Article pubs.acs.org/ac
Controlled Rotation and Vibration of Patterned Cell Clusters Using Dielectrophoresis Rebecca Soffe,† Shi-Yang Tang,† Sara Baratchi,†,‡ Sofia Nahavandi,§ Mahyar Nasabi,† Jonathan M. Cooper,∥ Arnan Mitchell,† and Khashayar Khoshmanesh*,† †
School of Electrical and Computer Engineering, RMIT University, Melbourne, Victoria 3001, Australia Health Innovations Research Institute, RMIT University, Melbourne, Victoria 3083, Australia § Faculty of Medicine, Dentistry, & Health Sciences, The University of Melbourne, Melbourne, Victoria 3010, Australia ∥ The Bioelectronics Research Centre, Department of Electronics and Electrical Engineering, University of Glasgow, Glasgow G12 8LT, United Kingdom ‡
S Supporting Information *
ABSTRACT: The localized motion of cells within a cluster is an important feature of living organisms and has been found to play roles in cell signaling, communication, and migration, thus affecting processes such as proliferation, transcription, and organogenesis. Current approaches for inducing dynamic movement into cells, however, focus predominantly on mechanical stimulation of single cells, affect cell integrity, and, more importantly, need a complementary mechanism to pattern cells. In this article, we demonstrate a new strategy for the mechanical stimulation of large cell clusters, taking advantage of dielectrophoresis. This strategy is based on the cellular spin resonance mechanism, but it utilizes coating agents, such as bovine serum albumin, to create consistent rotation and vibration of individual cells. The treatment of cells with coating agents intensifies the torque induced on the cells while reducing the friction at the cell−cell and cell−substrate interfaces, resulting in the consistent motion of the cells. Such localized motion can be modulated by varying the frequency and voltage of the applied sinusoidal AC signal and can be achieved in the absence and presence of flow. This strategy enables the survival and functioning of moving cells within large-scale clusters to be investigated.
C
Microfabricated devices enable quick and accurate positioning of cells to be stimulated through micropipette aspiration,19 atomic force microscopy tip,20 or optical,21 acoustic,22 and magnetic tweezers23 under precisely controlled microenvironment conditions. Additionally, unique features of microfabricated/microfluidic systems have led to the development of a variety of novel techniques for the mechanical stimulation of cells, which could not be achieved using conventional macro-scale systems.5,8 This include dilation of cells in response to flow-induced shear stress,24 deformation of cells by flowing them through narrow channels,25 stretching of cells using deformable substrates such as silicone elastic membranes,26 and squeezing of cells using pneumatically actuated membranes.27 Furthermore, microfluidic devices enables the automation of multiple processes, such that the cells can be cultured, treated, and separated, before being dynamically analyzed during mechanical stimulation.28−32
ells can both sense and respond to the mechanical and chemical cues from their microenvironment.1,2 The signaling pathways triggered by mechanical cues can impact different cellular processes including proliferation, transcription, and organogenesis.1,3,4 A variety of techniques have been developed for the mechanical stimulation of cells,5−8 which, according to the extent and number of affected cells, can be divided into three groups. The first group includes cell poking,9 atomic force microscopy indentation,10 magnetic bead microrheometry,11 and magnetic twisting cytometry techniques,12 which enable local deformation of single cells. The second group includes micropipette aspiration,13 optical tweezing,14 and acoustic tweezing techniques,15 which enable deformation of an entire cell. The third group includes the application of hydrostatic pressure16 or inducing osmotic pressure, also known as hypotonic cell swelling techniques,17 which enable the mechanical stimulation of multiple cells at a time. Recent advances in microfabrication technologies, and in particular microfluidics, has realized devices that are smaller, simpler, cheaper, and more accurate than their macro-sized predecessors for the mechanical stimulation of cells.5−8,18 © 2015 American Chemical Society
Received: November 19, 2014 Accepted: January 22, 2015 Published: January 22, 2015 2389
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry
modulated by varying the frequency and voltage of the applied sinusoidal AC signal and can be achieved in the absence and presence of flow.
The majority of the above-mentioned techniques rely on passive immobilization of cells over the predetermined locations of the microdevice. The ability to actively immobilize the cells and to mechanically stimulate them using the same mechanism can accelerate and simplify the experimental process. Dielectrophoresis, the induced motion of polarizable particles in a nonuniform electric field, has been extensively used for sorting, trapping, and microscopic analysis of cells.33,34 However, dielectrophoresis has not been utilized for mechanical stimulation of cells. A closer look at the literature of dielectrophoresis reveals that cells can be rotated using electrorotation (ROT) or cellular spin resonance (CSR) mechanism. ROT refers to the rotation of cells when subjected to a rotating electric field produced by four-pole microelectrodes35 and has been extensively used for measuring the dielectric properties of cells. However, ROT is mostly used for rotation of single cells and not a cluster of cells, and, more importantly, it cannot be applied in the presence of flow, as the rotating cells will be washed away by the hydrodynamic drag force. In contrast, CSR refers to rotation of adjacent cells when subjected to an alternating electric field produced by two-pole microelectrodes. CSR was first reported by Pohl et al. in 1971,36 based on experiments conducted by Teixeira-Pinto et al. in the 1960s.37 However, an examination into the literature of CSR provides no complete experimental regime to produce consistent mechanical stimulation.36,38,39 Therefore, revisiting CSR may provide opportunities to develop a mechanism capable of clustering and mechanically stimulating large numbers of cells in the presence of flow. In this article, we present a strategy for the mechanical stimulation of customized cell clusters using dielectrophoresis (Figure 1). This strategy is based on CSR mechanism but
■
MATERIALS AND METHODS Fabrication of Microelectrodes. Gold on chrome films were deposited onto a 100 μm thick glass substrate at thicknesses of 1500 and 500 Å, respectively, using physical vapor deposition. The microelectrodes were then patterned using standard photolithography and etching techniques. Fabrication of Microchannel. The microchannel was fabricated from poly(dimethysiloxane) (PDMS) utilizing soft lithography and replica molding techniques.40 To create the PDMS structure, PDMS base and curing agent (Sylgard 184, Dow Corning Corporation, MI) were mixed with a ratio of 10:1 w/w and degassed until all of the bubbles disappeared. The dimensions of the PDMS microchannel were set to 500 × 80 μm (W × H). Cell Suspension Preparation. Saccharomyces cerevisiae yeast cells were used in this work to investigate the potential of using dielectrophoresis to invoke dynamic movement. These cells are an important model organism to understand the biology of eukaryotic cells at the cellular and molecular levels and have been widely used for studying cell biology, genetics, and genomics.41 In addition, these cells can be easily cultivated, treated, and used in bioengineering laboratories and thus are widely used in cellular studies involving dielectrophoresis.33 For preparation of cell suspension, 25 mg of S. cerevisiae yeast powder (Sigma-Aldrich) was suspended in an isotonic low electrical conductivity buffer of 100 mL deionized water, 8.5% w/v sucrose, and 0.3% dextrose to yield a cell concentration of 0.025% w/v (corresponding to 1.5 × 105 cells/mL). To avoid agglomeration, the suspension was placed in an ultrasonic bath for 15 min. The resulting cell suspension had an electrical conductivity of approximately 0.006 S/m, which was adjusted accordingly using phosphate buffer saline (PBS) (SigmaAldrich). The electrical conductivity was measured using an ECTester11+ conductivity meter (Eutech Instruments, Singapore). Treating Cells with Coating Agents. The dynamic response of cells was examined by coating them with four coating agents, including bovine serum albumin (BSA), poly-Llysine, poly-L-ornithine, and whey protein concentrate, as explained below. The liquid BSA suspension (50 mg/mL, Life Technologies) was diluted initially with either PBS or deionized water. This solution was then added to the cell suspension such that the final concentration of BSA was 0.2% w/v. The existence of BSA over the cells was confirmed by treating cells with labeled BSA (Alexa Fluor 488 conjugate, Life Technologies) solution and exciting them by a 488 nm laser using an inverted microscope (Nikon Eclipse Ti) (Supporting Information S1). The poly-L-lysine and poly-L-ornithine solutions (0.1% w/v, Sigma-Aldrich) were diluted in deionized water and added to the cell suspension to yield a final concentration of 2.0 × 10−5% w/v. Whey protein concentrate manufactured from sweet whey was purchased from a commercial supplier (80% w/w, Eatme supplements). The concentrate was added to deionized water and diluted in the cell suspension to yield a final concentration of 0.2% w/v. The viability of treated cells was examined using of a membrane-impermeant nucleic acid stain such as propidium iodide (PI) (Life Technologies). Viability assays indicated that treatment of cells with coating agents does not
Figure 1. Three-dimensional schematic of the localized motion of patterned cell clusters under the influence of dielectrophoresis. Cells can exhibit rotational and vibrational movements according to their location within the cluster. Cells located at the free ends of pearl chains have the highest occurrence of rotation, which are indicated as blue cells in the schematic. Alternatively, the cells packed along the long chains bridged between the microelectrodes have the highest occurrence of vibration. The arrows indicate the direction and strength of the vibrational movement of the cells.
utilizes coating agents, such as bovine serum albumin (BSA), to produce consistent mechanical stimulation of cells. The incubation of cells with coating agents increases the torque induced on the cells, whereas it reduces the friction at the cell− cell and cell−substrate interfaces, resulting in the rotation and vibration of the cells. Such mechanical motions can be 2390
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry have an adverse effect on the viability of cells (Supporting Information S2).
■
RESULTS AND DISCUSSION Cluster Formation. The first series of experiments was conducted using an open-top dielectrophoresis platform, which consisted of a PDMS cell chamber integrated onto the glass slide accommodating the microelectrode arrays. This enabled us to study the response of BSA-treated cells in the absence of flow. In doing so, 100 μL of the BSA-treated yeast suspension was transpired onto the PDMS chamber. The electrical conductivity of the suspension was set to 0.02 S/m through the addition of PBS. The microelectrodes were then activated using a 10 MHz, 5 Vpk−pk AC sinusoid signal, which led to the immobilization of cells along the microelectrode edges. Next, the frequency of the applied signal was sequentially increased in 5 MHz increments to 50 MHz, in 5 min intervals. This significantly changed the overall configuration of the BSA-treated cell clusters patterned along the microelectrodes, as discussed below. At low frequencies (f < 10 MHz), dense clusters of cells formed along the both internal and external edges of microelectrodes, similar to the case of untreated cells (Figure 2a). At medium frequencies (10 < f < 30 MHz), the cells patterned along the internal edges of microelectrodes left the surface of microelectrodes and lined up along the previously patterned cells, resulting in the formation of pearl chains stemming at the external edges of microelectrodes (Figure 2b). At higher frequencies (f > 30 MHz), the chains were further elongated and the intercellular distance between the cells increased, leading to formation of loosely joined cell chains along the electric field (Figure 2c). In contrast, untreated cells exhibited negative dielectrophoretic (DEP) response and were repelled from the microelectrodes at such high frequencies. Interestingly, for the case of treated cells, increasing the applied voltage led to a reduction of the intercellular gap between adjacent cells, whereas the patterned cells exhibited rotational and vibrational responses, as described in the succeeding sections. Dynamic Cell Clusters. To further examine this dynamic response, the voltage was increased to 10 Vpk−pk while the frequency was set to 36 MHz. Under these conditions, the BSA-treated cells started to rotate (Supporting Information Movie 1). The captions for the movies are given in Supporting Information S3. Cells that exhibited a rotational behavior were generally close to the free end of a pearl chain comprising of several cells. The reference image, Figure 3a, indicates the location of three randomly selected cells within the electric field, positioned at various locations within a pearl chain. A time-lapse of snapshots taken at 10 ms intervals is used in Figure 3b to highlight the rotational response of cells diagrammatically. The cells identified as i, ii, and iii rotated at 1.5, 0.5, and 1.2 rev/s, respectively. Monitoring the rotational behavior of the cells indicated that rotation is sporadic and that each cell has a different rotational axis and velocity. The rotational axis varied between cells such that the axis could be aligned or unaligned with the electric field; also, cells would rotate in either clockwise or anticlockwise directions, as observed in Figure 3b. This sporadic behavior could be potentially attributed to various factors, including the heterogeneity of the patterned cells, the inconsistency of BSA layer coating the cells, the electric field experienced by the cells, and the number of cells located along the pearl chain. In
Figure 2. Response of BSA-treated yeast cell clusters, patterned onto a finger-shaped microelectrode array when operated with a 5 Vpk−pk AC sinusoid signal. (a-i, b-i, c-i)Response of BSA-treated cell clusters at 5, 20, and 40 MHz, respectively. (a-ii, b-ii, c-ii) Schematic representation of the response of patterned cell clusters at different frequencies; with BSA-treated cells, represented as green, being compared to untreated cells, represented as blue. Evidently, the BSA-treated cells exhibit a positive DEP response as the distance between the two adjacent cells increases at high frequencies. Alternatively, the untreated cells exhibit a negative DEP response at frequencies higher than 30 MHz. Scale bar is 50 μm.
addition, our experiments indicated that the ability of a cell to rotate was not significantly affected by its location within the electric field. To further examine the effect of BSA coating on the dynamic response of the cells, the cells were kept in the BSA suspension for 120 min before being applied to our DEP system. This increased the dynamic response of cells such that more cells within the cluster could rotate. However, the cells that were located in the middle of the cluster, were surrounded by several cells on different sides, thus, could not freely rotate due to an increased intercellular friction and instead exhibited a vibrational motion. The vibration of cells was more tangible along the cells bridging between the opposite microelectrodes, making pearl bridges (Supporting Information Movie 2). The elongation of the BSA incubation process did not affect the mortality rate of the cells significantly (Supporting Information S2). Through observation, it was apparent that the overall dynamic movement of the cell cluster was governed by both the cluster configuration and the extent of both the rotational and vibrational cells within the cluster (Figure 3c,d). The extent 2391
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry
(W × H) was integrated onto the glass substrate containing the microelectrode array. To pattern the cells along the microelectrodes, the cell suspension was applied through the microchannel at a flow rate of 2 μL/min while the microelectrodes were energized by applying a 10 Vpk−pk sinusoidal AC signal at 10 MHz. To enable cell rotation, the voltage was decreased to 8 Vpk−pk while the frequency was increased to 70 MHz. The presence of flow facilitated the formation of pearl chains concentrated around the tip of the microelectrodes, thus increasing the likelihood of cells exhibiting rotation when located at the free end of the pearl chains. Four cells presented in Figure 4 display the rotational response of the cells in
Figure 3. Rotational and vibrational movement of patterned cells in the absence of flow, when operating the microelectrodes with a 36 MHz, 10 Vpk−pk AC sinusoidal signal. (a) The reference image used to identify the location of three rotating cells. (b: i−iii) The extent of the rotational movement can be observed through the use of snapshots taken at 10 ms intervals, as presented for three cells in this case (Supporting Information Movie 1). The rotational velocity of cells can be obtained by measuring the displacement of the yellow lines along the yellow circles. (c) The reference image used to highlight the structure of a vibrating cell cluster. (d) The outline of the cells was obtained every 2 s to highlight the vibration of the cell cluster (Supporting Information Movie 2). Scale bar is 20 μm in a, 30 μm in c and d, and 5 μm in b.
Figure 4. Rotational and vibrational movement of patterned cells in the presence of flow, when operating the microelectrodes with a 70 MHz, 8 Vpk−pk AC sinusoidal signal. The extent of movement of individual cells can be observed through the use of snapshots taken at 10 ms intervals, as presented for four cells in this case (Supporting Information Movie 3). The yellow arrows indicate the direction of flow through the microfluidic channel. Scale bar is 30 μm.
snapshots taken every 10 ms from Supporting Information Movie 3. The identified cells rotated consistently at 1.8, 0.8, 1.0, and 1.5 rev/s, from left to right. Moreover, the cells that were bridged between the microelectrodes exhibited a vibrational motion following a 120 min treatment with BSA. The ability to induce rotational and vibrational movement onto the patterned cells in the presence of flow enables more control over the configuration and density of cell clusters and can facilitate several experimental scenarios. Modeling Characteristics. A theoretical model was developed to determine the rotational velocity of BSA-treated cells. The model consists of two identical cells in contact with each other, which are exposed to an external AC electric field, as depicted in Supporting Information S4. The torque exerted on a cell (τrotation) within an electric field in the presence of an adjacent cell is adopted from the equation developed by Holzapfel et al.38 Such torque, should overcome the retarding torques generated by the surrounding viscous liquid (τliquid) and the friction at the cell−cell (τcell−cell) and cell−substrate (τcell−substrate) interfaces, as given below:
of the movement of the vibrating cell cluster was best observed through comparing the outline of the cells within the electric field. Thus, snapshots acquired at 2 s intervals from Supporting Information Movie 2 were imported into Adobe Photoshop (CS6, CA) to obtain the outline of the cells. The outlines were then superimposed on top of each other, as presented in Figure 3d, which highlighted that vibration is considerably more dynamic across the pearl bridges compared to that in pearl chain clusters. The rotation or vibration of cells could be sustained for 45 min (longer periods were not examined in our experiments). The rotational behavior of individual cells could vary over time due to local displacement of cells, but no significant difference was observed in the overall dynamic response of the patterned cell cluster. The viability of cells was examined after each experiment, including the experiments carried out for 45 min by on-chip PI staining of cells.28 No significant difference was observed between the viability of the control (BSA-treated cells, without DEP) and dynamic (BSA-treated cells, with DEP) samples, as shown in Figure S2. The application of a lowamplitude, high-frequency sinusoidal AC signal minimized the cell damage associated with Joule heating of the medium and the effects of inducing transmembrane potential.33,42 In addition, we examined the dynamic response of BSAtreated cells in the presence of flow. In doing so, a PDMS microchannel with cross-sectional dimensions of 500 × 80 μm
⟨τrotation⟩ = τliquid + τcell − cell + τcell − substrate
(1)
Substituting the equations for each of the above torques, the following equation is obtained, using which the constant rotational velocity of the cells at equilibrium can be achieved: P2 × E −
⎛ P. D ⎞ 1 × ∇⎜ 1 3 ⎟ ⎝ D ⎠ 4πε
= 8πηωr 3 + ξcell − cellr 2ω + ξcell − substrater 2ω 2392
(2)
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry
Figure 5. Variations of the rotational velocity of BSA-treated cells with respect to the voltage and frequency of AC sinusoidal signal obtained in the absence of flow. Measured and modeled data are presented as solid and dashed lines, respectively. (a) The frequency response indicated that the highest rotational velocity of cells occurs at a frequency range of 30−40 MHz when operating at 10 Vpk−pk (Supporting Information Movie 4). (b) The voltage response of cells indicated that the highest rotational velocity of cells occurs at a voltage range of 12−14 Vpk−pk when operating at 36 MHz (Supporting Information Movie 5).
The first component of τrotation appearing on the left-hand side of this equation represents the torque that acts to align the dipole of the rotating cell (cell-2, as shown in Figure S4) with the field,42 whereas the second component represents the torque contributed from the dipole of the stationary cell (cell-1, as shown in Figure S4) on the rotating cell (cell-2).38 P1 and P2 are the dipole moments of the stationary and rotating cells, respectively, D is the vector connecting the center points of the adjacent cells; E is the external electric field; ε and η are the permittivity and viscosity of surrounding medium, respectively; r and ω are the radius and rotational velocity of the cell, respectively; and ξcell−cell and ξcell−substrate are the friction coefficient at the interface of cell−cell and cell−substrate, respectively. However, both ξcell−cell and ξcell−substrate are difficult to obtain due to the nature of the system; thus, to enable simulated results to be obtained, τrotation was multiplied by a factor of 0.8 to account for these frictional losses. A detailed explanation of the resulting torque and the induced cell rotation is presented in Supporting Information S4. Characterization of the Dynamic Response. The dynamic response of BSA-treated cells was characterized through measuring the rotational velocity of individual cells at different frequencies and voltages of the applied signal as well as applying various coating agents and cell concentrations, as described below. Rotational velocity is defined as the number of cell rotations per second, ignoring the direction of rotation. Frequency Characteristics. The rotational response of BSAtreated cells was examined at a frequency range of 10−70 MHz, with the voltage of the sinusoid signal kept at 10 Vpk−pk. Figure 5a shows the variations of the average rotational velocity of the cells with respect to the applied frequencies (Supporting Information Movie 4). A resonant frequency was observed to occur at approximately 32 MHz according to the balance between the rotational and resisting torques acting on the cells, as described in eq 1. The rotational velocity of cells obtained from our theoretical model is shown by dashed line. A 10 MHz difference is observed between the experimental and simulated resonance frequencies (Figure 5a). This can be attributed to several
factors, including the possibilites that the pearl chains may contain more than two cells in experiments, the yeast cells may have an ellipsoidal shape rather than the spherical shape assumed in our model, and the intercellular gap may vary at different frequencies while it is assumed to be constant in our model (Figure 2). Thus, the actual dipole moments from adjacent cells are more complicated than the simulated scenario. Although our model is simplified without considering these parameters, it was able to predict the trend of the rotational response of the cells with respect to the frequencies of the applied signal. Voltage Characteristics. The rotational characteristics of BSA-treated cells influenced by voltage were determined at a frequency of 36 MHz (Figure 5b), very close to the resonance frequency observed in Figure 5a. Surprisingly, the rotational velocity reached its peak at 13 Vpk−pk and is reduced at higher voltages (Supporting Information Movie 5). The rotational velocity of cells obtained from our theoretical model is shown by a dashed line (Figure 5b). The model predicted the increase of rotational velocity with respect to the applied voltage, but it failed to predict the decrease of rotational velocity at higher voltages. This contradiction could potentially be attributed to strengthening of the DEP force at high applied voltages, which results in a reduced intercellular gap. This not only can disrupt the condition of loosely joined cells, which is essential for the rotation of cells, but also can increase the intercellular friction. This can also change the magnitude of the torque induced from the dipole of stationary cells on the rotating cells located at the free end of pearl chains, as given in eq 2. However, in our model, the intercellular gap was assumed to be constant across the entire voltage range. Gap between Microelectrode Pairs. In addition, we investigated the effect of the distance between the microelectrode pairs on the rotational response of the patterned cells. We used a DEP array with variable gap between its consequential microelectrode pairs. Our experiments indicated that the gap between the microelectrode pairs can influence the rotational characteristics of the patterned cells, as presented in Supporting Information Movie 6. This behavior was observed 2393
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry
in a reproducible manner. The movement of cells can be achieved in the absence and presence of flow and using different coating gents. Future work includes assessing the capability of our approach for the controlled rotation and vibration of mammalian cells. This can lead to a versatile microfluidic platform for studying the functioning (proliferation, metabolism, and death) of dynamic cells. It also enables the uptake and response of dynamic cells to chemicals (e.g., drugs) or particles (e.g., nanoparticles) to be investigated.43 Considering that cells constituting several tissues and organs of human body might be dynamic, our approach enables more realistic in vitro models. Furthermore, the rotation of cells can induce a moderate shear stress over their membrane (see Supporting Information S6), which can be used to study the shear-induced calcium signaling of cells.
because the distance between the microelectrode pairs determines the induced electric field; thus, in turn, can change the balance between the DEP force and the rotating torque acting on the cells, as further explained in Supporting Information S5. Coating Agents. In addition to BSA, additional coating agents were investigated, including poly-L-lysine, poly-Lornithine, and whey protein concentrate. We found two major differences in the capability of different coating agents to induce dynamic motion when operating the system using a 10 Vpk−pk 36 MHz sinusoid signal (Supporting Information Movie 7). Firstly, incubating of cells with poly-L-lysine, poly-Lornithine, and whey protein concentrate facilitated rotation of 16 ± 3%, 17 ± 4%, and 21 ± 4% of the cells, respectively, which was considerably less than what was achieved with BSA-treated cells, where around 36 ± 6% of the cells could rotate. Secondly, incubation of cells with BSA for more than 120 min intensified the dynamic response of cells and enabled the cells to rotate or vibrate according to their location within the cluster (Supporting Information Movies 2 and 3); whereas, elongating the incubation process with poly-L-lysine, poly-L-ornithine, and whey concentrate protein neither intensified the dynamic response of treated cells nor led to vibration of cells. To further understand the interaction between the yeast cells and the coating agents, the cells were incubated in BSA for 10 min; then, they were washed twice and suspended in PBS before being transpired onto the DEP platform. The movement observed proved to be indifferent to the coating agents present in the suspension medium, which we conjecture could be, in fact, that the coating agents potentially modify the surface properties of the cells. In addition, our experiments suggested that to enable the occurrence of dynamic cell clusters the concentration of the coating agent should result in an overall cell suspension conductivity within 0.01 and 0.03 S/m. However, the upper conductivity limit can be extended through the addition of PBS up to 0.15 S/m. The lower conductivity was determined as the threshold to induce the dynamic response of cell clusters, whereas the higher conductivity is the threshold to avoid the formation of bubbles at the surface of microelectrodes. Yeast Cell Concentration. The optimal concentration of yeast cells was determined to be 1.5 × 105 cells/mL. The time required to immobilize cells significantly increased at a lower cell concentration. In contrast, increasing the cell concentration led to the formation of dense cell clusters between the microelectrodes. This reduced the possibility of rotation; however, it increased the percentage of cells vibrating within the cluster and resulted in a large dynamic cell cluster.
■
ASSOCIATED CONTENT
S Supporting Information *
Section 1: Coating yeast cells with BSA. Section 2: Viability assays. Section 3: Captions for seven supplementary movies. Section 4: Theoretical model for cell rotation. Section 5: Rotation of BSA-treated cells using a DEP array with different microelectrode gaps. Section 6: Comparing the shear stress induced by cell rotation. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS K. Khoshmanesh acknowledges the Australian Research Council for funding under an Discovery Early Career Researcher Award (DECRA) (project DE120101402).
■
REFERENCES
(1) Polacheck, W. J.; Li, R.; Uzel, S. G.; Kamm, R. D. Lab Chip 2013, 13, 2252−2267. (2) Chingozha, L.; Zhan, M.; Zhu, C.; Lu, H. Anal. Chem. 2014, 86, 10138−10147. (3) Bukoreshtliev, N.; Haase, K.; Pelling, A. Cell Tissue Res. 2013, 352, 77−94. (4) Bastock, R.; Strutt, D. Development 2007, 134, 3055−3064. (5) Delmas, P.; Hao, J.; Rodat-Despoix, L. Nat. Rev. Neurosci. 2011, 12, 139−153. (6) van Loon, J. W. A. Microgravity Sci. Technol. 2009, 21, 159−167. (7) Guck, J.; Lautenschläger, F.; Paschke, S.; Beil, M. Integr. Biol. 2010, 2, 575−583. (8) Bao, G.; Suresh, S. Nat. Mater. 2003, 2, 715−725. (9) Daily, B.; Elson, E. L.; Zahalak, G. I. Biophys. J. 1984, 45, 671− 682. (10) A-Hassan, E.; Heinz, W. F.; Antonik, M. D.; D’Costa, N. P.; Nageswaran, S.; Schoenenberger, C.-A.; Hoh, J. H. Biophys. J. 1998, 74, 1564−1578. (11) Bausch, A. R.; Ziemann, F.; Boulbitch, A. A.; Jacobson, K.; Sackmann, E. Biophys. J. 1998, 75, 2038−2049. (12) Wang, N.; Ingber, D. E. Biochem. Cell Biol. 1995, 73, 327−335. (13) Hochmuth, R. M. J. Biomech. 2000, 33, 15−22. (14) Lim, C.; Dao, M.; Suresh, S.; Sow, C.; Chew, K. Acta Mater. 2004, 52, 1837−1845. (15) Kundu, T.; Bereiter-Hahn, J.; Hillmann, K. Biophys. J. 1991, 59, 1194−1207.
■
CONCLUSIONS AND FUTURE OUTLOOK In this article, we have demonstrated a new approach for the mechanical stimulation of large-scale cell clusters. In this approach, the cells are treated with coating agents such as BSA before being applied into a dielectrophoresis platform. Dielectrophoresis facilitates the rapid immobilization of cells and formation of cell chains along the microelectrodes. Increasing the frequency of the applied signal to ∼30 MHz leads to formation of loosely joined cell chains. Under these conditions, the cells that are located along the free end of the cell chains can rotate, whereas the interaction of multiple rotating cells can lead to vibration of the entire cell chain. The dynamic response of the patterned cells can be modulated through varying the voltage and frequency of the applied signal, 2394
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry (16) Kim, S. H.; Choi, Y. R.; Park, M. S.; Shin, J. W.; Park, K. D.; Kim, S. J.; Lee, J. W. J. Biomed. Mater. Res., Part A 2007, 80, 826−836. (17) Cook, B.; Hardy, R. W.; McConnaughey, W. B.; Zuker, C. S. Nature 2008, 452, 361−364. (18) Baratchi, S.; Khoshmanesh, K.; Sacristán, C.; Depoil, D.; Wlodkowic, D.; McIntyre, P.; Mitchell, A. Biotechnol. Adv. 2014, 32, 333−346. (19) Lee, L. M.; Liu, A. P. Lab Chip 2015, 15, 264−273. (20) Rosenbluth, M. J.; Lam, W. A.; Fletcher, D. A. Biophys. J. 2006, 90, 2994−3003. (21) Honarmandi, P.; Lee, H.; Lang, M. J.; Kamm, R. D. Lab Chip 2011, 11, 684−694. (22) Fan, Z.; Sun, Y.; Chen, D.; Tay, D.; Chen, W.; Deng, C. X.; Fu, J. Sci. Rep. 2013, 3, 2176 . (23) Zhao, R.; Boudou, T.; Wang, W.-G.; Chen, C. S.; Reich, D. H. J. Appl. Phys. 2014, 115, 172616. (24) Baratchi, S.; Tovar-Lopez, F. J.; Khoshmanesh, K.; Grace, M. S.; Darby, W.; Almazi, J.; Mitchell, A.; McIntyre, P. Biomicrofluidics 2014, 8, 044117. (25) Lam, W. A.; Rosenbluth, M. J.; Fletcher, D. A. Blood 2007, 109, 3505−3508. (26) Gilchrist, C. L.; Witvoet-Braam, S. W.; Guilak, F.; Setton, L. A. J. Biomech. 2007, 40, 786−794. (27) Xu, T.; Yue, W.; Li, C.-W.; Yao, X.; Cai, G.; Yang, M. Lab Chip 2010, 10, 2271−2278. (28) Khoshmanesh, K.; Akagi, J.; Nahavandi, S.; Skommer, J.; Baratchi, S.; Cooper, J. M.; Kalantar-Zadeh, K.; Williams, D. E.; Wlodkowic, D. Anal. Chem. 2011, 83, 2133−2144. (29) Lapizco-Encinas, B. H.; Simmons, B. A.; Cummings, E. B.; Fintschenko, Y. Anal. Chem. 2004, 76, 1571−1579. (30) Li, S.; Guo, F.; Chen, Y.; Ding, X.; Li, P.; Wang, L.; Cameron, C. E.; Huang, T. J. Anal. Chem. 2014, 86, 9853−9859. (31) Zhan, M.; Chingozha, L.; Lu, H. Anal. Chem. 2013, 85, 8882− 8894. (32) Hou, H. W.; Warkiani, M. E.; Khoo, B. L.; Li, Z. R.; Soo, R. A.; Tan, D. S.-W.; Lim, W.-T.; Han, J.; Bhagat, A. A. S.; Lim, C. T. Sci. Rep. 2013, 3, 1259 . (33) Khoshmanesh, K.; Nahavandi, S.; Baratchi, S.; Mitchell, A.; Kalantar-zadeh, K. Biosens. Bioelectron. 2011, 26, 1800−1814. (34) Becker, F. F.; Wang, X. B.; Huang, Y.; Pethig, R.; Vykoukal, J.; Gascoyne, P. R. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 860−864. (35) Hölzel, R.; Lamprecht, I. Biochim. Biophys. Acta, Biomembr. 1992, 1104, 195−200. (36) Pohl, H. A.; Crane, J. S. Biophys. J. 1971, 11, 711−727. (37) Teixeira-Pinto, A. A.; Nejelski, L. L., Jr.; Cutler, J. L.; Heller, J. H. Exp. Cell Res. 1960, 20, 548−564. (38) Holzapfel, C.; Vienken, J.; Zimmermann, U. J. Membr. Biol. 1982, 67, 13−26. (39) Mischel, M.; Pohl, H. J. Biol. Phys. 1983, 11, 98−102. (40) Whitesides, G. M.; Ostuni, E.; Takayama, S.; Jiang, X.; Ingber, D. E. Annu. Rev. Biomed. Eng. 2001, 3, 335−373. (41) Botstein, D.; Fink, G. Science 1988, 240, 1439−1443. (42) Morgan, H.; Green, N. G. AC Electrokinetics: Colloids and Nanoparticles; Research Studies Press Ltd.: Philadelphia, PA, 2003. (43) Selimović, Š.; Dokmeci, M. R.; Khademhosseini, A. Curr. Opin. Pharmacol. 2013, 13, 829−833.
2395
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Controlled Rotation and Vibration of Patterned Cell Clusters Using Dielectrophoresis Rebecca Soffe,† Shi-Yang Tang,† Sara Baratchi,†,‡ Sofia Nahavandi,§ Mahyar Nasabi,† Jonathan M. Cooper,∥ Arnan Mitchell,† and Khashayar Khoshmanesh*,† †
School of Electrical and Computer Engineering, RMIT University, Melbourne, Victoria 3001, Australia Health Innovations Research Institute, RMIT University, Melbourne, Victoria 3083, Australia § Faculty of Medicine, Dentistry, & Health Sciences, The University of Melbourne, Melbourne, Victoria 3010, Australia ∥ The Bioelectronics Research Centre, Department of Electronics and Electrical Engineering, University of Glasgow, Glasgow G12 8LT, United Kingdom ‡
S Supporting Information *
ABSTRACT: The localized motion of cells within a cluster is an important feature of living organisms and has been found to play roles in cell signaling, communication, and migration, thus affecting processes such as proliferation, transcription, and organogenesis. Current approaches for inducing dynamic movement into cells, however, focus predominantly on mechanical stimulation of single cells, affect cell integrity, and, more importantly, need a complementary mechanism to pattern cells. In this article, we demonstrate a new strategy for the mechanical stimulation of large cell clusters, taking advantage of dielectrophoresis. This strategy is based on the cellular spin resonance mechanism, but it utilizes coating agents, such as bovine serum albumin, to create consistent rotation and vibration of individual cells. The treatment of cells with coating agents intensifies the torque induced on the cells while reducing the friction at the cell−cell and cell−substrate interfaces, resulting in the consistent motion of the cells. Such localized motion can be modulated by varying the frequency and voltage of the applied sinusoidal AC signal and can be achieved in the absence and presence of flow. This strategy enables the survival and functioning of moving cells within large-scale clusters to be investigated.
C
Microfabricated devices enable quick and accurate positioning of cells to be stimulated through micropipette aspiration,19 atomic force microscopy tip,20 or optical,21 acoustic,22 and magnetic tweezers23 under precisely controlled microenvironment conditions. Additionally, unique features of microfabricated/microfluidic systems have led to the development of a variety of novel techniques for the mechanical stimulation of cells, which could not be achieved using conventional macro-scale systems.5,8 This include dilation of cells in response to flow-induced shear stress,24 deformation of cells by flowing them through narrow channels,25 stretching of cells using deformable substrates such as silicone elastic membranes,26 and squeezing of cells using pneumatically actuated membranes.27 Furthermore, microfluidic devices enables the automation of multiple processes, such that the cells can be cultured, treated, and separated, before being dynamically analyzed during mechanical stimulation.28−32
ells can both sense and respond to the mechanical and chemical cues from their microenvironment.1,2 The signaling pathways triggered by mechanical cues can impact different cellular processes including proliferation, transcription, and organogenesis.1,3,4 A variety of techniques have been developed for the mechanical stimulation of cells,5−8 which, according to the extent and number of affected cells, can be divided into three groups. The first group includes cell poking,9 atomic force microscopy indentation,10 magnetic bead microrheometry,11 and magnetic twisting cytometry techniques,12 which enable local deformation of single cells. The second group includes micropipette aspiration,13 optical tweezing,14 and acoustic tweezing techniques,15 which enable deformation of an entire cell. The third group includes the application of hydrostatic pressure16 or inducing osmotic pressure, also known as hypotonic cell swelling techniques,17 which enable the mechanical stimulation of multiple cells at a time. Recent advances in microfabrication technologies, and in particular microfluidics, has realized devices that are smaller, simpler, cheaper, and more accurate than their macro-sized predecessors for the mechanical stimulation of cells.5−8,18 © 2015 American Chemical Society
Received: November 19, 2014 Accepted: January 22, 2015 Published: January 22, 2015 2389
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry
modulated by varying the frequency and voltage of the applied sinusoidal AC signal and can be achieved in the absence and presence of flow.
The majority of the above-mentioned techniques rely on passive immobilization of cells over the predetermined locations of the microdevice. The ability to actively immobilize the cells and to mechanically stimulate them using the same mechanism can accelerate and simplify the experimental process. Dielectrophoresis, the induced motion of polarizable particles in a nonuniform electric field, has been extensively used for sorting, trapping, and microscopic analysis of cells.33,34 However, dielectrophoresis has not been utilized for mechanical stimulation of cells. A closer look at the literature of dielectrophoresis reveals that cells can be rotated using electrorotation (ROT) or cellular spin resonance (CSR) mechanism. ROT refers to the rotation of cells when subjected to a rotating electric field produced by four-pole microelectrodes35 and has been extensively used for measuring the dielectric properties of cells. However, ROT is mostly used for rotation of single cells and not a cluster of cells, and, more importantly, it cannot be applied in the presence of flow, as the rotating cells will be washed away by the hydrodynamic drag force. In contrast, CSR refers to rotation of adjacent cells when subjected to an alternating electric field produced by two-pole microelectrodes. CSR was first reported by Pohl et al. in 1971,36 based on experiments conducted by Teixeira-Pinto et al. in the 1960s.37 However, an examination into the literature of CSR provides no complete experimental regime to produce consistent mechanical stimulation.36,38,39 Therefore, revisiting CSR may provide opportunities to develop a mechanism capable of clustering and mechanically stimulating large numbers of cells in the presence of flow. In this article, we present a strategy for the mechanical stimulation of customized cell clusters using dielectrophoresis (Figure 1). This strategy is based on CSR mechanism but
■
MATERIALS AND METHODS Fabrication of Microelectrodes. Gold on chrome films were deposited onto a 100 μm thick glass substrate at thicknesses of 1500 and 500 Å, respectively, using physical vapor deposition. The microelectrodes were then patterned using standard photolithography and etching techniques. Fabrication of Microchannel. The microchannel was fabricated from poly(dimethysiloxane) (PDMS) utilizing soft lithography and replica molding techniques.40 To create the PDMS structure, PDMS base and curing agent (Sylgard 184, Dow Corning Corporation, MI) were mixed with a ratio of 10:1 w/w and degassed until all of the bubbles disappeared. The dimensions of the PDMS microchannel were set to 500 × 80 μm (W × H). Cell Suspension Preparation. Saccharomyces cerevisiae yeast cells were used in this work to investigate the potential of using dielectrophoresis to invoke dynamic movement. These cells are an important model organism to understand the biology of eukaryotic cells at the cellular and molecular levels and have been widely used for studying cell biology, genetics, and genomics.41 In addition, these cells can be easily cultivated, treated, and used in bioengineering laboratories and thus are widely used in cellular studies involving dielectrophoresis.33 For preparation of cell suspension, 25 mg of S. cerevisiae yeast powder (Sigma-Aldrich) was suspended in an isotonic low electrical conductivity buffer of 100 mL deionized water, 8.5% w/v sucrose, and 0.3% dextrose to yield a cell concentration of 0.025% w/v (corresponding to 1.5 × 105 cells/mL). To avoid agglomeration, the suspension was placed in an ultrasonic bath for 15 min. The resulting cell suspension had an electrical conductivity of approximately 0.006 S/m, which was adjusted accordingly using phosphate buffer saline (PBS) (SigmaAldrich). The electrical conductivity was measured using an ECTester11+ conductivity meter (Eutech Instruments, Singapore). Treating Cells with Coating Agents. The dynamic response of cells was examined by coating them with four coating agents, including bovine serum albumin (BSA), poly-Llysine, poly-L-ornithine, and whey protein concentrate, as explained below. The liquid BSA suspension (50 mg/mL, Life Technologies) was diluted initially with either PBS or deionized water. This solution was then added to the cell suspension such that the final concentration of BSA was 0.2% w/v. The existence of BSA over the cells was confirmed by treating cells with labeled BSA (Alexa Fluor 488 conjugate, Life Technologies) solution and exciting them by a 488 nm laser using an inverted microscope (Nikon Eclipse Ti) (Supporting Information S1). The poly-L-lysine and poly-L-ornithine solutions (0.1% w/v, Sigma-Aldrich) were diluted in deionized water and added to the cell suspension to yield a final concentration of 2.0 × 10−5% w/v. Whey protein concentrate manufactured from sweet whey was purchased from a commercial supplier (80% w/w, Eatme supplements). The concentrate was added to deionized water and diluted in the cell suspension to yield a final concentration of 0.2% w/v. The viability of treated cells was examined using of a membrane-impermeant nucleic acid stain such as propidium iodide (PI) (Life Technologies). Viability assays indicated that treatment of cells with coating agents does not
Figure 1. Three-dimensional schematic of the localized motion of patterned cell clusters under the influence of dielectrophoresis. Cells can exhibit rotational and vibrational movements according to their location within the cluster. Cells located at the free ends of pearl chains have the highest occurrence of rotation, which are indicated as blue cells in the schematic. Alternatively, the cells packed along the long chains bridged between the microelectrodes have the highest occurrence of vibration. The arrows indicate the direction and strength of the vibrational movement of the cells.
utilizes coating agents, such as bovine serum albumin (BSA), to produce consistent mechanical stimulation of cells. The incubation of cells with coating agents increases the torque induced on the cells, whereas it reduces the friction at the cell− cell and cell−substrate interfaces, resulting in the rotation and vibration of the cells. Such mechanical motions can be 2390
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry have an adverse effect on the viability of cells (Supporting Information S2).
■
RESULTS AND DISCUSSION Cluster Formation. The first series of experiments was conducted using an open-top dielectrophoresis platform, which consisted of a PDMS cell chamber integrated onto the glass slide accommodating the microelectrode arrays. This enabled us to study the response of BSA-treated cells in the absence of flow. In doing so, 100 μL of the BSA-treated yeast suspension was transpired onto the PDMS chamber. The electrical conductivity of the suspension was set to 0.02 S/m through the addition of PBS. The microelectrodes were then activated using a 10 MHz, 5 Vpk−pk AC sinusoid signal, which led to the immobilization of cells along the microelectrode edges. Next, the frequency of the applied signal was sequentially increased in 5 MHz increments to 50 MHz, in 5 min intervals. This significantly changed the overall configuration of the BSA-treated cell clusters patterned along the microelectrodes, as discussed below. At low frequencies (f < 10 MHz), dense clusters of cells formed along the both internal and external edges of microelectrodes, similar to the case of untreated cells (Figure 2a). At medium frequencies (10 < f < 30 MHz), the cells patterned along the internal edges of microelectrodes left the surface of microelectrodes and lined up along the previously patterned cells, resulting in the formation of pearl chains stemming at the external edges of microelectrodes (Figure 2b). At higher frequencies (f > 30 MHz), the chains were further elongated and the intercellular distance between the cells increased, leading to formation of loosely joined cell chains along the electric field (Figure 2c). In contrast, untreated cells exhibited negative dielectrophoretic (DEP) response and were repelled from the microelectrodes at such high frequencies. Interestingly, for the case of treated cells, increasing the applied voltage led to a reduction of the intercellular gap between adjacent cells, whereas the patterned cells exhibited rotational and vibrational responses, as described in the succeeding sections. Dynamic Cell Clusters. To further examine this dynamic response, the voltage was increased to 10 Vpk−pk while the frequency was set to 36 MHz. Under these conditions, the BSA-treated cells started to rotate (Supporting Information Movie 1). The captions for the movies are given in Supporting Information S3. Cells that exhibited a rotational behavior were generally close to the free end of a pearl chain comprising of several cells. The reference image, Figure 3a, indicates the location of three randomly selected cells within the electric field, positioned at various locations within a pearl chain. A time-lapse of snapshots taken at 10 ms intervals is used in Figure 3b to highlight the rotational response of cells diagrammatically. The cells identified as i, ii, and iii rotated at 1.5, 0.5, and 1.2 rev/s, respectively. Monitoring the rotational behavior of the cells indicated that rotation is sporadic and that each cell has a different rotational axis and velocity. The rotational axis varied between cells such that the axis could be aligned or unaligned with the electric field; also, cells would rotate in either clockwise or anticlockwise directions, as observed in Figure 3b. This sporadic behavior could be potentially attributed to various factors, including the heterogeneity of the patterned cells, the inconsistency of BSA layer coating the cells, the electric field experienced by the cells, and the number of cells located along the pearl chain. In
Figure 2. Response of BSA-treated yeast cell clusters, patterned onto a finger-shaped microelectrode array when operated with a 5 Vpk−pk AC sinusoid signal. (a-i, b-i, c-i)Response of BSA-treated cell clusters at 5, 20, and 40 MHz, respectively. (a-ii, b-ii, c-ii) Schematic representation of the response of patterned cell clusters at different frequencies; with BSA-treated cells, represented as green, being compared to untreated cells, represented as blue. Evidently, the BSA-treated cells exhibit a positive DEP response as the distance between the two adjacent cells increases at high frequencies. Alternatively, the untreated cells exhibit a negative DEP response at frequencies higher than 30 MHz. Scale bar is 50 μm.
addition, our experiments indicated that the ability of a cell to rotate was not significantly affected by its location within the electric field. To further examine the effect of BSA coating on the dynamic response of the cells, the cells were kept in the BSA suspension for 120 min before being applied to our DEP system. This increased the dynamic response of cells such that more cells within the cluster could rotate. However, the cells that were located in the middle of the cluster, were surrounded by several cells on different sides, thus, could not freely rotate due to an increased intercellular friction and instead exhibited a vibrational motion. The vibration of cells was more tangible along the cells bridging between the opposite microelectrodes, making pearl bridges (Supporting Information Movie 2). The elongation of the BSA incubation process did not affect the mortality rate of the cells significantly (Supporting Information S2). Through observation, it was apparent that the overall dynamic movement of the cell cluster was governed by both the cluster configuration and the extent of both the rotational and vibrational cells within the cluster (Figure 3c,d). The extent 2391
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry
(W × H) was integrated onto the glass substrate containing the microelectrode array. To pattern the cells along the microelectrodes, the cell suspension was applied through the microchannel at a flow rate of 2 μL/min while the microelectrodes were energized by applying a 10 Vpk−pk sinusoidal AC signal at 10 MHz. To enable cell rotation, the voltage was decreased to 8 Vpk−pk while the frequency was increased to 70 MHz. The presence of flow facilitated the formation of pearl chains concentrated around the tip of the microelectrodes, thus increasing the likelihood of cells exhibiting rotation when located at the free end of the pearl chains. Four cells presented in Figure 4 display the rotational response of the cells in
Figure 3. Rotational and vibrational movement of patterned cells in the absence of flow, when operating the microelectrodes with a 36 MHz, 10 Vpk−pk AC sinusoidal signal. (a) The reference image used to identify the location of three rotating cells. (b: i−iii) The extent of the rotational movement can be observed through the use of snapshots taken at 10 ms intervals, as presented for three cells in this case (Supporting Information Movie 1). The rotational velocity of cells can be obtained by measuring the displacement of the yellow lines along the yellow circles. (c) The reference image used to highlight the structure of a vibrating cell cluster. (d) The outline of the cells was obtained every 2 s to highlight the vibration of the cell cluster (Supporting Information Movie 2). Scale bar is 20 μm in a, 30 μm in c and d, and 5 μm in b.
Figure 4. Rotational and vibrational movement of patterned cells in the presence of flow, when operating the microelectrodes with a 70 MHz, 8 Vpk−pk AC sinusoidal signal. The extent of movement of individual cells can be observed through the use of snapshots taken at 10 ms intervals, as presented for four cells in this case (Supporting Information Movie 3). The yellow arrows indicate the direction of flow through the microfluidic channel. Scale bar is 30 μm.
snapshots taken every 10 ms from Supporting Information Movie 3. The identified cells rotated consistently at 1.8, 0.8, 1.0, and 1.5 rev/s, from left to right. Moreover, the cells that were bridged between the microelectrodes exhibited a vibrational motion following a 120 min treatment with BSA. The ability to induce rotational and vibrational movement onto the patterned cells in the presence of flow enables more control over the configuration and density of cell clusters and can facilitate several experimental scenarios. Modeling Characteristics. A theoretical model was developed to determine the rotational velocity of BSA-treated cells. The model consists of two identical cells in contact with each other, which are exposed to an external AC electric field, as depicted in Supporting Information S4. The torque exerted on a cell (τrotation) within an electric field in the presence of an adjacent cell is adopted from the equation developed by Holzapfel et al.38 Such torque, should overcome the retarding torques generated by the surrounding viscous liquid (τliquid) and the friction at the cell−cell (τcell−cell) and cell−substrate (τcell−substrate) interfaces, as given below:
of the movement of the vibrating cell cluster was best observed through comparing the outline of the cells within the electric field. Thus, snapshots acquired at 2 s intervals from Supporting Information Movie 2 were imported into Adobe Photoshop (CS6, CA) to obtain the outline of the cells. The outlines were then superimposed on top of each other, as presented in Figure 3d, which highlighted that vibration is considerably more dynamic across the pearl bridges compared to that in pearl chain clusters. The rotation or vibration of cells could be sustained for 45 min (longer periods were not examined in our experiments). The rotational behavior of individual cells could vary over time due to local displacement of cells, but no significant difference was observed in the overall dynamic response of the patterned cell cluster. The viability of cells was examined after each experiment, including the experiments carried out for 45 min by on-chip PI staining of cells.28 No significant difference was observed between the viability of the control (BSA-treated cells, without DEP) and dynamic (BSA-treated cells, with DEP) samples, as shown in Figure S2. The application of a lowamplitude, high-frequency sinusoidal AC signal minimized the cell damage associated with Joule heating of the medium and the effects of inducing transmembrane potential.33,42 In addition, we examined the dynamic response of BSAtreated cells in the presence of flow. In doing so, a PDMS microchannel with cross-sectional dimensions of 500 × 80 μm
⟨τrotation⟩ = τliquid + τcell − cell + τcell − substrate
(1)
Substituting the equations for each of the above torques, the following equation is obtained, using which the constant rotational velocity of the cells at equilibrium can be achieved: P2 × E −
⎛ P. D ⎞ 1 × ∇⎜ 1 3 ⎟ ⎝ D ⎠ 4πε
= 8πηωr 3 + ξcell − cellr 2ω + ξcell − substrater 2ω 2392
(2)
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry
Figure 5. Variations of the rotational velocity of BSA-treated cells with respect to the voltage and frequency of AC sinusoidal signal obtained in the absence of flow. Measured and modeled data are presented as solid and dashed lines, respectively. (a) The frequency response indicated that the highest rotational velocity of cells occurs at a frequency range of 30−40 MHz when operating at 10 Vpk−pk (Supporting Information Movie 4). (b) The voltage response of cells indicated that the highest rotational velocity of cells occurs at a voltage range of 12−14 Vpk−pk when operating at 36 MHz (Supporting Information Movie 5).
The first component of τrotation appearing on the left-hand side of this equation represents the torque that acts to align the dipole of the rotating cell (cell-2, as shown in Figure S4) with the field,42 whereas the second component represents the torque contributed from the dipole of the stationary cell (cell-1, as shown in Figure S4) on the rotating cell (cell-2).38 P1 and P2 are the dipole moments of the stationary and rotating cells, respectively, D is the vector connecting the center points of the adjacent cells; E is the external electric field; ε and η are the permittivity and viscosity of surrounding medium, respectively; r and ω are the radius and rotational velocity of the cell, respectively; and ξcell−cell and ξcell−substrate are the friction coefficient at the interface of cell−cell and cell−substrate, respectively. However, both ξcell−cell and ξcell−substrate are difficult to obtain due to the nature of the system; thus, to enable simulated results to be obtained, τrotation was multiplied by a factor of 0.8 to account for these frictional losses. A detailed explanation of the resulting torque and the induced cell rotation is presented in Supporting Information S4. Characterization of the Dynamic Response. The dynamic response of BSA-treated cells was characterized through measuring the rotational velocity of individual cells at different frequencies and voltages of the applied signal as well as applying various coating agents and cell concentrations, as described below. Rotational velocity is defined as the number of cell rotations per second, ignoring the direction of rotation. Frequency Characteristics. The rotational response of BSAtreated cells was examined at a frequency range of 10−70 MHz, with the voltage of the sinusoid signal kept at 10 Vpk−pk. Figure 5a shows the variations of the average rotational velocity of the cells with respect to the applied frequencies (Supporting Information Movie 4). A resonant frequency was observed to occur at approximately 32 MHz according to the balance between the rotational and resisting torques acting on the cells, as described in eq 1. The rotational velocity of cells obtained from our theoretical model is shown by dashed line. A 10 MHz difference is observed between the experimental and simulated resonance frequencies (Figure 5a). This can be attributed to several
factors, including the possibilites that the pearl chains may contain more than two cells in experiments, the yeast cells may have an ellipsoidal shape rather than the spherical shape assumed in our model, and the intercellular gap may vary at different frequencies while it is assumed to be constant in our model (Figure 2). Thus, the actual dipole moments from adjacent cells are more complicated than the simulated scenario. Although our model is simplified without considering these parameters, it was able to predict the trend of the rotational response of the cells with respect to the frequencies of the applied signal. Voltage Characteristics. The rotational characteristics of BSA-treated cells influenced by voltage were determined at a frequency of 36 MHz (Figure 5b), very close to the resonance frequency observed in Figure 5a. Surprisingly, the rotational velocity reached its peak at 13 Vpk−pk and is reduced at higher voltages (Supporting Information Movie 5). The rotational velocity of cells obtained from our theoretical model is shown by a dashed line (Figure 5b). The model predicted the increase of rotational velocity with respect to the applied voltage, but it failed to predict the decrease of rotational velocity at higher voltages. This contradiction could potentially be attributed to strengthening of the DEP force at high applied voltages, which results in a reduced intercellular gap. This not only can disrupt the condition of loosely joined cells, which is essential for the rotation of cells, but also can increase the intercellular friction. This can also change the magnitude of the torque induced from the dipole of stationary cells on the rotating cells located at the free end of pearl chains, as given in eq 2. However, in our model, the intercellular gap was assumed to be constant across the entire voltage range. Gap between Microelectrode Pairs. In addition, we investigated the effect of the distance between the microelectrode pairs on the rotational response of the patterned cells. We used a DEP array with variable gap between its consequential microelectrode pairs. Our experiments indicated that the gap between the microelectrode pairs can influence the rotational characteristics of the patterned cells, as presented in Supporting Information Movie 6. This behavior was observed 2393
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry
in a reproducible manner. The movement of cells can be achieved in the absence and presence of flow and using different coating gents. Future work includes assessing the capability of our approach for the controlled rotation and vibration of mammalian cells. This can lead to a versatile microfluidic platform for studying the functioning (proliferation, metabolism, and death) of dynamic cells. It also enables the uptake and response of dynamic cells to chemicals (e.g., drugs) or particles (e.g., nanoparticles) to be investigated.43 Considering that cells constituting several tissues and organs of human body might be dynamic, our approach enables more realistic in vitro models. Furthermore, the rotation of cells can induce a moderate shear stress over their membrane (see Supporting Information S6), which can be used to study the shear-induced calcium signaling of cells.
because the distance between the microelectrode pairs determines the induced electric field; thus, in turn, can change the balance between the DEP force and the rotating torque acting on the cells, as further explained in Supporting Information S5. Coating Agents. In addition to BSA, additional coating agents were investigated, including poly-L-lysine, poly-Lornithine, and whey protein concentrate. We found two major differences in the capability of different coating agents to induce dynamic motion when operating the system using a 10 Vpk−pk 36 MHz sinusoid signal (Supporting Information Movie 7). Firstly, incubating of cells with poly-L-lysine, poly-Lornithine, and whey protein concentrate facilitated rotation of 16 ± 3%, 17 ± 4%, and 21 ± 4% of the cells, respectively, which was considerably less than what was achieved with BSA-treated cells, where around 36 ± 6% of the cells could rotate. Secondly, incubation of cells with BSA for more than 120 min intensified the dynamic response of cells and enabled the cells to rotate or vibrate according to their location within the cluster (Supporting Information Movies 2 and 3); whereas, elongating the incubation process with poly-L-lysine, poly-L-ornithine, and whey concentrate protein neither intensified the dynamic response of treated cells nor led to vibration of cells. To further understand the interaction between the yeast cells and the coating agents, the cells were incubated in BSA for 10 min; then, they were washed twice and suspended in PBS before being transpired onto the DEP platform. The movement observed proved to be indifferent to the coating agents present in the suspension medium, which we conjecture could be, in fact, that the coating agents potentially modify the surface properties of the cells. In addition, our experiments suggested that to enable the occurrence of dynamic cell clusters the concentration of the coating agent should result in an overall cell suspension conductivity within 0.01 and 0.03 S/m. However, the upper conductivity limit can be extended through the addition of PBS up to 0.15 S/m. The lower conductivity was determined as the threshold to induce the dynamic response of cell clusters, whereas the higher conductivity is the threshold to avoid the formation of bubbles at the surface of microelectrodes. Yeast Cell Concentration. The optimal concentration of yeast cells was determined to be 1.5 × 105 cells/mL. The time required to immobilize cells significantly increased at a lower cell concentration. In contrast, increasing the cell concentration led to the formation of dense cell clusters between the microelectrodes. This reduced the possibility of rotation; however, it increased the percentage of cells vibrating within the cluster and resulted in a large dynamic cell cluster.
■
ASSOCIATED CONTENT
S Supporting Information *
Section 1: Coating yeast cells with BSA. Section 2: Viability assays. Section 3: Captions for seven supplementary movies. Section 4: Theoretical model for cell rotation. Section 5: Rotation of BSA-treated cells using a DEP array with different microelectrode gaps. Section 6: Comparing the shear stress induced by cell rotation. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS K. Khoshmanesh acknowledges the Australian Research Council for funding under an Discovery Early Career Researcher Award (DECRA) (project DE120101402).
■
REFERENCES
(1) Polacheck, W. J.; Li, R.; Uzel, S. G.; Kamm, R. D. Lab Chip 2013, 13, 2252−2267. (2) Chingozha, L.; Zhan, M.; Zhu, C.; Lu, H. Anal. Chem. 2014, 86, 10138−10147. (3) Bukoreshtliev, N.; Haase, K.; Pelling, A. Cell Tissue Res. 2013, 352, 77−94. (4) Bastock, R.; Strutt, D. Development 2007, 134, 3055−3064. (5) Delmas, P.; Hao, J.; Rodat-Despoix, L. Nat. Rev. Neurosci. 2011, 12, 139−153. (6) van Loon, J. W. A. Microgravity Sci. Technol. 2009, 21, 159−167. (7) Guck, J.; Lautenschläger, F.; Paschke, S.; Beil, M. Integr. Biol. 2010, 2, 575−583. (8) Bao, G.; Suresh, S. Nat. Mater. 2003, 2, 715−725. (9) Daily, B.; Elson, E. L.; Zahalak, G. I. Biophys. J. 1984, 45, 671− 682. (10) A-Hassan, E.; Heinz, W. F.; Antonik, M. D.; D’Costa, N. P.; Nageswaran, S.; Schoenenberger, C.-A.; Hoh, J. H. Biophys. J. 1998, 74, 1564−1578. (11) Bausch, A. R.; Ziemann, F.; Boulbitch, A. A.; Jacobson, K.; Sackmann, E. Biophys. J. 1998, 75, 2038−2049. (12) Wang, N.; Ingber, D. E. Biochem. Cell Biol. 1995, 73, 327−335. (13) Hochmuth, R. M. J. Biomech. 2000, 33, 15−22. (14) Lim, C.; Dao, M.; Suresh, S.; Sow, C.; Chew, K. Acta Mater. 2004, 52, 1837−1845. (15) Kundu, T.; Bereiter-Hahn, J.; Hillmann, K. Biophys. J. 1991, 59, 1194−1207.
■
CONCLUSIONS AND FUTURE OUTLOOK In this article, we have demonstrated a new approach for the mechanical stimulation of large-scale cell clusters. In this approach, the cells are treated with coating agents such as BSA before being applied into a dielectrophoresis platform. Dielectrophoresis facilitates the rapid immobilization of cells and formation of cell chains along the microelectrodes. Increasing the frequency of the applied signal to ∼30 MHz leads to formation of loosely joined cell chains. Under these conditions, the cells that are located along the free end of the cell chains can rotate, whereas the interaction of multiple rotating cells can lead to vibration of the entire cell chain. The dynamic response of the patterned cells can be modulated through varying the voltage and frequency of the applied signal, 2394
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
Article
Analytical Chemistry (16) Kim, S. H.; Choi, Y. R.; Park, M. S.; Shin, J. W.; Park, K. D.; Kim, S. J.; Lee, J. W. J. Biomed. Mater. Res., Part A 2007, 80, 826−836. (17) Cook, B.; Hardy, R. W.; McConnaughey, W. B.; Zuker, C. S. Nature 2008, 452, 361−364. (18) Baratchi, S.; Khoshmanesh, K.; Sacristán, C.; Depoil, D.; Wlodkowic, D.; McIntyre, P.; Mitchell, A. Biotechnol. Adv. 2014, 32, 333−346. (19) Lee, L. M.; Liu, A. P. Lab Chip 2015, 15, 264−273. (20) Rosenbluth, M. J.; Lam, W. A.; Fletcher, D. A. Biophys. J. 2006, 90, 2994−3003. (21) Honarmandi, P.; Lee, H.; Lang, M. J.; Kamm, R. D. Lab Chip 2011, 11, 684−694. (22) Fan, Z.; Sun, Y.; Chen, D.; Tay, D.; Chen, W.; Deng, C. X.; Fu, J. Sci. Rep. 2013, 3, 2176 . (23) Zhao, R.; Boudou, T.; Wang, W.-G.; Chen, C. S.; Reich, D. H. J. Appl. Phys. 2014, 115, 172616. (24) Baratchi, S.; Tovar-Lopez, F. J.; Khoshmanesh, K.; Grace, M. S.; Darby, W.; Almazi, J.; Mitchell, A.; McIntyre, P. Biomicrofluidics 2014, 8, 044117. (25) Lam, W. A.; Rosenbluth, M. J.; Fletcher, D. A. Blood 2007, 109, 3505−3508. (26) Gilchrist, C. L.; Witvoet-Braam, S. W.; Guilak, F.; Setton, L. A. J. Biomech. 2007, 40, 786−794. (27) Xu, T.; Yue, W.; Li, C.-W.; Yao, X.; Cai, G.; Yang, M. Lab Chip 2010, 10, 2271−2278. (28) Khoshmanesh, K.; Akagi, J.; Nahavandi, S.; Skommer, J.; Baratchi, S.; Cooper, J. M.; Kalantar-Zadeh, K.; Williams, D. E.; Wlodkowic, D. Anal. Chem. 2011, 83, 2133−2144. (29) Lapizco-Encinas, B. H.; Simmons, B. A.; Cummings, E. B.; Fintschenko, Y. Anal. Chem. 2004, 76, 1571−1579. (30) Li, S.; Guo, F.; Chen, Y.; Ding, X.; Li, P.; Wang, L.; Cameron, C. E.; Huang, T. J. Anal. Chem. 2014, 86, 9853−9859. (31) Zhan, M.; Chingozha, L.; Lu, H. Anal. Chem. 2013, 85, 8882− 8894. (32) Hou, H. W.; Warkiani, M. E.; Khoo, B. L.; Li, Z. R.; Soo, R. A.; Tan, D. S.-W.; Lim, W.-T.; Han, J.; Bhagat, A. A. S.; Lim, C. T. Sci. Rep. 2013, 3, 1259 . (33) Khoshmanesh, K.; Nahavandi, S.; Baratchi, S.; Mitchell, A.; Kalantar-zadeh, K. Biosens. Bioelectron. 2011, 26, 1800−1814. (34) Becker, F. F.; Wang, X. B.; Huang, Y.; Pethig, R.; Vykoukal, J.; Gascoyne, P. R. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 860−864. (35) Hölzel, R.; Lamprecht, I. Biochim. Biophys. Acta, Biomembr. 1992, 1104, 195−200. (36) Pohl, H. A.; Crane, J. S. Biophys. J. 1971, 11, 711−727. (37) Teixeira-Pinto, A. A.; Nejelski, L. L., Jr.; Cutler, J. L.; Heller, J. H. Exp. Cell Res. 1960, 20, 548−564. (38) Holzapfel, C.; Vienken, J.; Zimmermann, U. J. Membr. Biol. 1982, 67, 13−26. (39) Mischel, M.; Pohl, H. J. Biol. Phys. 1983, 11, 98−102. (40) Whitesides, G. M.; Ostuni, E.; Takayama, S.; Jiang, X.; Ingber, D. E. Annu. Rev. Biomed. Eng. 2001, 3, 335−373. (41) Botstein, D.; Fink, G. Science 1988, 240, 1439−1443. (42) Morgan, H.; Green, N. G. AC Electrokinetics: Colloids and Nanoparticles; Research Studies Press Ltd.: Philadelphia, PA, 2003. (43) Selimović, Š.; Dokmeci, M. R.; Khademhosseini, A. Curr. Opin. Pharmacol. 2013, 13, 829−833.
2395
DOI: 10.1021/ac5043335 Anal. Chem. 2015, 87, 2389−2395
