Differential responses of human liver cancer and normal cells to atmosphericpressure plasma

Bomi Gweon,1 Mina Kim,2 Dan Bee Kim,1,a) Daeyeon Kim,2 Hyeonyu Kim,2 Heesoo Jung,1

Jennifer H. Shin,2,3,b) and Wonho Choe1,b)

1Department of Physics, Korea Advanced Institute of Science and Technology, 291 Daehak-ro, Yuseong-gu,Daejeon 305-701, South Korea2Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology,291 Daehak-ro, Yuseong-gu, Daejeon 305-701, South Korea3Department of Bio and Brain Engineering, Korea Advanced Institute of Science and Technology,291 Daehak-ro, Yuseong-gu, Daejeon 305-701, South Korea

(Received 8 March 2011; accepted 22 May 2011; published online 10 August 2011)

When treated by atmospheric pressure plasma, human liver cancer cells (SK-HEP-1) and normal

cells (THLE-2) exhibited distinctive cellular responses, especially in relation to their adhesion

behavior. We discovered the critical threshold voltage of 950 V, biased at the electrode of the

micro-plasma jet source, above which SK-HEP-1 started to detach from the substrate while THLE-

2 remained intact. Our mechanical and biochemical analyses confirmed the presence of intrinsic

differences in the adhesion properties between the cancer and the normal liver cells, which provide

a clue to the differential detachment characteristics of cancer and normal cells to the atmospheric

pressure plasma. VC 2011 American Institute of Physics. [doi:10.1063/1.3622631]

Atmospheric pressure plasma (APP) has been suggested

as a biomedical tool for its simplicity in application and its

capability to generate abundant chemically reactive spe-

cies.1–3 While sterilization of medical devices and removal/

prevention of biofilm have been widely reported as the repre-

sentative applications of atmospheric pressure plasma,1–4

recent attempts have been made on the direct application of

plasma to wounded skin for enhancement of cellular healing

process5,6 and to dental cavities for sterilizing the infected

tissues.7 Recently, in vitro studies on a few mesenchymal

and epithelial cell types have indicated that APP induces cel-

lular necrosis, apoptosis, and detachment, suggesting its

potential use in cancer cell removal.8–11 However, induction

of such cellular changes would be meaningful in cancer ther-

apy practices if and only if removal was selective to cancer

cells leaving the normal cells unaffected.9 This requires a

parallel comparison of cancer and normal cells from the

same tissue under the controlled plasma condition. There-

fore, in the present study, the responses of cancer and normal

cells to the plasma treatment are compared to discover the

behavioral difference between two cell types in their detach-

ment characteristics from the substrate. We hypothesized

that either physical or bio-chemical stresses were responsible

for the plasma-induced changes in cells and performed sets

of experiments to elucidate the mechanism behind the dis-

tinct response between cancer and normal cells [Figs. 1(a)–

1(c)]. For instance, physical shear stress imposed by the gas

flow, as shown in Fig. 1(b), could possibly deliver a shearing

force to the adherent cells to peel them off from the sub-

strate. In addition, bio-chemical stress induced by reactive

oxygen species (ROS) can inflict damage on intracellular

proteins as well as surface proteins such as those involved in

focal adhesions (FAs), leading to the detachment of the cells

[Fig. 1(c)].

For the plasma applicator, we used a single pin electrode

type micro jet plasma, whose detailed description is reported

elsewhere12 and the additional information is provided in the

supplementary material.13

To study the effects of plasma on cellular viability, we

stained cells using live/dead assay13 following a brief expo-

sure of cells to APP (950-1200 V applied voltage, 50 kHz

driving frequency, 2 min treatment time with 2 slpm Helium

gas) and noticed the presence of differential responses

between two cell types. As shown in Figs. 2(a)–2(d), cells

begin to detach from the surface leaving a void above a criti-

cal voltage. At higher voltages, over 1000 V, three distinc-

tive regions are observed; the dead cell region at the center,

the live cell region at the periphery and the void region along

the interface between the live and dead zones [Figs. 2(c) and

2(d)]. The dead zone consists of necrotized cells, and the

void zone is the blank area of no cells, both of which collec-

tively refer to the plasma effective zone (PEZ) [Figs. 2(b)–

2(d)].

When the plasma above a critical voltage impinged on

the monolayer of the cells covered by phosphate buffered sa-

line solution, we observed cells starting to detach from the

FIG. 1. (Color online) The description of (a) an adherent cell to ECM

coated substrate through surface receptor protein integrins, (b) a cell detach-

ment forced by physical stress, and (c) by bio-chemical stress.

a)Present address: Korea Research Institute of Standards and Science, 267

Gajeong-ro, Yuseong-gu, Daejeon 305-340, South Koreab)Authors to whom correspondence should be addressed. Electronic

addresses: j_shin@kaist.ac.kr and wchoe@kaist.ac.kr.

0003-6951/2011/99(6)/063701/3/$30.00 VC 2011 American Institute of Physics99, 063701-1

APPLIED PHYSICS LETTERS 99, 063701 (2011)

Author complimentary copy. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp

substrate along the periphery of the plasma edge while cells

at the center region became necrotized.14 At the lowest

applied voltage of 950 V, the liver metastatic cancer SK-

HEP-1 cells started to detach from the substrate as shown in

Fig. 2(b). Unlike the case for SK-HEP-1, the normal liver

THLE-2 cells remained intact at the same applied voltage of

950 V [Fig. 2(a)]. At the highest applied voltage of 1200 V,

SK-HEP-1 exhibited a much enlarged plasma effective zone

including both the dead and void zones as shown in Fig.

2(d). To quantify the distinctive response of these two differ-

ent cell types, we plotted the fraction of live cells as depicted

in Fig. 2(e). SK-HEP-1 cells, having a larger PEZ and a

smaller fraction of live cells, seem more susceptible to the

plasma than THLE-2 at all applied voltages. Also, the quan-

titative results in Fig. 2(f) clearly show that the SK-HEP-1

void zone is distinctively larger than that of the THLE-2

cells by more than 1.5 times, which led us to question the

possible existence of fundamental differences in the adhe-

sion strength and ability to respond to external disturbance

among different cell types.11,14

In our results with metastatic cancerous SK-HEP-1 and

normal THLE-2 cells, the important question was whether

the differential detachment characteristics arise from the

intrinsic distinction between cancer and normal cell types in

the liver epithelium. Based on our observations, we hypothe-

sized that SK-HEP-1 cells would have weaker adhesion

strength than THLE-2 cells resulting in easier detachment,

which would also be consistent with the characteristics of

highly motile metastatic cells compared to that of sessile

type normal cells. In this scenario, when certain plasma

properties, either physical or biochemical, trigger cells to

detach from the substrate, cancer cells would detach more

readily due to their weaker nature in adhesion. To investigate

which of the physical or biochemical stress would have a

dominant effect on cellular detachment, we performed a cou-

ple of experiments. Although our previous study on THLE-2

cells has shown that the gas flow alone does not inflict intra-

cellular changes,14 there still remains a possibility that the

drag force of the liquid medium induced by the impinging

gas flow can impose physical shear stress on the adherent

cells. To test the possibility of functional dominancy of

physical shear stress in the plasma induced detachment of

cells, we designed a set of experiments using commercial

micro-channels (l-Slide VI0.4 ibidi GmbH) and investigated

whether SK-HEP-1 and THLE-2 responded differentially to

shear stress.13

The results shown in Fig. 3(a) indicate no significant dif-

ference in the adhesion strength against physical shear stress

between SK-HEP-1 and THLE-2 in all tested magnitudes

from 10 to 40 dyn/cm2. The critical shear stress at which

only 50% of the cells remain attached reflects the physical

adhesion strength of each cell. In our case, the critical shear

stresses for SK-HEP-1 and THLE-2 were measured to be

25.62 dyn/cm2 and 25.87 dyn/cm2. Since our shear experi-

ment confirms that the difference in cellular detachment

characteristics induced by physical shear stress is insignifi-

cant between the two cell types, we conclude that the

observed differential responses in SK-HEP-1 and THLE-2

upon plasma treatment are unlikely due to physical shearing

imposed by the gas flow.

The next candidate for cell detachment is the biochemi-

cal stresses from various plasma species. Since the chemical

species in the plasma have been shown to play critical roles

to induce changes in intracellular cytoskeletal proteins as

well as surface adhesion proteins,8,11,14 we designed a set of

experiments to test the susceptibility of cells to biochemical

stresses causing them to detach from the surface. Because

the chemical species originated from the plasma are likely to

alter membrane proteins causing cellular detachment in such

a short term treatment of 5 min, the detachment experiment

due to biochemical stress was limited to the digestion of

integrins serving as the linkage between the extracellular ma-

trix (ECM) on the substrate and the cells. To realize the

appropriate experimental conditions, we treated the cells

with trypsin-ethylenediaminetetraacetic acid (EDTA), which

is the most well known clipper of biotic anchors including

integrins. As shown in Figs. 3(b) and S1(c), when cells were

FIG. 2. (Color online) (a)-(d) The live (green)/dead (red) assayed samples

after plasma treatment. THLE-2 treated at the applied voltage of (a) 950 V

and (c) 1200 V and SK-HEP-1 treated at (b) 950 V and (d) 1200 V, respec-

tively (white circle: PEZ). (e) Proportion of live cells of SK-HEP-1 and

THLE-2 after plasma treatment where No is number of live cells of control

sample and N is number of live cells of plasma treated sample. (f) Areal pro-

portion of void for SK-HEP-1 (-�-) and THLE-2 (-*-) formed by plasma

treatment: Ao is total area from plasma treated sample, and A is void area

from plasma treated sample. “IMAGE J” software was used.

FIG. 3. (Color online) Red (right column) for SK-HEP-1 and blue (left col-

umn) for THLE-2 (a) Fraction of adherent cells brought about by different

shear force. (b) Schematic of de-adhesion process: time interval between i to

ii is s1, which represents the biological adhesion strength, and time interval

between ii to iii is s2, which reflects the elasticity of the cell. (c) De-adhesion

time s1 of SK-HEP-1 and THLE-2 at the biochemical stress condition.

063701-2 Gweon et al. Appl. Phys. Lett. 99, 063701 (2011)

Author complimentary copy. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp

treated with this enzyme, cells started to round up (i-ii with

s1) by decreasing the contact area between the cell and the

substrate and finally get detached by contractility (ii-iii with

s2). The de-adhesion time s1 was obtained by plotting the

cell’s areal changes over time using the following sigmoid

function [Figs. S3(a)–S3(c), Ref. 15]: Areanormalized

¼ 1� 1=½1þ expððt� T1Þ=T2Þ�:Here, s1 represents the adhesion strength of the cell rela-

tive to the substrate. The longer the s1, the higher the adhesion

strength while s2 characterizes the elasticity (or contractility)

of the cell.15 In our study, s1 can be considered as the toler-

ance for cells to biochemical stress inflicted by trypsin-EDTA.

As shown in Fig. 3(c), cancerous SK-HEP-1 featured much

shorter detachment time s1 than normal THLE-2 cells. The

differential de-adhesion time s1 between the two cell types is

consistent with our findings that SK-HEP-1 is more suscepti-

ble to plasma exposure than THLE-2, creating larger void

area [Fig. 2(g)]. These distinctive de-adhesion properties may

have originated from the difference in intrinsic properties of

the FA. Consistent to these results, the immunofluorescence

images indicate stronger and more pronounced paxillin dots in

the THLE-2 cells when compared to those in SK-HEP-1

[Figs. 4(a) and 4(b)], implying more stable FA in normal

THLE-2 than cancerous SK-HEP-1. From the confocal

images, only the paxillin dots, whose long axis is larger than 2

lm and the ends are connected to actin stress fibers, are selec-

tively counted as mature FA and their lengths are measured.

As shown in Fig. 4(c), we find that there exists almost twice

the number of mature FA in a single THLE-2 cell than those

in a single SK-HEP-1 cell averaged over 30 cells in 10 differ-

ent images. Interestingly, the distribution of the lengths of

focal dots is similar in both cell types except that only THLE-

2 cells feature very large FA whose length is longer than 7 lm

[Fig. 4(d)], possibly indicating more mature and stronger

FA.16 To test whether these phenotypic differences in FA cor-

relate with their activity, western blot assay was performed

for focal adhesion kinase (FAK) and a5 integrin protein [Figs.

S4(a) and S4(b)]. As shown in Fig. S4(a), more FAK proteins

exist in THLE-2 than in SK-HEP-1. Moreover, the quantita-

tive value of a5 integrin protein, which is a part of FA, consis-

tently reveals that THLE-2 has stronger coupling to the ECM

by showing more a5 integrins on THLE-2 [Fig. S4(b)]. These

results from liver cells are consistent with our aforementioned

hypothesis, where the presence of intrinsic difference in adhe-

sion properties between cancer and normal cells leads to dif-

ferential detachment behavior upon application of the

atmospheric pressure plasma, and the detachment is achieved

by biochemical disruption of anchorage proteins rather than

physical tearing off from the substrate.

To support our hypothesis, the similar experiments

were performed on another set of cancer and normal cells

from the mammary gland epithelium (MDA-MB-231 vs.

MCF10A) [Figs. S5–S7]. These two cells also exhibited

consistent detachment characteristics where the void zone

of cancer cells formed after plasma treatment was larger

than that of normal cells [Fig. S5(b)].

In summary, we found that the cancer cells (SK-HEP-1)

detached more readily compared to normal cells (THLE-2)

upon a short treatment by atmospheric plasma. Based on the

results from the biophysical and biochemical assays, SK-

HEP-1 was found to feature weaker adhesion strength than

THLE-2, exhibiting different responses against plasma treat-

ment, which seems to inflict biochemical stress to disrupt the

integrin anchorage of cells at the cell-ECM interface. This

difference in cellular adhesion property between two cell

types attributes to the intrinsic physiological difference of

focal adhesions between the two cancer and normal cells in

hepatocytes.

The authors thank Wonjong Song, Sunghyun Kim,

Unghyun Ko Sukhyun Song, Se Youn Moon, and Sunhee

Kim for their initial contribution in experiments. This work

was in part supported by KAIST.

1M. Moisan, J. Barbeau et al., Pure Appl. Chem. 74, 349 (2002).2M. Laroussi, J. P. Richardson et al., Appl. Phys. Lett. 81, 772 (2002).3B. Gweon, D. B. Kim et al., Curr. Appl. Phys. 9, 625 (2009).4X. T. Deng, J. J. Shi et al., J. Appl. Phys. 101, 074701 (2007).5G. Fridman, G. Friedman et al., Plasma Processes Polym. 5, 503 (2008).6S. U. Kalghatgi, G. Fridman et al., IEEE Trans. Plasma Sci. 35, 1559

(2007).7R. E. J. Sladek, E. Stoffels et al., IEEE Trans. Plasma Sci. 32, 1540 (2004).8I. E. Kieft, M. Kurdi et al., IEEE Trans. Plasma Sci. 34, 1331 (2006).9G. C. Kim, G. J. Kim et al., J. Phys. D: Appl. Phys. 42, 032005 (2009).

10G. Fridman, A. Shereshevsky et al., Plasma Chem. Plasma Process. 27,

163 (2007).11H. J. Lee, C. H. Shon et al., New J. Phys. 11, 115026 (2009).12D. B. Kim, J. K. Rhee et al., Appl. Phys. Lett. 91, 151502 (2007).13See supplementary material at http://dx.doi.org/10.1063/1.3622631 for the

text detailing for protocols of cell culture, live/dead assay, immunostaining,

Western blotting and other supporting experiments.14B. Gweon, D. Kim et al., Appl. Phys. Lett. 96, 101501 (2010).15S. Sen and S. Kumar, Cell. Mol. Bioeng. 2, 218 (2009).16J. M. Goffin, P. Pittet et al., J. Cell Biol. 172, 259 (2006).

FIG. 4. (Color online) Immunofluorescence image of paxillin dots (green)

and actin stress fibers (red) of (a) THLE-2 and (b) SK-HEP-1 (scale bar¼ 20

lm). Insets in (a) and (b) show the z axis-sectioned confocal microscopy

images (scale bar¼ 10 lm). (c) Number of FA of THLE-2 and SK-HEP-1

obtained by paxillin dots from the confocal images. (d) Distribution of focal

adhesion dots of THLE-2 and SK-HEP-1.

063701-3 Gweon et al. Appl. Phys. Lett. 99, 063701 (2011)

Author complimentary copy. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp