Characterizing Deformability and Electrical Impedance of Cancer

Subscriber access provided by READING UNIV

Article

Characterizing Deformability and Electrical Impedance of Cancer Cells in a Microfluidic Device Ying Zhou, Dahou Yang, Yinning Zhou, Bee Luan Khoo, Jongyoon Han, and Ye Ai Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b03859 • Publication Date (Web): 27 Nov 2017 Downloaded from http://pubs.acs.org on November 28, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Characterizing Deformability and Electrical Impedance of Cancer Cells in a Microfluidic Device Ying Zhou,†,∥ Dahou Yang,‡,∥ Yinning Zhou,‡ Bee Luan Khoo,† Jongyoon Han,*,†,§ Ye Ai*,‡

†

BioSystems and Micromechanics IRG (BioSyM), Singapore-MIT Alliance for Research and

Technology (SMART) Centre, Singapore 138602 ‡

Pillar of Engineering Product Development, Singapore University of Technology and

Design, Singapore 487372 §

Department of Electrical Engineering and Computer Science, and Department of Biological

Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

∥These

authors contributed equally to this work.

*Corresponding authors: Jongyoon Han ([email protected]); Ye Ai ([email protected]).

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract Mechanical properties of cells, reflective of various biochemical characteristics such as gene expression and cytoskeleton, are promising label-free biomarkers for studying and characterizing cells. Electrical properties of cells, dependent on the cellular structure and content, are also label-free indicators of cell states and phenotype. In this work, we have developed a microfluidic device that is able to simultaneously characterize the mechanical and electrical properties of individual biological cells in a high-throughput manner (> 1,000 cells/min). The deformability of MCF-7 breast cancer cells was characterized based on the passage time required for an individual cell to pass through a constriction smaller than the cell size. The total passage time can be divided into two components: the entry time required for a cell to deform and enter a constriction, which is dominated by the deformability of cells; and the transit time required for the fully deformed cell to travel inside the constriction, which mainly relies on the surface friction between cells and channel wall. The two time durations for individual cells to pass through the entry region and transit region have both been investigated. In addition, undeformed cells and fully deformed cells were simultaneously characterized via electrical impedance spectroscopy technique. The combination of mechanical and electrical properties serves as a unique set of intrinsic cellular biomarkers for single cell analysis, providing better differentiation of cellular phenotypes, which are not easily discernible via single marker analysis.


Page 2 of 22

Page 3 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction Cell deformability is related to many molecular changes and can be used to phenotype cell populations.1 It has been proved as a useful mechanical biomarker for the diagnosis of various diseases, such as cancer, malaria, diabetes mellitus and sickle cell anemia. For example, cancel cells are reported to be more deformable than healthy cells and thus are more likely to invade surrounding tissues.1 A variety of techniques have been developed to probe cell deformability at the single cell level, such as atomic force microscopy (AFM) and micropipette aspiration.2,3 The common issue of the above conventional tools is that they are of low throughput (less than one cell per minute), therefore unable to address cell populations with minority subpopulations of interest. Recently, optical stretchers using laser beams have been developed to measure the mechanical property of single cells.4 Compared with AFM and micropipette, although optical stretchers allow contactless measurements, its throughput is still quite low (about one cell per minute). In addition, a direct exposure to the laser beam may cause irreversible damages to biological cells.

Microfluidics has great potential to provide alternative approaches for probing cell deformability with high throughput and low complexity. The microfluidics-based techniques for cell deformability characterization can be mainly divided into two categories: hydrodynamics-based

techniques,5-10

and

constriction-based

techniques.11-19

The

hydrodynamic stretching technique employs the shear stress of fluid flow to deform cells, allowing high throughput cell deformability assessment.7,8,10 The deformation of the cell is usually recorded with high-speed cameras, and the cell deformability is determined by evaluating the elongation and compression of the cell as compared to its original diameter. However, this technique requires expensive high-speed imaging setup and intricate image processing algorithms. In addition, high-speed camera produces large amounts of imaging data for post analysis that requires high-end computation time, making this approach difficult to realize real-time cell detection and, more importantly, deformability-based cell sorting. Alternatively, constriction-based methods are based on the assessment of the passage time for a cell to pass through a constriction or a constriction array that is smaller than the diameter of the cell. Currently, the techniques used for detecting and extracting passage time can be classified mainly into three categories. The first approach involves using high-speed cameras to capture sequential images of cells passing through a constriction, following which the 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 22

passage time can be extracted by tracking the cell positions.14,19 Again, the utilization of highspeed camera makes the setup expensive and bulky, and increases the difficulty for real-time cell analysis and subsequent cell sorting. The second technique for time quantification is based on integrating a constriction near the end of a suspended microchannel resonator (SMR) and tracking the change in the resonant frequency of the cantilever while a single cell flows through the constriction.12 The shift in the resonant frequency can provide information on the velocities of a cell entering the constriction and transiting through the constriction. A few thousand cells can be measured per hour. However, the fabrication of this device is very intricate, and it is impossible to directly observe the cell movement as the microchannel embedded inside the cantilever is not transparent. The third technique for passage time measurement is based on electrical detection, for example, electrical impedance spectroscopy (EIS).13,15,16,18

This

label-free

technique

enables

real-time

and

high-throughput

measurements,15,20 and eases the process of data extraction and processing. Usually, in such devices, two electrodes are placed at each side of a constriction. An input voltage is applied to one electrode, and the current is measured from the other electrode to calculate the electrical impedance. As a cell passes through the constriction, it replaces the medium between the two sensing electrodes, causing a change in the electrical impedance. The passage time can be extracted by calculating the time duration of the impedance change. However, this two-electrode non-differential configuration is only able to measure the impedance of deformed cells (inside the constriction), and less suitable for measuring the impedance of undeformed cells (before entering the constriction) which can yield useful information on the size and electrical properties of the whole cell. It is known that the passage time of a single cell passing through a constriction exhibits a power law dependence on cell size.12 The lack of capability to measure the whole-cell impedance makes the twoelectrode design not capable for cell size calibration. Thus, the dependence of passage time on the cell size cannot be easily obtained by this method. Moreover, it has been reported that the total passage time contains two components: the time required for the cell to deform for entering the constriction (mainly dominated by cell deformability) and the time required for the cell to transit inside the constriction (mainly dependent on surface friction).12 These two distinct time durations cannot be accurately extracted from electrical impedance measurements using the two-electrode design, which is typically employed in the current impedance-derived cell deformability measurement.13,15,16,18,20-22 On the other hand, approaches involving differential impedance measurements have been widely used in biological cell analysis for cell counting and sizing.23-26 Compared with non-differential 4 ACS Paragon Plus Environment

Page 5 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


configuration, differential measurements can cancel out any common-mode drift caused by the change in electrode properties or surrounding environment (e.g. pH, temperature, conductivity) and thus can achieve better signal stability and higher signal to noise ratio.23,27 However, the differential measurement technique has not been used for cell deformability measurement yet.

What is clear, based on what has been achieved so far both in electrical (impedance) and mechanical (deformability) characterization of cells, is that either of the two physical biomarkers may not be sufficiently specific or differentiating enough to characterize complex cell populations. In this work, the differential impedance measurement technique and the constriction-based cell deformability measurement technique were integrated together to simultaneously probe the mechanical and electrical properties of cells. The impedance of undeformed cells and deformed cells were both investigated for the first time. The total passage time of cells passing through a constriction was studied. Furthermore, the time for a single cell to squeeze into the constriction (defined as entry time) and the time for a cell to transit inside the constriction (defined as transit time) were characterized using electrical impedance-based method for the first time. By combining mechanical properties and electrical impedance information, four populations of cells (red blood cells, MCF-7, PMAmodified MCF-7 and fixed MCF-7) could be clearly distinguished from each other to the extent that is not possible by either of the two biomarkers alone. The developed microfluidic device enables reliable and fast single cell analysis based on mechanical and electrical properties of cells, and offers great potential for real-time cell sorting as this device can be easily integrated with cell sorting techniques, such as dielectrophoresis (DEP),28 and SAW (surface acoustic wave)-based cell sorting techniques.29

Experimental Section Device Concept and Design A schematic diagram showing the concept and design of the microfluidic device is illustrated in Figure 1. When a cell is squeezing through a constriction smaller than itself, it deforms and affects the passage time through the constriction. This passage time is related to cell deformability and surface friction, therefore can be used to characterize the mechanical 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

properties of single cells. In this work, the electrical impedance measurement is used to replace high-speed camera for rapid and accurate measurement of the passage time of single cells passing through a constriction. Inside the microfluidic channel, four pairs of coplanar electrodes are patterned on the bottom substrate for electrical impedance measurements (Figure 1). These electrodes enable the measurement of different time durations for a single cell to transit through different regions along the constriction.

Figure 1. Schematic diagram showing the principle of the microfluidic impedance device. (A) Electrode connection for measuring total passage time of a single cell entering and transiting through a constriction. (B) Electrode configuration for the measurement of entry time, i.e., the time for a single cell to squeeze into the constriction. (C) Electrode connection for the measurement of transit time, i.e., the time for a single cell to transit inside the constriction.

Differential impedance measurement scheme is employed. To measure the total passage time of a single cell passing through a constriction, the first pair and last pair of electrodes are used, as illustrated in Figure 1A. An input voltage (𝑉!" ) is applied to the middle two electrodes and currents are measured from the other two electrodes (𝐼! , 𝐼! ) to calculate the differential 6 ACS Paragon Plus Environment

Page 6 of 22

Page 7 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


impedance (∆ 𝑍 ). When a single cell passes through the first pair of electrodes, the medium between the electrodes is replaced by the cell, leading to an instantaneous change in the electrical current and thus the electrical impedance (i.e. the positive peak in Figure 1A). This peak indicates the starting time point of a single cell entering the constriction region. After passing through the first pair of electrodes, the cell enters the constriction and undergoes deformation. During this period, both pairs of electrodes measure the impedance of the medium, thus resulting in no differential signal (i.e. ∆ 𝑍 = 0). After the cell flows out of the constriction and passes through the last pair of electrodes, the cell replaces the medium between that pair of electrodes, causing another change in current and differential impedance (i.e. the negative peak in Figure 1A). This peak is an indicator that the cell has passed through the constriction region and serves as an ending time point of the passage time measurement. The electrical circuit model of a cell in a microfluidic channel and the detailed derivation of the differential impedance are provided in the Supporting Information and Figure S1.

The differential measurement scheme utilizes positive and negative peaks as indications of the starting and ending time points, respectively, when individual cells pass through the constriction. One positive peak must pair with a negative peak to indicate that single cells have passed through the constriction region. This feature makes real-time cell detection and cell sorting possible, as the passage time of single cells can be rapidly extracted from electrical signals using a simple peak detection algorithm in a real-time manner. The differential measurement strategy also enables easy recognition of the case when multiple cells travel through a constriction simultaneously (Supporting Information Figure S2), and thus, can provide more accurate and reliable measurement of impedance signal and passage time (as discussed in the Supporting Information), compared with the conventional nondifferential impedance measurement employed in most exiting studies.13,15,16,21,22

The total passage time of a cell passing through a constriction can be divided into two parts: entry time (i.e. the time of a cell entering a constriction) and transit time (the time for a cell transiting inside the constriction). By using the first two pair of electrodes (Figure 1B), entry time can be measured based on the same differential impedance sensing principle as discussed previously for the measurement of total passage time. Similarly, if the second and third pair of electrodes are connected in a differential configuration (Figure 1C), the transit 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

time of a fully deformed cell travelling inside a constriction can be determined, as well as the electrical impedance of the fully deformed cell.

Experimental Setup MCF-7 breast cancer cells were cultured in standard conditions. Fixed cells were prepared by fixing the cells in 4 % methanol-free formaldehyde; modMCF-7 cells were prepared by treating MCF-7 in 500 nM PMA (Phorbol 12-myristate 13-acetate) for 18 hours. Cells were then suspended in PBS for measurement, with a concentration of 0.5 ~ 1×106 cells/ml. Devices were fabricated using standard photolithography and PDMS soft photolithography processes. The fluidic flow in the microfluidic device was controlled with a pressure control system (Fluigent MFCS-EZ). Electrical impedance data were recorded by an impedance analyzer (HF2IS, Zurich Instruments). A input voltage with 500 mV amplitude and 1 MHz frequency was used. Details of the experimental protocols are provided in the Supporting Information.

Results and Discussion Normal and mechanical property-modified MCF-7 breast cancer cells were used to validate our microfluidic device. As a result, the channel dimensions were designed to fit the size of MCF-7 cells. The constriction is 10 µm in width, 20 µm in height and 560 µm in length. A snapshot of the impedance signal measured from one experiment is given in Figure 2A (containing 451 cell transit events), with the inset showing an enlarged view of the impedance signal for three successive cells. Here, the electrode configuration in Figure 1A was used, in which the first pair and last pair of electrodes were connected for the measurement of the total passage time of single cells flowing through a constriction. A positive pressure of 500 mbar was applied to the channel inlet and was used to drive the cells to squeeze through the constriction. The device used in this work could achieve a throughput of more than one thousand cells per minutes. However, it is noted that the throughput is dependent on various experimental conditions and thus can be further improved, for example, by using a higher flow rate and optimizing the concentration of cells.


Page 8 of 22

Page 9 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Sequential microscopic images of the cell positions at different time instants recorded by a high-speed camera are shown in Figure 2B, and corresponding electrical impedance signal of this cell transit event is illustrated in the time-domain impedance plot below. At t = 0 ms, the cell is located between the first two sensing electrodes before the constriction, resulting in an impedance change (i.e. the positive peak as shown in the signal plot below). This impedance change corresponds to the impedance of an undeformed cell. When the cell enters the constriction, the differential impedance between the first and last pairs of electrodes is zero, giving a flat signal line in the constriction region. When the cell flows out of the constriction and enters the last pair of electrodes, it results in another impedance change (the negative peak as shown in the signal plot below). The time duration between the two peaks is defined as the total passage time of a single cell passing through a constriction.

Figure 2. Examples of the differential impedance signal measured using the microfluidic impedance device. (A) Differential impedance of undeformed MCF-7 cells measured at 500 mbar flow pressure (containing 451 events). (B) Time-lapse microscopic images of a single cell flowing through the constriction were recorded by a high-speed camera. Corresponding electrical impedance signal of this cell transit event is illustrated in the plot below.

Cell Deformability Analysis The deformability of normal MCF-7, PMA-modified MCF-7 (i.e. modMCF-7) and fixed MCF-7 cancer cells were compared. Apart from cancer cells, we also measured red blood cells (RBCs) that are smaller than the constriction and they do not deform in the constriction. Therefore, the measurement of RBC can serve as a reference group for comparison with



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 22

MCF-7 cancer cells. Furthermore, these measurements can also demonstrate the possibility of using the proposed device for detecting circulating tumor cells (CTCs) in whole blood.

Passage time, entry time and transit time were investigated for all the four cell populations. Figure 3A shows the total passage time for cells passing through a constriction with a flow pressure of 500 mbar applied to the inlet. RBCs are smaller than the constriction and undergo no deformation, hence, their passage time through a constriction (3.61 ± 0.2 ms) is mostly determined by the flow rate in the channel. MCF-7 breast cancer cells are larger than the constriction and are forced to deform when they travel through the constriction. As a result, the time needed for them to pass through a constriction is significantly longer than that needed for RBCs. For the three populations of cancer cells (MCF-7, modMCF-7 and fixed MCF-7), there are significant differences (p < 0.001) among them, based on Mann-Whitney test. To illustrate the time difference between MCF-7 and modMCF-7 more clearly, a zoomin view of Figure 3, focusing on MCF-7 and modMCF-7 only, is provided in the Supporting Information Figure S3. A high-speed video demonstrating and comparing the transitions of different cancer cells through the same constriction is also provided in the Supporting Information, for visualization of the time difference amongst cells. Fixed MCF-7 cells exhibit the longest passage time (15.91 ± 7.74 ms), whereas modMCF-7 cells require the shortest passage time (6 ± 0.53 ms) to flow through the constriction. The passage time required for normal MCF-7 cells is in between (6.97 ± 0.56 ms). A different flow pressure (200 mbar) was also used and the trend in time difference among cell populations is consistent (Supporting Information Figure S4).

Figure 3. Passage time (A), entry time (B) and transit time (C) of cells flowing through a constriction, measured at 500 mbar pressure. Four cell populations were demonstrated: red blood cells (RBCs), modMCF-7, MCF-7 and fixed MCF-7. Cell number of RBCs, modMCF-7, MCF-7 and fixedMCF-7


Page 11 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


is: (A) 495, 342, 885 and 225; (B) 711, 460, 666 and 389; (C) 505, 302, 747 and 338. The numbers in the box charts indicate the mean values of time. *** indicates a p-value of less than 0.001, based on Mann-Whitney test.

When a cell is forced to squeeze through a constriction, its motion can be divided into two regions: the entry region where the cell squeezes into the constriction and transforms from a spherical shape to a deformed state; and the transit region where the deformed cell moves inside the constriction. It has been reported that cell deformability plays a more important role in determining the entry time, while the cell surface friction dominates the transit time inside the constriction.12 To further investigate the two properties of cells (deformability and surface friction), the entry time and transit time were measured and analyzed. From the entry time data (Figure 3B), it is found that RBCs required 0.91 ± 0.11 ms to enter the constriction without deforming itself. For the three different types of MCF-7 cells, the fixed cells need the longest time (5.02 ± 2.81 ms) to deform and enter a constriction; the normal MCF-7 cells take a moderate time (1.68 ± 0.24 ms); and the modMCF-7 cells require the shortest time (1.58 ± 0.2 ms). Therefore, the deformation process can substantially retard the entry into the constriction, and the ability to deform can be characterized by the entry time. The fixation process of MCF-7 cells used formaldehyde to cross-link the proteins in the cells and thus made the cells less deformable. The decreased deformability of fixed cells accordingly resulted in a longer entry time. On the other hand, compared with normal MCF-7 cells, modMCF-7 cells require less entry time to squeeze into a constriction, demonstrating that modMCF-7 cells are more deformable than their normal counterparts. This finding verifies the result reported in other literatures that PMA-modified MCF-7 cells exhibit an increased metastatic efficiency and invasiveness.4

After the three types of MCF-7 cells squeezed into the constriction, they were fully deformed and moved along the constriction without further deformation until exited out of the constriction. The transit time of fully deformed MCF-7 cells inside a constriction is shown in Figure 3C. As the reference for comparison, RBCs remained undeformed and were not in physical contact with the constriction during the transit process. It took RBCs 1.65 ± 0.08 ms to complete the transit process inside the constriction. For the three different types of MCF-7 cells, when they passed through the constriction channel, the forces exerted on them were the driving hydrodynamic force from the fluid-cell interaction and the retarding friction force 11 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 22

from the wall-cell interaction. As compared with RBCs, the additional surface friction force acting on the three MCF-7 cells significantly increased the transit time. Fixed MCF-7 cells again took the longest time (6.24 ± 2.96 ms) to transit inside the constriction, based on which it is concluded that fixed MCF-7 cells were subjected to the greatest retarding friction force during the transit process. The modMCF-7 cells traveled faster (3.17 ± 0.39 ms) than untreated MCF-7 cells (3.39 ± 0.45 ms), inferring that modMCF-7 cells exhibit less retarding surface friction. The decrease in surface friction of modMCF-7 cells makes the cells more invasive as these cells encounter less resistance to invade surrounding tissues. It should be noted that the surface friction is a multiplied result of the friction coefficient and the normal force acting on the cell against the channel wall.12 The surface morphology of the cell and the surface characteristics of the channel wall may influence the friction coefficient. The normal force acting on the cell against the channel wall, on the other hand, is dependent on the elastic mechanical property of the cell.19 In general, the more deformable the cell, the less normal force exerted on the channel wall, and vice versa. Therefore, the transit time inside a constriction is collectively governed by the frictional properties of the cell surface and the mechanical properties of the cell.

To sum up, when MCF-7 cells are treated with PMA, they become more deformable and exhibit less surface friction, thereby requiring less entry time, transit time and total passage time to pass through a constriction. Cell fixation process causes protein cross-linking that makes cells less deformable and may also change the frictional properties, therefore increases the time required for the fixed cells to enter and transit through a constriction.

Electrical Impedance Analysis Electrical impedance measurements of undeformed cells (before cells enter the constriction) and deformed cells (when cells are inside the constriction) were, respectively, recorded by employing the electrode configuration in Figure 1A and Figure 1C. Before entering the constriction, cells are intact and retain nearly spherical shape between the impedance sensing electrodes. The electrical impedance measured in this case is related to the electrical properties, but however dominated by the size of the whole cell. When a single cell squeezes into the constriction, it deforms to nearly block the entire cross-section of the constriction, and occupies most volume between two coplanar sensing electrodes. In this case, the total 12 ACS Paragon Plus Environment

Page 13 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


electrical current consists of the current through the cell membrane and intracellular structures, and the leakage current between the cell surface and channel wall. The electrical impedance measured in the second configuration thus reflects more information on the electrical properties of the cell and is less dependent on the original size of the cell. To our knowledge, this is the first time that the electrical impedance of undeformed cells and deformed cells are both characterized.

Figure 4A shows the impedance of cells at their initial states (i.e. undeformed states). RBCs exhibit the lowest impedance magnitude due to their smallest size. MCF-7 cancer cells are much larger than RBCs, thereby resulting in much larger impedance magnitudes. The three types of MCF-7 cells (MCF-7, modMCF-7 and fixed MCF-7) have a very similar size distribution (Supporting Information Figure S5), and therefore cannot be easily distinguished or separated from each other by size-based method, such as inertial cell sorting approach. However, by using electrical impedance-based approach, there is a significant difference (p < 0.001 based on Mann-Whitney test) in electrical impedance among the three different cell populations. The impedance measurement was performed at 1 MHz, at which the cell size and the electrical properties of the cell membrane collectively determine the overall electrical impedance of a single cell. Figure S5 demonstrates that the size difference among the three cell populations is negligible, implying that the huge difference in the electrical impedance among MCF-7, modMCF-7 and fixed MCF-7 cells primarily results from the change in the electrical properties of the cell membrane. The PMA-modification process might have caused a decrease in the membrane capacitance, thus increased the effective impedance magnitude of the whole cell. The cell fixation process, on the other hand, cross-linked proteins in cells and effectively changed the cell membrane properties. The disrupted cell membrane may play a role in decreasing the impedance magnitude of the whole cell.

The electrical impedance of fully deformed cells inside the constriction, however, shows a different trend (Figure 4B). RBCs still have the lowest impedance magnitude. Unlike RBCs, cancer cells that are larger than the constriction can largely block the entire cross-section of the constriction. Interestingly, the electrical impedance of the fully deformed fixed cells seems to be the largest. Our hypothesis for this phenomenon is that fixed MCF-7 cells form the tightest contact with the channel wall and generate the least leakage current flowing 13 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 22

between the cell and the channel wall, compared with the other two cell populations (MCF-7 and modMCF-7). Fixed MCF-7 cells are the least deformable among the three cancer cell populations, implying the largest Young's modulus. As the cross-section of the constriction has fixed dimensions, the three cell populations are subjected to the same amount of compression inside the constriction. Apparently, the deformed fixed MCF-7 exerted the largest normal force to the channel wall, as well as the largest surface friction as discussed previously. As a result, the gap between fixed MCF-7 cells and the channel wall is the thinnest. In other words, the fixed MCF-7 cells produce the best sealing effect of the constriction, leading to the lowest leakage current, thus the largest impedance.

Figure 4. Impedance of (A) undeformed cells before entering the constriction, and (B) fully deformed cells inside the constriction. The impedance was measured at 1 MHz frequency. Cell number of RBC, modMCF-7, MCF-7 and fixedMCF-7 is: (A) 495, 342, 885 and 225; (B) 505, 302, 747 and 338. *** indicates a p-value of less than 0.001, based on Mann-Whitney test.

One would probably argue that untreated MCF-7 should produce a larger impedance magnitude than modMCF-7 for the impedance measurement of deformed cells inside the constriction, because untreated MCF-7 is less deformable than modMCF-7. However, Figure 14 ACS Paragon Plus Environment

Page 15 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4B shows an opposite trend that the impedance of untreated MCF-7 inside the constriction is less than the impedance of modMCF-7. A possible explanation of this phenomenon is that the electrical properties of the cell membrane may have changed when cells squeeze through a constriction. Unlike fixed MCF-7 cells, both MCF-7 and modMCF-7 are live cells and thus they can respond to mechanical stimuli. It has been reported that transient membrane disruptions or holes may form when cells undergo mechanical stimulation or rapid mechanical deformation,30 which is very similar to the situation when cells pass through a constriction with a dimension smaller than the cell size. This stimuli-responsive cellular behavior has been used for intracellular drug delivery. Here, we hypothesize that that the different responses of untreated MCF-7 and modMCF-7 to mechanical stimulations may be a factor of determining the difference in their impedance. The untreated normal MCF-7 cells are less deformable than modMCF-7. Under the same mechanical stimulations when squeezing into the same constriction, the less deformable cells (MCF-7) should undergo larger impacts caused by the mechanical forces. As a result, more holes and disruptions may form on the membrane of the untreated MCF-7 cells, as compared with modMCF-7 cells. As the holes formed on the membrane may serve as additional paths for electrical current to flow, the membrane of untreated MCF-7 is a less effective barrier to current, compared with the membrane of modMCF-7. Therefore, the impedance of untreated MCF-7 is smaller than the impedance of modMCF-7.

Combining Mechanical and Electrical Measurements Figure 5A portrays the passage time information of cells flowing through a constriction versus the electrical impedance of undeformed cells. Inside the constriction, the impedance of cells at their deformed states was also measured and plotted with the transit time of the cells travelling inside the constriction, as illustrated in Figure 5B. Four populations of cells were demonstrated: RBCs, untreated MCF-7, modMCF-7 and fixed MCF-7 cells. It can be seen from Figure 5A that these four cell populations can be clearly distinguished from each other, by combining both the mechanical properties (represented by the passage time) and electrical properties (represented by the impedance magnitude). Interestingly, if only one-domain parameter is considered (e.g., only passage time or only impedance is considered), untreated MCF-7 and modMCF-7 exhibit a substantial overlap (Figure 5A). However, if the passage time and impedance are both considered, the mixture of the two cell types could be recognized as two different cell populations. Electrical impedance reflects cellular electrical 15 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 22

properties, while passage time reflects cellular mechanical properties such as deformability and surface friction. Both mechanical and electrical measurements are label-free and noninvasive. The properties in mechanical and electrical domains yield different information on cells and their unique properties, thus the combination of these two properties can serve as a more sensitive and selective set of bio-markers for single cell detection and sorting.

Figure 5. Scatterplot of time vs. impedance. (A) Passage time vs. impedance of undeformed cells before entering the constriction. (B) Transit time vs. impedance of deformed cells inside the constriction. Cell number of RBC, modMCF-7, MCF-7 and fixedMCF-7 is: (A) 495, 342, 885 and 225; (B) 505, 302, 747 and 338.

Another phenomenon observed here is the relationship between the passage time and the cell impedance. The impedance shown in Figure 5A is the impedance of intact and undeformed cell. The passage time of RBCs exhibits no dependence on their impedance magnitude, as RBCs do not deform and their passage time mainly depends on the flow velocity. However, for all the three populations of MCF-7 cancer cells, there seems to be a relationship between the passage time and impedance magnitude that a larger impedance magnitude typically 16 ACS Paragon Plus Environment

Page 17 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


corresponds to a longer passage time. This is explainable by the fact that the electrical impedance of cells is dependent on the cell size. Given the same electrical properties, the larger the cell, the more medium between two sensing electrodes that the cell can replace, thereby resulting in a larger change in the impedance magnitude. In other words, the electrical impedance of undeformed cells yields information on the cell size distribution in some degree. On the other hand, the size of cells can influence the passage time required for the cells to transit through the constriction.12 Given the same mechanical properties, a larger cell usually takes longer to pass through a constriction. Summarily, for a given cell type, an occurrence of a larger impedance magnitude implies a larger cell has passed, and a larger cell requires a longer passage time to transit through a constriction. Therefore, a relationship has been observed between the passage time and electrical impedance in each cancer cell population.

The observed dependence of passage time on the cell size sets some limitations on the constriction-based cell deformability measurement. The difference in passage time when cells passing through a fixed constriction can result from two factors: cell deformability and cell size. However, using the current method, it is difficult to determine in what degree each factor influences the passage time. Therefore, finding approaches that can decouple cell deformability and cell size information requires further studies. The other limitation is that such constriction-based devices only work for a certain cell size range, in which the constriction dimensions should be slightly smaller than target cells. In this work, the device with a specific constriction is particularly developed for studying breast cancer cells. If cells with different sizes are to be characterized, the physical dimensions of the constriction channels may need to be modified and optimized. Nevertheless, the device developed in this work can still serve as a useful tool for mechanical and electrical phenotyping of single cells.

To characterize the sensitivity of the developed system, different PMA concentrations (100 nM, 300 nM and 500 nM) were used to treat MCF-7, and compared with the untreated MCF7 cells. In terms of deformability, applying a 100 nM PMA dose to cells results in a slight decrease in the mean passage time compared with untreated cells (Supporting Information Figure S6A), indicating that the deformability is slightly increased after PMA treatment. However, these two populations have substantial overlap. Cells treated with 100 nM PMA 17 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 22

and 300 nM PMA cannot be easily distinguished from each other. A relatively high PMA concentration (500 nM) can lead to a significant decrease in passage time, and thus an increase in deformability. On the other hand, if only electrical impedance is considered, the system is able to detect the PMA concentration as low as 300 nM (Supporting Information Figure S6B). Cells modified with 500 nM PMA exhibit an evident increase in impedance, compared with that of other populations. However, cells treated with 100 nM PMA and untreated cells show no significant difference. Most importantly, if the two parameters (impedance and passage time) are put together, the system is able to effectively distinguish all the cell populations examined, as demonstrated in Figure 6. Overall, an increase in PMA dose leads to an increase in impedance magnitude, and an increase in cell deformability, indicated by the decrease in passage time. These results further verify that the combination of mechanical and electrical markers can significantly improve the sensitivity and efficacy for cell detection.

Figure 6. Scatterplot of passage time vs. impedance, demonstrating the difference among untreated MCF-7 cells (control group) and MCF-7 cells that were treated with different doses of PMA (100nM, 300nM and 500nM).

Conclusions A microfluidic device was developed for simultaneous characterization and measurement of mechanical and electrical properties of single cells. Using the microfluidic impedance measurement, the mechanical properties of cells were studied based on the measurement of the passage time for a cell to pass through a constriction smaller than itself. Entry time and 18 ACS Paragon Plus Environment

Page 19 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


transit time, the two major components of the total passage time, were both studied. The differential impedance strategy described in this work enables the electrical measurements of undeformed cells as well as deformed cells, with high stability and reliability. It has been further demonstrated that combining mechanical properties and electrical impedance information together can significantly increase the ability and efficacy of distinguishing different cell populations. In summary, the impedance-based deformability cytometry described in this work provides great potential for real-time label-free single cell analysis. Future work includes the optimization of the device, as well as the integration of real-time cell deformability/impedance detection and cell sorting function on the same microfluidic device to achieve single cell level sorting based on mechanical and electrical phenotyping of cells.

Author Information Corresponding Authors *E-mail: [email protected]. *E-mail: [email protected]. Author Contributions ∥Ying

Zhou and Dahou Yang contributed equally to this work.

Notes The authors declare no competing financial interest.

Acknowledgements This work was supported by Singapore-MIT Alliance for Research and Technology (SMART) centre, BioSystems and Micromechanics (BioSyM) IRG, which was funded by National Research Foundation (NRF) of Singapore. Ying Zhou is also supported by the seed grant from BioSyM IRG.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 22

Supporting Information. Electrical circuit model of cells; details of experimental protocols; differential vs. non-differential impedance measurements; examples of multiple cells passing through a constriction simultaneously; zoom-in view of Figure 3; passage time, entry time and transit time measured at 200 mbar pressure; size distributions of cancer cells; comparison of untreated MCF-7 and MCF-7 treated with different PMA concentrations (PDF). Supporting video: a high-speed video demonstrating the transitions of different cancer cells through the same constriction (AVI).

References (1) Darling, E. M.; Di Carlo, D. Annu. Rev. Biomed. Eng. 2015, 17, 35-62. (2) Marrese, M.; Guarino, V.; Ambrosio, L. J. Funct. Biomater. 2017, 8. (3) Haase, K.; Pelling, A. E. J. R. Soc., Interface 2015, 12, 20140970. (4) Guck, J.; Schinkinger, S.; Lincoln, B.; Wottawah, F.; Ebert, S.; Romeyke, M.; Lenz, D.; Erickson, H. M.; Ananthakrishnan, R.; Mitchell, D.; Kas, J.; Ulvick, S.; Bilby, C. Biophys. J. 2005, 88, 3689-3698. (5) Otto, O.; Rosendahl, P.; Mietke, A.; Golfier, S.; Herold, C.; Klaue, D.; Girardo, S.; Pagliara, S.; Ekpenyong, A.; Jacobi, A.; Wobus, M.; Topfner, N.; Keyser, U. F.; Mansfeld, J.; Fischer-Friedrich, E.; Guck, J. Nat. Methods 2015, 12, 199-202, 194 p following 202. (6) Dudani, J. S.; Gossett, D. R.; Tse, H. T.; Di Carlo, D. Lab Chip 2013, 13, 3728-3734. (7) Tse, H. T.; Gossett, D. R.; Moon, Y. S.; Masaeli, M.; Sohsman, M.; Ying, Y.; Mislick, K.; Adams, R. P.; Rao, J.; Di Carlo, D. Sci. Transl. Med. 2013, 5, 212ra163. (8) Deng, Y.; Davis, S. P.; Yang, F.; Paulsen, K. S.; Kumar, M.; Sinnott DeVaux, R.; Wang, X.; Conklin, D. S.; Oberai, A.; Herschkowitz, J. I.; Chung, A. J. Small 2017, 13. (9) Che, J.; Yu, V.; Garon, E. B.; Goldman, J. W.; Di Carlo, D. Lab Chip 2017, 17, 14521461. (10) Gossett, D. R.; Tse, H. T.; Lee, S. A.; Ying, Y.; Lindgren, A. G.; Yang, O. O.; Rao, J.; Clark, A. T.; Di Carlo, D. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 7630-7635. (11) Hodgson, A. C.; Verstreken, C. M.; Fisher, C. L.; Keyser, U. F.; Pagliara, S.; Chalut, K. J. Lab Chip 2017, 17, 805-813. (12) Byun, S.; Son, S.; Amodei, D.; Cermak, N.; Shaw, J.; Kang, J. H.; Hecht, V. C.; Winslow, M. M.; Jacks, T.; Mallick, P.; Manalis, S. R. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 7580-7585. (13) Chen, J.; Zheng, Y.; Tan, Q.; Shojaei-Baghini, E.; Zhang, Y. L.; Li, J.; Prasad, P.; You, L.; Wu, X. Y.; Sun, Y. Lab Chip 2011, 11, 3174-3181. (14) Hou, H. W.; Li, Q. S.; Lee, G. Y.; Kumar, A. P.; Ong, C. N.; Lim, C. T. Biomed. Microdevices 2009, 11, 557-564. (15) Yang, X.; Chen, Z.; Miao, J.; Cui, L.; Guan, W. Biosens. Bioelectron. 2017, 98, 408-414. (16) Adamo, A.; Sharei, A.; Adamo, L.; Lee, B.; Mao, S.; Jensen, K. F. Anal. Chem. 2012, 84, 6438-6443. (17) Ren, X.; Ghassemi, P.; Babahosseini, H.; Strobl, J. S.; Agah, M. ACS Sens. 2017, 2, 290299. (18) Zheng, Y.; Nguyen, J.; Wang, C.; Sun, Y. Lab Chip 2013, 13, 3275-3283. 20 ACS Paragon Plus Environment

Page 21 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(19) Tsai, C. H.; Sakuma, S.; Arai, F.; Kaneko, M. IEEE Trans. Biomed. Eng. 2014, 61, 1187-1195. (20) Zheng, Y.; Shojaei-Baghini, E.; Azad, A.; Wang, C.; Sun, Y. Lab Chip 2012, 12, 25602567. (21) Wang, K.; Zhao, Y.; Chen, D.; Fan, B.; Lu, Y.; Chen, L.; Long, R.; Wang, J.; Chen, J. Sci. Data 2017, 4, 170015. (22) Zhao, Y.; Chen, D.; Luo, Y.; Chen, F.; Zhao, X.; Jiang, M.; Yue, W.; Long, R.; Wang, J.; Chen, J. Sensors 2015, 15, 2763-2773. (23) Gawad, S.; Schild, L.; Renaud, P. H. Lab Chip 2001, 1, 76-82. (24) Gawad, S.; Cheung, K.; Seger, U.; Bertsch, A.; Renaud, P. Lab Chip 2004, 4, 241-251. (25) Sun, T.; Holmes, D.; Gawad, S.; Green, N. G.; Morgan, H. Lab Chip 2007, 7, 1034-1040. (26) Cheung, K. C.; Di Berardino, M.; Schade-Kampmann, G.; Hebeisen, M.; Pierzchalski, A.; Bocsi, J.; Mittag, A.; Tarnok, A. Cytometry, Part A 2010, 77, 648-666. (27) Wu, X.; Kang, Y.; Wang, Y. N.; Xu, D.; Li, D.; Li, D. Electrophoresis 2008, 29, 27542759. (28) Hu, X. Y.; Bessette, P. H.; Qian, J. R.; Meinhart, C. D.; Daugherty, P. S.; Soh, H. T. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 15757-15761. (29) Ma, Z.; Zhou, Y.; Collins, D. J.; Ai, Y. Lab Chip 2017, 17, 3176-3185. (30) Sharei, A.; Zoldan, J.; Adamo, A.; Sim, W. Y.; Cho, N.; Jackson, E.; Mao, S.; Schneider, S.; Han, M. J.; Lytton-Jean, A.; Basto, P. A.; Jhunjhunwala, S.; Lee, J.; Heller, D. A.; Kang, J. W.; Hartoularos, G. C.; Kim, K. S.; Anderson, D. G.; Langer, R.; Jensen, K. F. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 2082-2087.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

for TOC only


Page 22 of 22

Characterizing Deformability and Electrical Impedance of Cancer

Recommend Documents