Simple Detection of Surface Antigens on Living Cells by Applying

Sep 16, 2012 - Masashi Watanabe , Ryou Hatano , Chiho Koizumi. Polymer ... Tomoyuki Yasukawa , Yusuke Yoshida , Hironobu Hatanaka , Fumio Mizutani ,...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/ac

Simple Detection of Surface Antigens on Living Cells by Applying Distinct Cell Positioning with Negative Dielectrophoresis Tomoyuki Yasukawa,*,†,‡ Hironobu Hatanaka,† and Fumio Mizutani*,† †

Graduate School of Material Science, University of Hyogo, 3-2-1 Kouto, Kamigori, Ako, Hyogo 678-1297, Japan JST-CREST, 5, Sanbancho, Chiyoda-ku, Tokyo 102-0075, Japan

‡

S Supporting Information *

ABSTRACT: We report the fabrication of two different cell patterns based on negative dielectrophoresis (n-DEP) and apply it to simple and rapid distinction of cells with specific surface antigens from a cell population. The DEP device for cell manipulation comprised a microfluidic channel with an upper indium tin oxide (ITO) electrode and a lower ITOinterdigitated band array (ITO−IDA) electrode modified with an antibody. Cells immediately accumulated on the surface in the gap area between both bands of the ITO−IDA electrode by n-DEP upon AC voltage between the upper ITO and both lower bands. Switching of the applied band electrode voltage resulted in the removal of accumulated cells to form another pattern because of the formation of a different electric field pattern in the device. Modifying the ITO−IDA surface with the antibody inhibited the removal of the cells with a specific surface antigen for irreversible capture by immunoreactions during the first accumulation. In this study, we targeted the CD33 surface antigen expressed on human promyelocytic leukemia cells (HL-60). The time required for the assay was substantially short: 60 s for forcing and 60 s for separating the unbound cells. Furthermore, the present method does not require pretreatment such as target labeling or washing of unbound cells. Moreover, the use of the swing technique considerably improved cell binding to the antibody-modified surface for cells with a specific surface antigen. The distinct integration of cells with n-DEP in the high conductivity medium provided higher cell binding efficiency compared to that obtained in our previous study (Hatanaka, H.; Yasukawa, T.; Mizutani, F. Anal. Chem., 2011, 83, 7207−7212) without loss of rapidity and simplicity.

T

were removed in the gap region between the band electrodes after the force direction was switched to negative DEP (n-DEP), which is a force that repels particles from high electric field regions. This immunoreaction at the surface was assisted by the rapid accumulation of p-DEP and was completed within 30 s. The time required for separating unbound cells by n-DEP, estimated by rapid discrimination systems, was approximately 30 s. The total time required for detecting antigens expressed on the cell surface was as short as 60 s. Moreover, no pretreatment such as cell staining with fluorescent molecules is required for this method. We demonstrated rapid and simple detection of a surface antigen on cells using both p- and n-DEP. Nevertheless, CD33 was expressed in almost all HL-60 cells, and cell binding efficiency was approximately half of that in a previous study.19 The lower binding efficiency was principally due to the use of a cell suspension medium with relatively low conductivity. The number of cells captured on the surface modified with antibodies in the low medium conductivity region (200 mS/m). However, no living cell experienced the attractive force of p-DEP in the higher conductivity region within the used frequency region (10 kHz−10 MHz). In this study, we employed only n-DEP to fabricate two different cell patterns by switching the strength of the applied voltage in a high conductivity solution and improved cell binding efficiency to distinguish cells with a target antigen. Figure 1

■

THEORY Particles in a suspended solution are polarized to an alternating electric field. DEP refers to the migration of a particle resulting from the interaction between induced polarization and the spatially inhomogeneous electric field.20−22 The quantitative description of the time-averaged DEP force, ⟨F̅DEP⟩ [N], is given by: 3 2 ⟨FDEP ̅ ⟩ = 2πεma Re[K̲ (ω)]∇Erms

(1)

where a is the particle radius [m], εm is the permittivity of the suspension medium [F m−1], Erms is the root-meansquare electric field [V m−1], and ∇ is the del vector operator. In living cells, the frequency dependence of the induced DEP force is given by the real part of the Clausius−Mossotti factor, Re[K(εeff, εs)], ε̲ − ε̲ s ̲ ε̲ eff , ε̲ s) = eff K( ε̲ eff + 2 ε̲ s (2) Figure 1. Cross-sectional view of the dielectrophoresis (DEP) device and the principle of the method for discriminating cells with specific surface antigens using negative-DEP (n-DEP) manipulation.

where εeff is the effective permittivity of the cell and εs is the permittivity of the suspension medium. The underlined parameters denote complex quantities. The complex permittivity of the suspension medium is given by: σ ε̲ s = εs − s j (3) ω

schematically depicts a cross-sectional view of the DEP device and the principle of the method for detecting the surface antigen. The cell suspension was introduced into the device, which comprised an upper indium tin oxide (ITO) electrode and a lower ITO-interdigitated band array (ITO−IDA) electrode, along with two combs of bands A and B. The surface of the ITO−IDA electrode substrate was modified with the anti-CD 33 antibody. When AC voltage in the n-DEP frequency region is applied between the upper ITO and lower ITO−IDA electrodes (both bands A and B) to form nonuniform electric fields, the cells accumulate at the gap regions between the IDA bands to form the line pattern, thereby accelerating the capture of cells with the target antigen via an immunoreaction. The dashed lines between the upper and lower electrodes display the electric flux line. After the applied voltage for band B was switched off, electric fields disappeared between the upper ITO and band B and appeared between bands A and B, thereby leading to the formation of relatively weak electric field areas above band B. Thus, unbound cells were removed from the gap region to the area above band B to form another line pattern. The ratio of the captured cell density (cell binding efficiency) could be calculated easily from K̲ ( ε̲ eff , ε̲ s) = −

where εs is the permittivity of the medium, σs is the conductivity of the medium [S m−1], ω is angular frequency [=2πf, where f is the applied frequency (Hz)], and j = (−1)1/2. We adopted the single-shell model to calculate the effective permittivity of the cell.23,24 In this model, we assumed that the cell is represented as a spherical particle filled with a linear ohmic dielectric fluid (cell cytoplasm) and is enclosed by a thin insulating shell (cell membrane). The effective permittivity of the cell is given by: ε̲ eff =

C̲ mr ε̲ c C̲ mr + ε̲ c

ε̲ c = εc −

(4)

σc j ω

(5) −2

where Cm is the membrane capacitance [F m ], εc is permittivity of cytoplasm, and σc is conductivity of cytoplasm. By substituting eqs 3 and 4 into eq 2, we obtain the expression for K(εeff, εs):

1 + ω 2(τmsτc − τmcτs − τsτc) + jω(τmc − τms + τc + τs) 2 − ω 2(τmsτc + 2τmcτs + τsτc) + jω(2τmc + τms + 2τc + 2τs)

The time constants are defined as τms = Cmr/σs, τmc = Cmr/σc, τs = εs/σs, and τc = εc/σc. Figure 2 shows the theoretical DEP spectra calculated from the different medium conductivities

(6)

based on the single-shell model. We employed nominal values for other parameters:24 Cm = 0.015 F m−2, εc/ε0 = 60, εs/ε0 = 80, and σc = 0.5 S m−1. According to microscopic measurements, the 8831

dx.doi.org/10.1021/ac302239k | Anal. Chem. 2012, 84, 8830−8836

Analytical Chemistry

Article

antibody for 16 h at 4 °C. After washing with phosphate-buffered saline (PBS, pH 7.4), the electrode was immersed in the carbonate buffer containing 10 mg mL−1 bovine serum albumin (Wako, Osaka, Japan) for 24 h at 4 °C to eliminate nonspecific cell binding. Finally, the electrode, modified with the antibody, was washed and stored in PBS at 4 °C before use. Modeling and Calculations. The distribution of the electric field strength for the patterning device was calculated by the finite element method solver (COMSOL Multiphysics, Stockholm, Sweden). The ITO electrode was mounted on the ITO−IDA electrode. The width of the bands and the distance from the adjacent bands were set to 12 and 50 μm, respectively. The distance between the ITO and ITO−IDA electrodes was set to 35 μm. For the first cell pattern, the potentials for both bands A and B and the ITO electrode were set to 10 and 0 V, respectively. The potential of band B was switched from 10 to 0 V for the second cell pattern. Patterning with Cells by n-DEP. HL-60 cells (CD33positive cells) were grown in RPMI 1640 medium (Invitrogen Japan KK, Tokyo, Japan), containing 10% fetal bovine serum (Invitrogen), 50 units mL−1 penicillin (Invitrogen), and 50 μg mL−1 streptomycin (Invitrogen) at 37 °C under 5% CO2. HL-60 cells (4 × 107 cells mL−1) were suspended in DEP medium consisting of 250 mM sucrose and 250 mM HEPES buffer (pH 7.4), and conductivity was adjusted to 400 mS m−1 with a phosphate buffer (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4, and 1.5 mM KH2PO4) for cell manipulation by DEP. AC voltage in the n-DEP frequency region (typically 15 Vpp and 100 kHz) was applied between the ITO electrodes and both band electrodes to accumulate cells in the gap region between the bands. Then, the voltage applied to band B was switched off to remove cells from the gap region to areas above band B. Air bubbles were sometimes formed at electrodes by the water electrolysis by applying AC voltage with the frequency below 10 kHz. However, no bubble was formed over 10 kHz within the intensity of applied voltage used in the study (10−25 Vpp). Cell manipulation by DEP was performed under an optical and fluorescent microscope (IX70, Olympus, Tokyo, Japan) equipped with a charged-coupled device camera (DP72, Olympus). A cell suspension (1.0−1.5 μL) was introduced into the patterning device. AC voltage was applied between the ITO and ITO−IDA electrodes to form the alternating electric field in the device (function generator 7075, Hioki E.E. Co., Ueda, Japan). Cell Binding Based on Cell Manipulation by n-DEP. HL-60 cells were manipulated by n-DEP in the patterning device consisting of the ITO−IDA electrode modified with anti-CD33 or antimouse IgG. We prepared nonspecific cells treated with antiCD33 and used them as CD33-negative cells. Furthermore, the antibody-treated nonspecific cells were labeled with fluorescent molecules by successive incubation in 10 μg mL−1 anti-CD33 and 1 μmol dm−3 carboxyfluorescein diacetate succinimidyl ester (Invitrogen) in PBS for 15 min at 37 °C under 5% CO2. Suspensions of specific and nonspecific cells were mixed at different ratios to study the relationship between the number of CD33-positive cells in the suspension and the number of captured cells. The cell mixture accumulated in the gap region because of n-DEP in 60 s. Cell binding efficiency was calculated at 60 s after the AC voltage of band B was switched off. Efficiency was defined as the ratio of the average cell density on the band electrode immediately before and 60 s after the AC voltage of band B was switched off. We calculated each cell number as an average of three data points.

Figure 2. Modeling of the Clausius−Mossotti factor spectra, representing cells with respect to the surrounding medium with the different conductivities. Conductivity of the DEP medium: (a) 2 × 10−4, (b) 1 × 10−3, (c) 1 × 10−2, (d) 1 × 10−1, (e) 2 × 10−1, (f) 4 × 10−1, (g) 5 × 10−1, and (h) 1.0 S m−1.

average radius of cells (r) was 6.7 × 10−6 m. The direction and magnitude of the DEP force acting on the cells depends on the Re[K] value. Cells undergoing electric fields with a positive number Re[K] migrate toward the highest electric field region because of p-DEP, whereas cells undergoing electric fields with a negative number migrate in the opposite direction because of nDEP. Medium conductivity predominately affects the midfrequency region. Cells experience p-DEP in relative low medium conductivities. However, the increase in medium conductivity results in a decrease in the Re[K] value and a shift in the crossover frequency to a high frequency region. When the value of the medium conductivity reaches that of the cytoplasm conductivity (0.5 S m−1), the DEP force of 1−100 MHz becomes zero. Therefore, n-DEP must be used to form cell patterns in a high conductivity medium.

■

EXPERIMENTAL SECTION Fabrication of the Device for Cell Manipulation. A dielectrophoretic patterning device was constructed with the ITO and ITO−IDA electrodes, which were fabricated by conventional photolithography and chemical etching using an etchant solution (ITO-02, Kanto Chemical, Tokyo, Japan) for 10−15 min under ultrasonication. Each band element was 2.0 mm long and 12 μm wide and placed at a distance of 50 μm from the adjacent bands. A 5-μm thick insulating layer of negative photoresist (SU-8 3005, MicroChem Corp., Newton, MO, USA) was fabricated on the ITO−IDA electrode to define the electrode areas (750 × 750 μm) exposed to the solution. A 30-μm thick polyester film (Nitto Denko, Osaka, Japan) was used as a spacer and sandwiched between the upper ITO and lower ITO−IDA to construct the patterning device. Modification of the ITO−IDA Electrode with Anti-CD33 Antibody. We modified the ITO−IDA electrode with the mouse anti-CD33 monoclonal antibody (anti-CD33, AbD Serotec, Kidlington, UK) to identify cells with the CD33 cell surface antigen. A goat antimouse IgG polyclonal antibody (antimouse IgG, AbD Serotec) was used as a negative control. The ITO−IDA electrode was immersed in anhydrous ethanol containing 2 v/v% 3-(glycidoxypropyl)trimethoxysilane (Sigma-Aldrich Chemical, St. Louis, MO, USA) for 1 h at 25 °C to introduce a terminal epoxy group that readily reacts with the amino groups of biomolecules. After washing with ethanol, the ITO−IDA electrode was treated with a carbonate buffer (pH 9.6) containing 100 μg mL−1 of the 8832

dx.doi.org/10.1021/ac302239k | Anal. Chem. 2012, 84, 8830−8836

Analytical Chemistry

Article

Improving Cell Binding Efficiency by the Swing of Captured Cells. We investigated the effect of swinging the captured cells to improve cell binding efficiency. During cell accumulation in the gap region by n-DEP, we slightly reduced the intensity of the applied voltage to band B at intervals of 5 s to swing the captured cells. Finally, the band B voltage was switched off 60 s after cells accumulated in the gap region to remove unbound cells. The number of captured cells was calculated using microscopic images.

after applying AC voltage (20 Vpp) to the lower part of band A and band B against the upper ITO used as a ground. The cells dispersed randomly in the channel and rapidly accumulated in the gap area between the bands within 5 s because of n-DEP (Figure 1 middle and 3A). After the intensity of the voltage applied to band B was switched to 0 Vpp, most of the cells moved toward the area above band B, resulting in the formation of another line pattern within 10 s (Figure 1 bottom and 3B). Moreover, the line pattern was reproduced as the voltage was switched. These results clearly indicate that different cell patterns could be easily created by applying different intensity voltages. Figure 3C,D shows the cross-sectional area of the electric field formed in the patterning device calculated from digital simulation in which voltages were applied to both bands A and B and band A, respectively. Regions with a low electric field were found in the bottom of the gap between the bands (areas labeled as N1 in Figure 3C). Thus, the suspended cells moved to area N1 when AC voltage was applied to both bands A and B. However, upon applying zero voltage to band B, regions with a low electric field were found in the middle above band B (areas labeled as N2 in Figure 3D). These results indicated that applying AC voltage in the n-DEP frequency region to both bands A and B or band A to move to areas N1 or N2 was in good agreement with the experimental results (Figure 3A,B). Capture of HL60 Cells with CD33 Antigen Based on Manipulation by n-DEP. We investigated cell binding efficiency in the gap area between bands using devices modified with different antibodies. Figure 4 shows the optical images of cells that accumulated in the area between bands A and B after applying an AC voltage of 20 Vpp (Figure 4A) and the captured cells after separating unbound cells by switching the band B intensity to zero (Figure 4B,C). Again, the uniformly dispersed cells initially started to move toward the gap area between bands to form clear line patterns due to the repulsive n-DEP force (Figure 4A). The number of cells accumulating in the area between the bands increased with the duration of voltage application and was saturated within 5 s. Surface modification by the antibody had almost no effect on cell accumulation behavior. Next, we removed the cells accumulated in area N1 by switching the voltage of band B to zero 60 s after cells had accumulated. When the cells accumulated in area N1 modified with anti-CD33, some cells moved to the area above band B, whereas others remained even after the band B voltage was switched to zero (Figure 4B). In contrast, almost all cells moved to the area above band B after voltage was switched to zero, and cells accumulated in the gap modified with nonspecific antimouse

■

RESULTS AND DISCUSSION Cell Manipulation by n-DEP. Dielectrophoretic patterning with HL-60 cells was studied using a patterning device consisting of the ITO and ITO−IDA electrodes without immobilizing the antibody. We introduced a suspension (conductivity, 400 mS m−1) of HL-60 cells in the patterning device and subsequently applied AC voltage of 15 Vpp at a frequency of 100 kHz to induce cell accumulation in the gap area between the ITO−IDA bands using n-DEP. In addition, we already investigated the operating direction of the dielectrophoretic force on HL-60 cells at various frequencies and derived the n-DEP repulsive force at 50−100 kHz. We also investigated that cells were sufficiently captured on the microwells in solutions with conductivities over 200 mS m−1 by the cell binding assay using a microtiter plate modified with anti-CD33.19 Figure 3A shows the optical microscopic images 5 s

Figure 3. Optical images of cells patterned by n-DEP. Cells accumulated (A) in the gap area between the bands because of n-DEP and (B) on band B because of n-DEP after switching the band B voltage to zero. (C) and (D) Cross-sectional view of the numerically calculated electric field formed in the patterning device calculated from a digital simulation. Upper and lower gray bars show the ITO and band electrodes, respectively. Applied voltages for calculating bands A and B and the upper ITO were (C) 10, 10, and 0 V and (D) at 10, 0, and 0 V, respectively.

Figure 4. Optical images of cells patterned by n-DEP. (A) Cells accumulated in the gap area between the bands modified with the anti-CD 33 antibody by applying AC voltage of same intensity (20 Vpp) and frequency (100 kHz) as bands A and B. (B) Cell pattern captured in the gap area modified with the anti-CD 33 antibody after separating unbound cells by switching off the band B AC voltage. (C) Cells removed from the gap area modified with the antimouse IgG antibody after the band B voltage was turned off. Conductivity of the medium: 400 mS/m; duration of voltage application for accumulation and removal with n-DEP: 60 and 60 s. 8833

dx.doi.org/10.1021/ac302239k | Anal. Chem. 2012, 84, 8830−8836

Analytical Chemistry

Article

intensity of 20 Vpp was the most suitable to capture cells via immunoreaction. Specific Cell Binding in a Mixture Containing Nonspecific Cells. Mixtures of HL-60 cells specific and nonspecific to anti-CD33 were used to determine the number of cells captured in the gap area modified with anti-CD33. Figure 6A

IgG (Figure 4C). These results suggest that the cells with CD33 surface antigens reacted with the antibodies to be irreversibly captured at that position. Therefore, removal of cells was markedly inhibited by the anti-CD33 immunoreaction, which corresponded to the CD33 surface antigen expressed on HL-60 cells. The number of cells accumulating in the N1 area decreased rapidly within 20 s after the voltage was switched on and then reached a steady state. Cell binding efficiency was estimated from the steady-state value and was 68.3 ± 3.2%. Slight undesired binding, originating from nonspecific adsorption, was observed on the bands modified with antimouse IgG (4.2 ± 1.4%). The binding efficiency and undesired binding obtained in the present study improved compared to the case in a previous study using a combination of p- and n-DEP in which binding efficiency and undesired binding were 46.9 ± 3.4% and 8.7 ± 1.5%, respectively. The improved binding efficiency and undesired binding may have been due to the high efficiency of the immunoreaction and suppression of nonspecific adsorption in the high conductivity (400 mS/m) buffer solution. Times as short as 60 and 30 s were required to capture the cells with immunoreactions and to remove unbound cells, respectively. Therefore, cells with CD33 cell surface antigens could be rapidly identified from the cell suspension by spatial separation based on an immunoreaction and manipulation by n-DEP. The effects of applied voltage intensity were studied to optimize the cell binding assay with manipulation by n-DEP. We applied different intensities of AC voltage in a range from 5 to 25 Vpp to accumulate cells for 60 s. The capture of cells proceeded in the gap between the band electrodes during this time. The applied voltages for bands A and B were set to 20 Vpp and zero, respectively, to remove unbound cells, and cell binding efficiency was measured. Figure 5 shows the cell binding

Figure 6. Cell binding in the gap area modified with the anti-CD33 antibody from mixed suspensions of specific and nonspecific HL-60 cells. Treated cells were stained with a fluorescent molecule (carboxyfluorescein diacetate succinimidyl ester). (A) Photographs were obtained by combining the optical and fluorescent images. Initial ratios in the original suspensions were set to 50% of specific cells. (B) Ratio of cells captured in the gap.

shows photographs obtained by combining optical and fluorescent images, which were obtained 60 s after the applied voltage to band B was switched to zero. The initial concentration ratio of specific cells was set to 50%. Almost all nonspecific cells with a fluorescent signal moved to band B, whereas the specific cells were captured in the gap between the band electrodes, even in the presence of treated nonspecific cells (Figure 6A). Figure 6B shows the ratio of cells captured in the gap between the band electrodes. The ratio of captured cells increased linearly with the increasing ratio of specific cells in the prepared mixture suspension. These results indicate that the presence of cells without the target antigen did not obstruct specific cell binding for detecting cells with surface antigens. The binding efficiency obtained in the present study improved compared to that obtained in our previous study using a combination of p- and n-DEP. Further Improvement in Cell Binding Efficiency by the Swing of Accumulated Cells. The swing of accumulated cells allowed us to further improve cell binding efficiency. Figure 7A shows the accumulated cells in the gap to capture the immunoreacted cells. We applied AC signals with the same voltage (20 Vpp), frequency (100 kHz), and phase to bands A and B. Therefore, cells accumulated in the center of the gap because of balance of the equivalent repulsive forces from bands A and B. We slightly reduced the intensity of the applied voltage for band B at intervals of 5 s during cell accumulation. Figure 7B shows the accumulated cells upon applying the AC signal with voltage of 18 Vpp to band B. The black dashed line points to the center of the gap region. The position of accumulated cells was shifted to approximately 4.5 μm left of center. When the 15 Vpp voltage was applied, cells accumulated at a position shifted 5.8 μm left from center (Figure 7C). This shift in accumulated position was reflected by voltage switching, which reversibly returned to the original position by applying 20 Vpp voltage to band B. Figure 7D,E shows the cross-sectional areas of the calculated electric field when voltages applied to band B were set to 18 and

Figure 5. Cell binding efficiencies obtained at different applied voltage intensities for cell accumulation by n-DEP. Cells accumulating in the gap area modified with the (a) anti-CD33 and (b) antimouse IgG antibodies.

efficiencies by applying different voltages. Applying AC voltage at 5 Vpp was insufficient to accumulate cells in the gap between the band electrodes. The cell binding efficiency in the gap areas modified with anti-CD33 increased with increasing intensity of AC voltage during the first n-DEP step (Figure 5a). In contrast, the number of cells captured in the gap modified with antimouse IgG slightly increased by increasing the applied voltage because of nonspecific adsorption (Figure 5b). The higher voltage pressed cells to the substrate surface in the gap due to the high repulsive n-DEP force, resulting in increased contact area between the cell and surface by cell deformation. Nonspecific adsorption increased drastically during the high voltage application (25 Vpp), whereas cell binding efficiency tended to be saturated in this voltage region. Therefore, an 8834

dx.doi.org/10.1021/ac302239k | Anal. Chem. 2012, 84, 8830−8836

Analytical Chemistry

Article

(Figure 8B). After turning off the voltage to band B, a few cells moved to band B using the swing technique with 20−18 Vpp (Supporting Information, ac302239k_si_001.zip), whereas approximately half the cells moved with 20−15 Vpp (Supporting Information, ac302239k_si_002.zip). In these movies, we applied voltage (20 Vpp) to bands A and B after 5 s. Next, the accumulated cells in the movies were swung during 60 s (5−65 s) and removed at 65 s. Cell binding efficiency and swing effectiveness using an application of 20−18 and 20−15 Vpp were estimated to be 83.9 ± 1.4% and 44.5 ± 6.0%, respectively, compared to the case without swing, in which cell binding efficiency was 68.3 ± 3.2%. These results suggest that the slight swing increased the opportunity to contact the surface antigen with the antibody immobilized on the substrate; however, an excessive shift of cell position lead to dissociation of the immunocomplex. In contrast, binding efficiencies of cells that accumulated in the gap modified with antimouse IgG using 20−18 and 20−15 Vpp were as low as 4.3 ± 1.3% and 3.6 ± 0.8%, respectively, which was nearly the same as the value obtained without swing. Therefore, swing of accumulated cells does not greatly affect nonspecific cell adsorption. Use of the swing technique considerably improved binding the efficiency of cells with the specific surface antigen.

Figure 7. Swing of the accumulated cells in the gap to capture the cells by immunoreaction. (A) Cells accumulated in the center of the gap area by applying AC voltage with the same intensity (20 Vpp) and frequency (100 kHz) to bands A and B. Cells accumulating by switching the band B voltage to (B) 18 and (C) 15 Vpp. Cross-sectional areas of the calculated electric field when voltages applied to band B were set to (D) 9 and (E) 7.5 V.

■

CONCLUSIONS We improved the efficiency of a cell binding assay based on a surface antigen by manipulating cells with DEP without losing rapidity and simplicity. The n-DEP repulsive force was dominant in cells in the higher conductivity region, which was advantageous for cell binding via specific immunoreactions. We achieved two different n-DEP cell patterns by switching the strength of the applied voltage. A cell suspension was introduced into the device comprising the ITO and ITO−IDA electrodes along with two combs of bands A and B. Upon applying the same AC voltage and frequency in the n-DEP region to bands A and B, cells were forced to accumulate in the gap region between the bands yielding line formations. Cells were repelled from the gap region and formed another line pattern on band B after the applied voltage to band B was switched to zero. Therefore, cells accumulated with n-DEP were easily directed to another site by controlling the cell pattern position. The formation of a second line pattern with cells was markedly inhibited by immunoreactions between surface antigens and a specific antibody immobilized on the microband array electrode. HL-60 cells with the CD33 surface antigen were irreversibly captured in the gap area modified with specific antiCD33 (68.3 ± 3.2%). Importantly, cell capture through immunoreactions was assisted by rapid accumulation with n-DEP in the gap area, and n-DEP provided the uniform removal of unbound cells. Significant rapidity (total time required, 90 s) and simplicity (no pretreatment for cell staining with fluorescent molecules) were sustained in the proposed method. The swing of accumulated cells allowed further improvement of cell binding efficiency (83.9 ± 1.4%). The present system is promising for applications to the distinction of various bio-objects with different target molecules on the surface. The preparation of the electrode with a micrometer order may allow the appearance of applicable devices to smaller cells (e.g., bacteria, yeast cells, and red blood cells).

15 Vpp, respectively. With the decreased intensity of applied voltage to band B, the regions with a strong electric field on band B, which are shown as white color, became narrower. Moreover, low electric field regions formed at the gap slightly shifted to band B. Thus, cells that accumulated in the center of the gap moved toward band B, which was in agreement with the experimental results (Figure 7A−C). After the cells accumulated in the gap, we successively switched the intensity of the applied voltage to band B from 20 Vpp to 18 or 15 Vpp at intervals of 5 s to swing the accumulated cells. Figure 8 shows optical images of the captured cells 60 s after

Figure 8. Effect of swing to capture cells by immunoreaction. Optical images of the captured cells 30 s after separating unbound cells from the gap modified with anti-CD33 by switching the intensity of band B to zero. Accumulated cells were swung for 60 s by applying (A) 20−18 Vpp and (B) 20−15 Vpp.

■

ASSOCIATED CONTENT

S Supporting Information *

separating unbound cells from the gap modified with anti-CD33 by switching the band B intensity to zero. Cells were swung for 60 s by applying 20−18 Vpp (Figure 8A) and 20−15 Vpp

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. 8835

dx.doi.org/10.1021/ac302239k | Anal. Chem. 2012, 84, 8830−8836

Analytical Chemistry

■

Article

AUTHOR INFORMATION

Corresponding Author

*Tel: +81-791-58-0171 (T.Y.). Fax: +81-791-58-0493 (T.Y.). E-mail: [email protected] (T.Y.); [email protected] (F.M.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We are grateful to Dr. Kuniaki Nagamine and Prof. Matsuhiko Nishizawa at Tohoku University for their help calculating the electric field strength. This study was partly supported by a Grant-in-Aid for Scientific Research on Innovative Areas “Bio Assembler” of The Ministry of Education, Culture, Sports, Science, and Technology, Japan.

■

REFERENCES

(1) Campana, D.; Behm, F. G. J. Immunol. Methods 2000, 243, 59−75. (2) Belov, L.; Huang, P.; Barber, N.; Mulligan, S. P.; Christopherson, R. I. Proteomics 2003, 3, 2147−2154. (3) Ko, I.-K.; Kato, K.; Iwata, H. Biomaterials 2005, 26, 687−696. (4) Liu, A. Y. Cancer Res. 2000, 60, 3429−3434. (5) Belov, L.; De la Vega, O.; Dos Remedios, C. G.; Mulligan, S. P.; Christopherson, R. I. Cancer Res. 2001, 61, 4483−4489. (6) Konagaya, S.; Kato, K.; Nakaji-Hirabayashi, T.; Arima, Y.; Iwata, H. Biomaterials 2001, 32, 5015−5022. (7) Kato, K.; Toda, M.; Iwata, H. Biomaterials 2007, 28, 1289−1297. (8) Bailey, R. C.; Kwong, G. A.; Radu, C. G.; Witte, O. N.; Heath, J. R. J. Am. Chem. Soc. 2007, 129, 1959−1967. (9) Fernandes, T. G.; Kwon, S.-J.; Lee, M.-Y.; Clark, D. S.; Cabral, J. M. S.; Dordick, J. S. Anal. Chem. 2008, 80, 6633−6639. (10) Dexlin, L.; Ingvarsson, J.; Frendéus, B.; Borrebaeck, C. A. K.; Wingren, C. J. Proteome Res. 2008, 7, 319−327. (11) Ogata, S.; Yasukawa, T.; Matsue, T. Bioelectrochemistry 2001, 54, 33−37. (12) Suzuki, M.; Yasukawa, T.; Mase, Y.; Oyamatsu, D.; Shiku, H.; Matsue, T. Langmuir 2004, 20, 11005−11011. (13) Suzuki, M.; Yasukawa, T.; Shiku, H.; Matsue, T. Langmuir 2007, 23, 4088−4094. (14) Suzuki, M.; Yasukawa, T.; Shiku, H.; Matsue, T. Biosens. Bioelectron. 2008, 24, 1043−1047. (15) Lee, H. J.; Yasukawa, T.; Shiku, H.; Matsue, T. Biosens. Bioelectron. 2008, 24, 1000−1005. (16) Lee, H. J.; Lee, S. H.; Yasukawa, T.; Ramón-Azcón, J.; Mizutani, F.; Ino, K.; Shiku, H.; Matsue, T. Talanta 2010, 81, 657−663. (17) Ramón-Azcón, J.; Yasukawa, T.; Lee, H. J.; Matsue, T.; SánchezBaeza, F.; Marco, M.-P.; Mizutani, F. Biosens. Bioelectron. 2010, 25, 1928−1933. (18) Ramón-Azcón, J.; Yasukawa, T.; Mizutani, F. Anal. Chem. 2011, 83, 1053−1060. (19) Hatanaka, H.; Yasukawa, T.; Mizutani, F. Anal. Chem. 2011, 83, 7207−7212. (20) Pohl, H. A. Dielectrophoresis; Cambridge University Press: Cambridge, UK, 1978. (21) Jones, T. B. Electromechanics of Particles; Cambridge University Press: New York, 1995. (22) Morgan, H.; Green, N. G. AC Electrokinetics: Colloids and Nanoparticles; Research Studies Press: Baldock, Hertfordshire, England, 2003. (23) Huang, Y.; Holzel, R.; Pethig, R.; Wang, X.-B. Phys. Med. Biol. 1992, 37, 1499−1517. (24) Kaler, K. V. I. S.; Jones, T. B. Biophys. J. 1990, 57, 173−182.

8836

dx.doi.org/10.1021/ac302239k | Anal. Chem. 2012, 84, 8830−8836