti-human Hb IgG labeled with Fc–COOH was added and incubated for 1 h. The plate was rinsed again and 100 l alkaline phosphatase (ALP)labeled anti-goat IgG antibody (1000-fold dilution in PBS; secondary antibody) was added and incubated for 1 h. After each binding reaction, the plate was washed three times with PBST. Lumiphos 530, the substrate of ALP, was then added to the each well and luminescence was measured using a luminescent reader (Lucy2, Anthos Labtec Instruments, Salzburg, Austria). 2.4. Electrophoresis of HbA1c and anti-hemoglobin IgG complexes Each fraction collected from the boronate-affinity column was subjected to polyacrylamide gel electrophoresis with PhastGel IEF 3-9. Lenti lectin (basic, pI 8.65), lenti lectin (middle, pI 8.45), lenti lectin (acidic, pI 8.15), horse myogloblin (basic, pI 7.35), horse myogloblin (acidic, pI 6.85), human carbonic anhydrase B (pI 6.55), bovine carbonic anhydrase B (pI 5.85), -lactoglobulin A (pI 5.20), soybean trypsin inhibitor (pI 4.55), and amyloglucosidase (pI 3.50) were used as pI markers. Electrophoresis, isoelectric focusing, and silver staining of the gels (GE Healthcare UK Ltd., Phastsystem) were performed according to the manufacturer’s instructions. 2.5. Electrochemical flow immunoassay system The layout of the electrochemical flow immunoassay system is shown in Fig. 1. The on-chip column (height, 1.5 mm; width, 3 mm) was constructed using polydimethylsiloxane (PDMS). The negative molds were fabricated on a polymethylmethacrylate (PMMA) plate using an End-Mill (Toki Co., Tokyo, Japan). The PDMS elastomer was poured onto the negative molds. Flat PDMS was also prepared as a lid. The microchamber was constructed using a Machining Star 25 (Toki Co.) with a computeraided design (CAD) system. Layout files in a CAD graphics format were electronically transferred and compiled to function as motion control files for the computer-aided modeling machine (CAMM) (PNC-300, Roland DG Co., Tokyo). Both parts of PDMS were treated with oxygen plasma (18 W, Plasma Cleaner, Harrick Scientific Corporation, USA) and placed into contact
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
2053
Fig. 1. Schematic diagram of an on-chip electrochemical flow immunoassay system. The eluted species pass through the flow cell and the current associated with the HbA1c –Fc–IgG antibody complex is monitored. The applied potential is 600 mV vs. Ag/AgCl.
with each other. The column was packed with boronate-activated agarose beads. A three-electrode flow-cell system, equipped with a glassy carbon electrode, a platinum counter electrode, and an Ag/AgCl reference electrode (BAS Inc.), was used in this study (Lim et al., 2003). A constant potential was applied with a potentiostat (model LC-4C; amperometric detector, BAS Inc.). Results were recorded on a chart recorder (SP-J5C, Riken Denshi, Japan). Eluent flow was produced using a microsyringe pump (Model 260; Muromachi Kikai Co., Japan). The electrochemical oxidation current of the HbA1c –Fc–IgG complex in the eluate was measured using a three-electrode flow cell. 2.6. Separation of HbA1c –anti hemoglobin IgG complexes using boronate-affinity chromatography Separation of HbA1c –anti hemoglobin IgG complexes using on-chip boronate-affinity chromatography was carried out according to the following procedure. HbA1c solution (0–2000 g/ml, 12.5 l in PBS) was mixed with 12.5 l (3200 g) anti-human hemoglobin IgG labeled with Fc–COOH. The mixture was incubated for 1 h at 37 ◦ C and diluted 5-fold with loading buffer (50 mM taurine, 10 mM MgCl2 , and 50 mM NaCl, pH 8.6). The immunocomplexes were injected onto an on-chip column at a flow rate of 100 l/min. HbA1c –IgG complexes were trapped on the boronate-affinity column by covalent bonds between the cis-diol groups of the boronic acid and the cis-diol groups of the hemoglobin-bound glucose (HbA1c ). Subsequently, the elution buffer (the loading buffer containing 50 mM sorbitol) was used to elute the immunocomplexes. The amount of immunocomplex in each fraction was evaluated by measuring the absorbance at 415 nm for hemoglobin and at 280 nm for total proteins. Furthermore, the detection of HbA1c from hemolysates (Tominaga, 2001) diluted 20-fold with PBS were examined in the same manner.
3. Results and discussion 3.1. Characterization of IgG labeled with Fc–COOH The conjugation of Fc onto IgG was performed using NHS, which was used to modify a carboxyl group to an amine-reactive ester. The NHS ester-activated Fc was reacted with the primary amines of IgG. The hydration process was accelerated in DMSO compared with aqueous solutions. This fact suggests that the carboxyl group of Fc was successfully activated and chemically conjugated to the amine group of IgG. A maximum number of ferrocene moieties was obtained for labeling when Fc–COOH was reacted with IgG at a molar ratio of 1:10 (IgG:Fc–COOH) in the reaction mixture. The average number of ferrocene moieties bound to a single IgG molecule was estimated to be 2.6 ± 0.5 per IgG. This number was similar to that found in a previously published report (Okochi et al., 2005). The electrochemical properties of Fc–IgG in PBS were investigated by cyclic voltammetry using a glassy carbon electrode. The oxidation potential of free Fc–COOH appeared at 395 mV versus Ag/AgCl (Fig. 2A). No significant signals were observed in control anti-human hemoglobin IgG solution (Fig. 2C). On the other hand, the oxidation potential of Fc–IgG was observed at 600 mV versus Ag/AgCl (Fig. 2B). The redox potential of ferrocene derivatives depends on the functional groups present in the side chain (Padeste et al., 2000). The shift in the oxidation potential is due to the chemical conjugation of IgG to the carboxylic acid of Fc–COOH. The applied potential was set at 600 mV for the detection of immunocomplexes between HbA1c and Fc–IgG. The binding capacity of Fc–IgGs against human HbA1c and Hb was evaluated by ELISA. No differences in the binding capacities of Fc–IgG against HbA1c and Hb were confirmed. Furthermore, Fc–IgG retained 93.3% of binding capacity compared with that of native IgG. These results suggest that Fc–IgG can be used for both immunoreaction and electrochemical detection.
2054
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
Fig. 2. Cyclic voltammograms of Fc–IgG dissolved in PBS (pH 7.4) at a scan rate of 20 mV/s using a glassy carbon electrode. (A) IgG (8.7 mg/ml); (B) Fc–IgG (8.7 mg/ml); (C) Fc–COOH (10 mg/ml).
3.2. Boronate-affinity separation of HbA1c –IgG immunocomplexes The binding and elution of HbA1c –IgG immunocomplexes was evaluated using boronate-affinity column chromatography by measuring the absorbance of hemoglobin at 415 and 280 nm, respectively. A peak was observed in the eluted fractions at 415 nm (Fig. 3a). The peak increased linearly with increasing
Fig. 4. Isoelectric focusing of eluted fractions when each concentration of HbA1c was used: lane 1, HbA1c (200 g/ml); lane 2, a mixture of HbA1c (400 g/ml); anti-human Hb IgG (800 g/ml); lane 3, anti-human Hb IgG (400 g/ml) and lanes 4–7, each eluted fraction (HbA1c : 2000, 1500, 1000, and 200 g/ml).
HbA1c concentration in the 0–1000 g/ml range. Two peaks were detected in the washing and eluted fractions at 280 nm. The peak in the washing fractions corresponds to IgG since a single peak in the washing fractions was obtained in the absence of HbA1c (Fig. 3b, open square). The IgG peak decreased with increasing HbA1c concentration, while the peak in the eluted fractions increased (Fig. 3b). Therefore, the peak in the eluted fractions corresponds to the immunocomplexes of HbA1c and IgG. The identification of HbA1c –IgG immunocomplexes was carried out by isoelectric gel electrophoresis of each eluted fraction (Fig. 4). The isoelectric points were 6.85 for HbA1c (lane 1) and 6.50 for anti-human hemoglobin IgG (lane 2). The isoelectric point of the immunocomplex was similar to that of free IgG (lane 3). Each fraction eluted from the boronate-affinity column was also subjected to isoelectric gel electrophoresis when various concentrations of HbA1c were used for immunoreaction (molar ratio of HbA1c to IgG; 4:3, 4:4, 4:6, and 4:30) (lanes 4–7). The bands of HbA1c (upper band) and the immunocomplex (lower band) were detected in lane 4 when the molar ratio of HbA1c to IgG was 4:3. The upper bands disappeared at molar ratios of 4:4 to 4:30 (lanes 5–7). The density of the immunocomplex bands increased with increasing HbA1c concentration. These results suggest that HbA1c concentration is quantitatively measured by monitoring the amount of the immunocomplexes. The currents of the immunocomplexes for each molar ratio of HbA1c (10–2000 g/ml) to Fc–IgG (400–6400 g/ml IgG) were examined. Electrochemical signals from the immunocomplexes were obtained at 3:2 and 3:4 molar ratios. Based on these results, 3200 g/ml of anti-human hemoglobin IgG labeled with Fc–COOH was used in subsequent experiments. 3.3. Calibration of HbA1c by a electrochemical flow immunoassay system
Fig. 3. Elution pattern of HbA1c –IgG complexes in boronate-affinity chromatography by measuring absorbance at 415 nm (a) and 280 nm (b). The various concentrations of HbA1c were reacted with anti-human Hb IgG.
The currents in the eluted fractions were detected by a flow immunoassay system when various concentrations of HbA1c (0–2000 g/ml) were reacted with 3200 g/ml Fc–IgG (Fig. 5a). The signal from the HbA1c –Fc–IgG complex increased
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
2055
Table 1 Relationship between HbA1c concentration in hemolysates and currents hemolysate
HbA1c concentration (g/ml)
Current (nA)
HbA1c -low HbA1c -medium HbA1c -high
270 530 834
4.0 ± 1.3 8.5 ± 1.1 12.0 ± 2.1
in 6600 g/ml total Hb). HbA1c -low is considered a healthy hemolysate and HbA1c -medium and -high are hemolysates from diabetic patients. The concentrations of total hemoglobin and HbA1c were determined by HPLC. The relationship between HbA1c concentration and current was obtained using these hemolysates. The currents increased with increasing HbA1c concentration in hemolysates independent of total Hb concentration (Table 1). The errors are shown as standard deviations from three experiments on individual chips. The current values from hemolysates decreased compared with those from a pure HbA1c sample (Fig. 5) when HbA1c concentrations in both samples were the same. The decrease in current value was approximately 45%. This phenomenon may be explained as the difference in total Hb concentration per sample. 4. Conclusions
Fig. 5. Electrochemical detection of HbA1c under various concentrations: (a) the peaks after elution by sorbitol were presented, the concentration of Fc–IgG was 3200 g/ml. HbA1c concentrations were 0 g/ml (A), 200 g/ml (B), 400 g/ml (C), 1000 g/ml (D), and 2000 g/ml (E) and (b) relationship between currents and HbA1c concentration (0–2000 g/ml).
with increasing HbA1c concentration and reached a plateau at more than 1000 g/ml (Fig. 5b). The relationship between current and HbA1c was linear up to an HbA1c concentration of 500 g/ml. A required range for clinical examination of HbA1c is 200–980 g/ml, which is estimated from the HbA1c concentration in 20-fold diluted hemolysates. Therefore, the obtained detection range by proposed methods can be used clinically. On the other hand, samples containing higher HbA1c concentrations require further dilution. 3.4. Electrochemical detection of HbA1c in hemolysates In practical application, HbA1c should be expressed as a percentage of total hemoglobin. Therefore, the detection of HbA1c in the presence of total Hb was examined. Hemolysates without interference from other oxidizable substances (which are negligible in the electrochemical detection of ferrocene) were used as blood samples. Hemolysates (20-fold dilution) contained each concentration of HbA1c (HbA1c -low: 270 g/ml HbA1c in 6750 g/ml total Hb; HbA1c -medium: 530 g/ml HbA1c in 7250 g/ml total Hb; HbA1c -high: 834 g/ml HbA1c
An on-chip electrochemical flow immunoassay system for the detection of HbA1c was constructed. In this assay format, non-glycated hemoglobin can also be measured by the electrochemical detection of Hb and anti-Hb IgG complexes. The HbA1c –anti-human Hb IgG complexes were separated from Hb and free antibodies using boronate-affinity column chromatography and electrochemically detected using ferrocene-conjugated antibody as a probe. A linear correlation between elution currents and HbA1c concentration was obtained at HbA1c concentrations up to 500 g/ml, fulfilling the required range for clinical examination. These results indicate that the electrochemical flow immunoassay system enables the construction of a novel point-of-care testing device for the monitoring of HbA1c in hemolysates. In future work, point-of-care testing devices will be constructed for the simultaneous detection of total Hb and HbA1c . Acknowledgement This work was funded by a support program for technology development on the basis of academic findings from The New Energy and Industrial Technology Development Organization. References Jenkins, M., Ratnaike, S., 2003. Clin. Chem. Lab. Med. 41 (6), 747–754. Jeppsson, J.O., Kobold, U., Barr, J., Finke, A., Hoelzel, W., Hoshino, T., Miedema, K., Mosca, A., Mauri, P., Paroni, R., Thienpont, L., Umemoto, M., Weykamp, C., 2002. Clin. Chem. Lab. Med. 40 (1), 78–89. Koyama, T., Terauchi, K., 1996. J. Chromatogr. B Biomed. Appl. 679 (1–2), 31–40. Lafferty, J.D., McFarlane, A.G., Chui, D.H.K., 2002. Arch. Pathol. Lab. Med. 126 (12), 1494–1500.
2056
T. Tanaka et al. / Biosensors and Bioelectronics 22 (2007) 2051–2056
Li, Y., Jeppsson, J.O., Jornten-Karlsson, M., Linne Larsson, E., Jungvid, H., Galaev, I.Y., Mattiasson, B., 2002. J Chromatogr. B Analyt. Technol. Biomed. Life Sci. 776 (2), 149–160. Lim, T.K., Imai, S., Matsunaga, T., 2002. Biotechnol. Bioeng. 77 (7), 758– 763. Lim, T.K., Matsunaga, T., 2001. Biosens. Bioelectron. 16 (9–12), 1063–1069. Lim, T.K., Ohta, H., Matsunaga, T., 2003. Anal. Chem. 75 (14), 3316–3321. Metus, P., Ruzzante, N., Bonvicini, P., Meneghetti, M., Zaninotto, M., Plebani, M., 1999. J. Clin. Lab. Anal. 13 (1), 5–8.
Okochi, M., Ohta, H., Tanaka, T., Matsunaga, T., 2005. Biotechnol. Bioeng. 90 (1), 14–19. Padeste, C., Grubelnik, A., Tiefenauer, L., 2000. Biosens. Bioelectron. 15 (9–10), 431–438. Rohovec, J., Maschmeyer, T., Aime, S., Peters, J.A., 2003. Chemistry 9 (10), 2193–2199. Tanaka, T., Matsunaga, T., 2001. Biosens. Bioelectron. 16 (9–12), 1089–1094. Tominaga, M., 2001. Rinsho Byori 49 (12), 1199–1204. Wang, J., Ibanez, A., Chatrathi, M.P., 2002. Electrophoresis 23 (21), 3744–3749.