Ph. D. Dissertation Changes of Cellular Structure and Behavior Induced by Atmospheric Pressure Plasmas ( Gweon, Bomi) Department of Physics KAIST 2011

Transcription

1 Ph. D. Dissertation Changes of Cellular Structure and Behavior Induced by Atmospheric Pressure Plasmas ( Gweon, Bomi) Department of Physics KAIST 2011

2 Changes of Cellular Structure and Behavior Induced by Atmospheric Pressure Plasmas

3 Changes of Cellular Structure and Behavior Induced by Atmospheric Pressure Plasmas Major Advisor : Professor Choe, Wonho Co-Advisor : Professor Shin, Jennifer Hyunjong by Gweon, Bomi Department of Physics KAIST A thesis submitted to the faculty of the KAIST in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Physics. The study was conducted in accordance with Code of Research Ethics. a November, 26, 2010 Approved by Professor Choe, Wonho (Seal or signature) a Declaration of Ethical Conduct in Research: I, as a graduate student of KAIST, hereby declare that I have not committed any acts that may damage the credibility of my research. These include, but are not limited to: falsification, dissertation written by someone else, distortion of research findings or plagiarism. I affirm that my dissertation contain honest conclusions based on my own careful research under the guidance of my dissertation advisor.

4 Changes of Cellular Structure and Behavior Induced by Atmospheric Pressure Plasmas Gweon, Bomi The present dissertation has been approved by the dissertation committee as a Ph.D. dissertation at KAIST November, 26, 2010 Committee head (seal or signature) Committee member (seal or signature) Committee member (seal or signature) Committee member (seal or signature) Committee member (seal or signature)

5 DPH Gweon, Bomi. Changes of Cellular Structure and Behavior Induced by Atmospheric Pressure Plasmas.. Department of Physics p. Advisors Prof. Choe, Wonho, Prof. Shin, Jennifer Hyunjong. Text in English. Abstract The biomedical applications using the atmospheric pressure plasma (APP) came into widespread in recent years. The APP has been applied to many different kinds of cells (from prokaryotic to eukaryotic cells) and come up with interesting results. Yet, those studies tend to incline toward the application outcomes and pay less attention to the underlying physical, chemical, or biochemical mechanisms of the plasma which induced the cellular changes. Needless to say, not only the plasma but also the cells have complicated systems, and thus, there are many issues to be addressed for the effective utilization of the APP as the biomedical application tools. In this thesis, starting from understanding the APP s of the radio frequency (RF) large area and the low frequency (LF) micro-jet, optimal plasma conditions for the bio-applications were found, and then the cellular changes induced by the plasma were investigated by adopting both prokaryotic, E. coli and eukaryotic human cells as bio-targets. The MHz RF and the 50 khz LF plasmas were produced in the ambient air using helium gas. The role of the plasma including heat, UV photon, and radicals induced by the plasma was investigated in the E. coli inactivation, and the importance of the oxygen radicals was revealed. After a comparative study, it was found that both LF and RF plasmas have their own merits and drawbacks for treating bio-medical targets. For instance the LF plasma exhibited some merits such as lower gas temperature and easier accessibility to the target with its longer plasma length. Based on the preliminary study, the LF plasma was applied to the eukaryotic cells of THLE-2 and human dermal fibroblast. A direct contact with the LF plasma, generated at the applied electrode voltage of 1200 V for only 2 minutes resulted in various phenotypic intracellular changes which in turn induced the cell death and detachment from the substrate. Using the immunofluorescence assay, the necrotized cells were observed to undergo severe cytoskeletal disruptions except the non-electrically assembled intermediate filaments. In addition, the cells treated by the plasma at lower electrode voltage (< 1000 V) avoiding direct contact for a longer treatment time (10-20 min) experienced long term cellular changes of apoptosis and morphology or motility alteration within (24-48 hours). The following immunofluorescence analysis showed that the increase of the driving power frequency from 20 khz to 80 khz enhanced those cellular morphology changes. Throughout the experiments, the cells were treated with the plasmas under various discharge conditions, and it was inferred that the different plasma conditions induced different cellular changes for both short and long terms by means of reactive oxygen radicals or charged particles in the plasma. The next study was aimed at assessing the APP ability as a cancer therapeutic biomedical tool. The APP was applied to human liver SK-HEP-1 and mammary gland epithelial MDA-MB-231 cacner cells. The cancer cells were detached more easily than the normal cells for the same plasma treatment, suggesting the selective cancer cell removal possibility. For example, SK-HEP-1 cells were detached from the extracellular matrix (ECM) substrate at the input voltage of 950 V, while the normal THLE-2 cells remained intact. Similarly, MBA-MD-231 cells were detached at the input voltage of 1000 V, while the

6 normal MCF-10A cells were still attached even at the higher input voltage of 1200 V. This differential cellular behavior of the cancer and the normal cells to the plasma were attributed from different intrinsic characteristics of the cells such as maturity of focal adhesion, protein complexes through which the cytoskeleton of a cell connects to the extracellular matrix. These selective detachment characteristics of the cancer cells from the normal cells under the same discharge condition suggest its high potential as a medical therapeutic tool. Various cells were treated by the APP, and cellular changes were provoked due to the APP s highly active chemical species and electrically charged particles. These changes were different depending on the discharge conditions controlled by many operating parameters including input voltage, driving frequency, supply gas and flow rate, and treatment time. Thus, enhancing the APP controllability would be an important research issue for successful biomedical application of APP.

7 Contents Abstract Contents List of Figures i iv vi 1 Introduction Atmospheric Pressure Plasma (APP) and its Biomedical Applications Application targets: Prokaryotic Cells and Eukaryotic Cells Prokaryotic cell Eukaryotic cell Motivation of Thesis Structure of Thesis APP Characteristics for Bio-Applications and Preliminary Tests on a Prokaryotic Cell E. coli APP Sources Large Area Radio Frequency (RF) Plasma Low Frequency (LF) Micro Jet Plasma Inactivation of E. coli by Plasma Treatment Experimental Setup Effects of Heat and UV from APP

8 CONTENTS v Role of Oxygen Radicals from APP Inactivation Using LF Micro Jet Plasma Summary APP Condition Effects on Human Epithelial and Mesenchymal Cells Experimental Setup Temporal Phenotypic Cellular Changes by the Direct Contact of APP Necrosis and Detachment of Cells on Epithelial Origin Necrosis of Cells on Mesenchymal Origin Long Term Cellular Response by Indirect Contact of APP APP Induced Apoptosis of Cells on Epithelial Origin Migrational and Morphological Changes of Cells on Mesenchymal Origin Summary Different Effects of APP on Different Types of Cells: Cancer vs. Normal Cells Experimental Setup Differential Responses of Liver Cancer and Normal Cells to APP Liver Cancer (SK-HEP-1) and Normal (THLE-2) Cells Different Responses Related to the Intrinsic Biological Difference APP Effects on Mammary Gland Epithelial Cells Mammary Gland Epithelial Cancer (MDA-MB-231) and Normal (MCF10-A) Cells APP Treatment on Cancer and Normal Cells in a Co-Cultured Field Summary Conclusions 74 Summary (in Korean) 76 Bibliography 79

9 List of Figures 1.1 Previous APP studies for biomedical applications: (a) bacteria inactivation [22], (b) living tissue treatment [23], (c) blood coagulation [24], (d) dental cavity care [25], and (e) cell detachment [26] The cellular structure of prokaryotic and eukaryotic cell [47] Schematic of the plasma treatment on the human skin Cellular structures of the eukaryotic cells: three cytoskeletons and the focal adhesion assembly including the integrins Schematic of the large area RF plasma scource [72] (a) Plasma current for the input voltage: to the bottom ground (- -) and to the side ground (grounded casing) (- -), and (b) dissipated power: to the bottom ground (- -) and side ground (- -) [73] (a) Uniformity of the large area RF plasma varied by the gas flow rate (P in = 70 W), (b) Gas temperature of argon plasma (- -) and helium plasma (- -) measured by a Fiber thermometer Schematic of LF micro-jet plasma source [4] Image of (a) RF, and (b) LF single pin electrode jet plasmas [4] Temperature measured by a fiber thermometer for varying V in (a) LF and (b) RF [4] (a) Discharge current and plasma dissipated power for various O 2 mixing rates, (b) line ratio of O /He lines for various O 2 mixing rates (a) Preparation of E. coli samples on an agar plate, (b) large area RF plasma image (bottom view) with the E. coli spread agar gel in a Petri dish

10 LIST OF FIGURES vii 2.9 E. coli colonies on (a) control, (b) plasma treated (P in = 70 W) agar plates, and (b) survival curves of E. coli for increasing P in, (- -) 50 W, (- -) 60 W, (- -) 70 W, (c) D-value, the time interval for population deduction to one tenth from the original population, for increasing P in (a) The emission spectrum including UV emission between 220 nm and 280 nm, (b) The UV emission directly from the plasma (dotted line) and that transmitted the quartz plate (solid line) Survival curve of E. coli treated only by UV which penetrated through the quartz plate (- -) and the survival curve of E. coli treated by plasma (- -), P in = 70 W (a) O I (777 nm) emission intensity normalized by the He I (706 nm), (b) survival curve of E. coli for additional 10 sccm of oxygen gas in 6 slpm of helium: inactivation with additional to oxygen (- -), inactivation with only helium (- -) Deactivation of the E. coli spread agar by LF micro jet plasma for 30 sec. The circled area is where the plasma was impinged. The treatment conditions were (a) 13 mm from gas tube end, V in = 1000 V, 50 khz, (b) 19 mm from gas tube end, V in = 1000 V, 50 khz and (c) 13 mm from gas tube end, V in = 1000 V, 30 khz, (d) control (a) Schematic of LF plasma cell treatment. When the plasma spread out on PBS surface, hydraulic jump was formed at the boundary of the plasma impinged zone (= R eff ). (b) Plasma treatment process for the cell seeded slideglass sample Direct plasma treatment: (a) cell are seeded on slideglass, (b) DPBS thickness are 0.15 mm and the plasma directly contact to cells. Indirect plasma treatment: (c) cells seeded on coverglass bottom dish, (d) DPBS thickness are 2-4 mm and the plasma does not contact to cells (a) Live/Dead image of plasma treated THLE-2 cells: dead zone (DZ) where cells are necrotized, void zone (VZ) where cells are detached from the substrate, and live zone (LZ) where cells are minimally affected and still alive, (b) Areal changes of plasma affected zone for varying V in and (c) areal change of LZ against the plasma treatment time (fixed V in = 1200 V) Immunofluorescence image of Cleaved caspase-3 on (a) control, (b) plasma treated cells Immunofluorescence image of plasma treated THLE-2 cells: AF (red) and MT (green) (a) Boundary region between the dead cells by plasma and live cells and the enlarged image of (b) plasma impinged center and (c) enlarged image of live cells Plasma emission spectrum from LF plasma, (a) with PBS under the plasma, (b) without PBS under the plasma Immunofluorescence image of plasma treated THLE-2 cells: AF (red) and IF (green) (a) Boundary region between the dead cells by plasma and live cells and the enlarged image of (b) plasma impinged center and (c) enlarged image of live cells (Right) Detached cells after plasma treatment: cells lumped together around the void expressing green signals for live (green)/dead (red) assay. (Left) Enlarged image of the detached cells

11 viii LIST OF FIGURES 3.9 (a) The immunofluorescence image of the plasma treated THLE-2 where the integrin β-3 (green) and actin (red) were double labeled (a) The immunofluorescence image of the gas blown THLE-2 where the integrin β-3 (green) and actin (red) were double labeled The live/dead assayed fibroblast cells treated by plasma of (a) V in = 950 V, (b) V in = 1200 V (center), (c) (boundary), the immunofluorescence image of (d) plasma treated center and (e) the live area boundary Apoptosis assayed (a) control THLE-2 cells and (b) THLE-2 cells incubated for 24 hours after detachment by plasma (white arrows: apoptotic cells with annexin-v binding, green) Image of control cells after (a) 2 hr and (b) 24 hr from the time wound was scratch on the slideglass, and images after (c) 2 hr, (d) 24 hr from the wounding and plasma treatment (thickness of DPBS = 4 mm), (e) Areal reduction of wound The cells were treated twice for 10 minutes every 12 hours by the plasma at V in = 970 V and with different fin. Cell images of (a) control 10x, (b) control 40x, (c) plasma treated (f in = 20 khz) 10x, (d) plasma treated (f in = 20 khz) 40x (e) plasma treated (f in = 50 khz) 10x, and (f) plasma treated (f in = 50 khz) 40x Aspect ratio of cells were measured for control cells and the plasma treated cells (with f in = 20 khz, 50 khz and 80 khz) (a) Schematic of the image taking process for the plasma treated field. The images were taken as the 3x3 panel in order of numbers (1 to 9). Plasma treated cell images for various input voltages: (b) SK-HEP-1 with 950 V, (c) SK-HEP-1 with 1200 V, (d) THLE-2 with 950 V, and (e) THLE-2 with 1200 V (a) Fraction of live cells in the 3x3 image panel 4.1(a): THLE-2 (- -) and SK-HEP-1 (- -). (b) Fraction of the void area where cells detached from the substrate: THLE-2 (- -) and SK-HEP-1 (- -) SK-HEP-1 cells were incubated for 24 hours and the apoptosis test kit was applied (Apoptotic signal: annexin-v, green). (a) control, (b) plasma treated cells and (c) the fraction of apoptotic cells with annexin-v binding Schematic of cell detachment process by biophysical or biochemical stress Fraction of adherent cells for applied shear stresses (10 dyne/cm 2-40 dyne/cm 2 [91] (a) The cells attached to the ECM by solid integrin-ecm binding (τ 1 : the time interval between a and b). (b) The cells started to be detached from ECM and started to shrink to 50% from the original cell area (τ 2 : the time interval between b and c). (c) The cells totally detached from the ECM using their own elastic force De-adhsion test by Trypsin-EDTA, images of (a) THLE-2, (b) SK-HEP-1 cells after Trypsin-EDTA (from right to left 0 s, 60 s, 120 s) (a) The Boltzmann sigmoidal fitting using the equation 4.1 to get the De-adhesion time τ 1 (THLE-1 ( ) and SK-HEP-1 ( )). (b) De-adhsion time τ 1 for THLE-1 ( ) and SK-HEP-1 ( )

12 LIST OF FIGURES ix 4.9 Immunofluorescence image taken by fluorescence microscope (a) THLE-2 and (b) SK-HEP- 1 [paxillin (green), actin stress fibers (red)], and the images taken by confocal microscope for (c) THLE-2 and (d) SK-HEP (a) The length of paxillin dots for THLE-2 ( ) and SK-HEP-1 ( ) and (b) the distribution of paxillin dot length for paxillin dots of THLE-2 ( ) and SK-HEP-1 ( ) (a) The FAK of each THLE-2 and SK-HEP-1 cells (the inset is the western blot band) (b) The α5-integrin of each THLE-2 and SK-HEP-1 cells (the inset is the western blot band) Plasma treated cell images for the input voltages of 1000 V: (a) MCF-10A and (b) MDA- MB (a) Fraction of live cells in the 3x3 image panel 4.1(a): MCF-10A (- -) and MDA-MB-231 (- -). (b) Fraction of the void area where cells detached from the substrate: MCF-10A (- -) and MDA-MB-231 (- -) (a) The Boltzmann sigmoidal fitting using the equation 4.1 to get the De-adhesion time τ 1 (MCF-10A ( ) and MDA-MB-231 ( )). (b) De-adhsion time τ 1 for MCF-10A ( ) and MDA-MB-231 ( ) Images of cells (THLE-2, SK-HEP-1, MCF-10A, and MDA-MB-231) before and after H 2 O 2 (2 M) treatment for 10 min (a) Immunostaining image of MDA-MB-231 and (b) MCF-10A labeled at 5 integrin in green color and actin in red color Co-cultured image of normal (blue) and cancer (red) cells: (a) before plasma treatment and (b) after plasma treatment

13 CHAPTER 1 Introduction Plasma is a gaseous substance consisting of charged particles and neutrals. The atmospheric pressure plasma (APP) is one of the plasma types which was first employed in the industrial processing field, expecting economical cut off by removing the complex vacuum systems. Although the initial goal of APP in industry was not fully achieved due to the physical barrier of the atmospheric pressure discharge conditions, many beneficial aspects of APP were revealed. For example, the abundance of the active radicals, the simple discharge system, and the non-thermal characteristic make the APP attractive. Even though APP is difficult to control, many physical characteristics were discovered aiding in controlling of the plasma and constructing various plasma discharge systems [1 12]. Lately, based on the studies of the basic APP characteristics, bio-applications have attracted much attention and shown high potential of expansion with many exciting results. In this chapter, a brief introduction to the physical characteristics of APP and their present status for biomedical applications are addressed. Additionally, a short introduction of biological backgrounds, the motivation, and the structure of this thesis are provided. 1.1 Atmospheric Pressure Plasma (APP) and its Biomedical Applications In the early studies regarding APP, the subjects about the industrial processing were studied to meet the industrial needs such as the plasma welding and the surface modifications [13 21]. For APP discharge sources in early studies, the radio frequency (RF, MHz) plasma has been extensively used in

14 2 Chapter 1. Introduction a a. E-coli [Laroussi et al. (2002)] b. Living tissue (animal model) [Fridman et al. (2008)] c. Blood cell [Fridman et al. (2007)] d. Tooth [Jiang et al. (2009)] e. Mammalian cell [Eva et al. (2008)] b c d e Figure 1.1: Previous APP studies for biomedical applications: (a) bacteria inactivation [22], (b) living tissue treatment [23], (c) blood coagulation [24], (d) dental cavity care [25], and (e) cell detachment [26]

15 1.1 APP & Biomedical Application 3 planar capacitively-coupled plasma (CCP) system as in low pressure plasma. Recently, the RF single pin electrode type corona source became one of the popular sources. In low frequency (LF) power range, one of the classical plasmas source is the dielectric barrier discharge (DBD) [2], which usually use direct current (DC), pulse, or LF power. As in RF plasma, one of the popular plasma sources is the single pin electrode type, or micro jet plasma, plasma which was used for the small size plasma generation and the localized surface treatments. In parallel to the developing the plasma sources, the physical and chemical characteristics of the plasmas were studied for decades to understand and control the APP. Using those CCP and pin type corona systems, the plasma mode transition depending on the input voltage and discharge pressure [6] and the change of the plasma characteristics depending on the plasma driving frequency [4, 12] were investigated. In addition, electrical and optical characteristics studies of the various plasmas were done [3, 10, 11]. The CCP generated at the moderate high pressure range (1-760 Torr) exhibited different discharge modes and transition voltages depending on the pressure [6]. For the micro-jet plasmas with the RF (13.56 MHz) driving frequency showed several distinctive discharge characteristics from the those of the LF. While the plasmas generated by the frequency under 2 MHz have different discharge modes for each half period: the negative glow and the positive streamer, the discharge was more continuous in time for the MHz RF plasma [4]. Plasma diagnostic methods were also needed to be developed because the well defined diagnostic tools for the low pressure discharge could not be applied to the high pressure discharges. There is a relatively simple and cost effective method for the plasma gas temperature measurement which is one of the optical emission spectroscopic (OES) methods relating the gas temperature to the dimolecular rotational temperature [7]. The high pressure (> 1 torr) characteristic makes the plasma smaller, but the plasma density is higher than the low pressure plasmas. The non-equilibrium plasma characteristic makes the plasma non-thermal so that the plasma gas temperature can be kept low [5]. The biomedical applications using APP can be categorized into two different types; the application to the biomedical devices and the direct application on the bio-organisms (e.g. cell, tissue and living body [Fig. 1.1]). The application to biomedical devices can be represented as the surface treatments of biomedical materials. There are already various types of conventional plasma surface treatment tools used at the laboratories [15]. Lately, the surface properties modification using plasma was adopted to make bio-compatible surfaces for the polymers of medical devices or the scaffolds for tissue engineering [16,17]. APP is a convenient method for modifying surface properties because it changes surface wettability and enables the functional groups and grafting of polymer chains on the surface [16]. Thereby, APP can improve the adhesive property for biological tools including titanium or polymer scaffolds [18]. In addition, the nano- fabrication has also adopted the plasma surface modification method following the current trend. There were attempts to make the nano pattern which guides the neuron cell growth [19]. For the industrial applications, there are also attempts to use the plasma on scaffolds modifying the scaffold surface for bio-compatible surfaces. Some of the plasma treatment can be used at surgical tools to avoid the blood or tissues to detach and make the coagulation on the tools. Plasma applications directly on the bio-organisms was started in 1980 s as the sterilization tool of microorganisms. The low pressure plasma was applied first at bio-organisms using its high energy particles, highly reactive radicals, and UV photon during the process [27,28]. Gradually, the sterilization process was expanded to the atmospheric pressure plasmas. Conventional sterilization tools, for instance,

16 4 Chapter 1. Introduction the autoclave device which only works at the high temperature and pressure conditions, make the nonthermal atmospheric pressure plasma more beneficial. Based on the favorable process conditions of low temperature, simple system, dry condition etc., the plasma stretches out its application to other biomedical tools. The dental cavity care [25, 29, 30], the cleaning of surgical instruments [31], the blood coagulator [24, 32, 33] are just a few examples. Beyond utilizing APP as the biomedical material processing tools, the target of interest has recently been expanded to the mammalian cells of human bodies. Gathering recent results together, the APP applications can be divided into two different ways considering the plasma effects to cells: the beneficial or the destructive influences to cell proliferation. At the beneficially effective regime, APP can help the wounded skins or unhealthy tissues to recover. The other destructively effective regime, APP can remove the tumorous or pathological tissues from the body. So far, many attempts were tried to achieve those medical goals such as the plasma applications to animal models, wounded skin, and cancer cells [23,34 39]. Interestingly, there are reports that plasma induces the suicidal cell death, or apoptosis, for cancer cells which originally is not programmed to die by itself [39, 40]. It was reported that the plasma alters the DNA structure of mitochondria result in the cell apoptosis [39]. Additionally, there are also reports that plasma helps to heal the wounded skins [23,24]. The exact mechanisms of the plasma were not yet discovered. 1.2 Application targets: Prokaryotic Cells and Eukaryotic Cells Prokaryotic cell The prokaryotic cells are the organisms that do not have nucleus and other organelles. This structural feature is the big difference from the eukaryotes which have the nucleus as well as organelles inside the cell membrane. Bacteria is one of the representative single-celled, prokaryotic cells. Bacteria are inhabitable almost every where, for example, in soil, acidic waters, radioactive waste, and even deep inside the Earth s crust [41]. The size and the shape of the bacteria is multifarious. The typical size is a few µm which is far smaller than the eukaryotic cells, and the type of shape ranges from rod to spiral [42]. Some species of bacteria kill or attack other microorganisms, which is called the predatory bacteria [41, 43]. Instead, certain bacteria species has beneficial function of survival of other microorganisms making the special associations, which is called the mutualists [44]. If bacteria form a parasitic habitat on other organisms, they are called the pathogens. Pathogenic bacteria are a major cause of human disease such as diphtheria, cholera, food-borne illness, leprosy and tuberculosis and so on. Furthermore in some case human can die of the pathogenic bacteria infections [45]. The bacterial cell is encapsuled by a lipid membrane as shown in Fig The cell membrane holds the contents of a cell and has a function as a barrier. As mentioned before, bacteria are one of the prokaryotic cells so that bacteria do not have membrane-bound organelles in their cytoplasm and thus contain few large intracellular structures. They consequently do not have a nuclear membrane, mitochondria, and the other organelles present in eukaryotic cells such as the Golgi apparatus and endoplasmic reticulum [46]. Bacteria were first considered as a very simple bags encapsuling cytoplasm, but elements such as prokaryotic cytoskeleton, and the proteins in specific locations within the cytoplasm

17 1.2 Application Targets 5 Prokaryotic cell Eukaryotic cell Plasma memebrane a Cytoplasm DNA Nucleoid region Nucleus Ribosome Figure 1.2: The cellular structure of prokaryotic and eukaryotic cell [47] have been found showing certain levels of complexity. In addition, the bacteria can be categorized to gram positive and gram negative cells by the membrane type Eukaryotic cell One of the representative eukaryotic cells is animal cells. The plants, fungi and protists are also categorized as the eukaryotic cells. Only the animal cells will be introduced in this chapter and in more specific, the human cells were exclusively employed for the plasma biomedical application experiments in this thesis. There are three kinds of cell categories for animal cells which are divided from the animal embryogenesis process: the endoderm, the mesoderm, and the ectoderm. The endoderm is formed by the inner layer of the embryo cells to become columnar and forms the epithelial lining of the most of the digestive tube. It also forms the lining cells of all the glands which open into the digestive tube. As a result, the endoderm forms the stomach, the liver, the pancreas, the urinary bladder, the colon, the intestines etc. On the other hand, during the embryogenesis, some of the cells move inward forming the layer between the endoderm and the ectoderm. That cell layer Which is the mesoderm. The mesoderm helps to forming the cardiac muscles (the heart), skeleton, skeletal muscle, red blood cells (lymph cells) etc. The ectoderm is the outmost layer of germ, later becoming the origin of a surface tissue. The ectoderm produces tissues within the epidermis, neurons within the brain, and constructs melanocytes. The ectoderm forms the central nervous system, the lens of the eye, cranial and sensory, the ganglia and nerves, pigment cells, head connective tissues, the epidermis, hair, and mammary glands [48, 49]. Cells from epithelial origin [50] Epithelial cells are cells that forms the inner and outer surfaces of the body in continuous sheet layer called as epithelial membranes or epithelia [51]. Epithelial tissues can develop from ectoderm, mesoderm and also endoderm. Epithelial cells are connected together by

18 6 Chapter 1. Introduction Figure 1.3: Schematic of the plasma treatment on the human skin

19 1.2 Application Targets 7 three cell junctions (tight junctions, adhering junctions and gap junctions) [52]. Not like other adhering and gap junctions, which found at other cells, tight junctions only found at epithelial so that the tight junctions can be a important marker to figure out the cell origin. The epithelial cell layer is attached to the underlying connective tissue. Epithelia do not contain blood vessels but instead are supplied with nutrients and oxygen by blood vessels in the connective tissue. The role of these cells are protection, sensation, absorption, secretion, excretion, diffusion etc. Epithelial cells can be subdivided by the shape and specialization. The types of epithelial cells are squamous, cuboidal, and columnar epithelial cells. Squamous epithelial cell have thin and irregular flat plate-like shape. They usually constitute lining of cavities. Cuboidal epithelial cells have cubical shaped and their nucleus usually placed in the center of the cell. Columnar epithelial cells have a elongated shape as a column. Columnar epithelial cells have microvilli which looks like a small hair. By this microvilli cells can increase their surface area. [53]. Cells from mesenchymal origin [54] Most of the mesenchymal cells are derived from the mesoderm and some mesenchymal cells are from neural crest cells which is originates from the ectoderm. However, many embryologists only call for the cells derive from the mesoderm to mesenchymal cell [55]. Mesenchymal cells have spindle like shape, not like polygonal shape of epithelial cells, forming linkage to other cells to build up a multi-dimensional cellular network [56]. In addition, mesenchymal cells have the character of stem cells and can differentiate into osteoblasts, adipocytes, myoblasts, or chondroblasts [57,58]. Thus the mesenchyme, a loosely organized embryonic tissue which is formed by unspecialized mesnchymal cells and a ground substance matrix of a reticular fibril aggregate, develop to all connective tissue type of the body: muscle, vascular and urogenital systems and so on. Mesenchymal cells can migrate easily, unlike epithelial cells, which do not have great mobility because they are organized into closely adherent sheets layers. Since, mesenchymal cells can migrate to a particular region of the germ layers that will give rise to a particular tissue and perform a essential function for the progression of tissue development. There are interesting reports about mesenchymal cells, that the epithelial cells can change into mesenchymal cells, the epithelial-mesenchymal transition (EMT). This change accompanied by the controlled of cell adhesion molecules [59]. This transition was found frequently during embryogenesis and also after invasion of metastatic cancer cells, where the distinctive characteristics of epithelial cells are lost [59]. In reverse, the change of mesenchymal cells into epithelial cells is also exists and called as mesenchymal-epithelial transformation (MET). Cell Structures and Functions Eukaryotic cells have much larger and complicated structure than prokaryotic cells. Since the cells have large body, they need the specialized internal membrane-bound organelles to support the structure, carry out metabolism, and transport chemicals throughout the cell. As shown in Fig. 1.4, there are three types of cytoskeleton structures which support the cellular shape, actin stress fiber (AF), microtubule (MT), and intermediate filament (IF). AFs and MTs are connected to the integrins which is crosslinked between cells and Extracellular Matrixes (ECM) and deliver the signal between cells or cell-ecm. Integrins are part of focal adhesions (FAs) or the cell-matrix adhesions [60,61]. FAs are specific types of macromolecular unit which transmit signals and mechanical force and mediate the effects of extracellular matrix (ECM) adhesion and regulates cell behavior [62]. Focal adhesions are large, dynamic protein complexes through which the cytoskeleton of a cell connects to the extracellular matrix, or ECM. They are limited to clearly defined ranges of the cell at which the plasma membrane closes to within 15 nm of the ECM substrate [63]. Focal adhesions are not stable system, thus the integrin associates continuously with one ligand and disassociates with it to meet

20 8 Chapter 1. Introduction Schematic of Intracellular Structures AF IF MT ECM Integrins Paxillin AF Talin Cell membrane Integrins Ligands ECM Figure 1.4: Cellular structures of the eukaryotic cells: three cytoskeletons and the focal adhesion assembly including the integrins

21 1.3 Motivation of Thesis 9 another ligand and make the integrin-ligand binding unit. As signals are transmitted to other parts of the cell, relating to anything from cell motility to cell cycle [64, 65]. Focal adhesions have proteins more than 100 different types, which suggests a considerable functional diversity [66]. They actually serve for not only the anchorage of the cell, but also can function beyond that as signal carriers (sensors) which inform the cell about the condition of the ECM and thus affect their behavior [67]. In sessile cells, focal adhesions are quite stable under normal conditions, while in moving cells their stability is diminished: this is because in motile cells, focal adhesions are being constantly assembled and disassembled as the cell establishes new contacts at the leading edge, and breaks old contacts at the trailing edge of the cell. One example of their important role is in the immune system, in which white blood cells migrate along the connective endothelium following cellular signals and to damaged biological tissue. Actin is a globular protein which is soluble in aqueous solution the monomeric subunit of filaments in cells. AF participates in many important cellular processes including muscle contraction, cell motility, cell division and organelle movement inside cell, cell signaling, and the establishment and maintenance of cell junctions and cell shape [68]. Microtubules are polymerized structure of α- and β-tubulin dimers. The tubulin dimers polymerize end to end in protofilaments. Then the protofilaments make bundles into hollow cylindrical shape to make the filamentous structure. Microtubules are nucleated and organized by the microtubule organizing centers (MTOCs) such as centrioles and basal bodies. MTs are part of the cytoskeleton and participate in many other processes in addition to structural support. For that, there are motor proteins which make organelles or other cellular factors move along the MT network. To support the mechanical part, they generate force by changing the length [69]. Intermediate filaments (IFs) are a family of related proteins that share common structural and sequence features. There are about 70 different genes coding for various IF proteins. However, different kinds of IFs share basic characteristics: they are all polymers that generally measure between 9-11 nm in diameter when fully assembled. IF are subcategorized into six types based on similarities in amino acid sequence and protein structure [70]. 1.3 Motivation of Thesis Despite the huge interests in the biomedical applications using APP, the mechanisms of which provokes the cellular changes have not been been clearly elucidated. APP has some favorable discharge conditions for the biomedical applications: sufficient radicals which makes it possible to treat the target effectively and the low gas temperature which avoids the thermal damage of the materials [71]. So far, many biomedical applications have been achieved by many researchers. Several corona based (needle/pin to plane or coaxial source structures) small size plasmas have been developed to treat living cells [34, 36, 38, 39]. However, most of the studies were limited to the superficial observations of the cellular changes brought about by plasma treatments under various conditions. The bottom mechanism causing the cellular changes is after all the most important issue for applying the plasma as biomedical tools. In this study, not only the plasma physics but also the biological research methods were carefully considered to understand the cellular changes led by the plasma. Most of the studies in this field has focused on the mere observations rather than elucidation of the biological significance. To utilize the APP for the biomedical applications, however, the biological significance should be understood. For example, decrease in the cell motility upon the plasma treatment may affect crucial biological processes such as wound healing and cancer metastasis. Since cellular changes are not only expressed by the

22 10 Chapter 1. Introduction visual means, just observing the cellular movement or the viability may not completely represent the intracellular changes. Hence, various biological analysis methods were used in this study to observe the intracellular changes including distribution or quantity of the proteins caused by APP. These are directly related to the fundamental processes of cells such as cell fate, adhesion properties, motility, proliferation, and so on. Moreover, the effects of plasma on different cell types and the cellular responses to different plasma conditions should also be investigated. Cells on different origins could experience the exposure to different condition of plasmas. In other words, plasma may affect the cells in different strengths depending on the cell position; for instance, when it comes to tissues epithelial cells, they will be directly exposed to plasma while the dermal cells will be indirectly exposed. In addition, since cells exhibit different biological and mechanical characteristics, the plasma treatment may have result in different responses. For example, the cancer cells might exhibit different cellular responses to normal cells and the characteristic cellular responses depending on the cell types is important to utilize the plasma as cancer treatment. Therefore, in this thesis study, the APP was applied on various types of cells with different plasma conditions to approach the goal of using the APP as biomedical tool. By finding the key working species of the plasma and by finding the biological significance, the plasma control and the concrete application target will be successfully established. 1.4 Structure of Thesis Motivation of Thesis In Chapter 2, two different types of the atmospheric pressure plasma sources are introduced: RF large area and LF micro-jet plasmas. The physical parameters of the plasmas were each measured, and the adequate plasma discharge conditions for bio-applications were investigated. The preliminary study on cells were carried by applying both plasmas to the prokaryotic cells, E. coli in this case. Heat, UV photon, and oxygen radicals of the plasma parameters were individually studied for finding the underlying inactivation mechanism. The cellular changes derived by the LF micro-jet plasma were studied in Chapter 3, where the short term responses like cell necrosis and detachment and the long term responses like cell apoptosis, motility changes, and differentiation changes were found. The cells were the epithelial origin (THLE-2) and the mesenchymal origin (human dermal fibroblast) cells. Additionally, many biological tools were used to find out the subcellular changes in order to explain the cellular changed induced by plasma. After understanding the cellular responses induced by the LF micro-jet plasma, the plasma was applied to various types of cells considering the medical application of the cancer therapy, and the details are given in Chapter 4. First, the cellular changes of the human liver cancer (SK-HEP-1) and the human liver normal (THLE-2) cells were compared to find out how the plasma differently affects each cell owing to the intrinsic cellular characteristics. In addition, another cancer and normal cell group of mammary gland epithelial cells were compared to check the general application purpose. The conclusions and possible further works are briefly mentioned in Chapter 5.

23 CHAPTER 2 APP Characteristics for Bio-Applications and Preliminary Tests on a Prokaryotic Cell E. coli As introduced in the previous chapter, the atmospheric pressure plasma (APP) source is easy to fabricate and there is a wide range of plasma sources that we can use. Among those atmospheric pressure plasma sources, the plasmas sources can be categorized based on their driving frequencies. Apart from the microwave plasma which uses high frequencies ranging from hundreds of MHz to a few GHz or some of the torch plasma, most of the plasmas are non-thermal plasma. This low temperature characteristic of non-thermal APP is useful for bio-applications, since most of the targets are thermally vulnerable. In this chapter, characteristics of two plasma sources, the radio frequency (RF) large area plasma and the low frequency (LF) single electrode jet type plasma were investigated, verifying the bio-compatibility of these plasma sources. As for a preliminary test of APP on organisms, the plasmas were applied to a prokaryotic cell, E. coli. Then the major plasma species that influence cells were examined by controlling the plasma parameters. 2.1 APP Sources Large Area Radio Frequency (RF) Plasma Considering the bio-target and the thermal sensitivity of the bio-surfaces, the gas temperature of the APP needs to be around the room temperature. Only the helium gas was used for generation of the

24 12 Chapter 2. APP Sources & Bio-application Intro plasma in this experiment. Since various radicals are expected to be present during the bio-applications, the plasma was produced in to an open system where oxygen, nitrogen and moisture exist. Two types of plasma sources were used for the experiments with bio samples. The large area RF plasma was used. In Fig. 2.1 the large area plasma is shown where in the bottom part of the image is 110 mm 15 mm. The plasma source has a powered rod electrode covered with a dielectric material inside a grounded case (side ground). To increase the discharge efficiency, the grounded casing was placed near the powered electrode in a 1 mm gap. With this electrode structure, the powered and grounded electrodes were sufficiently close together to decrease the discharge voltage. The dielectric cover on the powered electrode was adopted to prevent the arcing between the close powered electrode and grounded electrodes. The electrical characteristics of this large area RF plasma was measured by the voltage probes (1:1000, Tektronix P6015A), (1:10, Tektronix TCP202). While the voltage and current were measured, Indium tin oxide (ITO) glass was placed under the powered electrode [Fig. 2.2(a)]. As depicted in Fig. 2.2, the electrical discharge properties were measured. Discharge currents through the side- and the bottom-grounded electrodes were measured as shown in Fig. 2.2(a). Also shown in Fig. 2.2(a), the root-mean-square discharge current (I rms ) increased with increasing discharge voltage (V rms ), which indicates that the plasma is in the abnormal glow discharged condition. In Fig. 2.2(b), the average dissipated power was calculated by 1 T Vrms I rms dt over a period T. It is noticeable that the portion of dissipated power guided to the bottom-grounded electrode is very small compared to the portion guided to the side-grounded electrode. For example, as seen in Fig. 2.2(b), at an input power of 100 W, 75% of the power was consumed by the plasma and only 15% of the power consumed was dissipated to the bottom-grounded electrode, resulting in the low discharge current. This also indicates that the plasma was produced at the gap between the grounded casing and the powered electrode first and then diffused out via the bottom-grounded electrode. The spatial uniformity was measured by the plasma image processing. At atmospheric pressure when the plasma size or area is large, it is hard to maintain the plasma uniformly. Considering that the plasma size is 110 mm 15 mm, it is important to sustain the a uniform plasma treatment across the broad surface of the bio materials. The charged-coupled device (CCD) camera image was taken through the bottom-grounded electrode under the plasma source. Since the plasma condition was easily influenced by the gas flow for the atmospheric pressure plasma [9], the plasma uniformity was measured with variation of the gas flow rate. The uniformity was calculated by 1 (max min) (max+min) as plotted in Fig. 2.3(a). The plasma achieves its most uniform rate (0.9 arbitrary units) when the gas flow rate was over 6 slpm. With this information, the large area RF plasma source was used with a gas flow rate of 6 slpm to ensure the uniform plasma treatment. A spectrometer (Chromex 250is) with a CCD detector was employed to obtain the plasma emission spectrum. The spectrum was used to measure temperatures; this experimentally-obtained diatomic molecular spectrum was compared to the calculated spectrum to determine the rotational temperature [7]. At atmospheric pressure the plasma s rotational temperature is known to be similar to the plasma gas temperature. To compare the rotational temperature to the experimentally measured gas temperature, the gas temperature was measured using a fiber thermometer (FISO FOT-H), as shown in Fig. 2.3(b). As shown in Fig. 2.3(b), the plasma generated with argon gas tends to have higher gas temperature compare to the plasma generated with helium. Thus only the helium gas was used for generation of the plasma, so that the gas temperature of the plasma was maintained around room temperature in order

25 2.1 APP Sources 13 RF Power Supply Dielectric Powered electrode Current probe A A V Voltage probe Figure 2.1: Schematic of the large area RF plasma scource [72]

26 14 Chapter 2. APP Sources & Bio-application Intro a 3 Side discharge Bottom discharge Irms(A) 2 1 b Powerdiss(W) Voltage rms (V) 120 P diss at side discharge P diss at bottom discharge Power in (W) Figure 2.2: (a) Plasma current for the input voltage: to the bottom ground (- -) and to the side ground (grounded casing) (- -), and (b) dissipated power: to the bottom ground (- -) and side ground (- -) [73]

27 2.1 APP Sources 15 a Uniformity (a.u.) b Flow rate (lpm) Gas temperature ( o C) Input power (W) Figure 2.3: (a) Uniformity of the large area RF plasma varied by the gas flow rate (P in = 70 W), (b) Gas temperature of argon plasma (- -) and helium plasma (- -) measured by a Fiber thermometer

28 16 Chapter 2. APP Sources & Bio-application Intro Gas Powered electrode LF Power Supply V A Figure 2.4: Schematic of LF micro-jet plasma source [4] to the discharge voltage was maintained at a low level Low Frequency (LF) Micro Jet Plasma The LF micro jet plasma which is described in Fig. 2.4 was used. A copper pin of 360 µm in radius is covered with alumina except the pin end to generate the plasma from the pin head. This pin is placed inside a pyrex tube of 2 mm in radius for guiding the helium gas supply. The helium gas was supplied at a 2 slpm flow rate. The voltage, with a sinusoidal waveform, was applied to the powered pin electrode. For a wavelength resolved image, an ICCD camera was used. The discharge voltage and current were measured using appropriate probes (Tektronix P6015A, Pearson 4100, ProSys EI-100). The LF micro jet plasma type is one of the most popular types of atmospheric pressure plasma because of its simple system and that it is useful for localized plasma treatment. Although in the previous studies RF power is mainly used for this type of jet plasma, LF power was used in this study because it allows greater control of temperature and easier accessibility to the application target [4]. From the previous study, it was shown that the LF power is safer for bio targets, for instance human skin, because the low temperature and low discharge current of the LF plasma has lower risk of causing the electric shocks on the surface of the skin and thermal damage [74]. As mentioned earlier, since plasma formed inside the gas guiding tube protrudes out when using LF power, the LF plasma has a longer length while the plasma with RF power only forms the plasma at the pin head [Fig. 2.5]. It is known that the plasma produced using RF power also bulges out from the gas guiding tube when the grounded electrode is sufficiently close, however without the grounded electrode the plasma is localized to short lengths. These

29 2.1 APP Sources 17 a b d Figure 2.5: Image of (a) RF, and (b) LF single pin electrode jet plasmas [4] characteristic features occur because the electrical characteristics are not consistent between low and high driving frequency power due to the behavior of the ions. At high driving frequency, f in, around a few MHz, where the ions cannot respond to the E-field, electrons solely moved along the short path along the applied field so that the electron loss is reduced and the breakdown voltage is largely reduced. However, when the low frequency power is applied to the electrode, since the frequency is large enough, both electrons and ions can respond to the applied field and since the oscillation amplitude is large enough both electrons and ions can reach the target surface. As a result, both species are largely lost by hitting the boundary or the grounded surface, and the plasma is sustained by the secondary electrons [4, 37]. Because of this, the energy efficiency is higher with RF power than LF power, which, at the same time makes the plasma sustainable at a higher gas temperature. As depicted in Fig. 2.6, the jet type plasma with both LF and RF power frequencies does not have a high gas temperature. However, LF does not change much with increasing input voltage, but the gas temperature of the RF plasma increases, in total 10 o C when the input voltage was increased from 400 V to 500 V. The maximum gas temperature measured was around 310 K which is 20 degree higher than the that of LF plasma and exceeds the allowable gas temperature range for bio-organisms. Through this result, it was shown that the LF power for single pin electrode jet plasmas is more appropriate for the following biomedical application experiments. As mentioned from previous studies [40,75], it was shown that the oxygen radicals in the plasma play a crucial role at interacting with cells. Thus, in Fig. 2.7, to control the oxygen related radicals in APP, oxygen was mixed with the supplying helium gas: this enhances the oxygen radical effect. However, if the plasma is not stable enough the added oxygen atoms will disturb the plasma and decrease the power coupling rate to the plasma, so the change in electrical properties was monitored [Fig. 2.7]. As expected, with the additional oxygen supply to the helium, plasma dissipation power decreased along the discharge current, as shown in Fig. 2.7(a). When the oxygen is more than 20 sccm, the power

30 18 Chapter 2. APP Sources & Bio-application Intro a Gas temperature ( o C) b Gas temperature ( o C) V (V) rms V rms (V) Figure 2.6: Temperature measured by a fiber thermometer for varying V in (a) LF and (b) RF [4]

31 2.1 APP Sources 19 a I p-p (ma) b Line Ratio Current Power O2 ` flow rate (sccm) 0 sccm 10 sccm 20 sccm Power (mw) Vin (V) Figure 2.7: (a) Discharge current and plasma dissipated power for various O 2 mixing rates, (b) line ratio of O /He lines for various O 2 mixing rates

32 20 Chapter 2. APP Sources & Bio-application Intro coupling efficiency dramatically drops. In addition, the line intensity ratio of O /He which indirectly reflects the amount oxygen radicals in plasma were similar for the lines under 1300 V of electrode voltage [Fig. 2.7(b)]. Therefore, to get a greater number of radicals, the plasma input voltage should be higher than 1200 V where the plasma can be sustained in a stable condition and the oxygen supply should be higher than 20 sccm. 2.2 Inactivation of E. coli by Plasma Treatment The non-equilibrium atmospheric pressure plasmas have recently attracted much attention compared to the low pressure plasmas due to their many advantages including simple system for treatments and abundance of active radicals [27]. In fact, sterilization is an important issue in the bio-medical field because contaminated surgical tools or instruments used in experiments may put medical practices and bio-experiments in danger. Thereby, many sterilization tools have been developed, such as autoclaves, chemical drugs, and ultra-violet (UV) sterilizers. So far, most conventional sterilization techniques make it difficult to avoid damaging and aging the object being sterilized or show low efficiencies. Recently, to fulfill the demands of cold sterilization, many atmospheric pressure plasma sources were developed and tested on bio-surfaces and materials. Although several attempts to find the mechanism for deactivating microorganisms by plasma treatment were carried out, the exact mechanism and the major species in the plasma that cause deactivation have not yet been elucidated. There are many active species in plasma. For instance, highly active radicals, charged particles, and UV photons are those. However, their role in sterilization has not yet been defined. In the low pressure plasma regime, there are some reports that sterilization utilizes energetic charged particles and also UV photons from the plasma. However, at atmospheric pressure (at high pressure those species are easily absorbed and dissipated into the plasma) the energetic ions and the UV cannot work in the same way as they do at low pressure. There are instead some expectations that the active radicals in the plasma are the key factor in sterilization, but this has not yet been clearly revealed. The plasma inactivation mechanism and the principal plasma species in the inactivation process have been investigated for both RF large area plasmas and LF micro jet plasmas Experimental Setup Plasma Source As it was shown in previous studies, the gas supplied and discharge type influence the inactivation efficiency. In this work, a large area RF plasma source was used with the helium for safe and efficient treatment. With helium as the supply gas the plasma was produced at a relatively low temperature, so that the gas temperature was maintained within the bio-compatible range. The grounded electrode was placed under the plasma source, and used to support the petri dish with the E. coli-coated agar gel. The distance between the plasma source and the agar surface was 2 mm. A radio frequency (RF, MHz) power source (Young-Sin RF, YSE06F) was applied to the powered electrode from 10 to 100 W with helium gas supply at 6 slpm. This helium supply rate was fixed to maintain the plasma at a uniform condition as it was shown. Because the target object is a bio-material, the thermal effect to the cells should be removed in the first place to understand the bona fide plasma effect, so that only the plasma parameters including gas

33 2.2 E. coli Treatment 21 temperature, discharge voltage, current, power were investigated. Gas temperature was measured using a fiber thermometer (FISO, FOT-H) and the rotational temperature was measured by optical emission spectroscopy (OES) [7]. The fiber thermometer, which has a thin optical fiber at the end, was employed because conventional thermometer, thermocouples, have an exposed conductor for the probing unit and general mercury filled thermometer have a blunt end which will disturb the small size of the atmospheric pressure plasma. In Fig. 2.3, it is shown that the gas temperature measured by the fiber thermometer was a little higher than the OES which is consistent with the tendency for the gas temperature to be a little higher than the rotational temperature. The fiber thermometer seems to be influenced by the heat of the conductor which is only few millimeters from the thermometer head. Although the gas temperature increased with increasing input voltage, over the range of W for the input power, the gas temperature did not exceed 80 o C which is beneficial for this plasma source considering that the target objet is a biomaterial [Fig. 2.3(b)]. Cell Not only the plasma type itself but also the target configurations may affect the sterilization efficiency. For example, whether the bacteria is gram positive or gram negative will also affect the inactivation rate. Here, E. coli was used as a representative gram negative bacteria. E. coli is one of the easiest targets to inactivate, although it is the most abundant and sometimes varies to a dangerous pathogen. Since the agar plate should not be contaminated by other unknown microorganisms, Luria- Bertani (LB) agar ampicillin which is an anti-bacterial was added to the agar. Since the sterilization effect can vary depending on the cell condition, the bacteria preparation protocol is a important issues to get an exact result in this experiment. The E. coli transformed by plasmid to be ampicillin resistance, was incubated for 8 hours in 2 ml LB media at 37 o C and prepared on the agar plate where ampicillin was added to forbid the growth of any microbes except the E. coli. The E. coli was diluted with the LB media in ratio of 1 : 10 2, 1 : 10 3, 1 : 10 4, and 1 : 10 5, and then, 50 µl of the diluted E. coli was added to the LB agar plates and spread evenly. After the plasma treatment, the plates were incubated at 37 o C for 12 hours and used to count the E. coli colonies for quantification of the inactivation results. As depicted in Fig. 2.8, the E. coli spread LB agar plate was placed under the plasma source, which is 2 mm apart from the agar surface. Although the plasma size was 110mm 15mm, the width was 15 mm which is not enough to cover the whole agar plate so that the agar plate with a diameter of 40 mm was moved four times and each 1/4 of the agar plate was treated separately [Fig. 2.8]. For the following inactivation experiment using the LF single pin electrode jet type plasma, the E. coli sample were also prepared on the agar plate but in a monolayer Effects of Heat and UV from APP Figures 2.9(a) and (b) show the E. coli samples in a suspension diluted in a ratio of 1 : 10 2 and incubated for 12 hours before and after plasma treatment, respectively. The plasma was produced at 70 W of input power and each of 1/4 of the sample surface was treated for 30 seconds (in total 120 seconds). As shown in Fig.2.9(a), a high density colonies were formed on the control agar sample, and the image of the agar plate with colonies was taken after an incubation time of 12 hours for the E. coli suspension diluted in a 1 : 10 2 ratio. In contrast, in Fig. 2.9(b) lots of colonies had disappeared on the plasma treated agar surface. To quantify the inactivation effect for various plasma input power ranges, the E. coli inactivation effect was represented by a survival curve with a log scale. The survival curve shows the

34 22 Chapter 2. APP Sources & Bio-application Intro a b Figure 2.8: (a) Preparation of E. coli samples on an agar plate, (b) large area RF plasma image (bottom view) with the E. coli spread agar gel in a Petri dish.

35 2.2 E. coli Treatment 23 changes in the colony forming unit (CFU) as a function of plasma treatment time. As it is seen from the Fig. 2.9(c), raising the input power increases the E. coli inactivation effect. To compare the treatment efficiency for various input power levels, the D-value for each input power was obtained. D-value is the time constant of the survival curve where the CFU decreases to 1/10 from the initial value and this time constant represents the sterilization effect. In previous studies, there are reports that there are multiple values for the slope of the survival curve, where there are multiple inactivation processes attributed to the plasma. However, since the survival curve shown in Fig. 2.9(c) created from a single valued for each input power, it can be assumed that the inactivation mechanism for this plasma was simple. In Fig. 2.9(d) the changes in the D-value are presented that the D-value is 22 s for 50 W of input power and the D-value decreased to 11 s for 70 W meaning that the inactivation time reduces to half and the inactivation effect increased by 100% for higher input power of 70 W. One question that might occur is whether this increase in the inactivation effect may occur because of the increase in the gas temperature of the plasma. However, it was found that only for the temperatures over 70 o C does E. coli start to slow down its activities from the other study [76]. However, for this plasma condition, as investigated in the previous section [Fig. 2.3(b)] the gas temperature was only 47 o C which can rule out any heat effect in E. coli inactivation. This shows that the heat from the plasma was not playing a critical role in this work. Moreover, the E. coli was also inactivated at 50 W with a gas temperature of around 38 o C which is known to be the optimal temperature for the bacterial proliferation [76]. The role of UV photons emitted from the plasma in inactivating E. coli was investigated. As shown from previous studies, UV emission in the range from 220 nm to 280 nm is fatal to cells. Actually, the energy of the UV photon in this range is not the most aggressive but the membranes of cells have a the window to this UV range, resulting in UV photons being able to penetrate the membrane and harm the DNA of the cells [77]. In the large area RF plasma which was used in this experiment, the basic spectrum discharged by the helium includes nitrogen molecular lines, hydroxyl lines, and oxygen lines because the plasma produced in ambient air contains oxygen, nitrogen and moisture. As shown in Fig. 2.10(a), the plasma with 50 W of input power does not have the conspicuous UV emission in the range of germicide. On the other hand, at 70 W and 100 W of input power, the UV emission is stronger so that it becomes hard to neglect. At atmospheric pressure, UV photons are known to be easily reabsorbed so that UV does not play a critical role in the inactivation process Role of Oxygen Radicals from APP The role of highly active radicals in plasma in inactivating E.coli was investigated. For the helium plasma, the main emission spectral lines are from atomic helium. Because of the gas mixing from the air, oxygen, nitrogen and hydroxyl line also appear. In previous studies, oxygen radicals are were detected as one of the important plasma species inactivating the microorganisms [27,28,75]. To study the oxygen radical effect, oxygen was added to the helium gas supply (6 slpm). Since too large an amount of the oxygen gas being supplied affects to the power coupling of the plasma, and may influence the plasma conditions, only amount smaller than 0.1% (< 10 sccm) of oxygen were added. To find the relationship between the additional gaseous oxygen and the oxygen radicals produced inside the plasma, the spectral intensity of the O I line (777 nm) normalized with the He I line (706 nm) was observed. The intensity increased to 3 times and 5 times by adding 5

36 24 Chapter 2. APP Sources & Bio-application Intro a b c 10 5 N (ea) d D-value (s) Treatment time (s) Input power (W) Figure 2.9: E. coli colonies on (a) control, (b) plasma treated (P in = 70 W) agar plates, and (b) survival curves of E. coli for increasing P in, (- -) 50 W, (- -) 60 W, (- -) 70 W, (c) D-value, the time interval for population deduction to one tenth from the original population, for increasing P in

37 2.2 E. coli Treatment 25 a 3000 Relative intensity (a.u.) b Relative intensity (a.u.) NO γ OH γ N 2 N 2 + He I O I Wavelength (nm) Plasma + UV Only UV Wavelength (nm) Figure 2.10: (a) The emission spectrum including UV emission between 220 nm and 280 nm, (b) The UV emission directly from the plasma (dotted line) and that transmitted the quartz plate (solid line)

38 26 Chapter 2. APP Sources & Bio-application Intro N (ea) Treatment time (sec) Figure 2.11: Survival curve of E. coli treated only by UV which penetrated through the quartz plate (- -) and the survival curve of E. coli treated by plasma (- -), P in = 70 W sccm and 10 sccm of oxygen, respectively [Fig 2.12(a)]. O I lines occurring when only He is supplies are assumed to originate from the air because the discharge system was the open to ambient air. For direct examination of the role of the oxygen radical, the inactivation experiment was accomplished with varying the oxygen contents. The survival curves depicted in Fig. 2.12(b) show that the inactivation rate of E. coli with an input power of 70 W, one showing the results with oxygen added, and the other without oxygen; the latter was about 40% higher. This result strongly suggests that inactivation depends on the active radicals such as excited oxygen atoms Inactivation Using LF Micro Jet Plasma The feasibility of applying the LF micro jet plasma to cells was studied. Since the gas temperature of this LF plasma was [Fig. 2.6], only around room temperature, thermal damage can be safely negligible in the bacterial inactivation. Since basic interactions between the plasma and E. coli were investigated in the previous sections; the effectiveness of the plasma on bacteria was simply tested using the LF micro jet plasma. E. coli was prepared as a monolayer type on agarose gel so that the effectiveness of plasma treatment can and the area can easily be identified. The fully grown control sample looks like the sample displayed in Fig. 2.13(d) where the E. coli have grown as a monolayer. However as it is presented in Fig. 2.13, at the plasma treated surfaces, there are blank areas of around 2.6 mm diameter on the monolayered E. coli surface where cells were inactivated and thus no growth is visible. Interestingly, each of (a), (b), and (c) in Fig were treated for 30 seconds,however the profile of the dead zone was a little different for each plasma condition. Firstly, the distinctive thing among the samples is that there is a donut shape [Fig. 2.13(c)] which is different to the circular shape of other samples [Fig. 2.13(a)]. Secondly, comparing the Fig. 2.13(a) and (b) the circle is a little bit larger in (b) where a higher input

39 2.2 E. coli Treatment 27 a O I / He I 10 5 b N (ea) O 2 addition (sccm) Treatment time (s) Figure 2.12: (a) O I (777 nm) emission intensity normalized by the He I (706 nm), (b) survival curve of E. coli for additional 10 sccm of oxygen gas in 6 slpm of helium: inactivation with additional to oxygen (- -), inactivation with only helium (- -)

40 28 Chapter 2. APP Sources & Bio-application Intro voltage were applied to generate the plasma. By comparing (a) and (c) where different driving frequency were applied, it is clear that at higher frequencies the plasma profile changes and the intensity at the outside of the plasma becomes higher. As a result, the outside of the plasma will die and the E. coli in the center survives. In addition, it is clear that with the higher input voltages the plasma will spread in larger area and thus affects a greater area of E. coli. 2.3 Summary In order to optimize the plasma source for bio-organisms, the basic physical characteristics of the RF large area plasma and LF micro jet plasma were investigated. The temperature, electrical and optical features were measured. The temperature was controlled so that a plasma at non-thermal region could be applied to cells without causing any thermal damages. In addition, to find out which plasma species affects the cells most, the preliminary study on a prokaryotic cell, E. coli, was carried out by varying the plasma parameters. The maximum input power level was fixed at 70 W to avoid any thermal damage and to induce bona fide plasma effects on the cells. The D-value was reduced to 50% from the initial value, from 22 sec to 11 sec, as the input power was raised from 50 W to 70 W. Throughout the input power changes, from 50 W to 70 W, the plasma gas temperature, UV emission strength, and the oxygen radical production from the plasma were compared to identify the key player in E. coli inactivation. As a result, it was shown that both thermal heating and UV photons did not play a significant role in the plasma inactivation process, but that excited oxygen atoms are the dominant species in the inactivation. When additional oxygen was mixed in the helium supply, the E. coli CFU reduction rate increased by about 40% at a constant input power of 70 W. Throughout this experiment, the main active species were perceived to be an oxygen-related radical in APP. Additionally, the LF micro jet plasma was selected for the bio-compatible plasma source owing to its low plasma temperature and accessibility to the target surface. Thus the E. coli inactivation experiment was also performed using the LF micro jet plasma. Throughout the LF plasma treatment, it was verified that controlling the power driving frequency and the distance are important in ensuring that the target cells are affected uniformly.

41 2.3 Summary 29 a b c d Circle diameter (mm) Std. Err. (±) Figure 2.13: Deactivation of the E. coli spread agar by LF micro jet plasma for 30 sec. The circled area is where the plasma was impinged. The treatment conditions were (a) 13 mm from gas tube end, V in = 1000 V, 50 khz, (b) 19 mm from gas tube end, V in = 1000 V, 50 khz and (c) 13 mm from gas tube end, V in = 1000 V, 30 khz, (d) control

42 CHAPTER 3 APP Condition Effects on Human Epithelial and Mesenchymal Cells The atmospheric pressure plasma (APP) has been suggested as a useful biomedical tool due to its abundant chemically reactive species and its non-thermal characteristics. Recently, the direct application of APP on wounded skin and cancer cells, reported by several reports, has fulfilled the expectations of biomedical applications of APP. As shown in the previous chapter, APP application on prokaryotic cells or bacteria, has been studied for many years and some of the inactivation mechanisms were investigated for several plasma types. However, the study of the effects of plasma on cells or tissues had only just begun and much of the mechanism is not yet clearly understood. Therefore, the effects of plasma on cells will be investigated in detail in this chapter. This study may be very significant for the future application of APP on human skin or organs because plasma can either have advantageous or disadvantageous effects in treatment. Therefore, the probable effects on a cellular level were scanned to utilize plasma in medical practice; human hepatocytes (THLE-2) originating from epithelial and human skin fibroblasts originating from mesenchyme were employed as the target cells. These cells were selected since they have a high possibility of being exposed to atmospheric pressure plasma when plasma impinges on the skin. Both the epithelial cells and mesenchymal cells are located at the outermost layer of tissue, which makes them our first target for plasmas.

43 3.1 Experimental Setup Experimental Setup In this study, LF pin type jet plasma was used. In previous studies, RF power was mainly used to discharge the pin type jet plasma for cell treatment. However, LF AC power was used in this study, which has relatively higher discharge voltage but lower discharge current and lower temperature than RF plasma, so preparing the cell samples was simpler. The plasma discharge system and the cell samples were designed considering the cell condition as a priority. The temperature of the cell was carefully maintained by applying warm Dulbecco s Phosphate-buffered saline (DPBS, , Gibco) of 283 µl, an ion balanced buffer which will keep the cells from suffering osmotic shock, at 37.5 o C on the monolayered cell layer. The saline, at the same time, protects the cells from dehydration by the gas flow from the the plasma source. The DPBS thickness was calibrated to 0.15 mm, which was the appropriate condition where the cells can be directly affected by the plasma but not harmed by other external factors like cooling effects or gas flow. Helium gas of 2 slpm was applied to generate the plasma and the treatment time varied from 2 to 10 min. The input voltage also varied from 950 V to 1200 V. Since the plasma jet is near room temperature with the low frequency power supply at 50 khz, thermal effects on cells can safely be neglected. The monolayer of THLE-2 (ATCC, CRL2706) and human dermal fibroblast cells was grown to 70% confluency and 90% confluency respectively. Cells were seeded on the slideglasses (25 mm 75 mm) or coverglass-bottom dishes (dish diameter = 35 mm, hole diameter = 13 mm, , SPL). As shown in Fig. 3.1(a), these slideglasses (blue in color in the figure) were placed 15 mm below the pyrex tube (black in color in the figure) end. After the plasma treatment cells were observed by using various biological assays, the live/dead assay, and immunostaining. After labeling the cells or the intracellular part in a specific color, the cells were observed by a microscope (Axiovert 200M, Carl Zeiss) or a confocal microscope (LSM510 META NLO, Carl Zeiss). Cells without the labeling color were observed with the phase image using the microscope (NIKON TS100F, Nikon) Cell culture As mentioned before, THLE-2 and Human dermal fibroblast cells were used in this experiment. The cells were cultivated at 75 cm 2 flask, however, for the experiments, cells were seeded on the slideglass or the coverglass bottom dish. Before culturing THLE-2, the liver epithelial cell, the flasks or slideglasses or coverglass bottom dishes were precoated with a extracellular matrix (ECM) protein mixture of 0.01 mg/ml fibronectin ( , Invitrogen), 0.03 mg/ml bovine collagen type 1 (C8919, Sigma) and 0.01 mg/ml bovine serum albumin (BSA, aa8806, Sigma). For fibroblast cells only mg/ml of fibronectin was used to form the ECM on the surface of slideglass or coverglass bottm dish. The subcultivation ration was 1:3 for both cells. Bronchial epithelial basal medium (BEBM, Lonza) along with bronchial epithelial growth medium (BEGM, CC-3170, Lonza) kit was used for the growth medium of THLE-2, and Dulbecco s modified medium (DMEM, Lonza) was used for fibroblast cells. Usually, when the cells reached to 70-80% of confluency, the cells were subcultured. First, the culture medium was removed and cells were rinsed with DPBS. After that, 2 to 3 ml of Trypsin-EDTA solution ( , Gibco) was added to flask and waited for 5 to 7 min. Then, 0.1% of Soybean Trypsin inhibitor ( , Gibco) was added to neutralize Trypsin-EDTA. Cell suspension were transfer to conical tube and contrifuged at 125 xg for 5 min to remove Trypsin-EDTA. Fibroblast cells and THLE-2 cells were cultured by the similar protocol but the time of trypsinization was different because the cellular response to trypsin-edta was different; fibroblast cells detached faster.

44 32 Chapter 3. APP Effects on Human Dermal Cells a Gas Powered electrode LF Power Supply V R eff A b Figure 3.1: (a) Schematic of LF plasma cell treatment. When the plasma spread out on PBS surface, hydraulic jump was formed at the boundary of the plasma impinged zone (= R eff ). (b) Plasma treatment process for the cell seeded slideglass sample

45 3.1 Experimental Setup 33 Live/dead assay With this live/dead assay (Molecular Probes), the cell viability can be tested. The constituents of this live/dead assay are Calcein-acetoxymethryl (Calcein-AM, 4 µm, green) and Ethidium Homodimer-1 (EtdH-1, 2 µm, red). Calcein-AM is a membrane-permeable dye and at live cells Calcein-AM will be hydrolyzed by intracellular esterases which will express a strong green fluorescent color. On the other hand, EtdH-1 binds to DNA inside the nucleus when the cells die and the membrane is disrupted. Because EtdH-1 is a membrane-impermeable dye, it will be only expressed on the dead cells. Immunofluorescence This technique is a form of fluorescence microscopy that is used when the intracellular structures are needed to be observed. We can conjugate the fluorescent object to the target molecules by using the antibodies for specific antigens. First, the cells were fixed with 3.7% formaldehyde (533998, Sigma) solution for 20 min to retain the cell shape and location of the proteins inside the cells. Second, the cells were permeabilized by using a detergent, 0.2% TritonX-100 (T8787, Sigma) dissolved in DPBS solution, for 15 min to dissolve the membranes. After this permeabilizing step, antibodies could get into the cell through the dissolved membrane. After permeabilization, all steps were processed on an ice pack. For the third step, cells were blocked by a 3% bovine serum albumin (BSA, 0332, Amresco) solution so that the non-specific expression could be blocked out. The blocking step was conducted three times and each time a 20 min wait was applied. Next, primary antibody dissolved in BSA solution was loaded on to the cells overnight. After the over night incubation, cells were incubated with the secondary antibody dissolved BSA solution to conjugate the fluorochromes. The nuclei was stained by DAPI (4,6-diamidino-2-phenylindole, 1:50000, D1306, Molecular probes). Then, 1 or 2 droplets of the mounting solution, Vectashield (H-1000, Vector laboratory), were added on the sample surface before covering by coverglass to avoid the fast bleaching during the imaging. At last, the coverglass was sealed airtight on the slideglass by nailpolish to prevent cells from drying out. In this chapter, for the primary antibody, cleaved caspase-3 antibody (1:400, 9661S, Cell signalling), the anti α tubulin (1:200, A11126, Molecular probes), the anti-cytokeratin 18 (1:250, ab52948, Abcam), and the anti-integrin-β3 (1:100, ab7167, Abcam) were all used. For the secondary antibody, Alexa series (1:200, A1101 (anti-mouse), A1104 (anti-rabbit), Molecular probes) were used for fluorescence labeling. Apoptosis test There are two types of cell death. One is necrosis which is cell death that occurs because of membrane disruption, which arises due to external impact over a short time interval. The other type of cell death is apoptosis which is programmed cell death triggered by an inner signal that comes out from the nucleus passing through the cell cycle. Usually, when there are abrupt environmental changes or stresses, cells decide to die and step into the apoptosis process. There are various stresses which may trigger apoptosis. While the signalling pathway of the dying process is very complicated, there are some markers which can detect the apoptotic sign from cells. An apoptosis test kit (Propidium Iodine, Annexin-V, Biovision) will show that the apoptosis process is proceeding from cells by binding annixin-v at the membrane which is disturbed by the apoptosis process [78,79]; this is a simple procedure to test whether cell death is apoptotic or not. Additionally, there are other markers and another test will also be used in this study: caspase-3 was immunostained, this is a direct marker which shows early stage of cell signalling for apoptosis. Plasma treatment: direct and indirect treatment As seen in Fig, 1.3, there are different kind of cells forming the skin tissue. Especially, the cells from epithelial origins and mesenchymal origins is in different position inside the skin tissue. Epithelial cells which covering the skin as forming an outmost

46 34 Chapter 3. APP Effects on Human Dermal Cells a b PBS thickness: 0.15 mm c d PBS thickness: 2-4 mm Figure 3.2: Direct plasma treatment: (a) cell are seeded on slideglass, (b) DPBS thickness are 0.15 mm and the plasma directly contact to cells. Indirect plasma treatment: (c) cells seeded on coverglass bottom dish, (d) DPBS thickness are 2-4 mm and the plasma does not contact to cells

47 3.2 Temporal Phenotypic Cellular Changes by the Direct Contact of APP 35 layer may directly affected by plasma [Fig. 1.3]. In addition, the bacteria may also exist near the surface layer. However, the mesenchymal cells forming the connective tissue in dermis tissue so that the plasma may not directly contact with the mesenchymal tissue. Considering this physiological condition, the plasma treatment was proceeded in two different ways. The direct and indirect plasma treatment as described in Fig Cells, especially epithelial cells, were prepared on the slideglass and the saline buffer were added only having 0.15 mm of thickness [Fig. 3.2(a) and (b)]. The mesenchymal cells were prepared on the coverglass bottom dish where the dish wall exist to support the thicker layer of saline buffer on cells [Fig. 3.2(c) and (d)]. 3.2 Temporal Phenotypic Cellular Changes by the Direct Contact of APP Necrosis and Detachment of Cells on Epithelial Origin As reported in the previous studies, there are various cellular responses when the plasma treatment is applied to cells. Cell necrosis [38], apoptosis [39, 40] and detachment [34, 38] from the substrate are detected after plasma treatment. Those different cell reactions are induced by different plasma conditions or strengths and also by exposure time to the plasma [34]. Likewise, all those events were also observed in this study with the LF micro jet plasma. Especially, for the LF micro jet plasma, a unique multiple zones phenomenon was observed when cells were treated with the plasma, as shown in Fig 3.3(a) and (b). The image of Fig 3.3(a) was taken for the cell sample treated by the plasma with 1200 V of V in. Then, the area of DZ and the plasma effective area (the total area of DZ and VZ) in Fig 3.3(b) was obtained by varying the electrode voltage from 950 V to 1200 V. There are necrotized cells that are circular in shape which coincide with the plasma shape on the DPBS surface at the center surrounded by the void area in donut shape, and there are alive cells outside the void area which seem not to be affected by the plasma or minimally influenced by the plasma over short periods. As described in Fig. 3.1(a), when the plasma impinged on the DPBS barrier of the cells, this saline layer was pushed away out of the circle of radius R eff [see enlarged part of Fig. 3.1(a)] leaving the minimum amount of water, which protected the cells from drying. This effect provides direct plasma contact to the cells inside the the circle of radius of R eff. In contrast to the area outside of this circle, where cells will be protected by the hydraulic barrier, cells are left alive. This multiple zone seems to be formed after 1 min of plasma treatment Fig 3.3(b). Cell Necrosis Since there was no distinctive disruption at the cell membrane in the phase image from microscopy, except a small portion of cells shown to be aggregated and disrupted by the plasma, which could link the cell death directly to necrosis, the cell should be proven to be necrotized at the initial step. Therefore, the apoptosis test were carried out first to confirm whether the cell death is apoptosis or necrosis. Immunofluorescent labeling of cleaved Caspase-3 was used because the expression of cleaved Caspase-3 is known to be the most accurate way of sorting out the apoptotic cells. In Fig. 3.4(b) the immunofluorescent image of plasma treated THLE-2 is presented where the green color shows cleaved caspase-3 which has no distinctive difference compare to the control of THLE-2 [Fig. 3.4(a)]. There is a small difference in

48 36 Chapter 3. APP Effects on Human Dermal Cells a LZ VZ b Plasma effective area (mm 2 ) c Plasma effective area (mm 2 ) mm DZ Input voltage (V) Treatment time (sec) Dead zone area (mm 2 ) Figure 3.3: (a) Live/Dead image of plasma treated THLE-2 cells: dead zone (DZ) where cells are necrotized, void zone (VZ) where cells are detached from the substrate, and live zone (LZ) where cells are minimally affected and still alive, (b) Areal changes of plasma affected zone for varying V in and (c) areal change of LZ against the plasma treatment time (fixed V in = 1200 V)

49 3.2 Temporal Phenotypic Cellular Changes 37 a b Figure 3.4: Immunofluorescence image of Cleaved caspase-3 on (a) control, (b) plasma treated cells the location of the expression, however this difference does not show that the cells have highly apoptotic signals. In conclusion, it is affirmative that the cells directly died by plasma streamer, indicating that they did not die by apoptosis but by necrosis. In previous studies there are explanations that the cells froze or were desiccated by the plasma [38], however this is not sufficient explanation because there are distinctive differences between the drying out and death, as shown in Fig To find out the exact reason for cell death, the intracellular structures were observed closely using immunofluorescence. Especially, three cytoskeleton proteins were visualized, actin filaments (AFs), microtubules (MTs), and intermediate filaments (IFs), these support the cell body and the binding between the substrate and cellular skeleton. Interestingly, as shown in Fig. 3.5, the actin structure, which is one of the cytoskeleton proteins which supports and forms the cell structure shows total disruption. At the plasma-impinged center, the red signals of actin disappear which shows a big difference from the vivid red signals at the outside of the plasma impact zone where the cells were alive in the immediate post-plasma treated moment. Since actin was labeled by a phalloidin stabilizing agent which binds to the F-actin (the fiber structure), not directly on to actin monomers themselves, the disappearance of the signal does not means that the actin monomers were denatured but the disassembly of the fibrous structure which was clearly confirmed by observing the MTs. In Fig. 3.5(a) and (b), the microtubule, which was stained with a green color, also shows total disruption of the fibrous shape. When looking at the enlarged image shown in Fig. 3.5(b) the monomers of the MTs were scattered around the plasma treated dead cells. Unlike the staining on AF, the MT antibody directly conjugated to the MT monomers. It was verified in previous studies that the reactive oxygen species (ROS) causes the disruption of cytoskeleton structures especially AF. One piece of research suggested that AF fragmentation was caused by the addition of H 2 O 2 to the growth medium [82 84]. From those results the cytoskeletal disruption may explained by the ROS generated from APP. The Fig. 3.6(b) shows possible radicals from the LF

50 38 Chapter 3. APP Effects on Human Dermal Cells a b c 100 μm 50 μm Figure 3.5: Immunofluorescence image of plasma treated THLE-2 cells: AF (red) and MT (green) (a) Boundary region between the dead cells by plasma and live cells and the enlarged image of (b) plasma impinged center and (c) enlarged image of live cells

51 3.2 Temporal Phenotypic Cellular Changes 39 2 a N 2 Intensity I /I He (a.u.) b OH N2 + Hel O Wavelength (nm) Figure 3.6: Plasma emission spectrum from LF plasma, (a) with PBS under the plasma, (b) without PBS under the plasma micro jet plasma, specifically the red marks on the spectrum show the oxygen-related radicals. There is a possible explanation for this cell death, that the collapse of the cytoskeleton structures shift the position of the organelles inside the cells and induce enough stress to result in cell death [85, 86]. However, the sequentiality of cell death and the cytoskeletal disruptions is still debateable, and there is another hypothesis that cell death occurs for other reasons, for instance the ROS themselves [82]. The important thing is that the ROS from the LF micro jet plasma are somehow related to the cell death and the cytoskeleton disruption. Another interesting thing was found for the third type of cytoskeleton structure in cells. The IF also has a fibrous structure [Fig. 3.7(c)], however as seen in Fig. 3.7(a) and (b) the fibrous structures are still preserved where the AFs are totally disturbed. The basic building principle of IFs is different from that in AFs or MTs, in which the negatively charged monomers are assembled together [87,88], instead, in IFs the electrically neutral rod domain monomers form the dimer block by entangling together as filaments which then assemble into longer filament-like structures. As a result, there exists distinctive difference between the plasma effects of IFs compared to the effects on AFs and MTs. This discrepancy suggests another factor of plasma on cellular changes. As known from previous DBD studies, charge accumulation on dielectric surfaces by plasma application is certain [89]. In this experimental structure, the cell membrane is a dielectric surface where the plasma is targeting, so the charged particles from the plasma will clearly be stacked upon the cell membrane [22]. This electrostatic force may interfere with cytoskeleton formation to alter the fibrous cytoskeleton structures. Interestingly, AFs and MTs formed by the negatively charged monomers are disrupted while the IFs not formed by charged monomers are intact. This result can suggest one possibility that the electrostatic force interferes the intracellular structures. However, considering

52 40 Chapter 3. APP Effects on Human Dermal Cells a b c 100 μm 50 μm Figure 3.7: Immunofluorescence image of plasma treated THLE-2 cells: AF (red) and IF (green) (a) Boundary region between the dead cells by plasma and live cells and the enlarged image of (b) plasma impinged center and (c) enlarged image of live cells

53 3.2 Temporal Phenotypic Cellular Changes 41 VZ Detached cells Figure 3.8: (Right) Detached cells after plasma treatment: cells lumped together around the void expressing green signals for live (green)/dead (red) assay. (Left) Enlarged image of the detached cells the intrinsic biological structural difference between three different cytoskeletons the selective disruption cannot explain the strong relation between the cytoskeletal disruption and electrostatic force which may caused by plasma. Combining these results, the intracellular changes such as cytoskeletal disruptions were observed at the necrotized cells by plasma. Especially, the biochemical factor, ROS, can be suggested one of the major factors of the intracellular changes. Changes of Cell Adhesional Properties After the APP treatment, the detached cells were observed as shown in Fig The detached cells were lumped together around the boundary of the void area and the live area after the plasma impinging. The viability of the plasma treated cells were checked using the live/dead assay and, as shown in the Fig. 3.8, the detached cells were still alive which was shown by the presence of a green color which floats near the live cells. To find out the reason for the cell adhesion property changes, the cellular changes of the plasma treated cells were looked into. Since the detached cells washed away, these cells could not be fixed and immunostained instead the periphery of the cell detached area was observed. To find out the plasma effects on cell adhesional property, plasma-induced alterations to the cell membrane were observed using integrin-β3 immunofluorescence because the structure related to cell adhesion are focal adhesion proteins, especially the integrins that are the cross linked structure between the cell s cytoskeleton structure and the cell structure and extracellular matrix (ECM), which is briefly introduced in Chapter 1. The intensity of the integrin-β3 signal may indicate the cell s adherent strength, so that in Fig. 3.9(a) the intensity of the green fluorescence signal of the cells in the yellow box were quantified and depicted as sown in the graph Fig. 3.9(b). According to the image in Fig. 3.9(a), there is a green, integrin-β3, signal faded zone at the periphery of the plasma affected DZ. At the center, the green signals are rather stronger, however near VZ where the cells are detached, the cells lost their signals and show disrupted integrins. As seen in the graph, the integrin signal gradually fades from the plasma impinged center to the plasma impinged periphery, which is adherent to void area. To confirm the fact that these cellular changes were not caused by the gas flow from the plasma, the same

54 42 Chapter 3. APP Effects on Human Dermal Cells a b Intensity (I/IO) x(µm) Figure 3.9: (a) The immunofluorescence image of the plasma treated THLE-2 where the integrin β-3 (green) and actin (red) were double labeled.

55 3.2 Temporal Phenotypic Cellular Changes 43 a b 1.0 Intensity (I/IO) x(µm) Figure 3.10: (a) The immunofluorescence image of the gas blown THLE-2 where the integrin β-3 (green) and actin (red) were double labeled.

56 44 Chapter 3. APP Effects on Human Dermal Cells immunofluorescence imaging was carried out for gas-dried cells. As seen in Fig. 3.10(a), there is no sign of changes to the integrin or even the actin. In Fig. 3.10(b), the intensity of the integrin signal from the gas-dried cells are presented where there are no alterations to or fadedness of the cells. As a result, it was elucidated that the integrin disruption caused the decrease in the adherent strength of the cells, and those cellular changes were caused by the plasma species, not the gas flow or the drying effect. There are previous reports that the reactive oxygen species (ROS) influences the adhesional properties to cause the cellular detachment [82 84]. From the emission spectrum obtained from the LF micro jet plasma, there are O I and enhanced OH peaks associated with the PBS layer under the plasma [Fig. 3.6(b)]. The species are ROS which may decrease the adhesion strength between the ECM substrate and the cell Necrosis of Cells on Mesenchymal Origin Cell Necrosis Depending on the Plasma Input Voltage The same cell death and detachment of the human dermal fibroblast cells by the plasma was observed. The plasma electrode voltage was varied from 950 V to 1200 V. Except for the electrode voltage of 950 V [Fig. 3.11(b)] the necrotic cell death was observed for fibroblast cells [Fig. 4.1]. In Fig. 3.11(b), the necrotized fibroblast cells stained by live/dead assay are shown. The plasma-treated cells are detached, as was the case with THLE-2, in other words, VZ appeared at periphery of DZ. Interestingly, some of the living cells near VZ shrank [Fig. 3.11(c)]. This cell shrinkage seems to be the early step of cell detachment for the fibroblast cells and since fibroblast cells are larger than THLE-2 cells, this shrinkage is exhibited more clearly than for the THLE-2 cells. In this experiment, the cellular non-destructive plasma power regime (< 1000 V) were found by scanning the electrode voltage. This regime In addition, structural disruption of AFs and MTs were also shown for these fibroblast cells. In the living cells, AFs labeled by red color and MTs by green color as shown in Fig. 3.11(e). Examining Fig. 3.11(e) carefully, the AFs and MTs exist and form strong stress fibers across the cell body. However, the fibrous structures of AFs and MTs disappeared at necrotized cells by plasma [Fig. 3.11(d)]. Thus, it can be concluded that The disruption of the cytoskeleton is not the feature presented on the cell only from epithelial origin but is a global cellular feature attributed by plasma. 3.3 Long Term Cellular Response by Indirect Contact of APP Plasma can affect the cells in two ways, one is a direct effect which results in cellular changes in a short time interval, a few minutes after plasma treatment. On the contrary, plasma also influences the cell indirectly, where cellular changes will be observed over a a longer time interval. For example, changes which take more than a few hours, apoptotsis or cellular morphology or motility changes. The following experimental results were obtained from observations obtained over a longer time interval, a few hours or even days.

57 3.3 Long Term Cellular Response 45 a b 100 µm c 50 µm d e Figure 3.11: The live/dead assayed fibroblast cells treated by plasma of (a) V in = 950 V, (b) V in = 1200 V (center), (c) (boundary), the immunofluorescence image of (d) plasma treated center and (e) the live area boundary.

58 46 Chapter 3. APP Effects on Human Dermal Cells APP Induced Apoptosis of Cells on Epithelial Origin Delayed Cell Death (Apoptosis) So far the short term cellular responses were observed; the cell necrosis or the cell detachment which can be observed in a few minutes. However, there are also cellular changes in long term scale (a few hours). Since a cell signaling process takes more than a few minutes, observing the cellular changes in long term scale can help to understand plasma effects in cell signaling. The behavior of cells forced to become detached from the substrate by plasma was monitored after the plasma treatment. As shown from several other reports [39] plasma induces apoptosis; plasma does not only detach cells from the substrate, but it also affect to the cell viability and induces cells to apoptotize. Here, the cells detached from the substrate were assumed to be permanently changed by the plasma, so the cell might show other unusual behaviors. After the plasma treatment, the detached cells were collected and incubated in a medium containing 10% serum, in which many kinds of growth factors are mixed. However, not all the cells are re-attached to the ECM-coated substrate and most of the cells instead floated on the surface of the medium. Since the floating cells cannot be immunostained anti-caspase-3 could not be detected from the cells, however, the other apoptosis test using annexin-v was used. Fig shows an image of collected cells separated from the substrate after the plasma treatment. The cells were incubated for 24 h in an incubator at 37 o C. For the annexin-v binding solution, the cells showed the green signal characteristic of apoptotic cells. The experiment for detecting apoptotic cells were also carried out on the SK-HEP-1 cells, the cancer cell the endothelial origin cells which appear at the epithelial tissues [data presented on next chapter]. The cancer cells are designed not to apoptotize, however the floating cells detached by the plasma showed also massive apoptotic signals following the plasma treatment [Fig. 4.3(b)]. From these results, it can be concluded that the plasma can interfere the cell signaling process and affect cells in long term time scale. In addition, some of the cells which were not altered phenotypically may still damaged to become dead Migrational and Morphological Changes of Cells on Mesenchymal Origin Cellular Changes on Migration by APP The plasma was applied to fibroblast cells and the input voltage was fixed to 950 V, where the cell does not show necrosis or detachment. The treatment time was 10 min. This time scale were decided considering the cellular response time for extracellular chemical species [94, 95]. Since the fibroblast cells were treated, the plasma was applied indirectly as seen in Fig. 3.2(c) and (d). The medium layer on the cells took varying thicknesses of 2 mm, and 4 mm. The wound was made by swiping the pipet at the cell layer on the slideglass to observe the motility change of the plasma treated cell area. Plasma effect to cell motility could be measured by observing the wound closing area. Live cell were monitored immediately after the plasma treatment and every 2 hr for the following 24 hr. After the wounding, unhealthy cells appeared near the wound, which were damaged when the pipet had swiped the cell layers to make a wound. Since the damaged cells were detached, the wound had became larger in 2 hr after

59 3.3 Long Term Cellular Response 47 a b 100 μm Figure 3.12: Apoptosis assayed (a) control THLE-2 cells and (b) THLE-2 cells incubated for 24 hours after detachment by plasma (white arrows: apoptotic cells with annexin-v binding, green)

60 48 Chapter 3. APP Effects on Human Dermal Cells plasma treatment. However, after 2 hr the wound started to close. As seen in 3.13(a) - (d), the control cells and the plasma-treated cells does not show any significant difference at the wound closing behavior. In 3.13(e), the quantified wound reduction area consistently shows the similarity between the samples; control, plasma treated wound with 2 mm of DPBS, and 4 mm of DPBS. In conclusion, the cellular motility does not seem to be significantly affected by the indirect plasma treatment, at least by plasma treatment within 10 min. Cellular Morphology Changes by APP The LF plasma with a weak input voltage (< 1000 V) was used to treat cells covered by a medium layer (3 mm) that was thick enough that the plasma cannot directly come into contact with the cell to induce immediate cellular changes such as cell necrosis or detachment [Fig. 3.2(d)]. The treatment time, the discharge frequency, and the gas constitution was changed to determine the key species which influence the cells through the medium. The discharge frequency enhances the plasma current which influences the cells, on the other hand, in terms of the gas constitution, adding more O 2 will enhance the ROS activity on the cells. In this experiment the driving frequency was varied to find out the plasma effect and the working species. As shown in Fig. 3.14(a), the fibroblast cells showed morphological changes after 48 hr of incubation, followed by plasma treatment for 10 min. The 48 hr of time were decided to give the fibroblast cells enough time to differentiate into myofibroblast and to observe if the plasma treatment affected the cellular differentiation process. The most distinctive feature is that the fibroblast which has a rather longish body that is the signature of the mesenchymal cells became more square in shape compared to control cells in Fig. 3.14(b). To quantify a morphological change, the aspect ratio of cells were measured by dividing the length of the short axis from the length of the long axis of the cells [Fig. 3.15]. In the fluorescence image the boundary of the control cells were not clear so that the phase image was used to measure the aspect ratio. The aspect ratio gradually decreased for the increasing driving frequency. Compare the control cell to the cells treated by plasma (f in = 50 khz), the aspect ratio of the control cells are three times larger than the plasma treated cells. In addition, from the immunofluorescence image the actin stress fibers at the cytoplasm decreased. On the other hand, the cortical actin was shown to be stronger compared to the controls. Interestingly, epithelial cells are known to have slower cellular mobility compared to the mesenchymal cells. Since those cellular characteristics, the polygonal shape and the strong cortical actin, are epithelial-like features, the motility change observed in the previous section may be explained. Interestingly, the cell incubated in the plasma treated media does not changed at the cellular shape. The ph of the plasma treated media was also checked so that we can be sure that the change of the ph does not affect the cells. In conclusion, the ph does not changed and only degree changed which returned to the appropriate level inside the incubator. The increasing driving frequency of the power will increase the input current and also may affect the concentration of the ROS species in plasma. In Fig the image cells are also labeled with - smooth muscle actin ( - SMA) which are green In color. - SMA is the actin which appears in the myofibroblast cells and from the signal strength and location the development of the differentiation in the cells can be observed. However, the SMA signal or the location are not changed at all by plasma treated cells, from which it can be concluded that the plasma does not affect the cellular differentiation process.

61 3.4 Long Term Cellular Response 49 a b c d e Reduced area (a. u.) 0.4 Plasma + PBS (4mm) Plasma + PBS (2mm) control Time (hours) Figure 3.13: Image of control cells after (a) 2 hr and (b) 24 hr from the time wound was scratch on the slideglass, and images after (c) 2 hr, (d) 24 hr from the wounding and plasma treatment (thickness of DPBS = 4 mm), (e) Areal reduction of wound.

62 50 Chapter 3. APP Effects on Human Dermal Cells a b c d e f 100 μm Figure 3.14: The cells were treated twice for 10 minutes every 12 hours by the plasma at V in = 970 V and with different fin. Cell images of (a) control 10x, (b) control 40x, (c) plasma treated (f in = 20 khz) 10x, (d) plasma treated (f in = 20 khz) 40x (e) plasma treated (f in = 50 khz) 10x, and (f) plasma treated (f in = 50 khz) 40x.

63 3.4 Summary Control 20kHz 50kHz 80kHz Figure 3.15: Aspect ratio of cells were measured for control cells and the plasma treated cells (with f in = 20 khz, 50 khz and 80 khz). 3.4 Summary The atmospheric pressure plasmas of the LF (50 khz) single pin electrode jet plasma were used and applied to cells of epithelial and mesenchymal origin. Both of these cells form the dermal layer of the human body, which is why the those cells were selected for plasma targets. First, the cell necrosis and detachment by the plasma were observed for both epithelial cells (THLE-2) and mesenchymal cells (human dermal fibroblast) within a certain plasma input voltage range ( 1000 V). The cell death by the plasma was confirmed to be traumatic accidental cell Death, not programmed apoptosis. The intracellular structures were observed by immunofluorescence where the total disruption of the cytoskeleton structures were found. Interestingly, the actin filaments (AFs) and the microtubules (MTs) were disassembled to monomers however the intermedicate filaments (IFs) still remained and showed the filamentous structure even after cell death by plasma impinging. The IFs are the only cytoskeleton which do not have the charged monomers, so it can be assumed that those AFs and MTs may be interfered with by the charged particles from the plasma which may accumulated in the cell membrane. Additionally, the long term response of the cells was observed after plasma treatment. Cell apoptosis were observed for THLE-2 and also SK-HEP-1, the epithelial cancer cell. The cells were incubated over 24 hr for observation.

64 CHAPTER 4 Different Effects of APP on Different Types of Cells: Cancer vs. Normal Cells So far, the various effects of atmospheric pressure plasma jets were studied in detail, especially in terms of their dependence on the plasma strength. In this chapter, considering the application-wise aspect, the effects of the plasma are investigated with various cell types. In addition, the feasibility of application using the single pin electrode jet type plasma was performed. Interestingly, plasmas are known to induce cellular necrosis, apoptosis, and detachment, suggesting its potential use in cancer cell removal [39]. However, induction of those cellular changes would only be meaningful in cancer therapy if it is able to remove cancer cells selectively while leaving normal cells intact. Additionally, the plasma needs to be controllable for practical use, and the performance of the plasma on different cells needs to be investigated further. Therefore, in the present study, the responses of cancer and normal cells to the plasma treatment are compared to find out about the behavioral differences in cell necrosis and detachment. The cancer cells (SK-HEP-1) and normal cells (THLE-2) of liver epithelial and also the cancer and normal cells (MDA-MB-231 vs. MCF-10A) from another cell line, mammary gland epithelial cells, were used for the comparison study. 4.1 Experimental Setup Plasma setting The same low frequency single pin electrode jet plasma was used for both the cancer and normal cell treatment. AC power at 50 khz driving frequency was used and the input voltage was

65 4.1 Experimental Setup 53 varied from 950 V to 1200 V. The helium gas supply was maintained as 2 slpm. The distance between the tube end and the slide surface was 15 mm, where the cell samples were placed. Data analysis To quantify the plasma effects on cells the number of dead cells and live cells were counted from the images recorded. Live/dead assayed cell images were taken using the nine images panel from the plasma impinged center. As described in Fig. 4.1(a), the images were taken in numerical (1 to 9), from the plasma impinged center to the periphery. This image area covered most of the plasmaaffected zone and also the neighboring cells of the dead and detached cell zones. The green pixels and the red pixels in the live/dead assayed cell image were considered to be the live cells and the dead cells, respectively. The control image was taken from the same slideglass but far away from the plasma impinged site. Cell Culture Cells were seeded on the sterilized and extracellular matix (ECM) proteins-coated slideglasses after the air plasma treatment. The ECM proteins medium was made by the conventional protocol for THLE-2 (fibronectin 10 µg/ml, collagen µg/ml, BSA 10 µg/ml solution in serum free media). Each cell was cultured with different types of media and was incubated inside the 37 o C incubator where the CO 2 concentration was controlled at at 5% for the appropriate ph level of the culture media. Totally 4 cell types were used, THLE-2, SK-HEP-1, MCF-10A and MDA-MB-231 with the media BEBM (added with BEGM kit), DMEM, MEBM (added with MEGM kit and 500 µl of choleratoxin) and Leivobitz s L-15 Medium (ATCC) were used respectively. THLE-2 is human liver epithelial cell and SK-HEP-1 is human liver cancer cell which has the endothelial origin but form the tumor on epithelial tissue and has epithelial like features. MCF-10A and MDA-MB-231 cells are mammary glend epithelial cells and normal and cancer cells respectively. Especially, MDA-MB-231 cells were sealed to keep out the growth medium from the CO 2 to avoid calcifying by the ph change during the culture in incubator. Additionally, Immuno-fluorescence and Western blotting were performed to identify the distinctive biological characteristics between the cancer and the normal cells. Immunofluorescence The immunofluorescence process was introduced in the previous chapter. In this chapter, the same protocol was used with different primary antibodies, the Anti-paxillin (1:150, , BD Tranduction Lab.), and anti-α5 integrin(1:100, AB1928, Millipore). This paxillin and α5 integrin were labeled on cell focal adhesions (FA) to find out the relation between the adhesion strength and cellular structures. Each of the cells was co-labeled with phalloidin stabilizing agent (1:50, A12380, Molecular probes). Western Blotting Cells scraped from the slideglass surface were lysated and the cell extractions were prepared with SDS sample buffer and heated for 3 min at 100 o C by heatblock. Based on the protein density information measured by UV-Spectrophotometer (Nano drop ND-1000) equal amounts of cell lysates of different kind of cells were loaded on to 12% SDS-PAGE gel. After electrophoresis, the separated proteins were transferred to a nitrocellulose membrane using Gel casette sandwiches in electrode assembly (Mini-PROTEIN 3, , , Bio-Rad). The membrane was blocked with Tris-buffered saline (50 mm Tris, ph 7.5, and 100 mm NaCl) containing 0.1% skim milk several times with SNAP i.d. Protein Detection System (Snap i.d., Millipore) and after that the primary antibody was loaded. The blots were subsequently washed with Tris-buffered saline. Then incubated with an appropriate horseradish peroxidase-conjugated secondary antibody from Pierce Biotechnology (Rockford, IL) in saline buffer. Proteins were visualized HRP detection reagent (Amersham ECL T M Western Blotting Detection Reagents, GE Healthcare). Anti-try397 FAK (1:1000, , Millipore) and Anti-α5 inte-

66 54 Chapter 4. Plasm Treatment on Cancer and Normal Cells grin (1:1000, AB1928, Millipore) were used as the antibody for Western blotting. For the housekeeping protein Actin-HRP (1:1000, AB1928, Molecular probes) was used. 4.2 Differential Responses of Liver Cancer and Normal Cells to APP Liver Cancer (SK-HEP-1) and Normal (THLE-2) Cells When plasma impinged to the monolayered cells covered by Dubecco s phosphate buffered saline (DPBS) of 37 o C, which prevented the cells from drying and cooling, cells started to be detached from the ECMcoated substrate and necrotized [36]. Since plasma spreads in a circular shape on the dielectric surface, the plasma impacts the cells circularly to leave the necrotized cells at center and a void at the peripheral region where the cells are separated from the ECM substrate. For clear denotation, the cell necrotized area will be called as dead zone (DZ), the cell detached blank as void zone (VZ), and the combined area of dead and void area as plasma effective zone. The electrode voltage of the plasma have been changed from 950 V to 1200 V. The plasma treated cancer (SK-HEP-1) and normal (THLE-2) cells were presented in Fig. 4.1(b)-(e) where the cells were live/dead assayed, some of the cells were detached revealing the void area and some of the cells were intact. As mentioned before, the live cells are stained in green color and dead cells in red color by the live/dead assay. At low electrode voltage of 950 V, SK-HEP-1, the liver endothelial metastatic cancer cells, started to be detached from the substrate as shown at Fig. 4.1(b). However, unlike SK-HEP-1, THLE-2, the normal liver epithelial cells were not detached and still intact at the electrode voltage of 950 V, as shown in Fig. 4.1(d). At the higher electrode voltage of 1200 V, SK- HEP-1 has a larger plasma effective zone, indicating the dead and the void zone together [Fig. 4.1(c)]. Comparing the dead zone only, the DZ of THLE-2 in Fig. 4.1(e) was a little larger compared to the DZ of SK-HEP-1 in Fig. 4.1(e), the total plasma effective zone of THLE-2 is smaller than SK-HEP-1 due to the larger VZ. Along with these results, it was clear that SK-HEP-1, the cancer cells, seems to be more easily affected by the plasma at the same plasma input voltage. To clarify the differential response of cancer to normal cells quantitatively, we plotted the fraction of living cells numerically. Interestingly, as was shown from the further images in Fig. 4.2(a), SK-HEP-1 cells are more susceptible to the plasma than THLE-2, because at the same input voltage the plasma effective zone is larger for SK-HEP-1. When carefully comparing Fig. 4.1(c) and (e), where the SK-HEP-1 sample has a wider void zone than the THLE-2, the area difference arising due to the plasma treatment is most likely caused by the different cell detachment characteristics. To make sure of this hypothesis, the void zone alone was plotted in Fig. 4.2.(b) to show that the SK-HEP-1 void zone is distinctively larger than THLE-2 void zone. The following subject was to find the reason by which the cells are detached from the substrate with different responses for the same plasma treatment. As presented in the previous chapter, the detached cells shows apoptotic behavior so that these cellular detachment is meaningful for the cancer therapy. To be sure for the apoptosis induction in cancer cells, the experiment for detecting apoptotic cells were also carried out for the liver cancer cell (SK-HEP-1) which is designed not to apoptotize. The floating cells detached by the plasma showed

67 4.2 Differential Responses of Cancer vs. Normal Cells 55 a Plasma b 7 6 Image 5 frame c d e Figure 4.1: (a) Schematic of the image taking process for the plasma treated field. The images were taken as the 3x3 panel in order of numbers (1 to 9). Plasma treated cell images for various input voltages: (b) SK-HEP-1 with 950 V, (c) SK-HEP-1 with 1200 V, (d) THLE-2 with 950 V, and (e) THLE-2 with 1200 V.

68 56 Chapter 4. Plasm Treatment on Cancer and Normal Cells a Live cells (N / N o ) b Void area (A / A o ) Applied Voltage (V) Applied Voltage (V) Figure 4.2: (a) Fraction of live cells in the 3x3 image panel 4.1(a): THLE-2 (- -) and SK-HEP-1 (- -). (b) Fraction of the void area where cells detached from the substrate: THLE-2 (- -) and SK-HEP-1 (- -).

69 4.2 Differential Responses of Cancer vs. Normal Cells 57 a b c 30 0 Napop/Ni (%) Incubation time (hours) Figure 4.3: SK-HEP-1 cells were incubated for 24 hours and the apoptosis test kit was applied (Apoptotic signal: annexin-v, green). (a) control, (b) plasma treated cells and (c) the fraction of apoptotic cells with annexin-v binding

70 58 Chapter 4. Plasm Treatment on Cancer and Normal Cells Cell Integrins ECM Biophyical Stress Biochemical Stress Peeling off by Shear shear Dissociation by ROS Figure 4.4: Schematic of cell detachment process by biophysical or biochemical stress massive apoptotic signals after the plasma treatment [Fig. 4.3(b)], which did not appear at the nontreated cells at Fig. 4.3(a). This result is medically meaningful considering that the cancer cells are originally not suicidal and can develop as tumors. The quantified values of the apoptotic SK-HEP-1 cells were observed by the apoptosis test kit and were plotted in Fig. 3.12(c). The green apoptotic signals (annexin-v binding) gradually increased and more than 60% of the cells started to die by apoptosis. Furtherly, the mechanism of the cellular detachment by the plasma treatment were investigated followed by two different assays, the bio-physical and bio-chemical assay. Cells will be forced to detach from substrate leaving integrins either at the substrate or at cell body by the biophysical stress, for instance shear force [Fig. 4.4]. On the other hand, by the biochemical stress condition, integrins will be dissociated first and then cells will detach after [Fig. 4.4]. Both of these biophysical and biochemical stress may induce to cells by plasma treatment. As was described in a further study [36], it was evinced that the gas flow did not affect cellular changes along the plasma treatment, yet, the gas flow and which pushed away the DPBS and may have applied physical shear stress on the cells. To investigate the role of this physical shear stress along the plasma processes, a shear stress was applied to the SK-HEP-1 and THLE-2 cells. The SK-HEP-1 and THLE-2 were separately seeded in micro channels where the cocktail of ECM proteins were precoated. Commercial micro channels (ibidi, µ-slide VI 0.4 ) were used to generate shear flow. The shear rate was varied from 10 dyne/cm 2 to 40 dyne/cm 2. After 3 min of shear stress, the adherent cells were counted and compared to the initial adherent cell number. These results for SK-HEP-1 to THLE-2 were compared to verify whether the physical adherent strength of the cells is different, or not. According to Fig. 4.5, there is no meaningful difference between SK-HEP-1 and THLE-2 in the fraction of adherent cells for the same shear stress [Fig. 4.5]. This implies that SK-HEP-1 and THLE-2 do not display a difference in their adhesion strengths in terms of physical force conditions. The critical shear stress, the shear stress at which only 50% from the initial population of cells will remain after 3 min, which reflects the physical adhesion strength of each the cells, was also

71 4.2 Differential Responses of Cancer vs. Normal Cells 59 Fraction of adherent cells (n/n o ) 0.9 THLE-2 SK-HEP Shear force (dyne/cm 2 ) Figure 4.5: Fraction of adherent cells for applied shear stresses (10 dyne/cm 2-40 dyne/cm 2 [91] calculated based on the date in Fig. 4.5 and it was concluded that the two cell types have similar values, as SK-HEP1 and THLE-2 was dyne/cm 2 and dyne/cm 2 respectively. At the same time this also infers that the shear stress, the physical force, is not a dominant factor which causes the difference in the VZ area between the two cell types. Since, for atmospheric pressure plasma the chemical species is the key player for the interaction with cells [36], the relation between the biochemical stress and cell detachment was investigated. First, Trypsin-EDTA was applied as a biochemical stress because it is known to clip of cellular anchors, including integrins. As was shown in Fig. 4.6, when the bio-chemical stress was applied to the cells, the cells started to shrink a - b (τ 1 ) and finally became detached using their own elasticity b - c (τ 2 ). In previous de-adhesion dynamics studies it was shown that the de-adhesion time (τ 1 ) can represent the adhesion strength of the cells [92]. Here, the de-adhesion time (τ 1 ) was obtained by plotting the change in the cell s area over time and fitting it to: 1 Area norm = exp( t τ 1 τ 2 ). (4.1) In our study, τ 1 can be considered as the tolerance to bio-chemical strength. When a bio-chemical stress, Trypsin-EDTA, was applied, the cancer cells, SK-HEP-1, detached more easily compare to the normal cells, THLE-2 [Fig. 4.7(b)]. This means that the SK-HEP-1 is weaker against the bio-chemical stress than THLE-2. This tendency of de-adhesion time (τ 1 ) is consistent with the prior findings that the cancer cells (SK-HEP-1) are weaker to plasma exposure 4.2(b). As was mentioned before, plasmas are known to induce bio-chemical stress, by disrupting the integrin binding [36].

72 60 Chapter 4. Plasm Treatment on Cancer and Normal Cells a Cell Integrins ECM b τ 1 c τ 2 Figure 4.6: (a) The cells attached to the ECM by solid integrin-ecm binding (τ 1 : the time interval between a and b). (b) The cells started to be detached from ECM and started to shrink to 50% from the original cell area (τ 2 : the time interval between b and c). (c) The cells totally detached from the ECM using their own elastic force.

73 4.2 Differential Responses of Cancer vs. Normal Cells 61 a b Figure 4.7: De-adhsion test by Trypsin-EDTA, images of (a) THLE-2, (b) SK-HEP-1 cells after Trypsin- EDTA (from right to left 0 s, 60 s, 120 s).

74 62 Chapter 4. Plasm Treatment on Cancer and Normal Cells a 1.0 Area(t) - Area(i) Area(f) - Area(i) b De-adhesion time (min) Time (min) 0 THLE-2 SK-HEP-1 Figure 4.8: (a) The Boltzmann sigmoidal fitting using the equation 4.1 to get the De-adhesion time τ 1 (THLE-1 ( ) and SK-HEP-1 ( )). (b) De-adhsion time τ 1 for THLE-1 ( ) and SK-HEP-1 ( ).

75 4.2 Differential Responses of Cancer vs. Normal Cells 63 a b 50 µm c d Figure 4.9: Immunofluorescence image taken by fluorescence microscope (a) THLE-2 and (b) SK-HEP-1 [paxillin (green), actin stress fibers (red)], and the images taken by confocal microscope for (c) THLE-2 and (d) SK-HEP-1

76 64 Chapter 4. Plasm Treatment on Cancer and Normal Cells a b N/N tot N/N tot Length (nm) Figure 4.10: (a) The length of paxillin dots for THLE-2 ( ) and SK-HEP-1 ( ) and (b) the distribution of paxillin dot length for paxillin dots of THLE-2 ( ) and SK-HEP-1 ( )

77 4.2 Differential Responses of Cancer vs. Normal Cells 65 a FAK / Actin (a. u.) b α5 / Actin (a. u.) FAK Actin 5 Actin 0 THLE-2 SK-HEP-1 Figure 4.11: (a) The FAK of each THLE-2 and SK-HEP-1 cells (the inset is the western blot band) (b) The α5-integrin of each THLE-2 and SK-HEP-1 cells (the inset is the western blot band) Different Responses Related to the Intrinsic Biological Difference This distinction between the de-adhesion properties of cancer and normal cells, or SK-HEP-1 and THLE- 2, are expected to arise from the different biological characteristics related to their different focal adhesions (FAs). The immuno-fluorescence images taken by confocal microscopy labeled at paxillin in green fluorescence and actin in red fluorescence was observed by normal fluorescence microscopy and also by confocal fluorescence microscopy for a clear view of the paxillin dots within the sectioned image. From the z-stacked image obtained under normal fluorescence microscopy, stronger paxillin dots, the protein conjugated with FAs, were detected in the THLE-2 cells compared to SK-HEP-1, so THLE-2 seems to have more secure FAs compared to SK-HEP-1. This was verified quantitatively, as shown in Fig. 4.10(a), and the visible paxillin dots expressed on THLE-2 were two times more than those on SK-HEP-1. To obtain the number shown in Fig. 4.10(a), the paxillin dots were counted from the confocal images and only the dots conjugated at the end of the actin stress fibers were counted to be sure that the paxillin

78 66 Chapter 4. Plasm Treatment on Cancer and Normal Cells a b Figure 4.12: Plasma treated cell images for the input voltages of 1000 V: (a) MCF-10A and (b) MDAMB-231. dots are part of the FAs 1.4. In addition, paxillin dots of less than 2 nm were ignored, because, firstly those dots were not clear enough to be sure that they were FA dots, and secondly, from the reference only those FAs larger than 2 nm seem to be mechanically meaningful. The stronger FA conjugation on the THLE-2 cells than on SK-HEP-1 cells can also be confirmed by Fig. 4.10(b), where the distribution of the paxillin dot lengths is shown. On the THLE-2 cells the paxillin dots were longer which means that the FA are more matur and stronger. This tendency seems to be caused by the different focal adhesion kinase (FAK), the kinase maturate FA, expression on each cell. As shown in Fig. 4.11(b) more FAK proteins are exist in THLE-2 compared to SK-HEP-1. Moreover, the quantitative value of α-5 integrin protein, which is the part of FA, obtained by the Western blotting consistently reveals that THLE-2 has stronger coupling to the ECM by showing that more ±5 integrins exists on THLE-2 as shown in Fig. 4.11(b) APP Effects on Mammary Gland Epithelial Cells Mammary Gland Epithelial Cancer (MDA-MB-231) and Normal (MCF10A) Cells The same plasma treatment was carried on to another normal and cancer cell set, in order to assist the above arguments. For the consistency, cells from the same epithelial origin were needed so that the mammary gland epithelial cells MDA-MB-231 and MCF-10A were used. MDA-MB-231 is the matastatic cancer cell like SK-HEP-1, and MCF-10A is a normal epithelial cell. The same scanning of the plasma effects for different plasma input voltages were performed for mammary gland epithelial cells, as in the liver epithelial cell cases Fig shows an image of viability tested cells using live/dead assay, where the cells were treated using the plasma with 1000 V input voltage. There are discrepant responses between the normal cell, MCF-10A [Fig. 4.12(a)] and the cancer cell, MDA-MB-231 [Fig. 4.12(b)]. There are distinctive VZ at the center of the plasma impinged area at MDA-MB-231 sample, but not at the MCA-10A sample. As shown in Fig. 4.13(b), in some cases for the

79 4.3 APP Effects on Mammary Gland Cells 67 a Live cells (N/No) b Input Voltage (V) Void area (A/Ao) Input Voltage (V) Figure 4.13: (a) Fraction of live cells in the 3x3 image panel 4.1(a): MCF-10A (- -) and MDA-MB-231 (- -). (b) Fraction of the void area where cells detached from the substrate: MCF-10A (- -) and MDA-MB-231 (- -).

80 68 Chapter 4. Plasm Treatment on Cancer and Normal Cells plasma with a 950 V input voltage, some MDA-MB-231 samples presented detachment from the substrate and show void zones (VZs), on the other hand none of the MCF-10A cases showed VZs or dead zones (DZs) but for all cases the cells were intact. Also for the plasma with an input voltage of 1000 V after plasma treatment, as was presented in Fig. 4.13(a), MDA-MB-231 shows VZs at the center where the plasma was impinged. Unlike MDA-MB-231,MCF-10A did not show any different physiological changes for the same plasma strength in the same treatment time. Especially, for this mammary gland epithelial cell line, the void zones for the cancer cells formed after plasma treatment were larger compared to normal cells, where none of the MCF-10A cells had Fig. 4.13(b), even at 1200 V, the strongest plasma input voltage, which implies this these have a stronger adhesion strength for the same plasma condition. As shown from the previous liver cell cases, the chemically active plasma species, like reactive oxygen species (ROS), seem to be important in cell detachment. The same cellular detachment process produced by the plasma was performed using one of the ROS, hydrogen peroxide (H 2 O 2 ), and the cellular responses was observed afterwards. From the previous studies the cellular changes induced by H 2 O 2 had already been studied to find out that the cell adhesion strength got weaker after tens of minutes [82, 83, 90]. However, the plasma seems to produce extreme stress where the cell could also be dead. Even though, as shown before, the plasma induces mixed stress to the cells for cell detachment, it seems to be dominantly initiated by the chemically active species. To check this hypothesis, the same concentration of H 2 O 2 DPBS solution (2 M) was applied to SK-HEP-1, THLE-2, MDA-MB-231 and MCF-10A. The concentration was decided after applying different concentrations of H 2 O 2 DPBS solutions to cells to find the concentration at which cells detached in a few minutes, to mimic the way the cells respond to the plasma. As shown in Fig. 4.15, SK-HEP-1 and MDA-MB-231 shows the immediate response to the reactive oxygen species, H 2 O 2, where cells shrink in 10 min and some of the cells are already detached from the substrate, which was similar to what was detected after plasma treatment. However, for THLE- 2 and MCF-10A, the normal cells in Fig. 4.15, the cells did not show huge changes like for cancer cells. Some cells not adjoined to other cells and exhibiting exposed edges show membrane shrinkage, however, the shrinkage is not as dramatic as with SK-HEP-1 and MDA-MB-231. Overall the changes seem to be minimal and the cell adhesion strength was as strong as before the H 2 O 2 and oxygen stress application. Next the intrinsic biological characteristics related to the FA were observed by immunofluorescence. As was seen in Fig. 4.16(a) for MDA-MB-231, there are no clear integrin dots clustered together, but the green integrin signals show all over the memebrane cytosol, in contrast for MCF-10A Fig. 4.16(b) there are cluster of α 5integrin dots in the image. From the previous studies it was proven that the cell adhesion strength was created from the clustered integrins which make the FA-ECM binding stronger. Moreover, though both MDA-MB-231 and MCF-10A are epithelial cells only MCF-10A shows the stronger cell to cell contact where cells are attached close together and may form cell groups. In Fig. 4.16(b), the cell to cell contact is shown from the cortical actin at the edge of the cell membrane, or at the cell boundary. Plasma-induced cell detachment behavior appears differently for cancer and normal cells as in the previous liver epithelial cell cases. Clearly, for mammary gland epithelial cells and liver epithelial cells, that is, cancer cells especially for metastatic cancers, the cell adhesion strength against chemical stress is weaker. However, as revealed from this section for different cell types, the cell responses also differ. Another important fact that was found in this chapter is that the chemically active species play a crucial role in cell detachment as was shown from the comparison study between H 2 O 2 and the plasma.

81 4.3 APP Effects on Mammary Gland Cells 69 a 1.0 Area(t) - Area(i) Area(f) - Area(i) b De-adhesion time (min) Time (min) MCF-10A MDA-MB-231 Figure 4.14: (a) The Boltzmann sigmoidal fitting using the equation 4.1 to get the De-adhesion time τ 1 (MCF-10A ( ) and MDA-MB-231 ( )). (b) De-adhsion time τ 1 for MCF-10A ( ) and MDA-MB-231 ( ).

82 70 Chapter 4. Plasm Treatment on Cancer and Normal Cells THLE-2 (Normal) Before 10 min H2O2 SK-HEP-1 (Cancer) Before 10 min H2O2 MCF10A (Normal) Before 10 min H2O2 MDA-MB-231 (Cancer) Before 10 min H2O2 Figure 4.15: Images of cells (THLE-2, SK-HEP-1, MCF-10A, and MDA-MB-231) before and after H 2 O 2 (2 M) treatment for 10 min

83 4.4 Summary 71 a b 50 µm Figure 4.16: (a) Immunostaining image of MDA-MB-231 and (b) MCF-10A labeled at 5 integrin in green color and actin in red color APP Treatment on Cancer and Normal Cells in a Co-Cultured Field The cancer and normal cells were co-cultured in the same field for a feasibility study of cancer treatment using plasma. Since the cancer cell is a metastatic cancer which transfers from an original tumorous site to another tissue, the co-cultured field was mimicking the scattered cancer cells surrounded by normal cells. Here, MCF-10A and MDA-MB-231 were used as the normal and metastatic cancer cells, respectively. Each type of cell was stained using different colors of cell tracker, the MCF-10A with green and MDA-MB-231 with orange, for clear distinction between the cells. There are some diffculties in co-culturing cells in the same field because the growth rates of the cells were different and also different media were required by each cell type, so cells were cultured in different petri dishes before they reached the proper confluency for the experiment, then both cell types were stained individually and seeded on the same field. The plasma input voltage was controlled to 1000 V, which was based on the result from the last section where the cancer cells started to detach but the normal cells remained intact. In Fig. 4.17(a), there are cells stained in green and orange color. Interestingly, after the plasma treatment, as shown in Fig. 4.17(b), the MDA-MB-231 cells detached from the substrate where the MCF-10A still remained intact. These cell samples are only the monolayered model established with cells and ECM substrate, not the bulky tissue, but this this result still shows the possibility of selective cancer removal. 4.4 Summary It was found that, for the same plasma treatment, the plasma effect on cancer cells (SK-HEP-1) and normal cells (THLE-2) was different. The plasma effective zone was larger for SK-HEP-1 than for THLE- 2, at the same time it was found that the differentiated formation of the void zone by the plasma is the key factor which causes this difference. By following the shear application experiment on normal and cancer cells where the physical adhesion strength of cancer cells and normal cells turned to be alike, the shear stress generated by the gas and plasma flow is not seen to be the reason for the void area

84 72 Chapter 4. Plasm Treatment on Cancer and Normal Cells Figure 4.17: Co-cultured image of normal (blue) and cancer (red) cells: (a) before plasma treatment and (b) after plasma treatment.