Dielectrophoretic Separation of Prostate Cancer Cells

Abstract

Separation of cancer cells from other biological materials is significant for circulating tumor cell detection in cancer diagnosis and treatment. However, separation of one type of cancer cell from other types of cancer cells can be difficult, since they share similar morphology and biomarkers. In the present work, we have successfully manipulated and isolated LNCaP prostate cancer cells from HCT116 colorectal cancer cells, by dielectrophoresis (DEP) in a microfluidic platform in a continuous operation. In this cell sorter, the prostate cancer cells were treated as target cells and were deflected to a side channel from a main channel as they experienced a negative DEP force, when an AC electric field at the cross-over frequency of the HCT116 cells was supplied. This motion consequently led to the separation of the prostate cancer cells from the colorectal cancer cells. In this manuscript, we report the flow conditions, DEP spectra of the cancer cells and the isolation of LNCaP cells from HCT116 cells. The separation and enrichment factor have been investigated as well.

Introduction

Prostate cancer is a slow growing tumor but the second-leading cause of cancer death in men (1). A man’s lifetime risk of prostate cancer is one in six, and the chance of death due to this disease is 3.4% (2). The long latent period of prostate cancer has made early diagnosis possible. Usually if the development of a tumor can be detected early, it can be cured and the spread prevented. Currently, the most common screening methods for prostate cancer are digital rectal examination (3) and serum prostate specific antigen (PSA) testing (4). Digital rectal examination detects the palpable tumors in the peripheral zone of the prostate. Serum PSA is an FDA-approved biomarker for the early detection of prostate cancer. However, it remains controversial as a marker for population screening due to the lack of enough specificity and sensitivity. Some men may have elevated serum PSA but do not develop lethal prostate cancer. As a result, PSA testing tends to cause over-diagnosis and over treatment (5). Therefore, developing a simple, fast and accurate method for early cancer diagnostics at a low cost is highly desirable.

Existing of circulating tumor cells (CTCs) in peripheral blood could provide another opportunity for cancer diagnosis and prognosis (6–8). The major cause of cancer-associated death is metastasis that involves rare events such as cell dissemination and invasion. Although cell dissemination was thought to be a late step in cancer progression, this concept has recently been seriously challenged. Increasing evidence indicates that tumor cell dissemination and invasion occur at the very early stage of tumor development, even before the tumor can be detected clinically at the primary site (9). To provide information for locating the origin of the cancerous material, isolation of CTCs from other cells is often required (10). Since most cancer cells are epithelia and have common biomarkers, similar size and morphology, separation of the prostate cancer cells from other epithelial cancer cells can be difficult. Thus, to specifically determine which cancer cell exists in the blood sample, separation of different types of cancer cell is crucial. Hence, the major challenges for CTC detection are the enrichment and isolation of rare cancer cells from blood cells and the identification of source of tumor cells.

Several methods have been developed to capture the circulating tumor cells such as immunomagnetic separation (10–20). However, this method may lead to a high false positive rate due to the non-specific binding of the antibody. A microchip based CTC capture and detection from whole blood has also been developed (7,11–14). Benefits that microfluidic devices could offer include high specificity, contactless separation process, high efficiency and low cost (15–18).

There currently are several methods that can enable separation of different cell types (19). For prostate cancer cell separation, some methods have been developed, such as PSA based separation methods (20), prostate-specific membrane antigen (PSMA) based separation methods (21), size based methods (22), membrane microfilter device (23), magnetic based methods (19), etc. However, these methods face some challenges. Size based methods and membrane microfilter device have the limitation of separating cancer cells from the cells with similar size, such as normal epithelial cells. Antibodies are required for cells to bind to magnetic particles for magnetic based methods, and this method is limited by the specificity of antigen-antibody reaction. If the cell cannot be identified by antigen-antibody reaction, then it cannot be separated from another cells by magnetic based cell separation methods (24).

Dielectrophoresis (DEP) is one of the most widely used methods to separate different types of cell (25). Compared with other methods for manipulating biological and non-biological particles in a microfluidic platform, for instance, optophoresis (26), magnetic (24), acoustics (27), and dielectrophoresis, etc (28), DEP has emerged as a promising label-free method for a variety of engineering applications involving manipulation of micro- and nano-particles (13,28–34). DEP manipulation of biological particles such as mammalian cells (35–38), bacteria (39,40), viruses (33,41–43), yeast (S. cerevisiae) (32,44) and even breast cancer cells (29,30) has already been achieved by researchers. In our previous work, colorectal cancer cells HCT116 have been successfully separated from the mixture with Human Embryonic Kidney 293 (HEK 293) cells and Escherichia coli (E. coli) (45). The advantages that DEP can offer include label-free, easy operation and high specificity, efficiency and throughput at a low cost. DEP can also avoid nonspecific binding and has no moving part to increase system reliability. Compared with devices that use other electrokinetic approaches to move particles, such as electrophoresis or electroosmosis (46), DEP systems can be easily combined with electronic detection technologies (e.g. resistive and/or capacitive sensing), to give a real fully-electronic lab-on-a-chip (47). However, surprisingly, to our knowledge, DEP has not been used for prostate cancer separation.

DEP separation depends on the cell size and dielectric property. Since most cancer cells have similar size and morphology in solution, distinguishing different types of cancer cell according to their DEF property may have a potential application in isolating and identifying CTCs. In this work, we isolate LNCaP prostate cancer cells from other epithelial cancer cells, i.e. HCT 116 colorectal cancer cells as an example to demonstrate that these two types of epithelial cancer cell can be distinguished by DEP microfluidics despite the fact that they have the similar cell size and shape in the solution.

Materials and Method

Principle of the method

DEP is a phenomenon in which a force is exerted on a dielectric particle when it is subjected to a non-uniform electric field. It is not necessary for particle to be charged. This is due to the fact that when an electric field is applied to systems consisting of particles suspended in liquid, a dipole moment is induced on the particle, due to the electrical polarization at the interface between the particle and suspending liquid (13,14). The DEP force on the particle depends not only on the electrical properties of the particle and the medium, but also on the magnitude and frequency of the applied electric field. The polarizability of living cells depends strongly on their composition, morphology, and phenotype as well as on the frequency of the applied electrical field (9,48). This means that for a given type of cell and suspending medium, cells can experience, at a certain alternating current (AC) frequency applied to the electrodes, a translational force which directs the cells to the regions of the highest electric field strength (this phenomenon is called positive DEP, i.e. pDEP). Alternatively, by simply changing the frequency, they may experience a force that will direct them away from the highest electric field strength regions (this phenomenon is called negative DEP, i.e. nDEP).

In Fig. 1(a), DEP forces acting on two different kinds of particle in a non-uniform electric field. Grey particles are attracted toward the strongest field area. The driven dielectrophoresis force they experience refers to pDEP. Meanwhile, the white particles are directed away from the strongest field region due to their low polarization. The dielectrophoresis force they experience refers to nDEP.

Fig.1.

(a) DEP force acting on two different particles in a non-uniform electric field. (b)A schematic of a cell suspended in the electric field generated by thin ITO electrodes in the DEP sorter.

For a homogeneous sphere of radius a , the DEP force is described as:

$${\overrightarrow{f}}_{DEP}=2\pi a^3{\epsilon }_0{\epsilon }_m\text{Re}[K(\omega )]\nabla {\left|{\overrightarrow{E}}_{rms}\right|}^2$$ [1]

where ${\epsilon }_0$ is the vacuum permittivity, ${\epsilon }_m$ is the surrounding media dielectric constant, σ denotes the electrical conductivity, ${\overrightarrow{E}}_{rms}$ is the root mean square value of the electric field. The subscript p refers to the particles suspended in a medium and m represents the medium. K(ω) is the Clausius-Mossotti (CM) factor.

In many practical systems, at frequencies below 100 kHz, the CM factor can be approximated in terms of the real conductivities (49):

$$K(\omega )=\frac{{\sigma }_p-{\sigma }_m}{{\sigma }_p+2{\sigma }_m}$$ [2]

pDEP occurs when the polarizability of the particle is larger than the suspending medium ( Re[ K(ω) ]> 1) and the particle moves towards regions of high electric field strength (50). Negative DEP occurs if the polarizability of the particle is smaller than the suspending medium (Re[ K(ω) ]<1) and the particles are repelled from regions of high field strength.

Under this low frequency AC electric fields, the DEP force is described as:

$${\overrightarrow{f}}_{DEP}=2\pi a^3{\epsilon }_0{\epsilon }_m\text{Re}\left[\frac{{\sigma }_p-{\sigma }_m}{{\sigma }_p+2{\sigma }_m}\right]\nabla {\left|{\overrightarrow{E}}_{rms}\right|}^2$$ [3]

Under high frequency AC electric fields, dielectrophoresis is under the dielectric regime and the polarization of the particles will be dominated by the differences between the permittivity of the particle and that of the surrounding medium. In dielectric regime, the DEP force is given as:

$${\overrightarrow{f}}_{DEP}=2\pi a^3{\epsilon }_0{\epsilon }_m\text{Re}\left[\frac{{\epsilon }_p-{\epsilon }_m}{{\epsilon }_p+2{\epsilon }_m}\right]\nabla {\left|{\overrightarrow{E}}_{rms}\right|}^2$$ [4]

Indicated by these equations, the DEP force mainly depends on a particle’s radius a, its complex dielectric constant ${\epsilon }_p^{\mathrm{*}}$, dielectric constant of surrounding medium ${\epsilon }_m^{\mathrm{*}}$, and the electric field ${\overrightarrow{E}}_{rms}$. Based on this, we could predict, when a constant electric field (constant ${\overrightarrow{E}}_{rms}$, $\omega$) is applied to a medium (with a constant ${\epsilon }_m^{\mathrm{*}}$), the particles with different a or ${\epsilon }_p^{\mathrm{*}}$ (i.e. different types of cancer cells in the present work) in this medium, may experience varying magnitudes and directions of DEP force depending on the frequency response of the dielectric permittivity of the particles versus the medium. . Hence, different types of cells will experience different DEP forces under the same electric field. Consequently, selective separation can be achieved by applying an additional force such as gravity or hydrodynamic force by fluid flow.

Based on the aforementioned concept, we have designed a DEP cell separation system, as shown in Fig. 1(b). It has a main channel and a side channel on one sidewall. In this design, there are two electrodes on the bottom surface of the microchannel. The electrodes are geometrically parallel to each other and have an angle of 45 degree to the streamwise direction in the main channels, as shown in Fig. 1(b). Two AC electric signals having a phase shift of 180° are applied to both electrodes, The operational principle is as follow. As the target particles driven by the hydrodynamic force approach the electrode pair, they experience a strong negative DEP, which repels them against the hydrodynamic force in the direction with a certain angle to the flow direction in the main channel. The hydrodynamic force is obtained from Stokes equation as (51):

$${\overrightarrow{f}}_h=6\pi \mu \stackrel{\rightharpoonup }{u}a$$ [5]

where μ is the coefficient of viscosity and $\stackrel{\rightharpoonup }{u}$ is the relative velocity of the fluid with respect to the particle. While most non-target particles will keep their motion to the end of the main channel by the hydrodynamic force $f_{h2}$, the sum force resulted from DEP $f_{\mathit{DEP}}$ and hydrodynamic force $f_{h1}$ push the target particles (i.e. prostate cells) to the side channel, and could be described by equation:

$${\overrightarrow{f}}_{\mathit{SUM}}={\overrightarrow{f}}_h+{\overrightarrow{f}}_{DEP}$$ [6]

Moreover, we can further divide this sum force into x direction and y direction:

$${\overrightarrow{f}}_{\mathit{SUM}}={\overrightarrow{f}}_{\mathit{SUM}x}+{\overrightarrow{f}}_{\mathit{SUM}y}={\overrightarrow{f}}_{hx}+{\overrightarrow{f}}_{hy}+{\overrightarrow{f}}_{DEPx}+{\overrightarrow{f}}_{DEPy}$$ [7]

We define the sum force along x direction as:

$${\overrightarrow{f}}_{\mathit{SUM}x}={\overrightarrow{f}}_{hx}+{\overrightarrow{f}}_{DEPx}$$ [8]

Fig.1(b), indicates that, when the target cells are experiencing nDEP force, the direction of ${\overrightarrow{f}}_{\mathit{DEPx}}$ which is opposite to the direction of ${\overrightarrow{f}}_{hx}$, and ${\overrightarrow{f}}_{\mathit{DEPx}}$ slows down the target cells moving with the flow. Meanwhile, cells are pushed up by ${\overrightarrow{f}}_{hy}$, to the side channel entrance. On the other hand, when ${\overrightarrow{f}}_{\mathit{DEPx}}$ is not strong enough, cells may be pushed through the electrodes pair by ${\overrightarrow{f}}_{hx}$, before they are pushed into the side channel. Hence, ${\overrightarrow{f}}_{\mathit{DEPx}}$ has to be sufficiently stronger compared to ${\overrightarrow{f}}_{hx}$. In addition, because the flow direction is along the x direction in our case, therefore, ${\overrightarrow{f}}_{hy}$ could be approximately considered as zero if the cells are sufficiently far away from the entrance of the side wall.

To quantify the presented DEP sorter performance for the target cell separation and isolation, we define the cell enrichment factor as the ratio of target particles blocked and flow into the side channel by the nDEP force to the total target particles or cells that flow into the sorter (52):

$$\eta =A/(A+B)$$ [9]

where A is the number of the total target cells that are blocked and flow into the side channel, B is the number of total target cells flowed through the gaps in the main channel. η is calculated based on movie obtained from the visualization.

Microfabrication of microfluidic cell sorter

The separation chip was made using lamination based microfabrication (53–55). Transparent acrylic plastic substrates of 24 mm x 20 mm x 1.25 mm were used as top (cap) layer. One inlet and two outlet wells (one for the main channel, and the other for the side channel) were drilled on the top layer. The diameter of all the wells is 1.6 mm. Channels were formed on a middle layer, and the widths of main channel and side channel were 550 μm and 150 μm, respectively, with the same height of 40 μm. A pair of indium tin oxide (ITO, or tin-doped indium oxide) thin film electrodes was formed on the ITO coated glass with a gap of 20 μm (two gray triangles in Fig.1(b)). It is optically transparent and colorless in a thin layer, thus, we can visualize the separation process. Pressure sensitive adhesive (Scotch^® Double Sided Tape 666, 3M) was used to bond these layers together.

Sample preparation

The human prostate cancer cell line LNCaP (ATCC) was used as the separation target. LNCaP cells are androgen-sensitive human prostate adenocarcinoma cells. They are adherent epithelial cells growing in aggregates (56). Human colorectal cancer cell line HCT 116 (ATCC) was selected for LNCap separation. Both LNCaP and HCT 116 cells were in the shape of sphere in suspension and have almost the same diameters of about 20 μm (57). Equation (1) indicates that the DEP force is highly affected by the radius of the cells, therefore this could be more challenge and difficult to separate these two types of cell. If we could separate these two types of cell, it would be more convincing that the technique could have a high potential to isolate prostate cancer cells from other cancer cells and normal epithelial cells. We mixed LNCaP cells with HCT116 cells, in a phosphate-buffered saline (PBS, TEKnova, Inc.) solution, with approximately the same concentration of 1.5×10⁶ cells/ml. To distinguish LNCaP from HCT 116 cells under a microscope, we labeled HCT116 colorectal cancer cells with dye Hoechst 33342 (Invitrogen), whose peak absorption and emission are around 350 and 461 nm, respectively) (58). After labeling, cells would emit fluorescence when excited by UV light. Particularly, the labeling of HCT 116 is only for observation of the separation process. In real applications, the HCT 116 cells are not necessary to be labeled. Our test indicated that at the frequency of 2.5 MHz used for the DEP separation, the labeling did not change the DEP behavior of HCT 116 very much, which could be seen in following DEP spectra in this paper (Fig. 4). In this work, different concentrations of PBS buffer were used to control medium conductivity.

Fig. 4.

DEP spectrum of LNCaP prostate cancer cell and HCT116 colorectal cells.

Experimental setup

A schematic of the setup used for the DEP separation is shown in Fig. 2. The DEP cell sorter was placed on the test bed of an inverted epi-fluorescent microscope (Olympus-IX70). A syringe pump (Harvard PHD 2000) was used to deliver the sample to the chip, and the flow rate was kept between 0.1μL/min to 0.8 μL/min in our experiments. A function generator (Tektronix, Model AFG3102) was used to supply AC electrical signal to these two ITO electrodes. In our experiments, amplitude of applied AC signal was kept between 0 V_p-p to 20 V_p-p. A 10X objective lens (NA = 0.25) was used for magnification of the imaging. UV light (350 nm in wavelength) was used to excite the labeled cells. The cell motion was captured by a high-resolution CCD camera (SensiCam-QE, Cooke Corp).

Fig. 2.

Schematic of the experimental setup.

Results and discussion

DEP spectrum of LNCaP prostate cancer cells and HCT116 colorectal cancer cells

DEP based cell separation strongly depends on the DEP property of the cells and the suspending medium electrical property. To effectively separate two different kinds of cell, we first need to understand the DEP spectra of the relevant cells.

By using a simple wedge microchip (45), the dielectrophoresis character of LNCaP prostate cancer cell and HCT116 colorectal cancer cells were measured at first. In another word, we investigated the DEP force’s relative strength and direction (positive or negative) along with the applied AC frequencies (DEP spectrum), for both kinds of the cancer cell.

The DEP spectrums of both kinds of cell were measured in the wedge chip, as shown in Fig.3. Two gold electrodes with thickness of 230 μm were used to form the wedge. The distance between their edges was 40 μm. A non-uniform electric field required for the DEP force was generated between the sharp edges of the two gold electrodes. Solution of each kind of cell was first loaded into the wedge chip. Since the solution in this chip was almost stationary, the effect of hydrodynamic force on the cells due to the pressure driven flow was neglected. Therefore, DEP force can become the major force on these cells, and can be observed and characterized, as long as it exists. For nDEP, the cells will move away from the tips, whereas for the pDEP, the cells will move towards the tips, once an electric field is switched on. We recorded the movement by the CCD camera at a fixed frame rate and a fixed exposure time, and then by measuring the distance of the cell’s movement in a fixed time period (50 frames for each measurement), we could calculate the mean velocity, which is directly related to the force. The highest velocity is used for normalization of the DEP force. The faster the motion, the stronger the DEP force. For each frequency, we repeat 3 measurements and the standard deviations of DEP force are shown in the spectra as error bars. With this wedge DEP chip, a very small amount of sample (about 5 μL) is sufficient for measuring the DEP spectrum.

Fig. 3.

Prostate cancer cells moving under nDEP force in a wedge chip. (a) Two cells locate near the electrodes tips are marked by red circle. (b) Driving by nDEP force, these two cells start moving away from the tips. (c) These two cells are keeping move away from tips. (d) After three seconds, these two cells are far away from the tips.

Specially, from equation (2) and (4), the Clausius-Mossotti factor K is affected by the conductivity of the suspending medium and the cell. Consequently, suspending medium’s conductivity could affect the DEP force on suspending cells. Therefore, DEP spectrums for each kind of cell suspended in medium with different conductivity have been also investigated.

Fig. 4 shows the DEP spectrums for both LNCaP cells and HCT116 cells, measured by the wedge DEP test chip. Three different concentrations of PBS buffer were used, 0.025X PBS, 0.05X PBS and 0.1X PBS, whose conductivity were 600 μS/cm, 900 μS/cm and 3000 μS/cm, respectively. However, further increasing conductivity of suspending medium will have high risk of generating bubbles in the solution when voltage is supplied, due to electrolysis. Result showed that, in our experiment, the DEP spectra were different in different suspending medium’s conductivities. By comparing spectra of LNCaP cell with that of HCT 116 cell, one could see that in 0.1X PBS medium, the cross-over frequency where the DEP force was zero, was 2.5 MHz for HCT 116 cells. Meanwhile, LNCaP cells experienced an nDEP force at the same frequency. On the other hand, when we compared spectra in the media with three different conductivities, the DEP force difference between LNCaP and HCT116 cells was largest in the medium of 3000 μS/cm (0.1X PBS). This suggested that by using this medium, LNCaP cells would experience the relatively stronger nDEP force at the cross-over frequency for HCT116 cells. Therefore, in the following experiments, we focused on the medium with conductivity of 3000 μS/cm.

In the medium with conductivity of 3000 μS/cm, , in the frequency band of 1 Hz ~ 20 MHz and 55 MHz ~ 70 MHz, the LNCaP cells experienced an nDEP force, whereas in the frequency range of 20 MHz ~ 55 MHz and 70 MHz ~ 100 MHz, the cells experienced a pDEP force. At the mean time, , in the frequency band of 1 Hz ~ 2.5 MHz and 60 MHz ~ 100 MHz HCT116 cells experienced an nDEP force, whereas in the frequency range of 2.5 MHz ~ 10 MHz, the cells experienced a pDEP force. By analyzing these two curves, we could see that at the frequency of 2.5MHz, LNCaP cells experienced a strong negative DEP force while the DEP force acting on HCT116 cells was around zero. This indicated that, by applying AC electric field with this particular frequency in the DEP sorter, LNCaP cells should be repelled by the nDEP force and flow to the side channel, and meanwhile the HCT116 cells would keep flowing to the end of main channel without blocking by either pDEP or nDEP force. Furthermore, DEP force of labeled HCT116 cells had also been measured to be almost zero at 2.5 MHz, which suggested the usage of dye on HCT116 cells had no influence on separation results for the current separation experiment at 2.5 MH. Therefore, 2.5 MHz, which was the cross-over frequency of the HCT116 cell, was selected for the DEP sorter to isolate the LNCaP cells from the HCT116 cells. Specially, although these is a spectral difference between HCT 166 cell and labeled HCT 116 cell, which could be caused by the permittivity change of the HCT116 cell, the DEP response are luckily very close at the frequency we use to separate LNCaP cells from HCT 116 cells. Base on this fact, we can use labeled HCT 116 cells just for visualization purpose. In real applications, the HCT 116 cells are not necessary to be labeled.

Separation of LNCaP cells and HCT116 cells

Fig. 5 shows the visualization of the separation of LNCaP cells from HCT116 cells due to nDEP force in the DEP cell sorter, with and without the electrical activation. The fluid flowed from the inlet at the left side to the outlet at the right side in the main channel (about 80% of the inlet flow rate) and the side channel (about 20% of the inlet flow rate) at a total inlet flow rate of 0.15 μL/min. In Fig. 5(a), approximately 80% of LNCaP cells flowed through the gap between the two electrodes in the main channel without AC electrode activation. However, with the electric activation of the electrodes at voltage of 12 V_p-p and frequency of 2.5 MHz, almost all prostate cancer cells migrated to the side channel under the nDEP force as shown in Fig. 5(b), whereas, HCT116 cells remained the same motion through the gap. Fig. 5 demonstrates that DEP can be used for isolation of LNCaP cells from HCT116 cells. Thus, DEP could provide an opportunity for isolation of one type cancer cell from other similar epithelial and cancer cells.

Fig. 5.

Experimental demonstration of the DEP deflection. (a) Without AC activation, most cells (LNCaP cells, HCT116 cells) flowed through the electrode gap along the main channel. (b) With AC activation, almost all LNCaP cells were repelled to the side channel due to negative DEP force, whereas only HCT116 cells could be seen retaining the same motion through the gap. The dark spheres are LNCaP cells, and the bright spheres are HCT116 cells.

Quantitative analysis of cells separation

Flow rate effect

Sample flow rate effect on η is shown in Fig. 6. In this experiment, applied voltage was kept at 13 V_p-p under selected frequency of 2.5 MHz. 100 captured frames and a fixed frame rate were used to calculate η . For each flow rate, we repeat 3 measurement and standard deviations were represented as error bars. Fig. 6 indicates that η decreases with the increase of the flow rate Q. The enrichment factor was almost a constant of around 96%, as long as Q was less than about 0.4 μL/min, indicating that the DEP force was much larger than the hydrodynamic forces. However, as Q was larger than 0.4 μL/min, η decreased rapidly with the increased Q. When Q reached 0.8 μL/min, the hydrodynamic force overwhelmed the DEP force, resulting in almost no deflection of LNCaP cells and the device had almost no effect of cell enrichment. Since the side channel must have a small portion of flow to carry the isolated target cells to a collection well, there will always be a small portion of the targeted cells (about 20%) that follow this portion of fluid and flow into the side channel. Therefore, the enrichment factor cannot be lower than 0.2, even the flow rate was further increased.

Fig. 6.

Enrichment factor and normalized ${f^{\ast }}_{\mathit{SUMx}}$ as a function of the sample flow rate Q.

As we know the total force ${\overrightarrow{f}}_{\mathit{SUM}}$ on cells is the sum of ${\overrightarrow{f}}_{\mathit{DEP}}$ and ${\overrightarrow{f}}_h$, as shown in equation (6). ${\overrightarrow{f}}_{hx}$ mainly pushes the cells to flow along the main channel, while ${\overrightarrow{f}}_{\mathit{DEPx}}$ acts against the hydrodynamic force and ${\overrightarrow{f}}_{\mathit{DEPy}}$ pushes the target cells to the entrance of the side channel. ${\overrightarrow{f}}_{\mathit{DEPx}}$ counterbalances ${\overrightarrow{f}}_{hx}$ so that there can be sufficient time for the target cells to be pushed to the entrance of the side channel by ${\overrightarrow{f}}_{\mathit{DEPy}}$. Once the cells arrive the entrance, they will move to the side channel due to the hydrodynamic force there. At high Q, due to the large hydrodynamic force, it is very difficult to deflect the targeted LNCaP cells to the side channel. Thus, hydrodynamic force and DEP force are competitive.

To explain the flow rate effect shown in Fig. 6 based on force, we normalize the $f_{\mathit{SUMx}}$ by following equation:

$${f^{\ast }}_{\mathit{SUMx}}=\frac{f_{DEPx}-f_{hx}}{f_{DEPx}}$$ [10]

According to this equation, when $f_{hx}$ is zero (Q is zero), ${f^{\ast }}_{\mathit{SUMx}}$ equals to 1. When $f_{\mathit{DEPx}}=f_{hx}$, ${f^{\ast }}_{\mathit{SUMx}}$ become zero. During the experiment of characterizing flow rate effect, the DEP force should be kept constant. When $f_{\mathit{DEPx}}$ is as large as $f_{hx}$, the target cells could be decelerated to zero along x direction from their initial velocity near the electrode pair, so that they can be pushed to the entrance of the side channel. However, since the cells are initially moving with the bulk velocity, $f_{\mathit{DEPx}}$ has to be larger than $f_{hx}$, in order to ensure that cells near the side opposite to the side wall can be pushed to the side channel entrance. Therefore, we assume that when Q is increased up to $f_{hx}=f_{DEPx}$, there will be no enrichment anymore, and when Q is decreased down to $f_{hx}=0$, the enrichment will be the maximum, i.e. 1. Furthermore, since ${f^{\ast }}_{\mathit{SUMx}}$ is linearly proportional to $f_{hx}$ as shown in the equation (10), we can get the linear relation between ${f^{\ast }}_{\mathit{SUMx}}$ and Q as shown in Fig. 6. For guidance in design and operation of the DEP sorter, Fig. 6 indicates approximately that if ${f^{\ast }}_{\mathit{SUMx}}$ is lower than 0.5, η will decrease rapidly with Q. However if ${f^{\ast }}_{\mathit{SUMx}}$ is higher than 0.5, η reaches nearly the highest, and further decrease of Q will not improve η significantly.

Voltage effect

The effects of applied voltage on the DEP separation under three different flow rates were investigated, which were 0.15 μL/min, 0.3 μL/min and 0.4 μL/min, respectively. 100 captured frames and a fixed frame rate were used to calculate η. For each applied voltage, we repeat 3 measurement and standard deviations were represented as error bars.The results are shown in Fig. 7, which indicates that, under each Q, η increases with increasing electric field (or equivalently voltage). On the other hand, it is also shown in Fig. 7, under the same applied voltage, the enrichment factor of a higher Q is lower than that of a lower Q. This is, as shown in Fig. 6, also because that a higher Q causes a stronger hydrodynamic force on cells, to push the LNCaP cells to flow along the main channel. Therefore at a higher Q, a stronger nDEP force is required to repel cells to the side channel.

Fig. 7.

Enrichment factor as a function of applied voltage under different flow rates.

At a relatively low flow rate, e.g. Q = 0.15 μL/min, Fig. 7 shows that η reaches a relative high enrichment factor 0.86 when applied voltage is 10 V_p-p. This means that for a low Q, a low voltage can be sufficient to achieve a high η. However, as shown in Fig. 7, for the higher Q of 0.3 and 0.4 μL/min, at the same applied voltage of 10 V_p-p, η reaches only to 0.68 and 0.5 respectively. When we further increase applied voltage to 14 V_p-p, η under these higher Q are increased to high level of 0.95 and 0.93 respectively. After all, Fig. 6 and 7 indicate that increasing the voltage and decreasing flow rate could be two easy ways to enhance the enrichment factor.

From equation (1), $f_{\mathit{DEP}}$ increases with the increased electric field (or equivalently voltage), the higher voltage, the stronger $f_{\mathit{DEP}}$. The stronger $f_{\mathit{DEP}}$ is, the more cells near the side wall opposite to the side channel will be repelled to the side channel. This is not only because a strong $f_{\mathit{DEPx}}$ can counterbalance the motion of cells due to $f_{\mathit{h}}$ before they achieve electrodes pair in x direction, but also because a strong $f_{\mathit{DEPy}}$ pushes up the cells to move faster to the side channel in y direction. Consequently, η rapidly increases with the increased voltage within the test range of the voltage for a given flow rate.

Conclusion

In present work, we successfully manipulate and isolate LNCaP prostate cancer cells from the mixture with HCT116 colorectal cancer cells, by a microfluidic cell sorter based on conventional negative dielectrophoresis. DEP spectra of the prostate cancer cell and the colorectal cancer cell were measured in media with different conductivity. Effect of parameters such as flow rate and applied voltage has been investigated as well. The present work indicates that prostate cancer cells can be separated and isolated from other type of cancer cells by DEP. This result gives a new opportunity for clinic diagnosis of prostate cancer for circulating tumor cell using a lab-on-a-chip device.

Abbreviations:

DEP

dielectrophoresis

CTCs

circulating tumor cells

PSA

prostate specific antigen

PSMA

prostate-specific membrane antigen

AC

alternating current

pDEP

positive DEP

nDEP

negative DEP

CM

Clausius-Mossotti

PBS

phosphate-buffered saline