Study of Single Cell Sonoporation in Real Time

STUDY OF SINGLE CELL SONOPORATION IN REAL TIME

USING ELECTROPHYSIOLOGY TECHNIQUES

Yun Zhou

Submitted in partial fulfilment of the requirements

For the degree of Doctor of Philosophy

Dissertation Adviser: Prof. Cheri X. Deng

Department of Biomedical Engineering

CASE WESTERN RESERVE UNIVERSITY

May 2008 CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

______

candidate for the ______degree *.

(signed)______(chair of the committee)

______

(date) ______

*We also certify that written approval has been obtained for any proprietary material contained therein.

Dedicated to my parents and my wife for their love, support and patience Table of Content

Table of Content ...... 1

List of Tables ...... 5

List of Figures ...... 6

Acknowledgements ...... 9

List of Abbreviations ...... 11

Abstract ...... 12

CH1 INTRODUCTION...... 14

1.1 Specific aims of this research ...... 14

1.2 Background and significance of this study ...... 16

1.2.1 Targeted drug delivery ...... 16

1.2.2 Targeted gene delivery ...... 17

1.2.3 Ultrasound mediated intracellular drug and gene delivery ...... 18

1.2.4 Significance of this study ...... 21

CH2 MEASUREMENT OF SINGLE CELL SONOPORATION USING VOLTAGE CLAMP TECHNIQUE ...... 25

2.1 Introduction ...... 25

2.1.1 Monitoring of sonoporation using voltage clamp technique in real time ...... 25

2.1.2 Microbubbles to facilitate sonoporation ...... 27

2.1.3 Effects of vitelline membrane in sonoporation of oocytes ...... 28

2.1.4 Effects of ultrasound parameters in sonoporation process ...... 29

2.2 Methods ...... 29

2.2.1 Experimental setup ...... 29

2.2.2 Data acquisition ...... 30

2.2.3 Xenopus oocyte preparation and animal protocol ...... 30

1 2.2.4 Voltage clamp procedure ...... 31

2.2.5 Procedure to measure bubble concentration as a function of time ...... 32

2.2.6 Experimental design to examine the effects of vitelline membrane in sonoporation ...... 33

2.2.7 Experimental design to examine the effect of duty cycle on sonoporation process ...... 35

2.3 Results ...... 36

2.3.1 Microbubble concentration in solution during experiment ...... 36

2.3.2 The effects of vitelline membrane of Xenopus oocytes in sonoporation ...... 37

2.3.3 Duty cycle effect on sonoporation ...... 38

2.3.4 Time constant of voltage clamp system ...... 41

2.4 Discussion ...... 41

2.5 Conclusion ...... 43

CH3 EFFECTS OF EXTRACELLULAR CALCIUM ON CELL MEMBRANE RESEALING IN SONOPORATION ...... 61

3.1 Introduction ...... 61

3.2 Methods ...... 62

3.3 Results ...... 62

3.3.1 Extracellular Ca2+ is required for resealing in sonoporation ...... 63

3.3.2 Multiple dynamic processes and [Ca2+] dependent early-stage resealing ...... 65

3.3.3 Ca2+ dependent late -stage resealing ...... 67

3.3.4 Effects of extracellular Mg2+ on resealing process ...... 69

3.3.5 The effects of TMC amplitude ...... 69

3.4 Discussion ...... 71

3.5 Conclusion and future work ...... 77

CH4 DYNAMICS OF SONOPORATION CORRELATED WITH ACOUSTIC CAVITATION ACTIVITIES ...... 87

2 4.1 Introduction ...... 87

4.2 Methods ...... 88

4.2.1 Animal protocol and cell preparation ...... 88

4.2.2 Experimental setup ...... 88

4.2.3 Data acquisition procedure and processing algorithms ...... 89

4.3 Results ...... 91

4.4 Discussion ...... 93

4.5 Conclusion and future work ...... 94

CH5 TRANSIENT PORE SIZE ESTIMATION IN SONOPORATION ...... 101

5.1 Introduction ...... 101

5.2 Methods ...... 102

5.2.1 Experiments where a single pore is generated ...... 102

5.2.2 Experiment where multiple pores are generated ...... 104

5.2.3 Electro-Diffusion model to estimate pore size from transmembrane current .104

5.2.4 Experimental materials and experimental setup ...... 107

5.3 Results ...... 107

5.4 Discussion ...... 108

5.5 Conclusion and future work ...... 111

CH6 FUTURE WORK ...... 118

6.1 Specific topics for future investigation ...... 118

6.1.1 Study of single cell sonoporation of mammalian cells using patch clamp technique ...... 118

6.1.2 Controlled microbubble interaction with ultrasound and cell membrane ...... 119

6.1.3 Deterministic measurement of pore size ...... 119

6.2 Benefits to future applications on drug and gene delivery ...... 120

3 REFERENCE ...... 122

4 List of Tables

Table II-1: Bubble concentration measured by method 1 ...... 50

Table II-2: Bubble concentration measured by method 2...... 51

Table II-3: Cell survival rate at different duty cycles ...... 59

Table IV-1: Data acquisition procedure in this study ...... 97

5 List of Figures

Figure II-1: The experimental setup of real time monitoring sonoporation process by voltage clamp technique. A 3D motion system is utilized to precisely adjust transducer field...... 45

Figure II-2: Diagram of bubble concentration measurement procedure...... 46

Figure II-3: Picture of hemocytometer...... 47

Figure II-4: An example of the TMC of a single Xenopus oocyte under voltage clamp during sonoporation. The ultrasound application is indicated by the horizontal bar shown in the plot...... 48

Figure II-5: 3D ribbons plots show the number of bubble at 8 different positions change with time. A, B & C represent three data set we recorded. The color of ribbons represents the position of bubble. The shape of ribbons represents the number of bubble at certain position change with time...... 49

Figure II-6: Maximum amplitude of TMCs with and without vitelline membrane. N=29 for normal and N=38 for devitellinized group. The box shows 25, 50 and 75 quartile of data, the cross is mean of data, whiskers represent the max and min of data...... 52

Figure II-7: Least square fit parameters with and without vitelline membrane. (A,B): Rate constant for slow and fast recovery; (C,D): Normalized amplitude of slow and fast recovery. N=29 for the normal group and N=38 for devitellinized group...... 53

Figure II-8: Scatter plots of the maximum amplitudes of TMCs with and without vitelline membrane. A: Normalized amplitude of slow recovery vs. maximum current amplitude; B: Normalized amplitude of fast recovery vs. maximum current amplitude. N=29 for the normal group and N=38 for devitellinized group...... 54

Figure II-9: TMC of a Xenopus oocyte during sonoporation induced by pulsed ultrasound...... 55

Figure II-10: TMC at different duty cycles...... 56

Figure II-11: Maximal current vs. duty cycle. n =7 (5%), 7 (10%) and 8 (15%)...... 57

Figure II-12: Curve length of TMC vs. duty cycle. The ultrasound parameters are the same as in Fig. 5. n =7 (5%), 7 (10%) and 8 (15%)...... 58

Figure II-13: Screen shot shows the step response of voltage clamp system with an already clamped oocyte from a 5ms, 50mv pulse...... 60

6 Figure III-1: Time-dependent inward TMC recorded in oocytes under voltage clamp exposed to tone burst ultrasound (0.2 s duration, 0.3 MPa pressure amplitude) in the presence of 1% Definity® at different extracellular [Ca2+]. The holding potential was − 50 mV. Each current curve is normalized by the absolute value of its maximum which occurred at the end of ultrasound tone burst (1.2 s)...... 79

Figure III-2: Cell resealing and survival at different extracellular [Ca2+]. A) Percentage of the TMC recovery from the maximum value in the presence of 0.02, 0.09, 0.36, 0.54, 1.08, 1.8, and 3.0 mM extracellular [Ca2+] with n = 13, 10, 19, 10, 12, 20, and 11, respectively. B) Percentage of survived cells at extracellular [Ca2+] of 0, 0.09, 0.36, 0.54, 1.08, 1.80, and 3.0 mM with n = 12, 8, 19, 19, 8, 21, and 11, respectively. Inset: Light micrograph images of Xenopus oocytes, showing a single oocyte before sonoporation (i), one oocyte with an un-sealed pore (arrow) where discoloration around the un-sealed pore is apparent (ii), a degenerating oocyte with leaking cytoplasmic contents (arrow) in the absence of Ca2+ (iii), and a recovered oocyte after sonoporation in the presence of 1.8 mM Ca2+ in solution (iv)...... 80

Figure III-3:Multiple dynamic processes in sonoporation resealing. An example of the normalized TMC measured at [Ca2+] = 1.8 mM was shown to fit well by a superposition of two exponential functions with different rate constants 0.094 and 0.67 respectively. . 81

Figure IV-1: Simultaneously monitoring of sonoporation and cavitation via voltage clamp and acoustic signal detection within the ultrasound focus. Inset shows the ultrasound focus centered at the edge of the cell, placed on top of a gel block in a dish...... 95

Figure IV-2: Dual element transducer assembly calibration information. The result is from an automatic scan by a 0.04mm diameter needle hydrophone. A&B show the lateral scan result for detection transducer (7MHz) and excitation transducer (1.5MHz) with 0.2mm step; C&D show the axial scan result for detection and excitation transducer with 0.5mm step...... 96

Figure IV-3: (A) All time sequences of detection and excitation transducer used in this study. (B) The driving signal for excitation transducer in ramping procedure...... 98

Figure IV-4: (A, C) Increased AE signals and broadband spectra during bubble destruction. (B, D) No acoustic signals detected without Definity®. (E) The AE spectrum shows bubble destruction occurred in 15 ms after activation of the excitation ultrasound pulses. (F) The inward TMC increases with the decrease of BS RMS (middle plot) and increase of spectral RMS (3-11 MHz) (bottom plot)...... 99

Figure IV-5: (A, B) Examples of the AE signals and power spectra in passive detection. (C) Change in TMC current corresponds with delayed AE increase when ramping

7 ultrasound pulses were applied. (D) Change in TMC current occurs immediately after excitation ultrasound with constant amplitude...... 100

Figure V-1: Circuit model of Xenopus oocyte with voltage clamp system. A & B are equivalent circuit models. Vc is the command voltage, Rm is membrane resistance, Cm is membrane capacitance, RNa, RK, RCl is active resistance of membrane due to Na, K and Cl, VNa, VK, VCl is potential from ion gradient by Na, K and Cl, INa, IK, ICl is TMC from ion gradient by Na, K and Cl ...... 112

Figure V-2: The relationship between maximum current amplitude and radius of pore 113

Figure V-3: Histogram of maximum current amplitude from single pore. Bin size is 0.05µA; sample size N=33...... 114

Figure V-4: Probability density of experiment data (n=85) and simulation result from single pore data...... 115

8 Acknowledgements

I would like to express my deep gratitude to Dr. Cheri Deng, my advisor, for her patience

and kindness, for her understanding and confidence in me, and for her willingness to roll up her sleeves and work together with me in experiments. The dissertation would not

have seen the light of day without her constant inspiring mentoring. My enormous debt of

gratitude can hardly be repaid to her for her guidance and support during every stage in my graduate studies.

I thank my dissertation committee for their inspiring guidance and support. These giants deserve special mention for their enlightening effort during my doctoral studies and research: Dr. David Wilson for his excellent thought provoking questions aimed at making my work thorough and pushing me till I got to the right answers; Dr. Miklos

Gratzl for sharing his incredible depth of expertise in aspect of sensor; Dr. Yiping Han for her constant encouragement and help with cell biology; Dr. Agata Exner for her insightful advices on drug delivery strategy.

I am indebted to Dr. Jianmin Cui, Dr. Jinyi Shi and their group members, especially Dr. Hua Pan, for educating me in electrophysiology techniques. I really enjoy having interesting discussions and working together with them.

Many thanks go to all colleagues in our group, past and present, for their support, help and friendship: Hesheng Wang, Grant Steyer, Paras Parikh, Olivier Izad, David

Sabens, Matthew Aehle, Yu Dong, Dr. Ronald Kumon, Kun Yang and Dr. Juyoung Park.

I would like to especially thank Ron for his invaluable assistance on all kinds of topics. I would like to thank all colleagues and friends from Dr. Saidel’s group, Dr. Wilson’s

9 group, Dr. Gratzl’s group, Dr. Yu’s group, Dr. Rollins’ group, Dr. Exner’s group, Dr.

Han’s group at CWRU, and Dr. Cain’s group and Dr. O’Donnell’s group at the U of

Michigan, for their kindness and friendship. It’s my pleasure to work with these talented students and researchers.

Most of all I thank my parents and my wife for their endless love, support and encouragement through my doctoral studies. Lu, thanking you in words will never do

justice to what you have contributed to my life. You are the whole world to me.

10 List of Abbreviations

AE: Acoustic Emission

ANOVA: Analysis of Variance

BS: Backscatter

CTCL: Current-Time Curve Length

CW: Continuous Wave

DNA: Deoxyribonucleic Acid

FFT: Fast Fourier Transform

GFP: Green Fluorescent Protein

MT: Microtubule

PRF: Pulse Repetition Frequency

PRT: Pulse Repetition Time

RMS: Root Mean Square

SCID: Severe Combined Immunodeficiency Disease

TMC: Transmembrane Current

UCA: Ultrasound Contrast Agents

Study of Single Cell Sonoporation in Real Time

Using Electrophysiology Techniques

Abstract

YUN ZHOU

Safe and efficient intracellular delivery of drugs and genes is critically important in targeted cancer treatment and gene therapy applications. Ultrasound has been

demonstrated to increase the cell membrane permeability and exploited as a promising

non-viral strategy for intracellular delivery of DNAs, proteins, and other agents. Despite

increasing interest and progress made recently, challenges remain to achieve controllable

outcome. Ultrasound energy mechanically creates nano- to micron- sized, non-specific

pores on the cell membrane to allow entry of extracellular agents into the cell. However,

the biophysical mechanisms of this process, often called sonoporation, has not been fully

understood due to the lack of appropriate techniques to study the transient and sub-

micron process. Sonoporation studies have been largely limited to static post-ultrasound

assays. While important knowledge can be obtained through such analysis, such methods

inevitably overlook the actual transient process of cell poration, therefore are inadequate

to deterministically correlate ultrasound parameters with sonoporation outcome and to

12 uncover cellular mechanisms of sonoporation in individual cells. This research aims to

study the sonoporation dynamics at the single cell level to obtain better understanding of

the mechanisms of sonoporation affected by relevant physical and chemical parameters.

Using Xenopus oocytes as a model system, sonoporation of a single cell is monitored in

real time via the voltage clamp techniques which provide a novel means to monitor the

dynamics of sonoporation with high temporal resolution and sensitivity. The specific

aims of this research are: 1) To investigate the roles of extracellular [Ca2+] in sonoporation; 2) To correlate sonoporation with acoustic cavitation activities at the single cell level; and 3) To estimate the pore size in sonoporation by using statistical analysis.

CHAPTER I

INTRODUCTION

This objective of this dissertation is to investigate the mechanism of sonoporation, the

process of cell membrane poration by ultrasound energy application. In reversible

sonoporation, small and transient pores are generated in the cell membrane as the result of the mechanic impact of ultrasound exposure, and subsequently reseal to ensure cell

survival. The transient pores effectively increase the membrane permeability which

allows entry of the otherwise impermeable extracellular agents into the cell. The technique can be used as a new strategy for application of non-viral drug and gene

delivery.

1.1 Specific aims of this research

Safe and efficient intracellular delivery of drugs and genes is critically important in applications such as targeted cancer treatment and gene therapy. Ultrasound has been used to transiently increase the cell membrane permeability and has been exploited as a promising non-viral strategy for intracellular delivery of DNAs, proteins, and other

agents [1-3]. However, despite increasing interest and progress made in ultrasound

mediated delivery, challenges remain to achieve controllable outcome. While it is

hypothesized that ultrasound energy mechanically creates nano- to micron- sized, non-

specific pores on the cell membrane to allow entry of extracellular agents into the cell,

the biophysical mechanisms of this process, often called sonoporation [2, 4, 5], has not

been fully understood.

14 In particular, due to the lack of appropriate techniques to study the transient and

sub-micron process, sonoporation studies have been largely limited to static post-

ultrasound assays. Selection and attempted optimization of sonoporation parameters have

mainly relied on empirical results of delivery outcome obtained after sonoporation. While

important knowledge can be obtained through such analysis, post-ultrasound assays

inevitably overlook the actual transient process of cell poration. It is likely that the

ultrasound parameters determined this way are only associated with specific experimental

conditions due to the complexity of the ultrasound-cell interaction and its coupling with

surrounding bubble activities. Given the statistical variance of such interactions in large

number of cells, post-ultrasound assays are usually inadequate to deterministically

correlate ultrasound parameters with sonoporation outcome and to uncover cellular

mechanisms of sonoporation in individual cells.

In this research, we aim to study the sonoporation dynamics at the single cell level

to reveal the mechanisms and processes of sonoporation affected by relevant physical and

chemical parameters. Using Xenopus oocytes as the model system, sonoporation of a

single cell is monitored in real time via the inward transmembrane current (TMC) of the single Xenopus oocyte under voltage clamp, as demonstrated in previous study conducted in our laboratory [4]. Before ultrasound application, the TMC current is constant at a fixed membrane holding potential (voltage clamped) in the absence of activation of endogenous ion channels, since the whole cell membrane is regarded as a resistor with constant resistance [6]. In sonoporation, the ultrasound generated non-specific pores on the cell membrane decreases the membrane resistance; the ions flowing through the pores results in changes in the TMC current, which is determined by the pore size and ion

15 concentration gradient across the cell membrane, therefore providing a novel means to

monitor the dynamics of sonoporation in a single cell with high temporal resolution and

sensitivity. Three specific aims are designed for this research:

1. To investigate the roles of extracellular [Ca2+] in pore resealing in sonoporation;

2. To correlate sonoporation with acoustic cavitation activities at the single cell level;

3. To estimate the pore size in sonoporation by using statistical analysis.

1.2 Background and significance of this study

1.2.1 Targeted drug delivery

Intracellular delivery of drugs, proteins, and genes into viable cells remains one of the greatest challenges to achieve targeted cancer treatment and gene therapy. Although chemotherapy provides one of the most effective treatments in cancer therapy, the high doses often necessary to successfully eliminate the tumors also adversely affect healthy tissues in the host. Conventional systemic chemotherapy suffers from low tumor targeting efficiency and dose-limiting toxicity, which often results in significant patient morbidity and less desirable outcome[7-9]. Drugs that can interact with selective molecular targets in cancer cells have been discovered at an unprecedented rate, however, lack of effective delivery of these therapeutic agents prevents them from achieving desirable therapeutic outcome. A series of physical barriers (e.g. high interstitial pressure in solid tumors) and biological barriers (e.g. cell and nucleus membranes) limit the effective delivery of therapeutic agents to their cellular and molecular targets in vivo.

Targeted drug delivery can be achieved through physical targeting or chemical targeting, where a local placement of drug in a desirable location through physical (e.g.

16 implantation)[10] or chemical strategies[11]. Although conceptually different, the two approaches typically involve enhanced local drug concentration through direct physical implantation of drug carrier devices or targeted local release of drug inside tumors for local therapy. Newer active targeting systems are currently under development to utilize receptor mediated drug targeting in vital tumor tissues, such as proliferating vasculature[12]. Despite their fundamental differences, both physical and chemical targeting offers promising new treatments for previously refractory cancers. However, effective targeted drug therapy goes beyond the first step of increasing local drug concentration in many aspects. Particularly, at the cellular and/or sub-cellular levels, effective delivery of therapeutic agents across cell and/or nuclear membranes is a major challenge. In addition, physical implantation of drug carriers devices are not always possible and chemical targeting may not be available.

1.2.2 Targeted gene delivery

Intracellular delivery of compounds plays vital role in gene therapy including cancer gene therapy [13]. Gene therapy involves artificial introduction of exogenous genes into cells and promises effective treatment of a wide variety of diseases, both inherited and acquired. Diseases such as severe combined immunodeficiency disease (SCID), cystic fibrosis and some forms of cancer have been treated with gene therapy to varying degrees of success [14]; but apart from these few examples, gene therapy is far from the panacea it was imagined to be. Gene therapy’s failure to live up to its potential can be attributed, in part, to the lack of a high efficiency, in vivo method of gene transfer that can be used throughout the body, not merely in anatomically isolated regions [15, 16].The lack of safe and effective method for gene delivery to a specific organ or cell type remains an

17 obstacle to clinical use. Although viral and related vectors are efficient, they have drawbacks that limit their use in humans, including the potential for immunogenic and cytotoxic effects. Plasmid delivery methods are safe, but have low transfection efficiency even after direct injection[17]. Direct injection strategies are invasive and often technically challenging.

Forms of non-viral transfection include electroporation, particle bombardment, and lipofection. Electroporation refers to the utilization of high-intensity electric fields to open small pores in the membrane of a cell allowing for the diffusion of DNA into the cell [15, 18-20]. Particle bombardment refers to the use of high speed projectiles coated with DNA to introduce mechanically the coated DNA into cells[21-23]. Lipofection refers to the use of cationic lipid microbubbles called liposomes to deliver foreign DNA to cells[24, 25]. Because of opposite electrical charges, the cationic lipid encircles and packages the anionic foreign DNA. When these lipid–DNA complexes are added to cells, the lipid fuses with the membrane of the cell and delivers the foreign DNA. Compared to other methods, the liposomal method generally produces a high transfection rate with very little cell mortality. In addition, of the mentioned nonviral methods of transfection, only lipofection has the potential for extensive in vivo use[16], but, like viral methods, it also suffers from the lack of site specificity and various other application problems[26].

1.2.3 Ultrasound mediated intracellular drug and gene delivery

Ultrasound mediated delivery of compounds such as drug and gene into cells is a relatively recent development in drug delivery and gene transfection techniques, and has gained attention only recently. Ultrasound has been demonstrated to enhance local delivery of chemotherapeutic compounds, genetic materials, and fluorescent dextran

18 molecules into viable cells. Ultrasound has shown to enhance gene transfer in vitro[2, 27-

31] and in vivo[32-35]. Recently, ultrasound methods have been used in a wide range of

applications including cancer drug delivery, in vivo cardiac gene delivery[36],

thrombolysis [37-41], and opening the blood brain barrier[42-44] for drug delivery.

Ultrasound mediated transfection can be controlled both spatially and temporally through the exposure volume distribution and the application time of the ultrasound energy. The unique mechanism for external control makes this method potentially suitable for in vivo, site specific transfection as a means of gene therapy, providing

notable advantages over other strategies. Ultrasound application can be completely non-

invasive and does not require the insertion of an ultrasound device into the patient.

Application of ultrasound energy can be very precise as the ultrasound field can be readily focused within a desired tissue volume (as small as approximately 20 mm3) without affecting the surrounding and intervening tissue. In addition, as the ultrasound treatment is non-ionizing, it is devoid of any long-term cumulative effects and can be used repeatedly if needed.

The mechanism by which gene transfection is enhanced by ultrasound-mediated destruction of carrier microbubbles is unknown. It is hypothesized that the extracellular compounds are transported into the cell through membrane disruption or temporary pores induced by ultrasound application, a process called sonoporation [2, 45]. In this process, ultrasound increases the porosity of the cell membrane, possible through formation of transient openings in the cell membranes which allowing entry of protein and other macromolecules through diffusion. The mechanism by which this occurs is incompletely understood, but may include cavitation [46] [27, 31, 47-49] and other effects that may

19 cause sonoporation. Cavitation refers to the formation and destruction of microbubbles of gas in acoustic fields. Cavitation begins as propagating pressure waves strike bubbles, which are preexisting or formed by the low-pressure portion of the acoustic wave as it passes through media rich in dissolved gases. These bubbles oscillate and can be destroyed under the influence ultrasound exposure. The destruction can concentrate the intensity of an acoustic field up to 11 orders of magnitude in very small and localized volumes [50], which, hypothetically, increases cell membrane permeability and allows the uptake of foreign DNA.

However, despite of the increasing interest and recent encouraging progress, it remains a challenge to achieve consistent, controllable delivery outcome. Ultrasound- mediated transfection currently achieves transfection efficiencies of 2.4% using b- galactosidase reporter gene [27] or 15% using green fluorescent protein (GFP) [51]. The low efficiencies make potential application in vivo improbable and even makes in vitro experiments difficult. It also has been difficult to obtain consistent delivery outcome results. The development of sonoporation as a robust drug/gene delivery strategy has been hindered because of the insufficient understanding of the sonoporation process and mechanism. Therefore, understanding of the mechanism of sonoporation is of critical importance to ultimately develop optimal ultrasound strategy for targeted drug and gene delivery.

Due to the lack of methods for real-time monitoring of sonoporation at the cellular level the efficiency of drug/gene delivery and sonoporation associated side effects such as loss of cell viability and enhanced apoptosis[52] have been studied only through post ultrasound exposure analyses. Furthermore, because microporation appears

20 to be transient [30], it was not possible to directly correlate transfection with

microporation on an individual cellular basis using analysis performed after ultrasound applications. These questions and problems need to be answered before ultrasound method can be utilized to its full potential.

In order to address this important yet unanswered question, the research conducted in our laboratory has concentrated on investigating the mechanisms and dynamics of sonoporation. For the first time we obtained real-time measurements of sonoporation process by innovatively employing two-electrode voltage clamp technique

[4, 53] and fluorescent Calcium imaging. These methods enable us to monitor dynamic cell membrane status in real time and directly investigate the effects of ultrasound parameters and cell signals in cells sonoporation. Our results revealed important information of sonoporation and directions for future research.

1.2.4 Significance of this study

As described in the background, sonoporation offers an advantageous intracellular delivery strategy compared to other techniques such as electroporation [15, 18-20, 54] or viral gene transfection since ultrasound method is safe and can be non-invasive, which are promising properties ideal for in vivo application [36, 55, 56]. However, challenges remain to achieve controllable and consistent delivery outcome because of the insufficient understanding of the sonoporation process. Although previous studies have shown that factors associated with microbubble contrast agents and ultrasound

parameters affect delivery outcome, the correlation of these factors with delivery

outcome is only phenomenally rather than mechanistically established. Furthermore,

large variability exists quantitatively in the reported studies, thus such correlations

21 inevitably lack generosity and are unable to provide meaningful information to guide optimization of sonoporation delivery.

In particular, challenges in several key aspects remain to be addressed before sonoporation can be used successfully in humans as an efficient and safe strategy: 1) lack of means to rationally determine optimal sonoporation parameters, including physical and biochemical factors, to ensure high delivery efficiency and consistent outcome; 2) lack of mechanistic understanding of the causes for the downstream, cellular bio-effects and organ-level impacts of sonoporation; 3) lack of valid correlation and capability for

translating in situ results to in vivo environment. Clearly, tackling these difficult yet

important tasks requires in-depth investigation of the sonoporation process and

identification of the major factors affecting it.

The observation of stably enhanced intracellular uptake of markers and the expression of intentionally delivered genes in viable cells via sonoporation indicate the transient nature and small scale of the membrane poration process. The ultrasound generated pores on the plasma membrane must reseal to prevent the loss of intracellular contents to ensure cell survival, thereby limiting efficient inward transmembrane passage of desired extracellular agents within a time window before the completion of pore resealing. Furthermore, repair of the membrane disruption is necessary to avoid intracellular overload of ions that might be toxic to the cell or serve as the triggering sources for other irreversible and reversible cellular processes such as apoptosis [52] and calcium oscillation [57], making the rate of resealing one of the key factors determining the uptake efficiency and post-ultrasound cell fate.

22 It is therefore of significance to understand the process of sonoporation resealing,

yet it is a task challenged by the lack of appropriate techniques to study the transient and

sub-micron process. Consequently, sonoporation studies have been largely limited to

static post-ultrasound assays. While important knowledge can be obtained through such

analysis, post-ultrasound assays inevitably overlook the actual transient process of cell poration. Selection and attempted optimization of sonoporation parameters have mainly relied on empirical results of delivery outcome obtained after sonoporation. However, it

is likely that the ultrasound parameters determined this way are only associated with

specific experimental conditions due to the complexity of the ultrasound-cell interaction

and its coupling with surrounding bubble activities. Given the statistical variance of such

interactions in large number of cells, post-ultrasound assays are usually inadequate to

deterministically correlate ultrasound parameters with sonoporation outcome and to uncover cellular mechanisms of sonoporation in individual cells.

To address these challenges and to understand the mechanism of sonoporation, we demonstrated previously for the first time the feasibility of studying sonoporation at the single cell level in real time using the voltage clamp techniques [4, 58]. The TMC of single cell under voltage clamp was measured to assess the change of cell porosity in sonoporation. Before ultrasound application, the TMC is close to zero at a constant membrane holding potential (voltage clamped) in the absence of activation of endogenous ion channels, as the whole cell membrane is regarded as a resistor with constant resistance [6, 53]. In sonoporation, ultrasound generates pores on the membrane

(the pores effectively reduce the membrane resistance), which allows ions to flow

through the pores and results in change in TMC. The TMC is determined by the pore size

23 and ion concentration gradient across the cell membrane. Therefore the TMC can be used

as a sensitive means to monitor the dynamics of ultrasound induced pores in a single cell

with high temporal resolution and sensitivity [4]. The novel application of such electrophysiological techniques enables a time-resolved measurement of sonoporation at single-cell level, providing a sensitive and quantitative means to investigate the sonoporation process.

CHAPTER II

MEASUREMENT OF SINGLE CELL SONOPORATION USING VOLTAGE

CLAMP TECHNIQUE

In this chapter, we describe in detail the method of using voltage clamp technique to monitor the dynamic sonoporation process via the TMC of a single cell under voltage clamp (fixed membrane potential). Specifically, effects of three aspects are discussed in this chapter: microbubble concentration, vitelline membrane of oocyte, and the duty cycle of ultrasound exposure.

2.1 Introduction

2.1.1 Monitoring of sonoporation using voltage clamp technique in real time

As described in Chapter 1, study of the dynamic sonoporation is a task challenged by the lack of appropriate techniques to study the transient and sub-micron process.

Consequently, sonoporation studies have been largely limited to static post-ultrasound assays. To address these challenges and to understand the mechanism of sonoporation, research in our laboratory has demonstrated previously for the first time the feasibility of studying sonoporation at the single cell level in real time using the voltage clamp techniques [4, 58]. The TMC of single cell under voltage clamp was measured to assess the change of cell porosity in sonoporation. Before ultrasound application, the TMC is close to zero at a constant membrane holding potential (voltage clamped) in the absence of activation of endogenous ion channels, as the whole cell membrane is regarded as a

25 resistor with constant resistance [6, 53]. In sonoporation, ultrasound generates pores on the membrane (the pores effectively reduce the membrane resistance), which allows ions to flow through the pores and results in change in TMC. The TMC is determined by the pore size and ion concentration gradient across the cell membrane. Therefore the TMC can be used as a sensitive means to monitor the dynamics of ultrasound induced pores in a single cell with high temporal resolution and sensitivity [4]. The novel application of such electrophysiological techniques enables a time-resolved measurement of sonoporation at single-cell level, providing a sensitive and quantitative means to investigate the sonoporation process.

Xenopus oocyte is chosen as the model cell for this study. Xenopus oocytes have been used to study membrane proteins such as ion channels and neurotransmitter receptors [59]. It is a commonly used cell model for the study of channels of cell membrane because of the large size of the cell. The cells are easy to obtain from Xenopus laevis which is an easy feed frog. In addition, the cell contains limited number of endogenous ion channels on its membrane. The two kinds of endogenous currents are the potassium leak current out of the cell and the currents which link to extracellular receptors such as muscarinic receptors (which turn on a calcium activated outward chloride current) and Beta-adrenergic receptors (which turn on an outward potassium current) [60].

Sonoporation in a single cell was initiated in the presence of microbubbles used to facilitate sonoporation via cavitation exposed to ultrasound application. In this chapter, we specifically address two basic questions regarding the voltage clamp measurement of

26 sonoporation: the distribution of microbubbles and the effect of the vitelline membrane

surrounding the oocyte.

2.1.2 Microbubbles to facilitate sonoporation

It has been demonstrated that microbubbles such as ultrasound contrast agent (UCA)

greatly increased the rate of sonoporation and the delivery efficiency of drug and genes

into cells [2, 31, 61-63]. The mechanism of the enhancement is regarded as acoustic

cavitation where rapid growth and collapse of gaseous bubbles. During the collapse of bubbles, cell membrane will experience mechanically impact involving shear stress [64-

67], shock waves, and microjets [68]. The subsequent cell damage will be either reversible which increases the intracellular uptake of molecules or irreversible that leads to cell death [62].

Microbubble UCAs are usually microbubbles for ultrasound imaging applications.

Such blood pool agents, when injected into the blood stream, enhance the ultrasound

backscattered signals for better image quality due to increased signal to noise ratio[69].

Due to the large acoustic impedance mismatch of gas and liquid (e.g. blood),

microbubbles interact with ultrasound very efficiently, exhibiting dynamic responses

under the influence of ultrasound such as oscillation and collapse. Because of the

efficient ultrasound interaction and the mechanical impact of microbubble dynamics, the

microbubble agents have been used to facilitate sonoporation. Two FDA approved

ultrasound imaging microbubble contrast agents, Optison™ (GE Healthcare Inc.,

Princeton, NJ) and Definity® (Bristol-Myers Squibb Medical Imaging, N.Billerica, MA)

were used in our study. Both of these agents contain stabilized microbubbles with

different sizes (Optison™: mean diameter is 3.0-4.5µm, 95% of bubbles is smaller than

27 10µm, and max diameter is 32µm; Definity®: mean diameter is 1.1-3.3μm, 98% of bubbles is smaller than 10μm, and max diameter is 20μm).

Statistically, multiple bubbles are present in the vicinity of the membrane in our experiments where a single oocyte was immersed in microbubble solution. Therefore multiple pores may be generated at different locations unevenly distributed on the cell membrane. Since the whole cell clamp configuration measures the ion movement through any pores on the whole membrane, the measurement should not be affected by, nor be used to distinguish the locations of the pores and/or the electrodes. Additionally, since the microbubbles were suspended in solution, and they might be slowly flowing up due to buoyancy, the change of bubble concentration will affect the number of pores formed in sonoporation.

2.1.3 Effects of vitelline membrane in sonoporation of oocytes

The vitelline membrane, which is outside of plasma membrane, is an extra layer of membrane only shown up on egg cell. Composed by proteins, the vitelline membrane works as a mechanical shell for the egg cell and therefore may affect the sonoporation process which relies on mechanical stress to open membrane. The standard voltage clamp procedure does not include devitellinization procedure because vitelline membrane is transparent to small molecules such as ions (though patch clamp procedure does include the devitellinization procedure because vitelline membrane will prevent glass pipette to form a giga seal on plasma membrane). Therefore for sonoporation study, the effect of the vitelline membrane in sonoporation process needs to be determined.

28 2.1.4 Effects of ultrasound parameters in sonoporation process

The effects of ultrasound parameters such as the acoustic pressure amplitude and duty

cycle of ultrasound exposures are important in sonoporation process. Both continuous

wave (CW) and pulsed wave were investigated. While CW usually deposits more energy

to initiate sonoporation, pulsed ultrasound can potentially decrease cell death rate and has

the potential to integrate with ultrasound imaging, where typically low duty cycle pulses

are used. Voltage clamp method was employed to examine for the first time the effects of ultrasound parameters such as acoustic pressure and duty cycle in sonoporation process.

2.2 Methods

2.2.1 Experimental setup

The schematic diagram of experimental setup is shown in Figure II-1. An oocyte

(0.9~1.1mm diameter) was sitting on a 35 mm polystyrene BD Falcon™ bacteriological

Petri dish (Fisher Scientific, Pittsburgh, PA) containing ND96 (contains (mM) 96 NaCl, 2

KCl, 1.8 CaCl2, 1 MgCl2, 5 HEPES, pH=7.6) solutions. Oocyte membrane potential was

clamped at -50mV by a two-electrode voltage clamp amplifier (Dagan CA-1B, Dagan

Corp., Minneapolis, MN). A microbubble UCA (Optison™ or Definity®) was diluted and added directly into the dish or through perfusion. A single element, circular planar piezoelectric ultrasound transducer (~5.1cm diameter, 0.96 or 1.075MHz center frequency) was vertically directed upward to the oocyte in the dish 1 inch away by a customized holder. The acoustic pressure amplitude was determined by two methods, either by measuring power output by ultrasound power meter (UPM-DT-10, Ohmic

Instrument Co., Easton, MD) and scanning ultrasound field by a 3D scanning system

29 (ESP300, Newport Corp., Irving, CA), or by a calibrated 1mm diameter needle

hydrophone (NP-4, Dapco, Ridgefield, CT). Because of standing wave formation in the

chamber from the solution-air interface, the actual pressure in the dish was measured using the hydrophone placed at the location of oocyte in the dish during experiment.

2.2.2 Data acquisition

The data acquisition of voltage clamp signal was controlled by a commercial program

(Pulse, Heka Inc, Port Washington, NY). The recording time of TMC was up to 15 min.

A 5ms TTL trigger signal was generated 1 second after the beginning of voltage clamp

recording. A function generator (33250A, Agilent, Santa Clara, CA) was activated by the

rising edge of this trigger signal. A preprogrammed 0.2s tone burst pulse or multiple

shorter pulses were generated by the function generator, and then amplified by a 75 Watt

RF power amplifier (75A250, Amplifier Research, Souderton, PA), eventually drove

transducer to initiate ultrasound exposure. A computer interface (ITC-16, Instrutech, Port

Washington, NY) was used to digitize voltage clamp signal and to generate trigger signal.

The sampling rate of voltage clamp recording was 1 KHz. In order to obtain all dynamic

information, no hardware filter was used.

2.2.3 Xenopus oocyte preparation and animal protocol

Similar animal protocol for Xenopus oocytes harvest and preparation as described in

previous published papers [4, 58] were used in our study. The animal protocol has been

approved by the Institution Animal Care and Use Committee at Case Western Reserve

University and the University of Michigan, where the experimental studies have been

conducted. Below is the summary of the procedures for Xenopus occytes preparation.

30 Xenopus laevis (Adult female frog, NASCO, Fort Atkinson, WI) was anesthetized

by immersing in 1g/L Tricaine methane sulfonate (MS222) solution with 0.5g/L

NaHCO3 as buffer for 15-20 minutes. Part of ovary was taken from a small incision

(~1cm) in the lower abdomen which was sutured right away. The frog was put into water

to recover and monitored in the laboratory until the end of the day. Oocytes were

harvested by digesting overy lobe in collagenase (1-2mg/mL in calcium free OR2

solution, contains (mM) 82.5 NaCl, 2.5 KCl, 1 MgCl2, 2.5 HEPES, pH=7.6) around 1

hour for defolliculation. They were used immediately in experiments or stored in ND96 solution at 18oC for 1-2 days before use.

2.2.4 Voltage clamp procedure

Microelectrodes (tip diameter ~1 µm) made from glass pipettes (Warner Instrument

Corp., Hamden, CT) using an electrode puller (Sutter Instrument Co., Novato, CA) were

filled with 3 M KCl and had resistances between 0.3-1.2 MΩ. Two microelectrodes

connected to a voltage clamp amplifier (Dagan CA-1B, Dagan Corp., Minneapolis, MN)

[6] were inserted into the oocyte membrane to measure the TMC. One was to measure potential inside of cell and the other was to inject current into cell to maintain the membrane potential set by command. In our experiment, we set the membrane potential of the oocyte at −50 mV during recordings which was around the normal membrane potential value. The purpose was to reduce the possibility of opening any endogenous ion channel. Statistically, multiple bubbles are present in the vicinity of the membrane in our experiments. Therefore multiple pores may be generated at different locations unevenly distributed on the cell membrane. Since the whole cell clamp configuration measures the

31 ion movement through any pores on the whole membrane, the measurement should not

be affected by, nor be used to distinguish the locations of the pores and/or the electrodes.

2.2.5 Procedure to measure bubble concentration as a function of time

Figure II-2 describes the measurement method for bubble concentration at different

locations inside the dish as a function of time. A Nikon inverted microscope (TE200)

with a 20X objective lens (NA=0.40) was used with digital camera to take light

micrograph of microbubbles in solution at different focal depths as a function of time.

The area of the field of view was calibrated to be 433X322µm2 and the depth of field was

5.8 µm [70]. The numbers of bubbles in each image were carefully counted.

Images were taken at individual time points (from 0-600 seconds, mean time interval 43s) after adding and mixing solution of microbubbles into a 35mm diameter

Petri dish (t=0 s) and from 8 vertical positions equally spaced from bottom of dish to 1 mm above the dish bottom (average diameter of oocyte about 1 mm). Ideally, the images at different vertical locations should be taken simultaneously at the same time point.

However, due to practical limitations, images were taken consecutively at these vertical

positions with a short time delay (e.g. ~ 2 s). This time delay (can be considered as fast

time) is much shorter compared to the time interval of 43 s (slow time), so the images

were considered approximately simultaneously for a given time. One way analysis of

variance, or one way ANOVA, was used to compare the difference of the numbers of

bubbles at different positions and at different time.

The original microbubble concentration was measured after dilution by two

methods. In the first method, the number of bubbles was counted from the images taken

32 with the microscope (Figure II-2), and then the concentration was estimated by this

number divided by the focal volume of the microscope. In the second method, a hemocytometer (PGC Scientifics, Frederick, MD) was used (Figure II-3) to count microscopic particles/bubbles. (The hemocytometer is a device normally used to count blood cells.) After the chamber of the hemocytometer was filled with microbubble solution, the chamber was then covered with a cover glass slide and examined via a microscope. The number of bubbles in the chamber was then determined by counting.

The volume of solution can be easily calculated base on how many squares included in the counting. The concentration of bubbles was calculated as the number of bubbles divided by the volume of solution.

2.2.6 Experimental design to examine the effects of vitelline membrane in sonoporation

Vitelline membrane is a layer of membrane on egg cells with specific functions relating to fertilization. In typical electrophysiology studies using Xenopus oocytes as model cells, the vitalline membrane is considered to have no effect on ion transport and current measurement. Thus no devitellinization procedure is performed for these studies.

Vitelline membrane is composed of proteins and may decrease ultrasound effect in sonoporation. The following tests were performed to ascertain the effects of the vitelline membrane on the dynamics of cell resealing process as measured by the transmembrane currents in sonoporation, by comparing the recovery of currents between oocytes with and without the vitelline membrane.

The devitellinization procedure was done by first placing oocytes in stripping solution (NMG-Aspartate 200mM, KCl 2mM, EGTA 10mM, MgCl2 1mM, HEPES

10mM, pH=7.4) for 2-5 minutes until part of vitelline membrane detached from

33 extracellular matrix. Then two forceps were used to manually peel off the vitelline

membrane. The devitellinized oocytes were used immediately for voltage clamp

recording.

A comparison experiment was done with two groups. One group used oocytes

with vitelline membrane intact, while the other group used oocytes without vitelline

membrane. All the other experimental conditions were kept the same for the two groups

(0.2 s tone burst ultrasound at 1.075 MHz, 0.3MPa with 4ml ND96 solution in a 35mm

diameter Petri dish, 0.1% Definity® concentration, 50+ s voltage clamp recording). The

sample size was determined by

2 ⎡ z σ ⎤ n = α / 2 , [2.1] ⎣⎢ E ⎦⎥

where zα / 2 is the critical value of α / 2 of the standard normal distribution, σ is the

population standard deviation and can be replaced by the sample standard deviation s if

n>30, E is the margin of error which describes the maximum difference between the

observed sample mean x and the true value of the population mean μ , and n is the

sample size. From our preliminary data, s=3.4e-7. Assume E=1.2e-7, and zα / 2 =1. 96

(α = 0.05 ), n=31.

The information we extracted from the TMC measurements included maximum

current amplitude and recovery dynamic parameters, representing the total pore area and

pore resealing dynamics, respectively.

The mean of maximum current amplitude (which relates to total area of pores)

were tabulated for the two groups of oocytes with and without vitelline membrane. Data were shown in a box plot (includes mean, median, 25 and 75 quartile, max and min

34 value) of the two groups. A t-test was performed to test the significant differences between groups with and without vitelline membrane. A p-value of less than 0.05 was considered as statistically significant.

The dynamic information of sonoporation process was shown by a least square fitting of TMC. For most data, a two exponential function was able to fit these experimental data, suggesting the existence of a fast recovery and a slow recovery process, respectively. The function can be described by

I(t') = −I 0 − I f exp(−k f t') − I s exp(−kst') , [2.2]

where t' is time after ultrasound pulses, I 0 the pre-ultrasound current (~0 at -50mv

holding voltage), I f and K f the amplitude and time constant of fast recovery process,

I s and K s the amplitude and time constant of slow recovery process. Then the amplitude of both recovery processes was normalized by being divided to maximum current amplitude. Finally four box plots of these parameters were calculated to compare any difference between the results of oocytes with and without vitelline membrane. A t-test will be performed respectively for significant differences between two groups. A p-value of less than 0.05 will be considered as statistically significant.

2.2.7 Experimental design to examine the effect of duty cycle on sonoporation process

The same experimental setup and procedure are used as mentioned in 2.2.1, 2.2.2 and

2.2.3, except ultrasound exposure protocol. In this study, pulsed ultrasound applications at three different duty cycles (5%, 10% and 15%) were employed.

35 2.3 Results

We recorded TMC of a single Xenopus oocyte under voltage clamp during ultrasound application in the absence and presence of microbubbles in the solution. No change in the

TMC was observed without ultrasound application or with ultrasound application in the absence of Definity®. On the other hand, we have consistently observed that ultrasound application (0.3MPa) induces increased inward TMC in the presence of Definity®.

Figure II-4 shows a typical example of the TMC change during sonoporation. The inward

TMC exhibits a rapid increase almost immediately after ultrasound activation, indicating the formation of pores on the cell membrane. After reaching a maximum value and after the ultrasound exposure, the TMC starts to undergo a decay or recovery process, relaxing back to the pre-ultrasound equilibrium value, indicating resealing of the pores. These results demonstrated that the voltage clamp technique can be used as a novel means to monitor cell membrane porosity in real time with high sensitivity and temporal resolution. The quantitative descriptions of TMC behavior in sonoporation were discussed in our early paper [4].

2.3.1 Microbubble concentration in solution during experiment

The results of bubble counting (n=3) are plotted in 3D ribbons plots (Figure II-5). The plots show the number of bubbles within the region of interest in each image. The increase of bubble numbers at higher vertical locations at later times indicates the rise of the gaseous bubbles due to buoyancy. Furthermore, results in Figure II-5A and 5B (data set C does not have information at time 0) show a rapid initial drop in the number of bubbles from 0-43s. There were more than 40 bubbles at the beginning and only about 30 at the second recording (at about 43s later) at most of the vertical positions. The bubble

36 numbers do not have significant fluctuation after first 43s, suggesting stabilized spatial distribution of bubbles in the solution. Except at locations near the bottom of the dish

(location 0.14 mm above the dish bottom or the dish bottom) with fewer bubbles, the numbers of bubble at most positions are similar.

One way ANOVA is performed on three sets of data. If the data recorded at the beginning (time 0 or/and 43) or data captured at the bottom of dish (height=0mm) are excluded, the test shows no significant difference between three datasets, indicating that the numbers of bubbles are distributed homogeneous and are not changing with time.

Using these stabilized bubble distribution, the actual bubble concentrations were calculated. Table II-1 shows the bubble counting result using the microscopic counting method (method 1). Table II-2 shows the bubble counting result using the hemocytometer

(method 2). The last column of the table, the bubble concentrations were calculated by the number of bubbles divided by the solution volume. The concentration measurement result from method 1 was 3.75±0.65E+7 bubbles/ml, and the result from method 2 was

1.10±0.31E+7 bubbles/ml. These results were from the dilution of 1000, and were close to the original concentration of Definity® (manufacturer packing slip) of up to 1.2E+10 bubbles/ml.

2.3.2 The effects of vitelline membrane of Xenopus oocytes in sonoporation

Maximum amplitude of TMC change is associated with the maximum total pore area in sonoporation. Figure II-6 shows box plots of the maximum TMC measurements for normal and devitellinized cells in sonoporation. The devitellinized occytes exhibited a much larger maximum current values compared to the normal oocytes under the same

37 ultrasound parameters (p<0.01), suggesting the membrane acted as a mechanical insulator to the ultrasound impact.

Effects on the recovery dynamics of pores were investigated to demonstrate whether different mechanisms are involved in sonoporation recovery for oocytes with and without vitelline membrane. Based on reported results in literature and the physical/mechanical nature of sonoporation pore formation process, we expect no statistical significant differences between these two groups if the membrane damage (or maximum TMC) is within the range to maintain normal cellular functions and membrane resealing process. Our experimental results confirmed the expectation. Figure II-7 shows four parameters fitted from least square fitting of current data which represent the slow and fast recovery time constants and normalized amplitudes. The decay/recovery time constants show no statistical significant difference (p<0.05) between the two groups.

Difference exists for the amplitude of the fast recovery process, which has a smaller mean value for devitellinized cells, indicating an overall slower resealing process. This could be due to the generally weakened cell without the mechanical support of the vitelline membrane, and the larger maximum current achieved without the vitelline membrane (Figure II-6). As further demonstrated in Figure II-8, there is statistical significant difference between normal and devitellinized data in terms of normalized amplitude of fast recovery vs. maximum TMC. This is most likely due to a larger TMC for the devitellinized group.

2.3.3 Duty cycle effect on sonoporation

Figure II-9 shows the TMC measured in an oocyte exposed to pulsed ultrasound at a duty cycle 10% and a pulse repetition frequency (PRF) of 1 Hz. The Optison™ concentration

38 was 10% and the acoustic pressure 0.4MPa. Marked difference in the dynamics of sonoporation in the cellular level between tone-burst ultrasound [4] and pulsed ultrasound is clearly demonstrated when the TMCs are compared. During the tone burst application the TMC increased continuously until ultrasound is terminated and then underwent a monotonic declination (recovery to the equilibrium level), indicating the resealing of the membrane. During the pulsed ultrasound exposure, the amplitude of the current increases at the beginning of most discrete pulses and started to recover during the silent interval until the next pulse, allowing the membrane to reseal partially.

In order to quantitatively characterize the sonoporation dynamics affected by pulsed ultrasound systematically, we exposed Xenopus oocytes to pulsed ultrasound of duty cycles of 5%, 10% and 15%. With a constant PRF of 1 Hz, the corresponding pulse length is 0.05s, 0.1s, and 0.15 s respectively. The total duration of ultrasound activation was kept at 10 s and the acoustic pressure at 0.4 MPa. Optison™ concentration (5%) was kept constant for these experiments.

From the measured TMCs resulted from pulsed ultrasound exposures of different duty cycles (Figure II-10), several parameters are computed to quantitatively assay the cell’s dynamic states during sonoporation. The maximum current change (Figure II-11) denotes the maximum amplitude of the ionic current through the membrane. Since the pulse repetition time (PRT=1/PRF) was fixed at 1 s, an increase in duty cycle creates longer duration for each discrete pulse, which generates larger transmembrane increase as reported previously [4]. Figure II-11 shows an expected increase of maximum current change with increasing duty cycle.

39 The current-time curve length (CTCL) measures the length of the curve representing the dynamics of TMC change over a period of time (e. g. 20 s) after ultrasound activation. The signal is normalized by the maximum current change to eliminate any ionic or other environment dependence, to obtain more relevant information associated with cell membrane dynamics. The CTCL is computed via the following algorithm

t 2 2 2 ⎛⎞It() ⎛⎞ t ⎜⎟+ ⎜⎟dt , [2.3] ∫ ()ΔI PRT t1 ⎝⎠max ⎝⎠ where I (t) is the TMC at a particular time t, the time elapsed after ultrasound activation.

Our result shows that the curve length increases with increasing duty cycle (Figure II-12), as more fluctuations are often observed with increasing duty cycle (Figure II-9).

Obviously, a constant current change should yield a small curve length. This parameter is recognized to indicate the dynamic plasma membrane disruption and might be related to more complex cellular processes as the cell membrane experiences repeated disruption.

As the duty cycle elevates, the cell is being obstructed more from recuperating and thus the recovery process is expected to experience more dynamics. While CTCL is not expected to affect delivery efficiency to a significant degree, whether it has implications on ultimate cell fate (survival) requires further study.

We investigated post ultrasound cell recovery affected by duty cycles and found that higher recovery percentage is observed at lower duty cycle initially, indicating that the oocytes recovers faster exposed by ultrasound at lower duty cycles[58]. As the membrane must reseal completely for cells to survival, immediate loss of cell viability or short term cell death can also be assessed from the TMC.

40 We examined cell death rate for different duty cycles (at 0.4 MPa) and noted a decrease in cell survival rate at higher duty cycle, shown in Table II-3.

2.3.4 Time constant of voltage clamp system

The time constant of the voltage clamp system was measured from the step response of a

5ms, 50mv pulse on a clamped oocyte. The step response of the voltage was the same as the step response of TMC [71]. An example of the recorded signals is shown in Fig. II-

13. The time constant of voltage clamp τ is obtained from the measurements as

1 0.211±0.048 ms (N=8). The -3dB bandwidth is therefore calculated: f = = 754Hz . 2πτ

2.4 Discussion

Voltage clamp technique enables a high temporal resolution sufficient for the real time measurement of sonoporation which has not been possible with any other methods. For example, the time resolution of the measurement shown in Figure II-4 is 1 KHz.

Although the sampling rate of our A/D board can be as high as 20 KHz, a sampling frequency higher than 6 KHz is not meaningful due to the limitation of the time constant of voltage clamp amplifier. As shown in 2.3.4, the time constant of voltage clamp with oocyte is around 200µs, which suggests that the voltage clamp system is sufficient to capture signals below 800Hz. Therefore, in our experiments, if the pore formation is slower than 200µs, the signals measured by the voltage clamp system represent the dynamics of this process. Since the cell resealing process is much slower than the voltage clamp system response, e.g. the time constant of pore resealing process can be around

20ms, about 100 times slower than the time constant of the voltage clamp, the voltage clamp system can measure accurately the maximum amplitude of TMC. One potential

41 improvement for higher temporal resolution is to use patch clamp technique, which is generally used on small cells containing less membrane capacitance and therefore with a shorter time constant (can be as short as 1µs).

Maintaining identical experimental condition is critical in our experiments.

Suggested by the measured results of bubble concentration (Figure II-5), the bubble concentration is statistically stable at different locations, except near the bottom of dish or the air-solution surface due to buoyancy. The results obtained from the microscope images and hemocytometer shown a 2-3 times difference between the results, but mostly within the same order of magnitude. The results from the same methods are very close and have a small variance. Since the coulter counter can be another independent way to count the bubble concentration, a future study can done by both two methods and coulter counter to measure the bubble concentration and compare the results.

The results of both amplitude of current and dynamics of resealing confirmed the largely mechanical nature of the vitelline membrane on sonoporation, serving as a mechanical shell and reduce the TMC amplitude. Additionally, the vitelline membrane does not change the dynamics of oocyte resealing process significantly. It is noticed that the devitellinization procedure will generally weaken the cell. These results suggest that the devitellinization procedure for oocyte experiment on sonoporation can be avoided in the study of the cellular process of pore resealing.

The duty cycle study demonstrated the application of voltage clamp method to investigate the effect of acoustic parameter. The real time monitoring capability can provide feedback to adjust ultrasound parameters in order to ensure reversible

42 sonoporation sufficient for desirable delivery outcome. The total pore area (maximum current amplitude), the time when pores are formed, and even the dynamics of pore resealing, are clearly shown in the recording.

2.5 Conclusion

We demonstrated the feasibility of using voltage clamp technique to real time monitor sonoporation process at the single cell level by measuring TMC through transient pores on the cell membrane under voltage clamp.

Under our experimental conditions, the bubble concentration reached a stabilized value after some initial delay (3-5 min) of solution mixing. This delay was used in all experiments later on to ensure constant bubble concentration in sonoporation. The vitelline membrane appears to act as an attenuation layer to sonoporation pore formation, decreasing the maximum current achieved during sonoporation, compared to a larger mean value of maximum currents in devitellinized cells. However, no affects on the resealing time constants were observed, validating the choice of normal cells in the later study of sonoporation resealing process using voltage clamp techniques.

We also investigated the effects of ultrasound duty cycle on sonoporation dynamics using Xenopus oocyte as a model system. Transducer with a center frequency of 0.96 MHz was used to generate pulsed ultrasound of desired duty cycle (5%, 10%, and

15%) at a PRF of 1 Hz and an acoustic pressure of 0.4MPa in our experiments. We characterized the sonoporation dynamics from these time-resolved recordings of TMC to indicate cell membrane status such as pore formation, extension, and resealing. We observed that the TMC amplitude increased with increasing duty cycle, while the

43 recovering process of membrane pores and cell survival rate decreased at higher duty cycles.

Voltage clamp amplifier

Im V m 3D motion 1mm controller

Water

Signal generator

Power amplifier US Transducer

Figure II-1: The experimental setup of real time monitoring sonoporation process by voltage clamp technique. A 3D motion system is utilized to precisely adjust transducer field. The left picture shows a clamped oocyte in bubble solution in Petri dish.

Figure II-2: Diagram of bubble concentration measurement procedure.

Figure II-3: Picture of a hemocytometer.

Figure II-4: An example of the TMC of a single Xenopus oocyte under voltage clamp during sonoporation. The ultrasound application is indicated by the horizontal bar shown in the plot.

Table II-1: Bubble concentration measured by method 1

Experime Date Sample Number of Volume of Dilution Bubble nt # size (N) bubbles focal zone times concentration (ml) (bubble/ml) 1 03/20/06 91 31.63±4.28 8.0867E-7 1000 3.91±0.53 E+7 2 03/20/06 91 31.35±5.35 8.0867E-7 1000 3.88±0.66 E+7 3 03/20/06 84 27.15±5.41 8.0867E-7 1000 3.36±0.67 E+7 Overall 3.75±0.65E+7

Table II-2: Bubble concentration measured by method 2.

# of Definity Counting Sample Nmber of Volum Dilution Bubble Test activation date size (N) bubbles e (ml) times concentration date (bubble/ml) 1 08/08/06 08/31/06 6 39.83±3.43 4E-6 1000 9.96±0.86E+6 2 08/08/06 09/08/06 25 36.24±4.73 4E-6 1000 9.06±1.18E+6 3 08/08/06 09/08/06 25 32.60±4.12 4E-6 1000 8.15±1.03E+6 4 09/06/06 09/08/06 12 58.92±7.12 4E-6 1000 1.47±0.18E+7 5 09/06/06 09/08/06 25 57.00±6.66 4E-6 1000 1.43±0.17 E+7 Overall 43.92±12.43 1.10±0.31E+7

-2

-4

-6

-8

-10

-12 Maximum current amplitude(uA) Normal Devitellined

-5

-10 current (uA) current -15

-20

-25

-30

-35 01234567891011 time (s)

Figure II-9: TMC of a Xenopus oocyte during sonoporation induced by pulsed ultrasound [58].

5% 10% 15% 0

-2 A)

μ -4

-6

-8 current (uA) current

-10

-12 Transmembrane current ( -14 US on

0 2 4 6 8 10 12 14 16 18 20 22 time (s)

Figure II-10: TMC at different duty cycles [58].

maximum current (uA) 20

0 51015 duty cycle (%)

Figure II-11: Maximal current vs. duty cycle. n =7 (5%), 7 (10%) and 8 (15%) [58].

40 path length path Curve length Curve

0 51015 duty cycle (%)

Figure II-12: Curve length of TMC vs. duty cycle. The ultrasound parameters are the same as in Fig. 5. n =7 (5%), 7 (10%) and 8 (15%) [58].

Table II-3: Cell survival rate at different duty cycles

Duty cycle 5% (n=7) 10% (n=9) 15% (n=11) Cell survival rate 100% 78% 73%

Figure II-13: Screen shot shows the step response of voltage clamp system with an already clamped oocyte from a 5ms, 50mv pulse.

CHAPTER III

EFFECTS OF EXTRACELLULAR CALCIUM ON CELL MEMBRANE

RESEALING IN SONOPORATION

Sonoporation can be regarded as a two-step process: pore formation due to ultrasound application and pore resealing. Both steps are important in delivery outcome. The ultrasound parameters control pore size which determines the size of agents that might be able to pass through the membrane into the cytoplasma. The extension or duration of pores controls the duration of the time window when molecules can be delivered. The delivery efficiency is determined by both of the maximal pore size and the duration of the transient pores. In this chapter, we focus on the resealing process in sonoporation affected by the extracellular calcium concentration.

3.1 Introduction

As described previously, the time dependent TMC of single cell under voltage clamp is related to the size of the transient pore. The TMC shows the change of the transient pore size including the cell resealing progress.

Our previous results demonstrated that calcium in the extracellular solution affects sonoporation resealing [4]. The current study focused on quantitative investigation of the effects of extracellular Ca2+ on the resealing of ultrasound induced membrane disruption; results of the study might provide important molecular insight into reversible sonoporation to guide optimal delivery outcome.

61 3.2 Methods

The oocyte preparation procedure is the same as described in 2.2.3.The experimental setup and procedures are the same as described in 2.2.1, 2.2.2 and 2.2.4 except that oocytes were placed into the ND96-based test solution with the desired [Ca2+] and [Mg2+] several (~3 to 7) min before sonoporation experiment. The [Ca2+] was varied from 0-

3mM (physiological extracellular [Ca2+] of Xenopus oocyte is 1.8mM, of human body is

2.2-2.6mM) and [Mg2+] was either 0 or 1mM.

3.3 Results

Similar to our previous results [4], no change in the TMC was detected without ultrasound application or when ultrasound was applied (duration 0.2 s, acoustic pressure

0.3MPa) without Definity® in the bath solution. The current stayed at a constant, equilibrium level close to 0 μA (−50 mV membrane holding potential), indicating the absence of endogenous channel activation by ultrasound application. On the other hand, when ultrasound was applied in the presence of Definity® (0.1%), the amplitude of the inward TMC increased rapidly after ultrasound, as shown by the examples in Figure III-

1, indicating the creation of non-specific pores on the cell membrane. Although no direct measurement was conducted in this study, sonoporation appeared to occur in these experiments as the result of inertial cavitation, or the rapid collapse of bubbles driven by an ultrasound field, as > 98% of the initially present Definity® bubbles were destroyed after ultrasound application based on post ultrasound bubble counting. It was also evident visually that the slightly milky solution became immediately clear after ultrasound exposure.

62 Figure III-1 includes examples of the inward (due to the -50 mV holding potential) TMCs recorded in single cells during sonoporation in the presence of different

[Ca2+]. Ultrasound tone burst of duration of 0.2 s was applied at 1 s after the start of the recording, indicated by the horizontal bar in the figure. Maximum change in the current occurred at the end of ultrasound pulse (at 1.2 s). For the convenience of comparison, each curve in the figure was normalized to the absolute value of the corresponding maximum current change (thereby the negative sign of the inward current is retained in the figure).

The maximal amplitude of the current change, usually on the order of 0.1–10 µA in our experiment, depends on the ionic concentration gradient across the cell membrane and the total area of the pores, which was determined by the ultrasound parameters used and the intrinsic properties of each individual cell in the experiment. For reversible sonoporation, where the cells survive, the TMC exhibited an initial rapid increase reaching a maximal value, followed by a recovery process, and eventually returned to its pre-ultrasound level, indicating complete resealing of the pores.

3.3.1 Extracellular Ca2+ is required for resealing in sonoporation

Pore resealing or recovery of membrane disruption in sonoporation, as indicated by the

TMC returning to its pre-ultrasound level in our experiment, was found to require extracellular Ca2+ in our study.

As shown in the Figure III-1, the inward TMCs recorded in single cells during sonoporation in solutions of different [Ca2+] indicate that the resealing process in sonoporation was affected by the [Ca2+]. In general, the TMC increased in amplitude

63 followed ultrasound exposure and underwent a recovery process thereafter. At 1.8 or 3.0 mM [Ca2+], the TMCs have similar dynamics and returned almost to their pre-ultrasound level by 15 s. In contrast, for cells in calcium-free solution (0 [Ca2+] with 1 mM EGTA added as Ca2+ chelator), the TMC exhibited minimal recovery with no sustained trend after sonoporation throughout the whole recording period in our experiment (data in later times up to 15 min not shown). However, when Ca2+ was added into the previously Ca2+- free solution ( ~2 mM final concentration), recovery of current was initiated within a few seconds after the addition of calcium and reached pre-ultrasound level within 15–40 s thereafter (data not shown), similar to those cells originally in the 1.8 or 3 mM Ca2+ solution.

In general, lower [Ca2+] resulted in slower, weaker resealing, and often partial recovery of current was observed. A threshold of 0.54 mM was found for complete recovery. Figure III-2A shows the final percentage of TMC recovery achieved (~ 15 min in our experiments) at different [Ca2+]. No complete recovery was observed for cells when [Ca2+] < 0.54 mM; in such cases, the TMC stayed at a level significantly below

100% (normalized by its maximum value), indicating failure of pore resealing. On the other hand, recovery of TMC continued beyond 15 s when [Ca2+] > 0.54 mM, and the cells were able to achieve complete recovery eventually. Furthermore, complete recovery was found to relate to cell survival (Figure III-2B). The cells that failed to achieve complete resealing died eventually after a period of time ranging from minutes to hours, determined by visible loss of intracellular contents and degeneration of cellular integrity, observed under the microscope (inset in Figure III-2B). Each of the images included in the inset in Figure III-2B shows a single oocyte with a diameter of about 1 mm taken

64 under a regular low magnification microscope (SM-4TY, American Scope Inc., Chino,

CA). The oocytes are shown with pigmented dark brown in one hemisphere (animal hemisphere) and the bright hemisphere (vegetal hemisphere). The pores that failed to reseal in the absence of Ca2+ can often be seen in the images (inset in Figure III-2B). Our previous study indicated that the resealing of the pores requires Ca influx into the cell

[18]. The electrochemical gradient that is generated by the –50 mV holding potential and high extracellular [Ca2+] drives extracellular Ca2+ into the cell to mediate the resealing process. The entry of Ca2+ into the cell cannot be through Ca channels that are endogenous in the oocytes membrane because these channels are open only at membrane potentials more positive than –50 mV [24]. At the holding potential of –50 mV, no Ca2+ currents can be or were recorded through the membrane of oocytes.

3.3.2 Multiple dynamic processes and [Ca2+] dependent early-stage resealing

It is observed in our experiments that rapid recovering of TMC occurred immediately after ultrasound exposure but became slower in a few second (Figure III-1), and that the recovery slowed down even further thereafter, suggesting the involvement of different processes in the recovery of the TMC. For the convenience of discussion, the resealing before 15 s is hereafter regarded as the early-stage recovery and the resealing afterwards, the late-stage recovery. The choice of 15 s was based on observation of the general time scales in the TMC recordings of cell recovery in the presence of 1.8 or 3 mM Ca2+.

The involvement of multiple dynamic processes in resealing can be illustrated by the example in Figure III-3, where the normalized recovering TMC after ultrasound was shown to be fitted successfully by a two exponential functions. Mathematically, the early-

65 stage (< 15 s) inward TMC after ultrasound I(t) (t > 1.2 s), can be expressed by a

superposition of two exponential functions with rate constants k f and ks (noted as fast

and slow process), respectively, I(t') = −I _15 − I f exp(−k f t') − I s exp(−kst') with the

associated amplitudes I f and I s for the fast and slow process. In the equation, t’=t-1.2s is

time after ultrasound application (for convenience, t’ is noted as t from here on), I _15 is

the TMC at 15 s after ultrasound application. A zero I _15 indicates complete recovery while a larger value correlates with weaker resealing. The maximal current change is

therefore represented by Imax = I _15 + I f + I s . A normalized expression of the inward

TMC is obtained by dividing the current

ˆ ˆ ˆ ˆ by Imax asI(t) = −I _15 − I f exp(−k f t) − I s exp(−kst ) .

ˆ 2+ In order to assess the early stage recovery, I _15 as a function of [Ca ] is shown in

ˆ Figure III-4, where the decreasing I _15 values clearly show inhibited recovery at lower

2+ ˆ 2+ [Ca ]. No I _15 for 0 Ca is included in the figure because the cells in calcium free solution exhibited no recovery and the TMC often became even larger and irreversible as

ˆ the cells degenerated gradually, making I _15 meaningless in this case. It is interesting to notice that the presence of even very low [Ca2+] (e.g. 0.02 mM) promoted some resealing which delayed membrane degeneration and rupture (Figure III-4).

ˆ ˆ ˆ The best least square fitting parameters ( k f , ks , I f , I s ) for I(t ) by the two exponential functions described above are shown in Figure III-5. No values at 0 Ca2+ are included because no recovery was observed for cells in Ca2+ free solution. Additionally,

66 ˆ 2+ neither ks or I s values are included for [Ca ] = 0.02 mM because the TMC for cells at

2+ [Ca ] = 0.02 mM can be fitted with one exponential function with rate constant k f , thus

inclusion of the process of ks was not necessary. As shown in Figures III-5A and III-5B,

−1 −1 both rate constants k f (0.79 ~ 1.19 s ) and ks (0.11 ~ 0.21 s ) show no definite trend

2+ with varying [Ca ]; values of ks are significantly lower than those of k f . Thus process

with rate constant ks is regarded as slow resealing and the process with rate constant k f the faster resealing process.

ˆ Figure III-5C shows that I f , the amplitude of the fast resealing process, increased with increasing [Ca2+], and was more than 3 times at 1.8 mM (~80%) than at 0.02 mM

(20–25%), indicating that the calcium- dependent fast resealing process is dominant at

2+ ˆ higher [Ca ]. The amplitude of the slow resealing process ( I s ) is relatively constant value but small (~ 20%) for all concentration values above 0.09 mM Ca2+ (Figure III.5D).

The slow resealing process appeared to have a higher threshold of initiation (0.09 mM) than the fast process, which was already present at 0.02 mM. These thresholds were lower than the threshold for the complete resealing (0.54 mM) necessary for the cell to survive (Figure III- 2).

3.3.3 Ca2+ dependent late -stage resealing

The continued recovery of the TMC after 15 s is regarded as a late-stage recovery, with a much slower rate constant and is also Ca2+ dependent. To better delineate the late stage

ˆ and other recovery processes, let I L (t) represent the recovering or decaying current after

67 ˆ ˆ ˆ ˆ ˆ ˆ ˆ 15 s, so I L (15) = I _15 . I L (t) can be expressed in two parts, I L (t) = I L '(t) + I0 , where I0

ˆ represents the non-recovered current at the end and I L '(t ) the dynamic part of the late- stage resealing. The dependence of the whole recovery process on extracellular [Ca2+] can be seen by the results in Figure III-6A, which shows the time duration required for the TMCs to reach various percentage of TMC recovery for 0.54 and 1.8 mM [Ca2+].

ˆ I L '(t) is seen in the recovery at later ( >15 s) time points (Figure III-6A). The time required for cells to complete up to 40% recovery showed no difference between 0.54 mM (2.68 ± 0.46 s) and 1.8 mM (2.46 ± 0.41 s), consistent with the observation of fast recovery before 15 s (Figure III-5). The time required for cells to complete more than

50% recovery of current showed increasing difference between 0.54 mM and 1.8 mM, with a considerably lower late-stage recovery rate at 0.54 mM than that at 1.8 mM. The time required for the cells to recovery 90% recovery at 0.54 mM calcium concentration

(98 ± 50 s, n=10) was about 6 times as long as that at 1.8 mM (17.6 ± 12.8 s, n = 20), although large variation existed among the cells.

Figure III-6B shows the components constitutes the maximum,

ˆ ˆ ˆ ˆ ˆ ˆ ˆ I max = I0 + I L '(∞) + I s + I f , including the early-stage fast ( I f ) and slow ( I s ) recovery,

ˆ and the late-stage slow recovery (I L '(∞ ) ). The infinity symbol is used to indicate the end of the measurement. It can be seen from Figure III-6B that the Ca2+ threshold for fast recovery is 0.02 mM and that the early-stage slow recovery and late-stage recovery had the same threshold at 0.09 mM. At sufficiently high [Ca2+] (e.g. 1.8 mM), robust early- stage fast and early-stage slow recovery achieve almost complete recovery, while the

68 cells in lower [Ca2+] (e.g. 0.54 mM) achieved complete recovery after a much longer time duration.

3.3.4 Effects of extracellular Mg2+ on resealing process

To test whether 1 mM [Mg2+], present in ND 96 solution, has antagonizing effect on cell membrane resealing in sonoporation, a set of experiments were conducted to measure sonoporation in oocytes in modified ND96 solutions without Mg2+ (0 Mg2+). Figure III-7

ˆ 2+ 2+ 2+ shows the comparison of I _15 for 0 [Mg ] and 1 mM [Mg ] and various [Ca ] (0.02 mM, 0.09 mM, and 0.54 mM). No statistically significant difference was observed between 0 [Mg2+] and 1mM [Mg2+] groups for the same [Ca2+], indicating no Mg2+ (1

2+ ˆ mM) antagonism in early-stage cell membrane resealing at these [Ca ]. I _15 shows

2+ ˆ 2+ similar dependence on [Ca ] as Figure III-4 with higher I _15 values at lower [Ca ]. The

ˆ ˆ other fitting parameters ( ks , k f , I s , I f ) from the two exponential function fit also showed no statistically significant difference between the 0 [Mg2+] and 1 mM [Mg2+] group (data not shown). There is also no statistically significant difference for the late-stage resealing for these groups (data not shown). All these suggest the absence of Mg2+ antagonism on the resealing in sonoporation.

3.3.5 The effects of TMC amplitude

The maximum TMC change measured in our experiment exhibited a range of values (0.1

~ 10 µA). The spread may have come from a number of sources including those associated with inherent cellular differences, variations from the micro-bubble concentration, size, and state, as well as the statistical nature of bubble activities in the

69 vicinity of the cell membrane. For example, the mean bubble concentrations in our experiment was (9.0 ± 3.2) ×106 / ml (n = 20) ranging from 6.1×106 / ml to

14.8×106 / ml . Under the experimental conditions used in this study, it is likely that multiple pores were generated. As the TMC has contributions from all pores, the variation in measured current amplitude suggests variation in pore size and number, in addition to the contribution from intrinsic cellular differences, which may have played a more important role than bubble concentration in the spread of measured currents.

However, this current study cannot deterministically conclude which of these factors dominated in the measurement variation.

In order to test whether membrane resealing depended on the maximum TMC resulted from sonoporation, the resealing characteristics are compared at different maximum TMC. Figure III-8A is a scatter plot in semi-logarithmic scale in horizontal axis showing the actual amplitude of the TMC (not normalized) at 15 s after ultrasound

exposure, I _15 , vs. the maximum TMC change ( Imax ) for 0.02 mM, 0.54mM, and 1.8 mM [Ca2+]. The data for each concentration was fitted by linear regression and plotted on

a semi-logarithmic scale in Imax . The excellent agreement of the measurements and linear

fitting curve demonstrates a linear relationship of I _15 with Imax , indicating that the

ˆ early-stage recovery percentage (i.e., normalized I _15 , Figure III-3) was not affected by

Imax , consistent with our earlier findings [4]. The late-stage recovery percentage was

slightly slower for larger Imax , as shown in Figure III-8B. When Imax was between 5 µA and 25 µA (0.75 ± 0.12, n=9), the recovery percentage was slightly lower than that for

cells with Imax below 5 µA (0.88 ± 14, n=28) (p < 0.007).

70 3.4 Discussion

The results presented here demonstrate quantitatively for the first time that extracellular

Ca2+ was required for reversible sonoporation and the resealing of pores involved two or more distinctive processes, all affected by extracellular [Ca2+]. This conclusion is made based on the following findings from the real time recordings of TMC of single Xenopus oocytes under voltage clamp: 1) no recovery of TMC was detected and incomplete recovery led to cell death in Ca2+ free solution; 2) complete recovery of TMC and cell survival required extracellular [Ca2+] > 0.54 mM and the recovery rate increased with increasing [Ca2+]; 3) the recovery dynamics exhibited several different [Ca2+]-dependent cellular processes during the course of recovery. Descriptions of these important aspects of the sonoporation mechanism are novel and are made possible by the unique ability of voltage clamp techniques to monitor in real time the dynamic change in the membrane permeability in sonoporation at the single cell level. These findings are significant in providing molecular insight to understand sonoporation mechanism and outcome, given the involvement of Ca2+ as a second messenger which plays an important role in both intracellular and intercellular signaling in various cellular processes.

Our results quantitatively characterize the involvement of different Ca2+- dependent processes in sonoporation recovery. First, the rate constant for the fast recovery process did not appear to be affected by [Ca2+] (Figure III-5A) but its amplitude increased significantly with increasing [Ca2+] (Figure III-5C), suggesting a cellular process with enhanced recruitment by higher Ca2+ concentrations. This process is observed in very low extracellular [Ca2+] (0.02 mM) (Figures III-5A and III-5C), and possibly even in 0 Ca2+ (Figure III-1), suggesting that the process may happen at the level

71 of intracellular [Ca2+] (0-20 µM). On the other hand, the late-stage slow recovery of current exhibited an increased rate constant with increasing [Ca2+] (Figure III-6A) but the amplitudes of the late-stage slow process and early-stage slow process did not change with [Ca2+] (Figure III-5D and Figure III-6B), suggesting a cellular process regulated by high concentration of Ca2+ (>0.09 mM). Although the rate constant of the slow process

2+ ks in the early-stage did not show a clear Ca dependence, this could be due to the contamination of the fast process kf in the fitting. The slow process may not be properly expressed in the presence of a dominant fast process. It is also possible that the early- stage slow process is a precursor of the late-stage slow process as they both have the threshold of 0.09 mM (Figure III-6B). It should be noted that while the choice of 15 s was somewhat arbitrary, it is a proper time scale representing well the experimental observations that complete recovery occurs by the time at 1.8 mM Ca2+ and the rate constant for late stage recovery process is more than an order of magnitude lower than the slow process before 15 s (Figure III-6A). The distinctive characteristics of the different processes have not been described in previous studies using less sensitive and static measurement techniques.

The absolute requirement of extracellular Ca2+ in sonoporation resealing and cell recovery is in contrast with the conventional notion regarding pore resealing in electroporation. Electroporation [72-75] is a technique that uses electric pulses to create pores in the cell membrane for delivery of desirable agents (e.g. genes or biologically active molecules) into cells. It is believed that pores created in electroporation are capable of self-resealing in the absence of calcium, driven by the tendency of the hydrophobic

72 end of the lipid molecules to move from the energy unfavored membrane state with pores

[73] to the lower energy resealed state.

On the other hand, work over the past two decades in the field of cell membrane disruption and repair has asserted that disruptions or pores in nucleated cell plasma membrane generally do not repair themselves spontaneously, but rather heal via an active, calcium-dependent process which most likely involves two functional components

[76-79]: plasma membrane resealing by fusion of intracellular membrane compartments with each other and with the plasma membrane at the damage site, and cyctoskeletal reorganization. As the first step in cell repair, plasma membrane resealing has been shown to require extracellular calcium and has been temporally, spatially, and functionally related with rapid homotypic and heterotypic membrane fusion events associated with a calcium-regulated exocytosis process [78, 80].

The resealing of plasma membranes is demonstrated necessary for cell survival in experiments where material is delivered to the cytoplasm by microinjection, chemical permeabilization, electroporation, and sonoporation. The process may be the same as that involved in the cell repair of the plasma membrane disruption occurring during normal physiological functions (cardiac contraction) and injury. The pores generated in electroporation reseal themselves in the absence of extracellular Ca2+ probably only because of the extremely small size of the disruptions observed, which usually are below

1 µm [81-83] and mostly in the range of 20–120 nm [19]. In such cases, resealing of the small holes, besides being driven by the energetically unfavourable situation, may also involve a process similar to the fast process observed in this study, which may bring complete resealing at very low extracellular [Ca2+] when the disruption is extremely

73 small. Furthermore, it is possible that cells with larger disruptions did not survive in the absence of Ca2+ in electroporation, which is often accompanied by high cell death rate.

Thus larger disruptions were not easily detected in post-electroporation assays.

The [Ca2+] threshold for cell membrane repair, below which repair fails to take place, has been recognized to vary by cell type and cell state. Recent work using sea urchin eggs and cultured cells has shown that plasma membrane repair requires [Ca2+] above 0.3 mM [78, 84]. Successful resealing required 10 to 30 s, but sometimes took as long as 90 to 120 s when the calcium concentrations were low, with a lower resealing rate at a lower level of [Ca2+] [85]. Membrane resealing durations in the range of 30 ~ 240 s for mammalian cells using fluorescent reporter molecules [77, 78] have also been reported.

These results are in general agreement with our results for sonoporation resealing.

However, no Mg2+ antagonism was observed in our study of sonoporation resealing for 1 mM [Mg2+] in the solution, in contrast to the studies of cell wounding using a laser and glass pipet tip in sea urchin eggs and cultured 3T3 cells, where much larger membrane disruptions were studied [76]. It is important to note that the measured threshold value of

[Ca2+] for resealing depends on cell type and stage, and possibly the means and extent of disruption, and it is also likely to be affected by the methods used to assess the resealing.

Previous studies typically utilized fluorescence techniques such as monitoring of dye loss to gauge membrane resealing, which usually has less sensitivity, accuracy, and temporal resolution. Further studies are needed to investigate the reasons for these differences and the roles that the other factors described above may play in sonoporation resealing.

74 To complete the cellular repair, the cells must also restore the functionality of the damage site, which involving the cortical cytoskeleton in addition to membrane healing

[79]. Study of wounding in amphibian eggs demonstrated that cytoskeleton repair response occurred 30–60 s after wounding and involves accumulation of F-actin and myosin-II around the wound site, which then coalesce into a circular array that contracts inward until the damage is covered. Importantly, the removal of extracellular calcium blocked healing and prevented recruitment of F-actin and myosin-II [86]. Studies in cultured fibroblast cells [87] showed similarly that membrane disruption stimulates cytoskeletal response that is Ca2+-dependent, including disassembly of microtubules

(MTs) around the wounding site, recruitment of end binding protein (EB1) to the MTs especially around the wound site, and subsequent elongation of MTs towards the wounding site. Such cytoskeletal response was found to require calcium at concentrations above 0.4 mM and has a higher rate at higher levels of [Ca2+]. Furthermore, a time lapse of about 15–20 s is necessary for MT reassembly to occur after membrane disruption and can last for 60–140 s [88, 89]. The range of the delays for these cytoskeletal responses to complete cell repair is comparable with the time scale of late-stage resealing of sonoporation observed in our experiment. It should be noted that it is unclear at this point whether the functional restoration of membrane can be precisely derived from the TMC measurement, although such functional integrity does require normal ability of iron exchange through the membrane.

The fact that cells require Ca2+ to reseal suggests that a molecular mechanism of extracellular Ca2+ is involved in regulating membrane resealing in sonoporation. The influx of Ca2+ might signal cell injury and trigger active mechanisms similar to those

75 involved in cell repair of membrane injuries; such process rapidly seal membranes and ensure cell survival in sonoporation. The in-rush of Ca2+ may also play role in signal transduction linked to downstream effects of sonoporation such as apoptosis[52] and calcium oscillation and calcium waves[57].

Our results demonstrate for the first time that Ca2+ is required for membrane resealing and a threshold exists for complete resealing. Our finding establishes a link between sonoporation and the almost omnipresent phenomenon of highly regulated cellular wound healing, indicating important implications. First of all, the effects of

[Ca2+] on the membrane resealing dynamics could be utilized practically by manipulating

Ca2+ (locally and transiently) to achieve a controlled delivery outcome (e.g. slower resealing of pores may result in higher intracellular delivery efficiency). Secondly, a diverse range of diseases result from membrane repair defects which are associated with

Ca2+ related abnormal cellular/genetic regulations or processes, for example, muscular dystrophy related cardiomyopathy [33], which is a life threatening disease. Patients with membrane repair defects related diseases may need to be carefully evaluated before being considered as candidates for sonoporation delivery treatment to avoid causing unwanted side effects. Our study reveals insight into sonoporation mechanism, which is the key to successfully bring the technology into human application ultimately as an efficient and safe delivery strategy.

The findings from studies of sonoporation resealing may shed new light on the study of cellular wound healing, which is of considerable intrinsic interest. As a rapid emergency response, resealing is the process that cells need to repair membrane disruptions from physical or chemical insults and therapeutic interventions. It is also a

76 normal cellular activity that commonly occurs as a natural consequence of tissue function. Failure to reseal leads to rapid cell death from loss of cytoplasm and consequential unabated calcium influx, resulting in a direct loss of tissue integrity.

Negative consequences can also ensue indirectly, from release of proteases that attack neighboring cells or even the provocation of an inflammatory response [79], which are relevant in the assessing the safety of sonoporation applications.

3.5 Conclusion and future work

Electrophysiology techniques, such as the voltage clamp method, capable of measuring the movement of ions through the cell membrane to gauge membrane porosity, provide a unique and sensitive means to investigate the sonoporation process at the single cell level.

In this study, the time-resolved TMC of single Xenopus oocytes under voltage clamp were used to investigate the effects of extracellular Ca2+ on pore resealing in sonoporation. Our results demonstrated that Ca2+ plays an important role in regulating membrane resealing in sonoporation. Pore resealing exhibited multiple Ca2+ dependent processes with different characteristics and rate constants, including an early-stage fast recovery, followed by a slower, late-stage recovery. Complete resealing required extracellular [Ca2+] above 0.54 mM and occurred 10–30s after sonoporation inception at

1.8mM [Ca2+], but took up to 130 s at lower [Ca2+]. These results reveal new aspects of the sonoporation mechanism and can provide the necessary knowledge basis for designing safe and efficient sonoporation techniques by tuning the extracellular [Ca2+] to control the outcome in ultrasound gene/drug delivery applications.

Since extracellular Ca2+ plays a significant role in pore resealing process in sonoporation, similar experiments are planned on other cell models to confirm this

77 observation. Meanwhile, experiment will be designed to adjust the extracellular [Ca2+] in order to control the existing time of formed pore(s) in membrane for better delivery outcome.

Figure III-1: Time-dependent inward TMC recorded in oocytes under voltage clamp exposed to tone burst ultrasound (0.2 s duration, 0.3 MPa pressure amplitude) in the presence of 1% Definity® at different extracellular [Ca2+]. The holding potential was − 50 mV. Each current curve is normalized by the absolute value of its maximum which occurred at the end of ultrasound tone burst (1.2 s) [90].

Figure III-3: Multiple dynamic processes in sonoporation resealing. An example of the normalized TMC measured at [Ca2+] = 1.8 mM was shown to fit well by a superposition of two exponential functions with different rate constants 0.094 and 0.67 respectively [90].

Figure III-4: Normalized TMC (at 15 s) recorded in individual oocytes in the presence of different [Ca2+]. The boxes at each concentration show 25th, 50th and 75th percentiles, respectively, cross symbols represent the mean value, the whiskers show the standard deviation, and the two dashes show the maximum and minimum values in the data in this and subsequent figures. The experiments with [Ca2+] = 0, 0.02, 0.09, 0.36, 0.54, 1.08, 1.8, and 3.0 mM had n = 12, 14, 13, 19, 19, 12, 21, and 11, respectively [90].

Figure III-5: Fitting parameters describing the early-stage dynamic recovery process of TMC as a function of extracellular [Ca2+]. A) Rate constant of fast recovery process vs. extracellular [Ca2+]. B) Rate constant of slow recovery process vs. extracellular [Ca2+]. C) Normalized amplitude of fast recovery process by the absolute value of maximum current change vs. extracellular [Ca2+]. D) Normalized amplitude of slow recovery process by the absolute value of maximum current change vs. extracellular [Ca2+]. The respective sample sizes for [Ca2+] = 0.02, 0.09, 0.36, 0.54, 1.08,1.80, and 3.0 mM are n = 14, 13, 19, 19, 12, 21, and 11 [90].

Figure III-6: Rate of TMC recovery in sonoporation as a function of extracellular [Ca2+]. A) Mean value of the time duration (and the standard error of the mean value) required for cell to reach current recovery at 0.54 mM and 1.8 mM Ca2+. B) Stack column plot of the mean values for the early-stage fast (Îf) and slow resealing processes (Îs), and late-stage slow resealing process (ÎL). The respective sample sizes for [Ca2+] = 0.02, 0.09, 0.36, 0.54, 1.08, 1.80, and 3.0 mM are n = 14, 13, 19, 19, 12, 21, and 11, respectively [90].

Figure III-7: Normalized TMC at 15 s as a function of [Mg2+] and [Ca2+]. The TMCs are shown for cells in the presence 0 or 1 mM [Mg2+] and 0.02, 0.09, and 0.09 mM [Ca2+]. The sampling sizes are n = 14, 18, 13, 15, 10 and 21 for the groups from left to right [90].

Figure III-8: Early- and late-stage recovery as a function of maximum TMC. A) Scatter plots and linear regression fits show I_15 vs. Imax plotted in logarithmic scale at 0.02 (n = 14), 0.54 (n = 10), and 1.8 mM [Ca2+] (n = 47). B) Scatter plot shows transmembrane recovery percentage in late-stage (47 s) vs. Imax at [Ca2+] = 1.8 mM (n = 37) [90].

CHAPTER IV

DYNAMICS OF SONOPORATION CORRELATED WITH ACOUSTIC

CAVITATION ACTIVITIES

Early studies on sonoporation mainly rely on post ultrasound assays which lack temporal and spatial specificity, inevitably leading to uncertainty in relating actual sonoporation parameters with outcome, given the complexity of ultrasound interaction with cells and bubbles. As such, though the sonoporation is hypothesized to be resulted from cavitation activities, the exact relationship between cavitation and sonoporation has not been obtained.

4.1 Introduction

This study investigated the impact of microbubble cavitation on the cell membrane by measuring in real time the co-localized and con-current cavitation activities and sonoporation in a single cell. Using Xenopus oocytes as a model system [4] and a focused ultrasound strategy, localized cavitation and sonoporation in a single cell were only generated and detected within the ultrasound focus (Figure IV-1). Sonoporation was measured, in real time, via the inward TMC current of the single Xenopus oocyte under voltage clamp, as demonstrated in our previous study [4]. Before ultrasound application, the TMC current is constant at a fixed membrane holding potential (-50mv, voltage clamped) in the absence of activation of endogenous ion channels, since the whole cell membrane is regarded as a resistor with constant resistance [6]. In sonoporation, the

87 ultrasound generated non-specific pores on the cell membrane decreases the membrane resistance; the ions flowing through the pores results in changes in the TMC current, which is determined by the pore size and ion concentration gradient across the cell membrane, therefore providing a novel means to monitor the dynamics of sonoporation in a single cell with high temporal resolution and sensitivity.

4.2 Methods

4.2.1 Animal protocol and cell preparation

Xenopus oocyte, the single egg cell from Xenopus laevis is used in this study. The animal protocol is approved by our Instititional Animal Case and Use Committee. The preparation procedure for defolliculated oocyte cell is the same as described in 2.2.3.

4.2.2 Experimental setup

Figure IV-1 shows the experimental setup. A single oocyte (diameter 1.0~1.2 mm) placed on a 2 mm thick acoustic gel block (Parker Laboratories, Fairfield, NJ) was immersed in

4 mL ND96 solution in a 35 mm Petri dish with a round hole at the bottom (φ=1mm)

(MatTek corp., Ashland, MA). The gel block (Aquaflex®, Parker Lab, Inc., Fairfield, NJ) with acoustic impedance similar to water and the solution, created a standoff distance to acoustically separate the dish bottom from the cell without disturbing the ultrasound fields.

A dual-frequency ultrasound transducer assembly, including two concentric ultrasound transducers confocally- and collinearly-aligned, was utilized in this experimental study. A 40 μm diameter calibrated needle hydrophone (HPM04/1,

Precision Acoustics, Dorchester, Dorset, UK) was used to measure the transducers’

88 beamprofile as shown in Figure IV-2. The donut-shaped, outer transducer (excitation transducer) (inner and outer diameters 14 and 30 mm), was used to generate focused ultrasound beam at 1.5-MHz (focal distance 48 mm, full 3-dB lateral beam width 0.9 mm) to induce cavitation and sonoporation. The circular, center transducer (detection transducer) (diameter 14 mm) is a broadband ultrasound transducer ( Figure IV-2, center frequency 7 MHz, 50% bandwidth, focal distance 48 mm, 3-dB beam width 0.45 mm). It was used to detect acoustic signals from the overlapped focal zone. The transducer assembly was immersed in a water tank aiming upward with its center of focus positioned at the equator of the cell (inset in Figure IV-1). Localization of cavitation and sonoporation was achieved at the intersection of the cell membrane with the ultrasound focus.

4.2.3 Data acquisition procedure and processing algorithms

The acoustic signal from microbubble and cavitation was acquired using a digital oscilloscope (54830B Infiniium Oscilloscope, Agilent Inc, Santa Clara, CA) at a sampling rate of 50 MS/s in conjunction with two function/Arbitrary waveform generators (33250A, Agilent Inc, Santa Clara, CA). A segmented memory acquisition mode was used for pulsed data which containing consecutive trigger events.

Four data acquisition procedures (Active/Passive, two Passive only, and

Ramping) were used in this study. Each of procedures was focusing on specific acoustic signals and experimental purpose as listed in detail in Table IV-1. The time sequence of detection transducer and excitation transducer were shown in Figure IV-3a. Coupled to an ultrasound pulser/receiver (Panametrics NDT, 5910R, Waltham, MA), the detection transducer was operated in pulse-echo mode for Active/Passive procedure and through 89 transmission mode for Passive only and Ramping procedure with a pulse repetition frequency (PRF) of 5.88 kHz. The target acoustic signals were the backscatter (BS) signals from existing bubbles (active cavitation detection) around oocyte and/or the acoustic-emission (AE) signals from collapsing bubbles (passive cavitation detection) within the focal zone. A 17 µs delay was used for every pulse-echo period between the detection pulse to the excitation pulse in order to separate the BS signals from the AE signals received by the detection transducer. Each excitation pulse included 5 cycles of oscillating acoustic pressure at 2.09 MPa (non-ramping procedure) or 0.16-3.74 MPa

(ramping procedure); the short pulse duration avoided the buildup of a standing wave and the effects of multiple reflections inside the dish. The amplitude of excitation pulse in ramping procedure was shown in Figure IV-3b. A total of 120 pulses were applied with

12 linearly distributed amplitude levels. Each level was repeated for 10 times.

The raw data was transferred to a computer for analysis using MATLAB

(MathWorks, Inc., Natick, MA). For active cavitation detection, root mean square

(RMS), a statistical measurement of the magnitude of a varying quantity, was calculated from BS signal to disclose any dynamics change on the existing or the number of microbubbles around oocyte during ultrasound exposure. For passive cavitation detection, Fast Fourier Transform (FFT) was used to estimate the spectrum of AE signal.

A hamming window was used to minimize spectral leakage for AE signal. RMS of the spectrum of AE signal between 3 to 11 MHz was also calculated for dynamic information during ultrasound exposure.

The methods of data acquisition and processing for TMC were described early in

2.2.2.

90 4.3 Results

Figure IV-4A and IV-4B show examples of the acoustic signals (in arbitrary unit) received by the detection transducer with and without Definity® (1.2E+7 bubbles/ml) in the solution. The horizontal axis represents the time duration for the ultrasound pulses to travel from the transducer to the scattering targets and back to the transducer at sound speed 1480 m/s (in water), corresponding to the spatial location of an acoustic source. As the cell was placed near the ultrasound focus (48 mm, equivalent to 64.9 µs, the time it takes for an ultrasound pulse propagates from the transducer to a target at the focus and then back to the transducer), the signal segments indicated in the plots correspond to the

BS signals (65-67 µs) and the AE signals (82-86.5 µs) from bubbles within the intersection zone of the ultrasound focus with the cell. The AE signals arrived 17 µs after the BS signals because of the delay of each excitation pulse from the detection pulse. The

AE signals also lasted longer because the 1.5 MHz excitation pulse has a longer duration than the detection pulse (7 MHz). The echoes (58 µs & 73 µs) are reflections from the dish bottom and the solution-air interface.

As the excitation ultrasound was on from 34-204 ms, AE signals were not present at 0 ms (blue curve), but showed marked increase at the start of ultrasound application

(red curve at 34 ms) before returning to noise level later (black curve at 221 ms).

Destruction of bubbles by the excitation pulses is clearly seen via the characteristic broadband spectrum of AE signals (Figure IV-4C and IV-4E), in contrast to the cases when the same ultrasound exposures were applied without bubbles in the solution (Figure

IV-4B and IV-4D), or when no excitation pulses were applied with bubbles present (data not shown). The dynamic evolution of the broadband AE (Figure IV-4E) shows that

91 cavitation lasted for only 15 ms after the ultrasound activation as no AE signals were detected beyond 49 ms even though more excitation pulses were applied, indicating rapid and complete destruction of bubbles within the ultrasound focal zone.

Correspondingly, the inward TMC current (recorded at sampling rate of 20 kHz) of the Xenopus oocyte under voltage clamp (at - 50 mV holding potential) exhibited a rapid increase correlated with the increase of AE signals, as shown by the spectral RMS calculated from 3-11 MHz (Figure IV-4F). Similarly, the TMC current also showed no further increase beyond 49 ms (Figure IV-4E and IV-4F). The change in the BS signals is less pronounced visually, nevertheless is evident from the change of root-mean-square

(RMS) of the signals (Figure IV-4F). Furthermore, the decrease of BS RMS after the application of the excitation ultrasound (middle plot, Figure IV-4F) has a similar time course as the AE signals, correlating to the decreased number of scattering bubbles in the focal zone due to bubble destruction by the excitation pulses. The BS RMS shows an initial decrease, as the radiation force of the detection pulses pushed bubbles out of the focus.

The correlation of sonoporation with cavitation is further demonstrated by the results shown in Figure IV-5, when the amplitude of the excitation ultrasound was linearly ramped from 0 to 3.74 MPa in a period of 21 ms. In these experiments, only the

AE signals from the collapsing bubbles in the focal zone were detected using passive cavitation detection (no detection pulses were used). Increase of the AE spectral RMS (3-

11 MHz) (Figure IV-5B), correlates well with the increased amplitude of the inward

TMC current (Figure IV-5C). The initial change of both AE and TMC current occurred at

~ 41 ms or 7 ms (at acoustic pressure 1.33 MPa) after ultrasound application (at 34ms),

92 reaching maximum around 47 ms or 13 ms after ultrasound activation (at 2.26 MPa), in contrast with the immediate increase when constant-amplitude excitation pulses were applied (Figure IV-5D, also Figure IV-4F). The delay was due to the low acoustic pressure amplitude early in the ramp. Furthermore, the amplitude of the TMC current correlated with AE spectral RMS values, both higher in the ramping exposure than in the constant-amplitude exposure (Figures IV-5C and IV-5D). These acoustic pressures are higher than reported values in sonoporation experiments in non-focal, larger volumes, but this may be related to the scarcity of cavitation and sonoporation events in a small focal volume.

Figure IV-6 is another demonstration of the correlation of sonoporation with cavitation, where excitation ultrasound is a 0.2s tone burst with a pressure at 0.4 MPa.

Only the AE signals from the collapsing bubbles in the focal sone were detected by passive cavitation detection. Since excitation pulses are continuous on, the signal noise ratio of detected result is less than early experiment. The increasing of the AE spectral

RMS (3-11 MHz) correlates well with the decreasing of TMC at multiple points within excitation ultrasound exposure (Figure IV-6A). The spectrum of the AE signal at

1041ms, 1100ms and 1201ms are plotted which correspond to a high cavitation level moment with ultrasound exposure, a low cavitation level moment with ultrasound exposure and a non cavitation moment without ultrasound exposure, respectively (Figure

IV-6B).

4.4 Discussion

The results from this study demonstrate quantitatively that membrane opening in sonoporation is always corresponding to a peak cavitation activity temporarily and

93 spatially. Multiple experiments (Figure IV-4F, Figure IV-5C, Figure IV-5D and Figure

IV-6A) show clear correlations on cavitation activities and the increase of TMC increasing with a time resolution about 1 ms. The spatial resolution is increased by using focus transducer, comparing to early study in which a planar transducer with a relative long exposure (0.2s) is used.

4.5 Conclusion and future work

The time-resolved measurements of TMC current and AE signals using confocally aligned ultrasound transducers demonstrate for the first time the spatiotemporal correlation of sonoporation with cavitation. Since the detected acoustic signals came from the collapsing bubbles within the overlapped ultrasound focal zone of the detection and excitation transducers, the precision of the correlation is limited by the small but finite volume of the intersection of the ultrasound focus with the cell membrane (inset in Figure

IV-1).

A future study need to do is to quantitatively correlate the acoustic emission signal and TMC measurements, which requires an ultrasound field with better spatial resolution. Plus high speed camera possibly combined to optical tweezer or microfluid device to manipulate microbubble, experiment result of bubble behavior and pore dynamics will give us more clues on sonoporation.

Detection transducer, lateral @focus, 2mm width, 0.2mm step Excitation transducer, lateral @focus, 2mm width, 0.2mm step 0 0 A -1 B -1 -5 -0.5 -0.5 -5 -10 0 0 -15 -10 distance (mm) 0.5 distance (mm) 0.5 -20

1 -25 1 -15 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 distance (mm) distance (mm) Detection transducer, axial, 40,60mm, 0.5mm step Excitation transducer, axial, 40,60mm, 0.5mm step 0 0 C D

-2 -2

-4 -4

-6 -6 hydrophone reading (dB) hydrophone reading (dB) hydrophone reading -8 -8 40 45 50 55 60 40 45 50 55 60 axial distance (mm) distance (mm)

Table IV-1: Data acquisition procedure in this study

Name Purpose All parameters Active/Passive Correlation between Data size: 4096pts by 1400 lines sonoporation result and Sampling rate: 50 Ms/s acoustic signal (BS, AE) at PRF: 5.88 KHz ( 170µs PRT) fixed acoustic pressure Excitation pulse duration: 170 ms (1000 pulses) Excitation pulse width: 5 cycles (3.33µs) Excitation pulse pressure: 2.09 MPa Excitation pulse delay: 17µs Excitation pulse begin: 34 ms (the 201st pulse) Excitation pulse end: 204 ms (the 1200th pulse) Pulser/Receiver mode: Pulse-Echo Passive only 1 Correlation between Data size: 4096pts by 1400 lines sonoporation result and Sampling rate: 50 Ms/s acoustic signal (AE only) PRF: 5.88 KHz ( 170µs PRT) at fixed acoustic pressure Excitation pulse duration: 170 ms (1000 pulses) Excitation pulse width: 5 cycles (3.33µs) Excitation pulse pressure: 2.09 MPa Excitation pulse delay: 17µs Excitation pulse begin: 34 ms (the 201st pulse) Excitation pulse end: 204 ms (the 1200th pulse) Pulser/Receiver mode: Through transmission Passive only 2 Correlation between Data size: 40000pts by 205 lines sonoporation result and Sampling rate: 40 Ms/s acoustic signal (AE only) 0.205 second continuous tone burst at fixed acoustic pressure Excitation pulse pressure: ? MPa Excitation pulse begin: 1000 ms Excitation pulse end: 1200 ms Pulser/Receiver mode: Through transmission Ramping Correlation between Data size: 8500pts by 240 lines sonoporation result and Sampling rate: 50 Ms/s acoustic signal (AE only) PRF: 5.88 KHz ( 170µs PRT) at various acoustic Excitation pulse duration: 170 ms (1000 pulses) pressure Excitation pulse width: 5 cycles (3.33µs) Excitation pulse pressure: 0.31-3.74 MPa Excitation pulse delay: 17µs Excitation pulse begin: 0 ms (1st pulse) Excitation pulse end: 20.4 ms (the 120th pulse) Pulser/Receiver mode: Through transmission

Figure IV-3: (A) All time sequences of detection and excitation transducer used in this study. (B) The driving signal for excitation transducer in ramping procedure.

Figure IV-5: (A, B) Examples of the AE signals and power spectra in passive detection. (C) Change in TMC current corresponds with delayed AE increase when ramping ultrasound pulses were applied. (D) Change in TMC current occurs immediately after excitation ultrasound with constant amplitude [91].

100

CHAPTER V

TRANSIENT PORE SIZE ESTIMATION IN SONOPORATION

The pore size in sonoporation determines the size of molecules or agents to be delivered.

Therefore, the transient pore size is critical for delivery outcome. Post ultrasound assays such as AFM, TEM or SEM studies are sometime used to obtain information of the pore size. However, such analysis techniques provide information that may not be relevant and usually takes time and much effort. Also, the conclusions from those results are base on individual setup and no general understanding can be obtained. From the amplitude of

TMC of cells under voltage clamp during sonoporation can potentially provide relevant information of the pore area through the ion flow through the membrane. The study described in this chapter focuses on the determination of single pore size from the measured TMC.

5.1 Introduction

Since whole cell clamp was employed in our study, the TMC is related to the total area of the pores on the cell membrane. Even if the pores behave similarly, it is still a question whether these pores have a uniform size or not. The distribution of the number of pores generated in sonoporation is unknown either. In this study, two experiemtns were done to estimate the pore information in sonoporation: measurement of TMC in the presence of a single pore, and the number of pores is estimated from the TMC containing multiple pores.

101 5.2 Methods

Pore information (TMC of single pore and the number of pores in TMC of multiple pores) was estimated from two experimental designs: single pore experiments and experiments where multiple pores were present. A diffusion model was derived for our experimental conditions to relate the TMC with pore size.

5.2.1 Experiments where a single pore is generated

As previously described, microbubbles were used as cavitation nuclei to facilitate sonoporation [4]. These microbubbles must be present within a certain range away from the cell membrane to be effective to induce transient pores on the membrane through cavitation [92]. Lower bubble concentration will result in fewer effective bubbles in the vicinity of the membrane. A single pore can be expected if the microbubble concentration is sufficiently low. We used bubble concentration of 6E+3 bubbles/ml in the experiment.

We calculated the probability of having single pore from experiment observation, described below.

The following assumptions were used to ensure single pore event in our experiment: 1) The UCA bubbles (i.e. Definity®) are homogeneously distributed in the bulk solution, 2) The UCA is only effective to induce sonoporation within certain range from the cell membrane, 3) One microbubble generates one pore, 4) The number of microbubbles within a given volume surrounding the cell follows Poisson distribution, and 5) The diameter of Xenopus oocyte is 1.1mm. Under these assumptions, the probability of having N pores (or N bubbles) is described by a Poisson distribution:

f (N;λ) = λN e −λ / N!, [5-1]

102 where λ is the mean of the number of pore, N. Therefore the probability of having no pore isf (0;λ ) , and the probability of having one pore is

P1 = f (1;λ) . [5-2]

Thus the probability of having more than 1 pore(s) (N>=1) is

P = 1− f (0;λ) pore . [5-3]

This probability (equation 5-3) can be obtained from experimental observation as the probability of observing any current change in the experiment. Therefore λ can be estimated from this probability (Ppore, equation 5-3). The calculated λ can then be used to

calculate the conditional probability P1, pore , i.e., the probability of any poration is caused by a single pore:

P1, pore = P1 Ppore = f (1; λ) /(1− f (0; λ)). [5-4]

In a set of experimental observations when the probability of observed current change Ppore is much less than 1, the likelihood of any observed current change being the

result of a single is large. Therefore a large P1, pore (e.g. >95%) indicates that a single pore is generated when a current change is observed with a significance level less than 0.05

(orα < 0.05) .

The sample size for single pore size estimation can be calculated from equation 2-

1. Assume error between population mean and sample mean is 30%, standard deviation of TMC is 80% of the mean current value, the sample size is then calculated to be m=28.

Considering central limit theorem, we use sample size m=30.

103 Therefore, estimation of the current values for a single pore can be obtained from experiments (n=30). The statistics of single pore can be obtained from the distribution of the single pore current.

5.2.2 Experiment where multiple pores are generated

Under the assumptions that the number of bubbles follows Poisson distribution and the

TMC of single pore follows the distribution obtained in the experiment in 5.2.1, we can find the mean number of pores λ by fitting the simulated current distribution with the

TMC distribution in the experiment. The simulated current distribution will be the summation of the contribution from all pores. Calculation of λ is obtained based on least square fit with experimental measurements of maximum current amplitudes (histogram of maximum current). Therefore, by fitting the simulated data with the TMC distribution from the experiment where multiple pores were generated, the best least square fit

λ which represents the mean number of pores in the multiple pores experiment.

5.2.3 Electro-Diffusion model to estimate pore size from transmembrane current

TMC change recorded from voltage clamp system during sonoporation process is a result of ion movement through non-specific pore(s) driven by electrical force and ion concentration gradient across the membrane. An electro-diffusion model specific for our experimental condition is developed to estimate the pore size from measured TMC. This study only focuses on a static model for the estimation of the pore size correlated to maximum current amplitude (maximum pore size).

104 Ions involved in our experiments include Na+, K+, Ca2+, Mg2+, Cl- and other organic molecules. Only Na+, K+, and Cl- are included in our model due to their significant contribution to TMC.

During voltage clamp experiment, Vc(t) (command potential) is fixed to -50mv.

The circuit model for the cell membrane is shown in Figure V-1. The current due to each ion can be considered either voltage source or current source and is determined by both electrical field and ion concentration gradient.

The equation for ionic current related to pore size and ion flux is described by

2 I = (∑ J Ci zi F + ∑ J Ei zi F) ⋅ S = (∑ J Ci zi F + ∑ J Ei zi F) ⋅π ⋅ r [5-5] i i i i where I is TMC which is measured using voltage clamp technique, JCi is ion flux of a particular ion due to its concentration gradient through membrane, JEi is ion flux of the ion due to electrical field through membrane, z is the valency of the ion, F is the Faraday constant, S is pore area and r is the radius of the pore.

We need to obtain JCi and JEi for each ion considered in our model. From Fick’s law, in one dimension condition, the ion fluxion due to ion concentration gradient is calculated by

C − C J = −D ⋅∇C = −D ⋅ iout iin , [5-6] Ci i i i h where Di is the diffusion coefficient of a certain ion, Ciin and Ciout are concentrations inside and outside the plasma membrane of this ion, h is the thickness of the plasma membrane.

105 The ion fluxion due to electrical field is computed from

zF zF φ J = −D C ∇φ = −D C ⋅ , [5-7] Ei i RT i i RT i h where Ci is the ion concentration within the plasma membrane, z is the valency of the ion, F is the Faraday constant, R is the gas constant (8.314 joule/mole K), T is the absolute temperature, φ is the membrane potential.

All parameters involved in equation 5-6 and 5-7 are listed below:

CNao=96mM; (ND96), CNai=4-23mM (use 13.5mM); [93, 94], CKo=2mM; (ND96)

CKi=76-148mM (use 112mM); [93, 95] CClo=100mM; (ND96)

CCli=24-62mM (use 43mM); [93, 95] DNa=1.33e-5cm2/s; [96]

Dk=1.96e-5cm2/s; [96] DCl=2.03e-5cm2/s; [96] h=5nm; [97]

F=9.65e4 C/mol; R=8.314 J K-1Mol-1.

Calculation of the ion flux due to ion concentration for Na, K and Cl are 2.19E-2,

-4.31E-2, 2.31E-2, respectively. The unit is molm-2s-1, position sign means from outside to inside. The ion flux due to electrical field for Na, K and Cl are 5.04E-2, 8.67E-2, -

8.01E-2, molm-2s-1.

The radius of pore can then be derived from equation 5-5 and is shown in Figure

V-2:

I I r = = , [5-8] (∑ J Ci zi F + ∑ J Ei zi F) ⋅π 1.668E + 4 ⋅π i i

106 (Note: The definition of direction of current recorded in voltage clamp: When the membrane voltage was clamped, a negative current means the current is going through the current injection electrode. Therefore, the cell has a positive current at this clamp voltage, meaning positively charged ions moving into the cell. Overall, a negative recorded current represents a positive current go from outside to inside, and vice versa

[98] ).

5.2.4 Experimental materials and experimental setup

The same animal protocol approved by our Institutional Animal Care and Use Committee was used in this study. The preparation procedure for obtaining defolliculated oocyte cell is the same as described in 2.2.3.

The same experimental setup in 2.2.1 will be used. Briefly, A 35 mm polystyrene

BD Falcon™ bacteriological Petri dish (Fisher Scientific, Pittsburgh, PA) will be used to house a single Xenopus oocyte (diameter 1.0 ~1.2 mm) containing 4 mL ND96 solution with Definity™ (Bristol-Myers Squibb Medical Imaging, N. Billerica, MA) concentration at 6000 bubbles/ml. All other conditions such as ultrasound parameters will be kept the same as described previously.

5.3 Results

Experimental results from 5.2.1 include a total of 281 tests with 33 tests having TMC change. The probability of having pore(s) formed in membrane is1− f (0,λ) = 33/ 281 = 0.1174 , which determines the mean of the Poisson distribution as λ = 0.1249 . Therefore, the probability of having single pore calculated from equation

5-4 is 93.9% in this set of experiments.

107 The distribution of the maximum current amplitudes obtained experimentally is shown in Fig. V-3. The mean value is 288.6nA with a standard deviation is 225.6nA. The overall distribution of the current amplitude shows a gradual decreasing trend from as small as 50nA to over 1µA. 15 of 33 data distributed within 100±100 nA area with a

Gaussian shape.

Fig. V-4 shows the fitting result of simulated data and experiment data with multiple pores. The experiment data (n=85) were obtained with the same experimental condition. The mean number of pores in the experiment data is estimated from the least square fitting to be 4.13.

The corresponding pore size distribution estimated from the diffusion model is obtained from the experimentally measured maximum current values, and the results are shown in Fig. V-5 in a box plot format. The mean value of the single pore size is

2.347µm with the standard deviation of 2.075 µm.

5.4 Discussion

Pore formation in sonoporation results from the interaction of ultrasound field with bubbles in the vicinity of cell membrane. A complete quantitative description of this process is not available yet due to the challenges associated with the complexity and the rapid temporal scale of this interaction phenomenon.

The method used to estimate the TMC value of a single pore has merits and limitations. It provides indirect method based on statistical analysis. It requires the bubble concentration in experiments to be sufficiently low to ensure a single pore event and relies on statistics of experiment measurements for further analysis. This method does not

108 make any assumption of uniform pore size and thus can provide information of the pore size distribution. Since the majority portion of the current comes from a single pore, the current distribution obtained from this method includes information of the pore size and possibly any effects of ultrasound interaction with the bubbles/cells. As demonstrated previously, bubble size distribution does follow a similar trend [99]. On the other hand, the result included a small percentage (6%) of data where multiple pores are present. It is tedious in practice to achieve a higher possibility of single pore event because it requires a large number of trials. For example, if we expect to decrease the probability of having multiple pores to 3%, we need to decrease bubble concentration so that the mean number of effective bubbles equals 0.06 which means at least 567 experiments are required.

Using the TMC distribution of single pore, the number of multiple pores can be estimated for experiments. Poisson distribution for the number of bubbles and uniform size of pores were assumed in the estimation, which was demonstrated in Fig. V-6. The result shows the distribution of bubbles numbers in experiments (i.i.d.) in a given volume

(see description in Chapter 2). The only limitation is that enough sample number

(depends on data variance which is usually large in our case) is required to establish a statistical stable TMC distribution of multiple pores in experiments. The fitting result in

Fig.V-4 shows the fitting is limited by the sample size.

Comparing the single pore and multiple pores experiments, the mean numbers of pores (0.1 vs.4.4) do not correspond with the given bubble concentration for the experiments (6E+3 vs. 1.2E+7). The number of pores only increased 44 times (0.1 vs.

4.4). One possible reason is the higher attenuation of ultrasound intensity with more

109 bubbles in solution, which affect the actual process of pore formation from ultrasound application.

The accuracy of the estimation of pore size from TMC current relies on an accurate model of ion electro-diffusion. The simplified electro-diffusion model presented in this chapter is only to illustrate the process to estimate the pore size using TMC measurements. This simplified model assumes one-dimensional (1D) diffusion and calculates the current contributed by both electrical field and ion concentration gradient across the cell membrane in voltage clamp. As a static model, it ignores any ion concentration change in and outside of the cell membrane. The result obtained from the diffusion model shows that the pore size is related to the square root of the TMC amplitude, which suggests the measurement of TMC amplitude is more sensitive than the estimation of pore size. Another limitation is the assumption of a 1D difussion model, which may not be suitable for small pores where experimental conditions may be better described by a cylindrical model (two-dimension diffusion). Furthermore, the quasi- steady state assumption may be a limitation on the estimation results. The pore size estimated using this simple model appears to be larger than values conventionally recognized, which ranges from the order of 10nm to 1 µm.

Pore size is case specific depending on properties of individual cells, including membrane composition, thickness, surface tension, and mechanical and other properties.

Compared with other cells, one factor is that the size and membrane properties of oocytes will affect the pore size formed in sonoporation. Other factors, such as the existing of vitelline membrane, the composition of membrane, also need to be considered.

110 The above methods provide estimation of single pore size based on statistical analysis of experimental measurements and the validity of the diffusion model of ions through the pores.

5.5 Conclusion and future work

The distribution of TMC corresponding to a single pore during sonoporation process was obtained from voltage clamp measurements. The mean number of pores was estimated from experiments containing multiple pores generated in sonoporation. All these methods rely on statistical assumptions and analysis. An electro-diffusion model relating the pore size and TMC was then developed to estimate the pore size. These results are the first estimation of pore size using voltage clamp techniques.

The limitations of these statistical methods dictate studies in future investigation.

Updates in the electro-diffusion model to include a 2D geometry (e.g. cylindrical) and the temporal changes due to transient changes of pore sizes can potentially provide more precise estimation of pore size. Additionally, incorporation of bubble distribution can provide estimation of pore size in a more realistic fashion.

A different, deterministic approach to measure pore size using techniques such as

EM techniques can provide validation of the statistical estimation methods described in this chapter.

111

VNa, VK, VCl is potential from ion gradient by Na, K and Cl, INa, IK, ICl is TMC from ion gradient by Na, K and Cl

112

4 m) μ 3

Radius of pore ( of pore Radius 1

0 0 0.2 0.4 0.6 0.8 1 TMC (μA)

Figure V-2: The relationship between maximum current amplitude and radius of pore

113

2 Number of data 1

0 0.0 2.0x10-1 4.0x10-1 6.0x10-1 8.0x10-1 Current amplitude (μA)

Figure V-3: Histogram of maximum current amplitude from single pore. Bin size is 0.05µA; sample size N=33.

114

0.12 exp data 0.1 simulation data fit, λ=4.1254

0.08

0.06

0.04

0.02 Probability densityfunction

0 0 10 20 30 40 TMC (μA)

Figure V-4: Probability density of experiment data (n=85) and simulation result from single pore data.

115

A) μ 0.6 4.368 ) m μ

0.5 ( 3.907 0.4 3.384 0.3

0.2 2.763

0.1 1.954

0.0 Estimation of pore radius radius of pore Estimation 0

Transmembrane current amplitude ( N=33

Figure V-5: Amplitudes of TMC and estimated radius from single pore are shown in box plot. The box shows 25, 50 and 75 quartile, the whiskers represent the standard deviation, and the outliers show max and min value. The sample size is 33.

116

5 Number of measurement

0 10 20 30 40 50 Number of bubbles

Figure V-6: Histogram of the number of bubbles within 8.09µm3 solution. The curve is a Poisson function fitted to the measurement with a mean equals 30.35. The sample size N=234.

117

CHAPTER VI

FUTURE WORK

The objective of this study is to obtain improved understanding of the mechanism of sonoporation in order to provide guidance for the successful development and optimization of ultrasound mediated drug or gene delivery strategy. Through systematic, innovative experimental and theoretical investigation, we obtained the first results that reveal several important aspects of sonoporation process by studying the dynamic sonoporation at the single cell level in real time. Specifically, effects of ultrasound parameters and extracellular calcium concentration on sonoporation were demonstrated.

A direct correlation of inertial bubble cavitation with the inception of sonoporation has been established. These results provide not only new insights into the molecular and biophysical mechanism of sonoporation, but also lay a solid foundation for future work to further expand the current knowledge of sonoporation.

6.1 Specific topics for future investigation

6.1.1 Study of single cell sonoporation of mammalian cells using patch clamp technique

Xenopus oocytes were used as a model system in this study to obtain. A future study using various types of mammalian cells will be needed to further reveal sonoporation process in different model systems. Same principal should applied and patch clamp

118 techniques should readily applicable to study single cell sonoporation of these mammalian cells.

Another benefit using patch clamp is its superior time resolution, i.e. 1µs, vs. 1ms for voltage clamp techniques. This improved temporal resolution in measurement could potential provide new information regarding the effect of the initial ultrasound impact which occurs at a fast speed (a few micro second and up).

6.1.2 Controlled microbubble interaction with ultrasound and cell membrane

In the studies included in this dissertation, microbubbles are considered as statistical events randomly and uniformly distributed in the solution to interact with the cell membrane. While we have developed statistical analysis methods to obtain single pore size, a deterministic method will provide a direct measurement. A future study will include a system to control the bubble location with respect to the cell membrane during voltage clamp measurement. Research has been already underway in our laboratory to utilize optical methods (optical tweezer to move bubbles and laser induced optical breakdown to generate bubbles) to control the location of bubbles.

Additionally, ultra-fast photography is also being pursued to investigate the bubble dynamics (oscillation and collapsing) with a frame rate > 1 Mf/s, which can be especially critical to understand the ultrasound-bubble-membrane interaction.

6.1.3 Deterministic measurement of pore size

Future studies will be conducted to measure deterministically the size of the pores and their distribution in sonoporation, using electron microscope (EM) technique or extracellular agents/markers such as fluorescent nano-particles with known size. The EM

119 study can be designed by rapid freezing and the fact that the pores do not reseal in the absence of extracellular calcium. A characterization can be obtained by correlating the direct pore size measurement with transmembrane current measurement.

6.2 Benefits to future applications on drug and gene delivery

The work presented in this dissertation focuses on the fundamentals of sonoporation processes. Because non-ionizing US exposure can be non-invasively controlled in application location and duration, sonoporation may provide an advantageous, safe delivery strategy for in vivo applications. However, progress in the field is hindered by a lack of mechanistic understanding of the sonoporation process and its outcome beyond demonstrations of initial feasibility. It is therefore of significance to understand the process of sonoporation, yet it is a task challenged by the lack of appropriate techniques to study the transient and sub-micron process.

Voltage clamp technique was used to disclose the dynamics of pores formed in membrane in sonoporation by measuring the TMC through pores. The novel application of this electrophysiological technique enables a time-resolved measurement of sonoporation at single cell level, providing a sensitive and quantitative means to investigate the sonoporation process. By obtaining general dynamics of sonoporation, optimal sonoporation parameters of both physical and biochemical factors, can be determined rationally for individual case; a new mean is established to obtain mechanistic understanding of the causes for the downstream, cellular bio-effects and organ-level impacts of sonoporation; and more in situ results can be translated to in vivo environment base on valid correlation in mechanism.

120 The observations and conclusions of this study give us a basic understanding of sonoporation and can be used as a general guidance for future applications on drug and gene delivery. For example, from the information in calcium study, if extracellular calcium concentration can be manipulated in a controlled fashion, then lower this concentration will be a choice to increase delivery efficiency. Pore estimation study gives us a way to determine the possibility of delivery particles in certain size without doing actual delivery experiment.

Though this study is done with Xenopus oocyte model, it is directly connected to delivery applications in vivo or in human body. First, single cell level study is the first step for application in vivo or in human body; and second, since mechanical properties of plasma membrane in Xenopus oocyte is very similar to most cells, it is straightforward to transfer conclusions of this study to other type of cells and adjust the result by considering the cell size effect.

The ultimate goal of this study is to establish a controllable, robust drug and gene delivery method via sonoporation for animal and human applications. Better understanding of the mechanisms involved can lead to the determination of key factors to achieve this goal. Mechanistic studies such as those presented in this dissertation can be directly utilized for future in vivo human applications such as drug and gene delivery through blood-brain-barrier, focal stimulation of neurons using ultrasound application.

121 REFERENCE

[1] S. Mitragotri, "Healing sound: the use of ultrasound in drug delivery and other therapeutic applications," Nat Rev Drug Discov, vol. 4, pp. 255-60, 2005. [2] D. L. Miller, S. V. Pislaru, and J. E. Greenleaf, "Sonoporation: mechanical DNA delivery by ultrasonic cavitation," Somat Cell Mol Genet, vol. 27, pp. 115-34, 2002. [3] C. M. Newman and T. Bettinger, "Gene therapy progress and prospects: ultrasound for gene transfer," Gene Ther, vol. 14, pp. 465-75, 2007. [4] C. X. Deng, F. Sieling, H. Pan, and J. Cui, "Ultrasound-induced cell membrane porosity," Ultrasound Med Biol, vol. 30, pp. 519-26, 2004. [5] P. Hauff, S. Seemann, R. Reszka, M. Schultze-Mosgau, M. Reinhardt, T. Buzasi, T. Plath, S. Rosewicz, and M. Schirner, "Evaluation of gas-filled microparticles and sonoporation as gene delivery system: feasibility study in rodent tumor models," Radiology, vol. 236, pp. 572-8, 2005. [6] W. Stuhmer, "Electrophysiologic recordings from Xenopus oocytes," Methods Enzymol, vol. 293, pp. 280-300, 1998. [7] A. Awada and M. Piccart, "Strategies offering protection from the toxic effects of anticancer treatments with a focus on chemoprotective agents," Curr Opin Oncol, vol. 12, pp. 289-96, 2000. [8] R. M. Lowenthal and K. Eaton, "Toxicity of chemotherapy," Hematol Oncol Clin North Am, vol. 10, pp. 967-90, 1996. [9] M. T. Meinardi, J. A. Gietema, D. J. van Veldhuisen, W. T. van der Graaf, E. G. de Vries, and D. T. Sleijfer, "Long-term chemotherapy-related cardiovascular morbidity," Cancer Treat Rev, vol. 26, pp. 429-47, 2000. [10] L. K. Fung, M. Shin, B. Tyler, H. Brem, and W. M. Saltzman, "Chemotherapeutic drugs released from polymers: distribution of 1,3-bis(2-chloroethyl)-1-nitrosourea in the rat brain," Pharm Res, vol. 13, pp. 671-82, 1996. [11] A. Gabizon, H. Shmeeda, and Y. Barenholz, "Pharmacokinetics of pegylated liposomal Doxorubicin: review of animal and human studies," Clin Pharmacokinet, vol. 42, pp. 419-36, 2003. [12] J. D. Hood, M. Benarski, R. Frausto, and a. et., "Tumor regression by targeted gene delivery to the neovasculature," Science, vol. 296, pp. 2404-7, 2002. [13] V. Martin, M. L. Cortes, P. de Felipe, A. Farsetti, N. B. Calcaterra, and M. Izquierdo, "Cancer gene therapy by thyroid hormone-mediated expression of toxin genes," Cancer Res, vol. 60, pp. 3218-24, 2000. [14] E. G. Hanania, J. Kavanagh, G. Hortobagyi, R. E. Giles, R. Champlin, and A. B. Deisseroth, "Recent advances in the application of gene therapy to human disease," Am J Med, vol. 99, pp. 537-52, 1995. [15] D. J. Wells, "Gene therapy progress and prospects: electroporation and other physical methods," Gene Ther, vol. 11, pp. 1363-9, 2004. [16] H. K. Lyerly and J. M. DiMaio, "Gene delivery systems in surgery," Arch Surg, vol. 128, pp. 1197-206, 1993. [17] S. Yla-Herttuala and J. F. Martin, "Cardiovascular gene therapy," Lancet, vol. 355, pp. 213-22, 2000.

122 [18] M. R. Prausnitz, "A practical assessment of transdermal drug delivery by skin electroporation," Adv Drug Deliv Rev, vol. 35, pp. 61-76, 1999. [19] D. C. Chang and T. S. Reese, "Changes in membrane structure induced by electroporation as revealed by rapid-freezing electron microscopy," Biophys J, vol. 58, pp. 1-12, 1990. [20] P. J. Canatella, J. F. Karr, J. A. Petros, and M. R. Prausnitz, "Quantitative study of electroporation-mediated molecular uptake and cell viability," Biophys J, vol. 80, pp. 755-64, 2001. [21] S. Wang, S. Joshi, and S. Lu, "Delivery of DNA to skin by particle bombardment," Methods Mol Biol, vol. 245, pp. 185-96, 2004. [22] W. C. Heiser, "Delivery of DNA to cells in culture using particle bombardment," Methods Mol Biol, vol. 245, pp. 175-84, 2004. [23] P. M. Sudha, S. Low, J. Kwang, and Z. Gong, "Multiple tissue transformation in adult zebrafish by gene gun bombardment and muscular injection of naked DNA," Mar Biotechnol (NY), vol. 3, pp. 119-25, 2001. [24] S. W. Gersting, U. Schillinger, J. Lausier, P. Nicklaus, C. Rudolph, C. Plank, D. Reinhardt, and J. Rosenecker, "Gene delivery to respiratory epithelial cells by magnetofection," J Gene Med, vol. 6, pp. 913-22, 2004. [25] M. R. Almofti, H. Harashima, Y. Shinohara, A. Almofti, W. Li, and H. Kiwada, "Lipoplex size determines lipofection efficiency with or without serum," Mol Membr Biol, vol. 20, pp. 35-43, 2003. [26] R. I. Mahato, Y. Takakura, and M. Hashida, "Nonviral vectors for in vivo gene delivery: physicochemical and pharmacokinetic considerations," Crit Rev Ther Drug Carrier Syst, vol. 14, pp. 133-72, 1997. [27] H. J. Kim, J. F. Greenleaf, R. R. Kinnick, J. T. Bronk, and M. E. Bolander, "Ultrasound-mediated transfection of mammalian cells," Hum Gene Ther, vol. 7, pp. 1339-46, 1996. [28] S. V. Pislaru, C. Pislaru, R. R. Kinnick, R. Singh, R. Gulati, J. F. Greenleaf, and R. D. Simari, "Optimization of ultrasound-mediated gene transfer: comparison of contrast agents and ultrasound modalities," Eur Heart J, vol. 24, pp. 1690-8, 2003. [29] P. A. Frenkel, S. Chen, T. Thai, R. V. Shohet, and P. A. Grayburn, "DNA-loaded albumin microbubbles enhance ultrasound-mediated transfection in vitro," Ultrasound Med Biol, vol. 28, pp. 817-22, 2002. [30] Y. Taniyama, K. Tachibana, K. Hiraoka, M. Aoki, S. Yamamoto, K. Matsumoto, T. Nakamura, T. Ogihara, Y. Kaneda, and R. Morishita, "Development of safe and efficient novel nonviral gene transfer using ultrasound: enhancement of transfection efficiency of naked plasmid DNA in skeletal muscle," Gene Ther, vol. 9, pp. 372-80, 2002. [31] S. Bao, B. D. Thrall, and D. L. Miller, "Transfection of a reporter plasmid into cultured cells by sonoporation in vitro," Ultrasound Med Biol, vol. 23, pp. 953-9, 1997. [32] Y. Manome, M. Nakamura, T. Ohno, and H. Furuhata, "Ultrasound facilitates transduction of naked plasmid DNA into colon carcinoma cells in vitro and in vivo," Hum Gene Ther, vol. 11, pp. 1521-8, 2000.

123 [33] R. Ogawa, G. Kagiya, L. B. Feril, Jr., N. Nakaya, T. Nozaki, H. Fuse, and T. Kondo, "Ultrasound Mediated Intravesical Transfection Enhanced by Treatment with Lidocaine or Heat," J Urol, vol. 172, pp. 1469-1473, 2004. [34] T. P. McCreery, R. H. Sweitzer, E. C. Unger, and S. Sullivan, "DNA delivery to cells in vivo by ultrasound," Methods Mol Biol, vol. 245, pp. 293-8, 2004. [35] P. E. Huber and P. Pfisterer, "In vitro and in vivo transfection of plasmid DNA in the Dunning prostate tumor R3327-AT1 is enhanced by focused ultrasound," Gene Ther, vol. 7, pp. 1516-25, 2000. [36] R. V. Shohet, S. Chen, Y. T. Zhou, Z. Wang, R. S. Meidell, R. H. Unger, and P. A. Grayburn, "Echocardiographic destruction of albumin microbubbles directs gene delivery to the myocardium," Circulation, vol. 101, pp. 2554-6, 2000. [37] J. M. Tsutsui, P. A. Grayburn, F. Xie, and T. R. Porter, "Drug and gene delivery and enhancement of thrombolysis using ultrasound and microbubbles," Cardiol Clin, vol. 22, pp. 299-312, vii, 2004. [38] W. C. Ou, J. C. Xiu, J. G. Xie, D. G. Zha, J. Liu, R. Hao, J. P. Bin, and Y. L. Liu, "[Ultrasound microbubbles promote thrombolysis of rabbit femoral artery thrombus]," Di Yi Jun Yi Da Xue Xue Bao, vol. 24, pp. 536-8, 2004. [39] J. Eggers, B. Koch, K. Meyer, I. Konig, and G. Seidel, "Effect of ultrasound on thrombolysis of middle cerebral artery occlusion," Ann Neurol, vol. 53, pp. 797- 800, 2003. [40] T. R. Porter and F. Xie, "Ultrasound, microbubbles, and thrombolysis," Prog Cardiovasc Dis, vol. 44, pp. 101-10, 2001. [41] C. W. Francis, "Ultrasound-enhanced thrombolysis," Echocardiography, vol. 18, pp. 239-46, 2001. [42] N. Sheikov, N. McDannold, N. Vykhodtseva, F. Jolesz, and K. Hynynen, "Cellular mechanisms of the blood-brain barrier opening induced by ultrasound in presence of microbubbles," Ultrasound Med Biol, vol. 30, pp. 979-89, 2004. [43] K. Hynynen, N. McDannold, N. Vykhodtseva, and F. A. Jolesz, "Non-invasive opening of BBB by focused ultrasound," Acta Neurochir Suppl, vol. 86, pp. 555- 8, 2003. [44] M. Shimamura, N. Sato, Y. Taniyama, S. Yamamoto, M. Endoh, H. Kurinami, M. Aoki, T. Ogihara, Y. Kaneda, and R. Morishita, "Development of efficient plasmid DNA transfer into adult rat central nervous system using microbubble- enhanced ultrasound," Gene Ther, vol. 11, pp. 1532-9, 2004. [45] D. L. Miller and J. Quddus, "Sonoporation of monolayer cells by diagnostic ultrasound activation of contrast-agent gas bodies," Ultrasound Med Biol, vol. 26, pp. 661-7, 2000. [46] S. Koch, P. Pohl, U. Cobet, and N. G. Rainov, "Ultrasound enhancement of liposome-mediated cell transfection is caused by cavitation effects," Ultrasound Med Biol, vol. 26, pp. 897-903, 2000. [47] D. L. Miller and J. Song, "Tumor growth reduction and DNA transfer by cavitation-enhanced high-intensity focused ultrasound in vivo," Ultrasound Med Biol, vol. 29, pp. 887-93, 2003. [48] D. L. Miller and R. M. Thomas, "The role of cavitation in the induction of cellular DNA damage by ultrasound and lithotripter shock waves in vitro," Ultrasound Med Biol, vol. 22, pp. 681-7, 1996.

124 [49] A. Lawrie, A. F. Brisken, S. E. Francis, D. C. Cumberland, D. C. Crossman, and C. M. Newman, "Microbubble-enhanced ultrasound for vascular gene delivery," Gene Ther, vol. 7, pp. 2023-7, 2000. [50] L. A. Crum, R. A. Roy, M. A. Dinno, C. C. Church, R. E. Apfel, C. K. Holland, and S. I. Madanshetty, "Acoustic cavitation produced by microsecond pulses of ultrasound: a discussion of some selected results," J Acoust Soc Am, vol. 91, pp. 1113-9, 1992. [51] D. B. Tata, F. Dunn, and D. J. Tindall, "Selective clinical ultrasound signals mediate differential gene transfer and expression in two human prostate cancer cell lines: LnCap and PC-3," Biochem Biophys Res Commun, vol. 234, pp. 64-7, 1997. [52] L. B. Feril, Jr., T. Kondo, Q. L. Zhao, R. Ogawa, K. Tachibana, N. Kudo, S. Fujimoto, and S. Nakamura, "Enhancement of ultrasound-induced apoptosis and cell lysis by echo-contrast agents," Ultrasound Med Biol, vol. 29, pp. 331-7, 2003. [53] B. Hille, Ion Channels of Excitable Membranes, 3 rd ed. Sunderland, MA: Sinauer Associates, Inc., 2001. [54] F. Ryttsen, C. Farre, C. Brennan, S. G. Weber, K. Nolkrantz, K. Jardemark, D. T. Chiu, and O. Orwar, "Characterization of single-cell electroporation by using patch-clamp and fluorescence microscopy," Biophys J, vol. 79, pp. 1993-2001, 2000. [55] J. P. Christiansen, B. A. French, A. L. Klibanov, S. Kaul, and J. R. Lindner, "Targeted tissue transfection with ultrasound destruction of plasmid-bearing cationic microbubbles," Ultrasound Med Biol, vol. 29, pp. 1759-67, 2003. [56] R. Bekeredjian, S. Chen, P. A. Frenkel, P. A. Grayburn, and R. V. Shohet, "Ultrasound-targeted microbubble destruction can repeatedly direct highly specific plasmid expression to the heart," Circulation, vol. 108, pp. 1022-6, 2003. [57] R. E. Kumon, M. Aehle, D. Sabens, P. Parikh, D. Kourennyi, and C. X. Deng, "Ultrasound-induced Calcium Oscillations and Waves in Chinese Hamster Ovary Cells in the Presence of Microbubbles," Biophys J, 2007. [58] H. Pan, Y. Zhou, O. Izadnegahdar, J. Cui, and C. X. Deng, "Study of sonoporation dynamics affected by ultrasound duty cycle," Ultrasound Med Biol, vol. 31, pp. 849-56, 2005. [59] L. Hu, J. Shi, Z. Ma, G. Krishnamoorthy, F. Sieling, G. Zhang, F. T. Horrigan, and J. Cui, "Participation of the S4 voltage sensor in the Mg2+-dependent activation of large conductance (BK) K+ channels," Proc Natl Acad Sci U S A, vol. 100, pp. 10488-93, 2003. [60] D. F. Slish, Xenopus Oocyte Electrophysiology. http://faculty.plattsburgh.edu/donald.slish/Main.html. [61] D. L. Miller, S. Bao, and J. E. Morris, "Sonoporation of cultured cells in the rotating tube exposure system," Ultrasound Med Biol, vol. 25, pp. 143-9, 1999. [62] M. Ward, J. Wu, and J. F. Chiu, "Experimental study of the effects of Optison concentration on sonoporation in vitro," Ultrasound Med Biol, vol. 26, pp. 1169- 75, 2000. [63] W. J. Greenleaf, M. E. Bolander, G. Sarkar, M. B. Goldring, and J. F. Greenleaf, "Artificial cavitation nuclei significantly enhance acoustically induced cell transfection," Ultrasound in Medicine and Biology, vol. 24, pp. 587-595, 1998.

125 [64] J. R. Wu, "Shear stress in cells generated by ultrasound," Progress in Biophysics & Molecular Biology, vol. 93, pp. 363-373, 2007. [65] M. Lokhandwalla, J. A. McAteer, J. C. Williams, Jr., and B. Sturtevant, "Mechanical haemolysis in shock wave lithotripsy (SWL): II. In vitro cell lysis due to shear," Phys Med Biol, vol. 46, pp. 1245-64, 2001. [66] J. Wu, "Theoretical study on shear stress generated by microstreaming surrounding contrast agents attached to living cells," Ultrasound Med Biol, vol. 28, pp. 125-9, 2002. [67] J. Wu, J. P. Ross, and J. F. Chiu, "Reparable sonoporation generated by microstreaming," J Acoust Soc Am, vol. 111, pp. 1460-4, 2002. [68] P. C. Prentice, Alfred; Dholakia, Kishan; Prausnitz, Mark; Campbell, Paul, "Membrane disruption by optically controlled microbubble cavitation," Nature Physics, vol. 1, pp. 107-110, 2005. [69] C. X. Deng and F. L. Lizzi, "A review of physical phenomena associated with ultrasonic contrast agents and illustrative clinical applications," Ultrasound Med Biol, vol. 28, pp. 277-86, 2002. [70] Nikon, Depth of Field and Depth of Focus. http://www.microscopyu.com/articles/formulas/formulasfielddepth.html. [71] I. Axon Instruments, The Axon Guide for Electrophysiology & Biophysics, 1993. [72] H. Potter, "Electroporation in biology: methods, applications, and instrumentation," Anal Biochem, vol. 174, pp. 361-73, 1988. [73] S. I. Sukharev, V. A. Klenchin, S. M. Serov, L. V. Chernomordik, and A. Chizmadzhev Yu, "Electroporation and electrophoretic DNA transfer into cells. The effect of DNA interaction with electropores," Biophys J, vol. 63, pp. 1320-7, 1992. [74] J. C. Weaver, "Electroporation theory. Concepts and mechanisms," Methods Mol Biol, vol. 55, pp. 3-28, 1995. [75] S. B. Dev, D. P. Rabussay, G. Widera, and G. A. Hofmann, "Medical applications of electroporation," IEEE Trans. Plasma Sci., vol. 28, pp. 206–223, 2000. [76] R. A. Steinhardt, "The mechanisms of cell membrane repair: A tutorial guide to key experiments," Ann N Y Acad Sci, vol. 1066, pp. 152-65, 2005. [77] P. L. McNeil and R. A. Steinhardt, "Plasma membrane disruption: repair, prevention, adaptation," Annu Rev Cell Dev Biol, vol. 19, pp. 697-731, 2003. [78] P. L. McNeil and T. Kirchhausen, "An emergency response team for membrane repair," Nat Rev Mol Cell Biol, vol. 6, pp. 499-505, 2005. [79] W. M. Bement, H. Y. Yu, B. M. Burkel, E. M. Vaughan, and A. G. Clark, "Rehabilitation and the single cell," Curr Opin Cell Biol, vol. 19, pp. 95-100, 2007. [80] P. L. McNeil and R. A. Steinhardt, "Loss, restoration, and maintenance of plasma membrane integrity," J Cell Biol, vol. 137, pp. 1-4, 1997. [81] K. J. Muller, M. Horbaschek, K. Lucas, U. Zimmermann, and V. L. Sukhorukov, "Electrotransfection of anchorage-dependent mammalian cells," Exp Cell Res, vol. 288, pp. 344-53, 2003. [82] K. J. Muller, V. L. Sukhorukov, and U. Zimmermann, "Reversible electropermeabilization of mammalian cells by high-intensity, ultra-short pulses of submicrosecond duration," J Membr Biol, vol. 184, pp. 161-70, 2001.

126 [83] R. Shirakashi, C. M. Kostner, K. J. Muller, M. Kurschner, U. Zimmermann, and V. L. Sukhorukov, "Intracellular delivery of trehalose into mammalian cells by electropermeabilization," J Membr Biol, vol. 189, pp. 45-54, 2002. [84] T. Togo, J. M. Alderton, G. Q. Bi, and R. A. Steinhardt, "The mechanism of facilitated cell membrane resealing," J Cell Sci, vol. 112 ( Pt 5), pp. 719-31, 1999. [85] R. A. Steinhardt, G. Bi, and J. M. Alderton, "Cell membrane resealing by a vesicular mechanism similar to neurotransmitter release," Science, vol. 263, pp. 390-3, 1994. [86] W. M. Bement, C. A. Mandato, and M. N. Kirsch, "Wound-induced assembly and closure of an actomyosin purse string in Xenopus oocytes," Curr Biol, vol. 9, pp. 579-87, 1999. [87] T. Togo, "Disruption of the plasma membrane stimulates rearrangement of microtubules and lipid traffic toward the wound site," J Cell Sci, vol. 119, pp. 2780-6, 2006. [88] A. M. Sokac, C. Co, J. Taunton, and W. Bement, "Cdc42-dependent actin polymerization during compensatory endocytosis in Xenopus eggs," Nat Cell Biol, vol. 5, pp. 727-32, 2003. [89] A. M. Sokac and W. M. Bement, "Kiss-and-coat and compartment mixing: coupling exocytosis to signal generation and local actin assembly," Mol Biol Cell, vol. 17, pp. 1495-502, 2006. [90] Y. Zhou, J. Shi, J. Cui, and C. X. Deng, "Effects of extracellular calcium on cell membrane resealing in sonoporation," J Control Release, vol. 126, pp. 34-43, 2008. [91] Y. Zhou, J. Cui, and C. X. Deng, "Dynamics of sonoporation correlated with acoustic cavitation activities," Biophys J, vol. 94, pp. L51-3, 2008. [92] H. R. Guzman, A. J. McNamara, D. X. Nguyen, and M. R. Prausnitz, "Bioeffects caused by changes in acoustic cavitation bubble density and cell concentration: A unified explanation based on cell-to-bubble ratio and blast radius," Ultrasound in Medicine and Biology, vol. 29, pp. 1211-1222, 2003. [93] W. M. Weber, "Ion currents of Xenopus laevis oocytes: state of the art," Biochimica Et Biophysica Acta-Biomembranes, vol. 1421, pp. 213-233, 1999. [94] M. Cougnon, P. Bouyer, P. Hulin, T. Anagnostopoulos, and G. Planelles, "Further investigation of ionic diffusive properties and of NH4+ pathways in Xenopus laevis oocyte cell membrane," Pflugers Archiv-European Journal of Physiology, vol. 431, pp. 658-667, 1996. [95] P. F. Costa, M. G. Emilio, P. L. Fernandes, H. G. Ferreira, and K. G. Ferreira, "Determination of Ionic Permeability Coefficients of the Plasma-Membrane of Xenopus-Laevis Oocytes under Voltage Clamp," Journal of Physiology-London, vol. 413, pp. 199-211, 1989. [96] E. L. Cussler, Diffusion : mass transfer in fluid systems, 2nd ed. New York: Cambridge University Press, 1997. [97] K. A. DeBruin and W. Krassowska, "Modeling electroporation in a single cell. II. Effects of ionic concentrations," Biophysical Journal, vol. 77, pp. 1225-1233, 1999. [98] A. Guia, Voltage-clamp Electrophysiology. http://avivabio.com/ephysbk.php?MID=1.

127 [99] D. Chatterjee, P. Jain, and K. Sarkar, "Ultrasound-mediated destruction of contrast microbubbles used for medical imaging and drug delivery," Physics of Fluids, vol. 17, pp. 100603, 2005.

128