Dielectrophoretic separation of mouse melanoma clones
Ahmet C. Sabuncu,1Jie A. Liu,2,3Stephen J. Beebe,2and Ali Beskok1,a?
1Department of Aerospace Engineering, Old Dominion University, Norfolk,
Virginia 23529, USA
2Frank Reidy Research Center for Bioelectrics, Norfolk, Virginia 23508, USA
3The Department of Biology, Old Dominion University, Norfolk, Virginia 23529, USA
?Received 16 February 2010; accepted 17 May 2010; published online 16 June 2010?
Dielectrophoresis ?DEP? is employed to differentiate clones of mouse melanoma
B16F10 cells. Five clones were tested on microelectrodes. At a specific excitation
frequency, clone 1 showed a different DEP response than the other four. Growth
rate, melanin content, recovery from cryopreservation, and in vitro invasive studies
were performed. Clone 1 is shown to have significantly different melanin content
and recovery rate from cryopreservation. This paper reports the ability of DEP to
differentiate between two malignant cells of the same origin. Different DEP re-
sponses of the two clones could be linked to their melanin content. © 2010 Ameri-
can Institute of Physics. ?doi:10.1063/1.3447702?
Dielectrophoresis ?DEP? is the motion of polarizable particles in nonuniform electric fields.
DEP has been extensively used for biological cell manipulations. There are several studies aimed
to separate cells by DEP, such as isolation of human leukemia,1stem,2breast cancer,3or malaria
infected cells4from blood. One recent study demonstrated the use of a DEP-based microfluidic
system for efficient isolation of circulating tumor cells from blood and also showed that the
isolated cells were viable and suitable for further analysis.5In another study, platelets were sepa-
rated from blood by DEP activated cell sorter that can perform size based flow-fractionation.6In
addition to cell separation, it has been shown that cell dielectric properties can be extracted by
DEP. In one study, morphology changes in stimulated Jurkat T-cells were investigated by calcu-
lating the cell membrane capacitance from the DEP crossover frequency.7
In this study, we present a method that can differentiate between two malignant cells of the
same origin, which differ from each other by small changes in their melanin content. This study is
a precursor for future efforts on DEP separation of metastatic human melanoma cells using mi-
II. MATERIALS AND METHODS
DEP force arises when a dipole is in interaction with a nonuniform electric field. The time
averaged DEP force is given as8
?F¯DEP?t?? = 2??mR3Re?fCM? ? Erms
where ?f?t?? represents the time average of the function f?t?, ?mis the real part of the permittivity
of the medium, R is the particle radius, Ermsis the root mean square of the electric field, and fCM
is the Clausius–Mossotti ?CM? factor. This factor states the effective polarizability of the particle,
and it is given as8
a?Author to whom correspondence should be addressed. Electronic mail: email@example.com.
BIOMICROFLUIDICS 4, 021101 ?2010?
4, 021101-11932-1058/2010/4?2?/021101/7/$30.00© 2010 American Institute of Physics
f˜CM?? ˜p,? ˜m? =
? ˜p− ? ˜m
? ˜p+ 2? ˜m
where ? ˜=?−i?/?; ? and ? are the permittivity and conductivity, respectively; ? is the angular
frequency; p and m indices are for particle and medium.
As previously used by Gascoyne et al. ?1997? for mammalian cells, we have utilized the
single shell model to represent biological cells. This model approximates the cell’s complex
? ˜p= ? ˜m??3+ 2?
? ˜cyt− ? ˜m
? ˜cyt+ 2? ˜m??/??3−?
? ˜cyt− ? ˜m
? ˜cyt+ 2? ˜m??,
where ?=R/?R−t?; t is the cell membrane thickness; m and cyt indices show the membrane and
B. Numerical calculations
CM factors of B16 cells were computed using Eqs. ?2? and ?3?. Cells were assumed as perfect
spherical particles. This is a reasonable approximation as cells become nearly spherical after
nonspherical adherent cells were harvested from the culture flask by trypsinization. Cell’s dielec-
tric data for simulation were taken from the work of Oblak et al. and shown in Table I. Although
it is obvious that the dielectric data of the cells used in this study will differ from the literature
values, the simulations provided an initial guess for the crossover frequency. In addition, CM
factor simulations were employed to investigate the effect of the cytoplasm conductivity. A nu-
merical simulation of electric field on castellated electrode geometry with arbitrary electric poten-
tials was done using electromagnetic module of finite element modeling software COMSOL
MULTIPHYSICS®?COMSOL AB, Stockholm, Sweden?.10Electrostatic limit is assumed for electric
field simulations, and hence, Laplace’s equation was solved for electric potential in two dimen-
sions. In simulations, electric potentials were prescribed on electrode boundaries and side bound-
aries were taken as symmetric. Electric field was used to distinguish between the positive and
negative DEP zones on the chip.
C. Clone isolation
B16F10 clones were grown in Dulbecco’s Modified Eagle Medium, ?ATCC, Manassas, VA?
with 10% fetal bovine serum ?FBS?, 1% L-glutamine, and 1% penicillin-streptomycin in an incu-
bator with 5% CO2and at 37 °C. For all the measurements done in this study, cells in log phase
were harvested by trypsinization. Like all cancer tumors and cancer cells lines, B16F10 melanoma
cells are a heterogeneous population of cells. In order to isolate different clones from this hetero-
geneous population, B16F10 cells were treated with ten pulses of 60 ns, 60 kV/cm electric field
exposures. Afterward, treated cells were plated at limiting dilutions, grown until individual colo-
nies could be isolated as clones using sterile glass cloning cylinders. From a total of 18 clones, five
clones numbered 1–5 were analyzed in this study.
TABLE I. Dimensions and dielectric parameters of B16 cells ?Ref. 9?.
4.4?10−11A s/V m
7.1?10−10A s/V m
021101-2Sabuncu et al.Biomicrofluidics 4, 021101 ?2010?
D. DEP separation of clones
The cells were washed two times prior to DEP experiments with a cell manipulation buffer.
The buffer consists of 8.6% ?w/w? sucrose, 0.3% ?w/w? dextrose, and 1.0 mg/ml Bovine Serum
Albumin ?Sigma-Aldrich, St. Louise, MO? and sufficient amount of modified eagle’s minimum
essential media ?ATCC? to reach 0.16 S/m.7The conductivity of the cell suspension media was
measured by a conductivity meter ?OMEGA, PHH-80PMS, Stamford, CT?. The cell motion was
observed under an inverted microscope ?Olympus IX71, Cenmter Valley, PA? and at least three
?Q Imaging, Surrey, BC, Canada?. Cell counts in the results and discussion part were averaged
over these snapshots. About 500 cells were counted at each frequency.
byacharge coupleddevice camera
The microelectrodes were fabricated by using standard photolithography techniques. The ex-
perimental setup is shown in Fig. 1. The electrodes have characteristic dimension of 50 ?m. The
microelectrodes were energized by a function generator ?BK Precision, Yorba Linda, CA? operat-
ing in sinusoidal mode.
F. Biological tests on clones
Melanin levels were determined from cell extracts as previously described by Kasraee et al.11
and Kinoshita et al.12Briefly, 5.0?106cells/ml cells were removed from the flask by trypsiniza-
tion and dissolved in 1N NaOH. After 1 h incubation at 37 °C, the solution was centrifuged and
the absorbance of the extract was measured at 405 nm. The melanin amount was obtained from
comparison to the standard curve of synthetic melanin ?SIGMA, M8631?. Cell invasion assay:
Cultrex®BME cell invasion assay ?R&D Systems, Cat No. 3455-096-K, Minneapolis, MN? was
used to measure cell invasion through an extracellular matrix. Procedures were carried out in
accordance with the manufacturer’s recommendations. For growth rate measurements, cells were
cultured in a complete medium in six-well plates for 24 h with a starting concentration of 0.1
?106cells/ml. After 24 h, cells were trypsinized and viable cell numbers were determined using
a hemocytometer by exclusion of trypan blue. Assays were carried out in triplicate and experi-
ments were repeated three times. Rate of recovery from cryopreservation was determined by
freezing 1?106cells in liquid nitrogen for 36 h. To determine viable cell numbers, these cells
were thawed and cultured in 25 cm2cell flasks. 28 h later, viable cell numbers were determined
using a hemocytometer by exclusion of trypan blue. The cell number was divided by 28 to
determine the increase in cell number per hour. Assays were carried out in triplicate and experi-
ments were repeated three times.
FIG. 1. ?a? Two nontouching gold arrays were polarized by a function generator. Fluid and cell motions were observed by
an inverted microscope. ?b? Schematic view of the electrode array with a cover glass and O-ring on it.
021101-3DEP separation of melanoma clonesBiomicrofluidics 4, 021101 ?2010?
III. RESULTS AND DISCUSSION
A. DEP response and melanin content
The electric field simulation of castellated electrode is shown in Fig. 2, where the low and
high electric field intensities correspond to darker and lighter regions, respectively. From Eq. ?2?,
a cell with a positive CM factor will move toward the lighter zones ?positive DEP?. Likewise, a
cell with a negative CM factor will travel toward darker zones. DEP separation of five B16F10
clones was tested under the applied electric potential of 5 Vp.p.at 200, 300, and 400 kHz fre-
quencies. The response of clone 1 at 300 kHz was opposite of the other four clones. In the rest of
this study the behavior of clone 1 is compared to that of clone 2, which was chosen randomly from
the remaining four clones. In Fig. 3, final positions of cells on the chip are shown. At 200 kHz
both clones showed negative DEP responses. However, as the frequency was increased the two
clones showed different crossover frequencies from negative to positive DEP. At 300 kHz, clone
1 exhibited negative DEP, whereas clone 2 showed positive DEP. At 400 kHz, both clones showed
positive DEP. Figure 4 shows the cell count of positive DEP response of B16 clones normalized by
the total number of cells at three different frequencies. The results are obtained by counting the
positive responses at each captured image and by dividing these to the total number of cells within
the same image. At 200 kHz, most of the clone 1 cells show negative DEP response, while 25%
of clone 2 cells show positive DEP. By 300 kHz, 90% of clone 2 cells and 20% of clone 1 cells
show positive DEP response. The normalized count ratios differ significantly at 300 kHz, which
indicates that the clones’ DEP responses are almost opposite at this frequency.
FIG. 2. Numerical simulation of electric field generated by castellated microelectrodes. Darker and lighter zones corre-
spond to lower and higher electric field regions, respectively.
FIG. 3. Dielectrophoretic behavior of clones 1 and 2 under different excitation frequencies. Crossover frequency of clone
1 is found between 300 and 400 kHz; for clone 2 it is between 200 and 300 kHz. Applied voltage is 5 Vp.p..
021101-4Sabuncu et al.Biomicrofluidics 4, 021101 ?2010?
In order to determine if these clones exhibited functional differences, further investigations
were performed on B16F10 clones 1 and 2, including growth rates at normal conditions, recovery
from cryopreservation, in vitro invasive capability, and melanin content. Among these tests for the
two clones, normal growth rate, which is a measure of cell proliferation and invasion capability,
which is an in vitro measure of metastasis, did not show significant difference ?Table II?. Thus,
within the B16F10 heterogeneous population, which is an invasive cell line, both clones are
equally proliferative and invasive. The rate of recovery from cryopreservation, which is a measure
of recovery from extreme temperature stress, was very near the cutoff to be significant, and the
melanin content was significantly different between the two clones. After melanin extraction, the
absorbance at 405 nm of clone 1 was significantly higher than the absorbance of clone 2, with the
p value being 0.001 ?Table II?. Interestingly, the rate of recovery from cryopreservation tended to
be slower for the clone with the highest melanin content. Different melanin contents might also
explain the different crossover frequencies of the two clones. Conductivity of melanin is lower
than the conductivity of the cytoplasm.13An increase in the percentage of melanin in cytoplasm
may lead to a decrease in overall polarizability of the cell. In Fig. 5, we utilized the single shell
model based on the data given by Oblak et al., and varied only the ?cytvalue to investigate the
effect of increased melanin content in cell cytoplasm. We varied the cytoplasm conductivity in
between 0.3 and 0.6 S/m, based on the data given by Pethig ?2006?.14The DEP crossover fre-
quency increases by decreasing the cytoplasm conductivity, which may explain the different cross-
over frequencies observed in Figs. 3 and 4. It is worth noticing that the crossover frequency in Fig.
5 is based on the dielectric cell properties given by Oblak et al., and starts at 400 kHz for ?cyt
=0.5 S/m, while our experiments have shown the crossover frequency between 200 and 400 kHz.
This slight mismatch could be induced by the differences in the original B16F10 cell line. In
addition, the clones can be different in terms of their membrane properties, which also affect the
FIG. 4. Normalized cell count of B16 clones. The ratio is calculated as DEP positive cells divided by total number of cells.
TABLE II. Biological tests on B16 clones.
Clone 1 Clone 2P value
Rate slope recovery from cryopreservation ?106cell/h?
Growth rate slope ?106cell/h?
Invasion percentage ?%?
aRepresents statistical significance.
021101-5DEP separation of melanoma clones Biomicrofluidics 4, 021101 ?2010?
Since the cancer cell genomes are unstable, it is possible that these two melanoma clones may
differ in ways other than their melanin content, which could explain the differences in their
crossover frequencies. However, these two clones did not deviate in growth rate, they were not
significantly different in recovery from cryopreservation and they did not vary in invasive capa-
bilities. Each one of these functions requires hundreds of gene products, suggesting that the
genomes of these clones were not significantly different. Based on these results, we suggest that
the difference in the melanin content of these clones is most likely responsible for the differences
in their crossover frequency.
In this study, the capability of DEP to separate two B16F10 clones is presented. Clones 1 and
2 have significantly different melanin content. In contrast to other studies, which separate malig-
nant cells from their healthy counterparts,1–5this study focuses on separating two malignant cells
that have the same origin but differ in melanin content. The difference in the melanin content can
be connected to increased MC1 ?melanocortin-1 hormone?/MC1 receptor system, which is sug-
gested to play a role in melanoma progression.15While in vitro growth rates and invasion capaci-
ties of these clones were found to be similar, they do not specifically determine metastatic poten-
tial. Therefore, it might be inferred that differences in melanin content may determine metastatic
potential of these clones and other melanomas, and DEP could be utilized to differentiate between
these cells. Consequently, this study can be extended to a DEP-based microfluidic device that can
separate malignant cells having different metastatic levels.
This work was supported by the ODU Office of Research. We thank Ms. Sandy Anderson for
isolation of B16F10 clones. The authors would like thank to Dr. Bayram Celik for the numerical
simulations. A.C.S. and J.A.L. contributed equally to the work presented here.
1F. F. Becker, X. B. Wang, Y. Huang, R. Pethig, J. Vykoukal, and P. R. C. Gascoyne, J. Phys. D: Appl. Phys. 27, 2659
2M. S. Talary, K. I. Mills, T. Hoy, A. K. Burnett, and R. Pethig, Med. Biol. Eng. Comput. 33, 235 ?1995?.
3P. R. C. Gascoyne, X. B. Wang, Y. Huang, and F. F. Becker, IEEE Trans. Ind. Appl. 33, 670 ?1997?.
4P. Gascoyne, C. Mahidol, M. Ruchirawat, J. Satayavivad, P. Watcharasit, and F. F. Becker, Lab Chip 2, 70 ?2002?.
5P. R. C. Gascoyne, J. Noshari, T. J. Anderson, and F. F. Becker, Electrophoresis 30, 1388 ?2009?.
FIG. 5. Real part of CM factor of B16 cells at ?m=0.5 S/m calculated from the dielectric data taken from the work of
Oblak et al. ?2007?. The inset figure shows the change in the real part of CM factor between 370 and 450 kHz.
021101-6Sabuncu et al.Biomicrofluidics 4, 021101 ?2010?
6M. S. Pommer, Y. Zhang, N. Keerthi, D. Chen, J. A. Thomson, C. D. Meinhart, and H. T. Soh, Electrophoresis 29, 1213 Download full-text
7R. Pethig and M. S. Talary, Inst. Eng. Tech. Nanobiotechnol. 1, 2 ?2007?.
8H. Morgan and N. G. Green AC Electrokinetics: Colloids and Nanoparticles ?Research Studies Press, Baldock, UK,
9J. Oblak, D. Krizaj, S. Amon, A. Macek-Lebar, and D. Miklavcic, Bioelectrochemistry 71, 164 ?2007?.
10See: www.comsol.com for commercial finite-element package COMSOL.
11B. Kasraee, A. Hügin, C. Tran, O. Sorg, and J. Saurat, J. Invest. Dermatol. 122, 1338 ?2004?.
12M. Kinoshita, N. Hori, K. Aida, T. Sugawara, and M. Ohnishi, J. Oleo Sci. 56, 645 ?2007?.
13T. Ligonzo, M. Ambrico, V. Augelli, G. Perna, L. Schiavulli, M. A. Tamma, P. F. Biagi, A. Minafra, and V. Capozzi, J.
Non-Cryst. Solids 355, 1221 ?2009?.
14R. Pethig, in Encyclopedia of Surface and Colloid Science, edited by S. Somasundaran ?Taylor & Francis, Boca Raton,
2006?, Vol. 3, p. 1728.
15R. Lazova and J. M. Pawelek, Exp. Dermatol. 18, 934 ?2009?.
021101-7DEP separation of melanoma clones Biomicrofluidics 4, 021101 ?2010?