Acquired Electrophysiological Remodeling and Cardiac

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By

Ingrid M. Bonilla Mercado

Graduate Program in: Pharmaceutical Sciences

The Ohio State University

2014

Dissertation Committee:

Dr. Cynthia A. Carnes, Advisor

Dr. Peter J. Mohler

Dr. Kari Hoyt

Dr. Sandor Györke

Abstract

Heart disease is the number one killer in the U.S. Some of the most common diseases are (MI), (HF), and atrial (AF).

Due to the high mortality and morbidity caused by heart disease, it is important to improve our understanding of the pathophysiology to identify new therapeutic targets with the ultimate goal of developing new drug therapies. Each type of heart disease has to be treated differently since differences in structural and electrophysiological remodeling occurs. The purpose of the research described in this dissertation is to provide a more in depth understanding of acquired electrophysiological remodeling in specific diseases to provide a basis for improved therapeutic approaches. Diseases studied include: post- MI animals susceptible to , chronic HF, and chronic

HF with AF superimposed. Additionally since acquired electrophysiological remodeling can also result as an adverse drug effect, a study of drug-induced mechanisms is included.

We found that training protects against lethal arrhythmias in the post-

MI setting, through normalization of ventricular and restoration of the repolarizing current, IKr. IKr block is a common cause of drug-induced cardiac arrhythmia, and we evaluated a drug suspected of causing drug-induced arrhythmias. Surprisingly, ii rather than IKr block, a synergistic combination of decreased transient outward current and abnormal myocyte calcium were the basis of this form of drug-induced arrhythmia.

The potential role of reactive species in the pathogenesis of arrhythmias was examined using SIN-1(a peroxynitrite donor) in atrial and ventricular myocytes. While repolarization is longer during SIN-1 exposure in both chambers, effects on repolarization stability were chamber-dependent and effects in the were arrhythmic.

In chronic HF we found that IKCa contributes to ventricular but not atrial repolarization, and that IKCa blockade in ventricular cells is arrhythmogenic. We also found that AF superimposed on chronic HF does not further change atrial cellular repolarization, but it increases fibrosis, a possible mechanism for AF perpetuation.

The main conclusion of this body of work is that it is important to study both chambers (atria and ventricle) when considering possible antiarrhythmic targets or new drugs, and that disease-specific remodeling is an important contributor to abnormal electrical function and arrhythmogenesis.

iii

Dedication

To my Family and Friends

iv

Acknowledgments

Contrary to many people that know what they want to be when they grow up, I can say, that having a PhD was not on my list. Where I come from there are no big universities that offer PhD programs thus, my lack of information about graduate school.

The most I knew of was to get a masters and then you don’t find a job. Even though my dream wasn’t always being a scientist while doing my bachelor’s degree I always knew that I wanted something more and somehow I knew that if I stayed in Puerto Rico I wasn’t going to find that thing that make me happy. That’s the reason why I applied to the summer research opportunity here in Ohio State at the College of Pharmacy (without even knowing where Ohio was on the map and without an idea what I was going to do during that summer). I had no idea of what summer research was, and therefore I had no expectations. But I have to say that to my surprise once I started working I realized that I just found what I was looking for. I had found a profession that makes me happy and that

I wanted to do for the rest of my life. Now the hard part was to be able to get into the graduate school in OSU. This was no easy task and took a lot of effort not only on my part but from my mentor also.

v

I can’t say that it was an easy task to finish my graduate degree. There were many times where I thought it was the end and if not because of my hard work and the help of many people (I would say an army of people) that stood there for me, helped me and guided me I would not be writing this today. Every day I thank God for putting the right people in my way and for giving me the strength to keep going even when the road seemed very rocky.

I am extremely thankful and forever in debt with my advisor Dr. Carnes. I know it was not an easy task to deal with me and I have to say that she had a lot of patience in order to deal with me. Also I want to thank her for believing in me, giving me the confidence that I was capable of doing this and for treating me as more than a student, rather as a family member. Without her help and guidance I would not be here.

I will also like to express my deepest and sincere thanks to all the past and present members of the laboratory: Arun Sridhar, Yoshinori Nishijima, Chun Li, Pedro Vargas-

Pinto, Jeanne Green and last but not least Victor Long, for their constant help even if they are not in the lab anymore and for putting up with me.

Additionally I will like to thanks our collaborators and their respective laboratory members for helping me: Dr. Gyorke, Dr. Billman, Dr. Hamlin and Dr. Mohler.

I thank my committee members: Dr. Sandor Gyorke, Dr. Peter J. Mohler and Dr.

Kari Hoyt for making this experience a memorable one and for sharing your opinion and guidance during this process.

Thanks to the pharmacology division members for guiding me and for making sure that I was progressing adequately.

vi

All this work wouldn’t have been possible without the help of NHLBI supplement for minorities who supported me and part of my investigations for the past 4 years.

I would also like to thank my past and present friends for giving me the support necessary to keep going. They were the ones that were there on every step of my way and always motivated me to keep going.

Last but not least I will like to express my deepest and sincere thanks to my family. For their constant support, taking care of me, loving me no matter what and being there for me unconditionally. My sister, who was always one phone call away and has always been there for me and my husband for making my goals part of his goals and riding along with me on this roller coaster. I know this is not the end, this is just the beginning of a very long road. I hope I can continue doing research and helping to contribute to the improvement of cardiac disease and one day being able to have my own laboratory, do my own research and teach other students all my acquired knowledge. I will also like to return the favor of being supported and helped by so many wonderful people by helping new students find and pursue their passion as I am doing.

vii

Vita

March 2004 ...... Colegio San Benito

2008...... B.S. Industrial Microbiology, University of

Puerto Rico, Mayagüez Campus

2008-2009 ...... Laboratory technician, The Ohio State

University

2009 - Present ...... Graduate Research Associate, Department

of Pharmacology, The Ohio State University

Publications

1. Sridhar A, Nishijima Y, Terentyev D, Terentyeva R, Uelmen R, Kukielka M,

Bonilla IM, Robertson GA, Györke S, Billman GE, Carnes CA. Repolarization

abnormalities and afterdepolarizations in a canine model of sudden cardiac

. Am J Physiol Regul Integr Comp Physiol. 2008 Nov;295(5):R1463-72.

2. Nishijima Y, Sridhar A, Bonilla I, Velayutham M, Khan M, Terentyeva R, Li C,

Kuppusamy P, Elton TS, Terentyev D, Györke S, Zweier JL, Cardounel AJ,

Carnes CA. Tetrahydrobiopterin depletion and NOS2 uncoupling contribute to

heart failure-induced alterations in atrial . Cardiovasc Res.

2011 Jul ;91(1):71-9. viii

3. Belevych AE, Terentyev D, Terentyeva R, Ho HT, Gyorke I, Bonilla IM, Carnes

CA, Billman GE, Györke S. Shortened Ca2+ signaling refractoriness underlies

cellular arrhythmogenesis in a postinfarction model of sudden cardiac death. Circ

Res. 2012 Feb 17;110(4):569-77.

4. Gudmundsson H, Curran J, Kashef F, Snyder JS, Smith SA, Vargas-Pinto P,

Bonilla IM, Weiss RM, Anderson ME, Binkley P, Felder RB, Carnes CA, Band

H, Hund TJ, Mohler PJ. Differential regulation of EHD3 in and

mammalian heart failure. J Mol Cell Cardiol. 2012 May;52(5):1183-90.

5. Bonilla IM, Sridhar A, Györke S, Cardounel AJ, Carnes CA. Nitric oxide

synthases and . Front Physiol. 2012 Apr 23;3:105.

6. Wang H, Bonilla IM, Huang X, He Q, Kohr MJ, Carnes CA, Ziolo MT.

Prolonged and After Are Not due to Changes in

Potassium Currents in NOS3 Knockout Ventricular Myocytes. J Signal Transduct.

2012 Aug;2012:645721.

7. Bonilla IM, Belevych AE, Sridhar A, Nishijima Y, Ho HT, He Q, Kukielka M,

Terentyev D, Terentyeva R, Liu B, Long VP, Györke S, Carnes CA, Billman GE.

Endurance exercise training normalizes repolarization and calcium-handling

abnormalities, preventing ventricular fibrillation in a model of sudden cardiac

death. J Appl Physiol. 2012 Dec;113(11):1772-83.

8. Bonilla IM, Sridhar A, Nishijima Y, Györke S, Cardounel AJ, Carnes CA.

Differential effects of the peroxynitrite donor, SIN-1, on atrial and ventricular

myocyte electrophysiology. J Cardiovasc Pharmacol. 2013 May;61(5):401-7.

ix

9. Belevych AE, Ho HT, Terentyeva R, Bonilla IM, Terentyev D, Carnes CA,

Gyorke S, Billman GE. Dietary omega-3 Fatty acids promote arrhythmogenic

remodeling of cellular Ca(2+) handling in a postinfarction model of sudden

cardiac death. PLoS One. 2013 Oct;8(10):e78414.

10. Bonilla IM, Vargas-Pinto P, Nishijima Y, Pedraza-Toscano A, Ho HT, Long VP

3rd, Belevych AE, Glynn P, Houmsse M, Rhodes T, Weiss R, Hund TJ, Hamlin

RL, Györke S, Carnes CA. Ibandronate and Ventricular Arrhythmia Risk. J

Cardiovasc Electrophysiol. 2013 Nov 20. doi: 10.1111/jce.12327.

Fields of Study

Major Field: Pharmaceutical Sciences/

Specialization: Pharmacology

x

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgments...... v

Vita ...... viii

Publications ...... viii

Fields of Study ...... x

Table of Contents ...... xi

List of Tables ...... xix

List of Figures ...... xx

Chapter 1 : Introduction ...... 1

History of electrophysiology ...... 1

From theory to the ECG discovery ...... 1

The birth of cellular electrophysiology ...... 3

The phases and function of the ECG and mammalian ...... 5

Potassium channels ...... 8 xi

Voltage gated potassium channels ...... 8

Inward rectifier channels ...... 11

K+ channel modulation and ventricular arrhythmias ...... 13

Acquired K+ channel remodeling ...... 14

K+ blockade and drug induced Torsade de Pointes (TdP) ...... 14

K+ channel remodeling in ischemic heart disease and its contribution to SCD ...... 16

K+ channels remodeling in HF ...... 18

Atria ...... 18

Ventricle ...... 19

K+ channel remodeling in atrial fibrillation ...... 20

Anti-arrhythmic drugs ...... 20

Figures ...... 22

Chapter 2 : Endurance Exercise Training Normalizes Repolarization and Calcium

Handling Abnormalities Preventing Ventricular Fibrillation in a Model of Sudden

Cardiac Death...... 24

Abstract:...... 25

Introduction: ...... 27

Materials and Methods ...... 29

Surgical Preparation ...... 29

xii

Exercise plus Ischemia Test: selection for susceptibility to malignant arrhythmia ...30

Exercise Training Protocol ...... 30

Myocyte Isolation ...... 32

Protein Expression ...... 33

Electrophysiological studies ...... 33

Ca2+ Handling measurements ...... 34

Statistical analysis...... 35

Results ...... 36

Confirmation of Exercise Training ...... 36

Effect of Exercise Training on Susceptibility to VF ...... 36

ECG analysis ...... 37

Neither QRS duration nor PR intervals were altered by the coronary occlusion in

either the sedentary or the exercise trained groups ...... 38

Electrophysiological recordings ...... 38

Protein Expression ...... 39

Calcium Handling ...... 39

Discussion ...... 41

Effect of exercise training on cellular electrophysiology: normalization of

repolarization abnormalities ...... 42

xiii

Effect of exercise training on intracellular calcium handling ...... 45

Limitations ...... 46

Conclusion ...... 47

Figures ...... 48

Chapter 3 : Ibandronate and Ventricular Arrhythmia Risk ...... 59

Abstract ...... 60

Introduction ...... 62

Methods ...... 63

Case Report: ...... 63

Myocyte Isolation ...... 64

Calcium handling ...... 66

Ibandronate concentrations: ...... 66

Computational modeling: ...... 66

Data analysis: ...... 67

Results ...... 68

Ibandronate effects on canine QT ...... 68

Cellular electrophysiology:...... 68

Potassium currents ...... 69

Calcium depletion and buffering experiments ...... 69

xiv

Calcium Handling ...... 70

Computational modeling ...... 70

Discussion ...... 71

Limitations ...... 74

Conclusion ...... 75

Acknowledgments: ...... 75

Figures ...... 76

Chapter 4 : Differential effects of the peroxynitrite donor, SIN-1, on atrial and ventricular myocyte electrophysiology ...... 86

Previously Published: Front Physiol. 2012 Apr 23;3:105. Abstract ...... 86

Introduction ...... 88

Methods ...... 89

Myocyte Isolation ...... 89

Electrophysiological Studies ...... 90

Statistical Analysis ...... 92

Chemicals ...... 92

Results ...... 92

Specificity of effects of SIN-1 on action potential duration ...... 92

Beat to beat variability ...... 93

xv

Potassium currents ...... 94

Thapsigargin experiments...... 94

Discussion ...... 95

Limitation ...... 99

Conclusions ...... 100

Acknowledgments ...... 100

Figures ...... 101

Chapter 5 : Calcium-activated Potassium Current Modulates Ventricular Repolarization in Chronic Heart Failure ...... 105

Abstract ...... 106

Introduction ...... 107

Methods ...... 108

Heart failure model ...... 108

Myocyte Isolation and Tissue Collection ...... 109

Action Potential Measurements ...... 110

Calcium transient Measurements...... 111

Immunoblots ...... 112

Data Analysis ...... 112

Chemicals ...... 113

xvi

Results ...... 113

In vivo cardiac remodeling ...... 113

IKCa inhibition in control ventricular myocytes ...... 113

IKCa inhibition and SK expression in failing ventricle ...... 114

IKCa inhibition and SK expression in atrial myocytes ...... 115

Discussion ...... 116

Limitations ...... 121

Conclusions ...... 121

Acknowledgments ...... 122

Figures ...... 123

Chapter 6 Atrial electrical remodeling in heart failure: modulation by sustained atrial fibrillation ...... 133

Abstract ...... 133

Introduction ...... 134

Methods ...... 135

Heart failure and heart failure + atrial fibrillation canine models ...... 135

Myocyte Isolation ...... 136

Electrophysiological recordings ...... 137

Interstitial fibrosis ...... 138

xvii

Data Analysis ...... 138

Chemicals ...... 139

Results ...... 139

Echocardiographic measurements ...... 139

Electrophysiological data ...... 140

Fibrosis ...... 142

Discussion ...... 143

Limitations ...... 147

Summary and Conclusions ...... 147

Figures ...... 148

Chapter 7 : Summary and Conclusions ...... 154

Reference List ...... 161

xviii

List of Tables

Table 2.1: Effect of exercise training on ECG parameters at baseline, during exercise and during coronary occlusion ...... 50

Table 4.1: Electrophysiologic effects of SIN-1 in cardiac myocytes ...... 103

xix

List of Figures

Figure 1.1: ECG tracing. Letters represent their respective waves...... 22

Figure 1.2: Representative atrial and ventricular action potential tracings ...... 23

Figure 2.1 The effect of 10-week exercise training (n=19) on the and responses to submaximal exercise in animals susceptible to VF...... 48

Figure 2.2 Exercise training normalized repolarization abnormalities in dogs susceptible to VF...... 49

Figure 2.3: Exercise normalizes repolarization abnormalities in myocytes from animals susceptible VF ...... 51

Figure 2.4: Effects of endurance exercise training on potassium currents...... 52

Figure 2.5: Western blot analyses of subunit expression...... 54

Figure 2.6: Effect of exercise training on calcium spark frequency...... 55

Figure 2.7: Effect of endurance exercise training on the incidence and amplitude of calcium alternans...... 56

Figure 2.8: Western blot analyses of calcium handling proteins...... 57

Figure 2.9 : Phosphorylation of ryanodine receptor protein...... 58

Figure 3.1: Patient ECG after syncopal episode, and nine months after discontinuation of ibandronate; ibandronate discontinuation normalized the QT interval...... 76

xx

Figure 3.2: Ibandronate prolongs the action potential, causes EADs and decreases Ito current density and conductance in left ventricular cardiomyocytes...... 77

Figure 3.3: Inhibition of calcium cycling prevents ibandronate-induced action potential prolongation and action potential variability...... 79

Figure 3.4 Ibandronate treatment increases SR calcium load and washout induces spontaneous calcium waves...... 80

Figure 3.5: Ibandronate washout increases calcium spark frequency...... 81

Figure 3.6 (Supplemental): Ibandronate increases repolarization instability in vivo...... 82

Figure 3.7 (Supplemental): Calcium chelation prior to ibandronate treatment prevents the ibandronate-induced increase in short term variability of repolarization...... 83

Figure 3.8 (Supplemental): Mathematical modeling of electrophysiological changes with ibandronate-mimicking conditions...... 84

Figure 4.1: SIN-1 prolongs the atrial and ventricular action potential...... 101

Figure 4.2: SIN-1 affects atrial and ventricular repolarization ...... 102

Figure 4.3 : Peroxynitrite-induced changes in atrial and ventricular repolarization and variability of repolarization are prevented with thapsigargin pre-treatment...... 104

Figure 5.1 : In Vivo data from 1 month (1 Mo), 4 months (4 Mo), and heart failure with sustained AF (4 Mo HF + AF) canine groups...... 123

Figure 5.2 : IKCa inhibition does not alter repolarization in control ventricular cells. .... 124

Figure 5.3: Apamin modulates ventricular repolarization during HF...... 125

Figure 5.4: IKCa contributes to ventricular repolarization stability in canine HF, and HF increases SK3 expression...... 126

xxi

Figure 5.5 : Apamin modulates ventricular repolarization in end stage human HF...... 127

Figure 5.6: IKCa block does not affect atrial repolarization in normal or diseased myocytes...... 129

Figure 5.7 : SK expression and calcium transients in chronic HF with and without AF.130

Figure 5.8 : Apamin does not modulate repolarization in end-stage human HF atrial myocytes...... 131

Figure 6.1: In Vivo data from one and four month HF and HF + AF canine groups ..... 148

Figure 6.2: Chronic but not acute HF causes a shortening of the atrial action potential, while AF superimposed on chronic HF doesn't cause any further change...... 149

Figure 6.3: Four month HF and HF+AF alter I and I ...... 150 K1 Ksus

Figure 6.4: Chronic HF causes a significant increase in I density while superimposed AF to significantly reduces I vs. control...... 151 to

Figure 6.5: Chronic heart failure alters I inactivation kinetics...... 152 to

Figure 6.6: Atrial fibrosis occurs early in HF, and superimposed AF causes more fibrosis than chronic HF alone...... 153

xxii

Chapter 1 : Introduction

History of electrophysiology

From Pulse theory to the ECG discovery

It is “common knowledge” today that the mammalian heart is composed of four chambers and that its main job is to pump blood to the rest of the body. However, the

“common knowledge” that we have today wasn’t so common centuries ago and it took wise and careful observation and experimentation by many people throughout many years to understand the basic function of the heart as we do today. In the following paragraphs I will discuss some points of interest in the history of the electrophysiology and some of the key points in the study of the heart that lead to the understanding that we have today of the heart and its function.

The first known description of the relationship between the heart beat and the peripheral pulse was given in chapter 99 of The Papyrus Ebers, among the oldest known medical documents, thought to be composed between 1553 and 1550 B.C. in Egypt. It was not until the 5th century B.C. in China that Pien Ts’io started using the pulse beat as a diagnostic and prognostic criterion of pathology, by comparing the to a stringed instrument, with the corresponding to the strings and tones making it

1 possible to detect harmony and dissonance. In ancient Greece, Herophilos of Chalkedon

(300 B.C.) extended this musical analogy and described the pulse and its variations with respect to its volume, rate and rhythm. In order to get an accurate measurement he measured the pulse beats of patients with a water clock and compared it to the count of a healthy subject of the same age. 1

Many years later Claudius Galen (129 A.D.), who believed that the pulse was a measure of the arterial and and originated from the expansion of the heart, described many types of pulses and believed that every organ of the body and every disease was associated with its own unique type of pulse. This view was unchanged until the Middle Ages when William Harvey described the in a publication titled Exercitatio anatomica de motu cordis et sanguinis in animalibus

(An Anatomical Exercise on the Motion of the Heart and Blood in Living Beings) in

1628. In that publication he refuted the previously held views regarding the function of the heart and venous system in experiments with cold and warm blooded animals. This work was one of the first steps toward our modern understanding about the circulation and the peripheral pulse. It was not until Santorio Santorio an associate of Galileo invented the pulsimeter or pulsilogium (around 1624) that the pulse became a quantitative instead of qualitative measure, making it a more accurate diagnostic tool. The pulsilogium consisted of a pendulum, where its string could be lengthened or shortened until the pendulum swung in time to the pulse beats. The length of the string was used as an objective measurement of the pulse.1

2

Modern efficient studies of cardiac rhythm began after the development of . Augustus Desiré Waller who dedicated his studies to cardiac electricity obtained the first human electrocardiogram (ECG) from the human body surface in 1887 by using a Lippmann capillary electrometer, giving rise to the foundation for modern electrophysiology.1, 2 Willem Einthoven advanced Waller’s capillary image by using the string and identified 4 waves in the ECG but after applying mathematical corrections he increased the number of distinct waves in the ECG to 5 and gave them the designations that we still use today (P, Q, R, S, T). 2 This work of

Einthoven marked the beginning of the study and understanding of cardiac rhythm disorders, cardiac impulse formation and the cardiac conduction system. However, it wasn’t until Karel Frederik Wenckebach published his work (1898-1901) on cardiac arrhythmias and premature beats that cardiac arrhythmias were viewed as a clinically significant event. Additionally, due to his report of the efficacy of the antiarrhythmic drug in a patient with atrial fibrillation, described in a work titled Irregular

Cardiac Activity and Its Clinical Significance, Wenckebach is also considered to be the founder of modern pharmacological therapies for arrhythmias.1 However, the use of antiarrhythmic substances has a long story as the use of a cardiac glycoside-containing plant was described in the Papyrus Ebers (circa 1550 B.C.).

The birth of cellular electrophysiology

Around 1780 the Bolognese physiologist Luigi Galvani took the first step towards the discovery of the action potential by demonstrating the presence of electricity in animal tissue.3 However it was not until 1865 that Julius Bernstein with the help of Emil 3 du Bois-Reymond first recorded the action potential.3 Later in 1939 and

Alan Hodgkin formed the collaboration that lead to the discovery of the role of voltage gated channels in electrophysiology. By inserting a fine capillary electrode inside the giant axon of a squid they were able to record the potential difference across the membrane leading to the first recording of an intracellular action potential.4 Then, in

1949 they modified the already existing voltage-clamp technique by using a dual electrode approach and were able to record the first family of currents.5 Later in 1963, they won the in or Medicine for their discoveries concerning the ionic mechanisms involved in excitation of the cell membrane. Not only that, but their work also laid the foundation for other Nobel Prize winning work.

In 1949 Silvio Weidmann and Edouard Coraboeuf working in the laboratory of

Hodgkin and Huxley recorded the first intracellular cardiac action potential from canine right ventricular endocardium false tendon (Purkinje fiber) using glass microelectrodes,6,

7 giving rise to the era of cardiac cellular electrophysiology. In further work by Silvio

Weidmann, he described the dependence of the cardiac action potential upstroke on sodium and the resting on potassium ions. Additionally,

Weidmann described the relationship between the action potential and ionic current flow

6, 8, described all or nothing repolarization 6, 9-11, cell to cell communication and showed that local anesthetics can depress the action potential upstroke leading to propagation block.

4

The phases and function of the ECG and mammalian cardiac action potential

The pumping function of the heart depends on the electrical excitation and subsequent contraction of cardiac myocytes which are translated to mechanical function of the whole organ allowing the heart to contract and subsequently relax in order to pump blood to meet the physiological needs of the body. The electrical activity of the heart can be measured at the body surface with an ECG as seen in Figure 1.1. Each normal electrical signal originates in the (SAN) or pacemaker, located in the right . The SAN of an average healthy adult human has spontaneous electrical depolarizations between 60-100 times a minute at rest. After the electrical signal is generated in the SAN it travels through the right and left atria causing the atria to contract and subsequently move blood from the atria to the ventricles. The movement of electrical excitation through the atria results in atrial and is recorded as the

P wave in the ECG.

After depolarizing the atria, the electrical signal passes to the ventricles through the atrioventricular (AV) node where the signal is relatively slowly conducted, permitting time for the ventricles to fill with blood. This process is seen on the ECG as a flat line between the end of the P wave and the beginning of the Q wave.

Upon exiting the AV node, the electrical signal then passes down the bundle of

His which divides into the right and left to excite the right and left ventricles, respectively. The latter step depolarizes the ventricles and allows them to contract (systole) and pump blood to the ventricles and the rest of the body and represented in the ECG as the QRS complex.

5

When the ventricles repolarize, this is seen as the on the ECG.

Ventricular diastole (relaxation) follows repolarization and the ventricles relax and refill with blood and the process starts all over again.

The ECG represents the summation of cardiac myocyte action potentials (AP) recorded at the body surface. The AP is propagated from cell to cell in the heart allowing the heart to function as a , causing coordinated electrical excitation and contraction. The action potential is formed by a complex interplay between membrane ionic currents and the ionic milieu. The action potential is described by different phases:

0, 1, 2, 3 and 4. (Figure 1.2)

In atrial and ventricular myocytes the rapid upstroke or rapid depolarization of the action potential (Phase 0) results from the activation of voltage-gated sodium channels giving rise to the inward current known as INa. Phase 0 is followed by a transient repolarization or Phase 1 which is caused by inactivation of the voltage gated sodium channels and activation of the transient outward potassium channels (responsible for the current known as Ito). This transient repolarization contributes to the notch in the action potential, which is more prominent in ventricular cells than atrial cells (Figure 1.2), and influences phase 2 or action potential plateau height.12 During phase 2 of the action

2+ potential, activation of the L-type Ca (ICaL) channels occurs. The opening of the L-type calcium channels during this phase allows calcium ions to move into the cell and subsequently trigger what is known as calcium induced calcium release from intracellular calcium stores, the . Calcium induced calcium release occurs by triggering the opening of the Ryanodine Receptor (RyR, calcium release channels) which

6 releases the stored intra-sarcoplasmic reticulum Ca2+ into the cytoplasm resulting in the activation of the contractile proteins and subsequent cell contraction. This process is known as excitation contraction coupling. As the L-type calcium channels inactivate late in phase 2, the delayed outward rectifier K+ currents activate resulting in repolarization or

+ Phase 3 of the AP. The main repolarization currents are outward K currents IKr (rapid),

IKs (slow) and IKur (ultra rapid). The contribution of IKur and IKr to repolarization of atrial and ventricular myocytes differs, with IKur being a significant repolarizing current in atrial cells, with IKr serving as the main repolarizing current in ventricular myocytes. The final repolarization during late phase 3 is due to the outward movement of K+ ions through the inward rectifier IK1 channels.

Phase 4 or the resting membrane potential of atrial and ventricular myocytes results from the high inward conductance of K+ ions through the inward rectifying current

IK1 keeping the resting membrane potential stable at (approximately -80 mV).The membrane potential of excitable cells, specifically cardiac myocytes, is established by the concentration difference of K+ ions across the semipermeable sarcolemmal membrane where the concentration outside of the membrane is less (~5.4 mM) than inside of the cell (~120 mM). The equilibrium potential refers to the voltage where the driving force to move potassium across the is balanced, that is, there is no net flux of ions across the membrane; and the equilibrium potential is very close to the as permeability to potassium is high compared to other ions during diastole. Thus, variations in extracellular potassium concentration have an important effect on the resting

7 membrane potential as increased extracellular potassium depolarizes the membrane and reductions in extracellular potassium hyperpolarizes it.

Potassium channels

Cardiac potassium channels can be structurally grouped into two main categories: voltage-gated (Kv) channels and inward rectifier channels (Kir). Inward rectification means that the channels preferentially conduct inward current rather than outward current.

Voltage gated potassium channels

Voltage gated potassium (Kv) channels are the primary determinants of action potential repolarization. They consist of pore forming α-subunits and accessory β- subunits. Each α-subunit contains six transmembrane segments (S1-S6) and each pore is formed by a tetrameric arrangement of α-subunits. The voltage sensing properties of these channels is dictated by the highly conserved positively charged fourth transmembrane segment (S4) of the α-subunit, whereas the pore formation and ion selectivity are conferred by the S5-S6 transmembrane domain of the α-subunit. Voltage gated potassium channels fluctuate between different conformational states, in voltage- and time-dependent manners. The voltage and time dependent transition between channel conformational states is called gating. For example, during the AP, closed voltage- sensitive K+ channels activate (open) in response to membrane depolarization and then enter an inactive state in a time-dependent manner. The transition from inactive to the closed state is voltage and time-dependent.

8

Kv channels can be categorized into two classes of repolarizing currents, based on

+ differential time and voltage dependent properties: transient outward K (Ito) currents, responsible for phase 1 of the AP and outward delayed rectifying K+ currents, responsible for phase 3 of the AP.13

The transient outward potassium current (Ito) is encoded by channels composed

14 of four Kv α-subunits from the Kv1.4 (Itos) or Kv 4.3 (Itof) subfamilies. In addition to being classified based on the molecular components Itos and Itof can also be classified based on the channels kinetics of activation and recovery from inactivation.

Generally, Ito activates and inactivates rapidly (transiently open) upon membrane depolarization, compared to other currents. Ito has two components, fast and slow, which are designated as Ito,f and Ito,s. The two components have comparable activation constants, but the time constants for inactivation and the recovery time from steady-state

15 inactivation differs, with Itos having a slower recovery from inactivation than Itof. Ito,s density is smaller than Ito,f in most mammalian species including ; therefore, it is assumed that Ito,f is the primary determinant of the action potential notch or phase 1 in both atrial and ventricular myocytes. Several studies have shown that Ito,f expression is higher in the epicardium, the right ventricle and the base of the ventricles compared to

14, 16 other parts of the heart. The regional variations in Itof expression contribute to the regional heterogeneity of the AP waveform, and contribute to coordination of ventricular repolarization. Additionally, since Ito,f is responsible for the early repolarization phase of the AP, changes in the current may also affect L-type Ca2+ current magnitude therefore contributing to modulation of excitation contraction coupling and cardiac contractility.17

9

Itof channel expression and current density in the heart can be modulated by multiple ancillary proteins such as KChIP2 and DPP6 and signaling molecules such as PKC, and reactive species, among others. 14, 18Additionally, several neurohumoral and hormonal factors have been shown to modulate the current magnitude and/or channel expression of Ito,f, for example: , II, α1- adrenergic agonists.14, 19

The composite cardiac delayed rectifier potassium current (IK) has a relatively slow or delayed activation compared to other cardiac potassium currents and is the main modulator of phase 3 of the cardiac AP. Ventricular IK was initially separated into IKr and

20 IKs using pharmacologic approaches. At the present time, there are three different known cardiac delayed rectifier currents, known as IKur (ultra-rapid), IKr (rapid) and IKs

(slow), each with different time and voltage-dependent activation properties.

As with Ito channel expression, regional variations and differential expression patterns of the delayed rectifier ion channels IKur (Kv1.5), IKr (hERG), and IKs (KvLTQ1 coassembled with MinK), play an important role in action potential heterogeneity.21, 22

The relatively shorter action potential in the atria is attributed to high levels of expression of Kv1.5/IKur (relative to the ventricle); IKur is “ultra-rapid” in activation and very slowly inactivating current and is a major repolarizing current in the atria. Meanwhile, in the ventricle the main repolarizing currents are IKr and IKs. IKr activates and inactivates very rapidly and due to its voltage sensitivity the channels has a relatively small conductance during the phase 2 of the ventricular AP but increases during the phase 3 (with a half activation voltage of -21.5 mV and a voltage range of -20mV to + 40mV 20). In contrast

10

IKs has slower activation and inactivation kinetics and its conductance increases gradually during phase 2 and remains activated during phase 3 of the AP(with a half activation voltage of +15.7 mV and a voltage range of -20mV to + 50 mV20). The contribution of

IKur to ventricular repolarization has been shown to be very small or sometimes undetectable.22, 23 However, its contribution to ventricular repolarization may become more evident in disease states where the major repolarizing currents are decreased.

Delayed rectifier channel expression and current magnitude are modulated by several hormonal and neurohumoral factors such as α and β adrenergic stimulation, PKA, and thyroid hormone.

Inward rectifier channels

Inward rectifier potassium (Kir) channels preferentially pass K+ current in the inward rather than the outward direction. The inward rectifier potassium channel family is composed of seven subfamilies (Kir1 to Kir7). The pore forming α-subunits of these channels form a tetrameric complex with each monomer formed by two transmembrane segments, M1 and M2 connected by a pore forming P loop. The Kir channels with the most pronounced inward rectification belong to the families (Kir 2.1-2.4) which underlie the cardiac IK1. Specifically, in the heart IK1 plays a major role in terminal repolarization and setting the resting membrane potential (phase 4) of the AP. Thus, blockade of IK1 with Ba2+ depolarizes the membrane and prolongs the AP, demonstrating the functional

24 role of IK1 in AP morphology. As occurs with other ion channels there are some chamber specific expression differences in IK1, with the ventricle having a higher density of IK1 compared to the atria. The inward rectification exhibited by these channels results 11 from blockade of the channel pore by intracellular Mg2+ and the endogenous intracellular polyamines (spermine, putrescine, spermidine).25 Similar to Kv channel-encoded current, some reports suggest that Kir channel-encoded current, IK1, can also be modulated by angiotensin II, β-adrenergic stimulation and PKA.26, 27

In addition to Kir 2.x channels, other weaker inward rectifier channels are also present in the myocardium. For example IKACh channels, encoded by Kir 3.x, have been shown to regulate heart rate and conduction in response to , and are thought

28 to be constitutively active in atrial fibrillation. IKATP channels (encoded by Kir 6.x) which respond to changes in metabolic state can become an important player during ischemia. ATP blocks the channels thus, under physiologic conditions where the cells

ATP concentration is normal, the channel is closed. However, in conditions where the cell has depleted intracellular ATP, such as occurs in ischemia, IKATP gets activated leading to shortening of the action potential duration.

The Calcium activated Potassium Channels (IKCa)

IKCa is a non-voltage gated calcium activated potassium current, and has been reported to modulate cardiac repolarization. It was not until the last decade that this

29 current was considered as a potential modulator of cardiac repolarization. IKCa current is encoded by small-conductance calcium activated potassium channels or SK channels.

These channels, as the name implies, have a small conductance (10-20pS) and are gated by submicromolar concentrations of intracellular calcium. There are three different types of SK channels SK1, SK2 and SK3. Functional SK channels assemble as homo-tetramers and are differentially blocked by the bee venom apamin with SK2 being the most

12 sensitive (EC50 ~40 pM), SK1 the least sensitive (EC50 ~ 10nM) and SK3 having

30, 31 intermediate sensitivity (EC50 ~ 1nM). The calcium gating properties of the SK channel are due to a highly conserved binding domain (CaMBD) which is found across the SK family.32 Each subunit of the tetrameric channel binds one CaM and the subsequent binding and unbinding of Ca2+ ions to CaM causes a conformational change that allows channel opening or closure.32 In chapter 5 I will discuss my work on the contribution of SK channels to cardiac repolarization in atrial and ventricular myocytes and explain their importance in heart failure.

K+ channel modulation and ventricular arrhythmias

Due to the role of potassium channels in regulating cardiac repolarization these channels play an important role in the formation of cardiac arrhythmias. Although, changes in potassium channels can occur by gain or loss of function most of the disease- associated changes in potassium channels are seen as decreased current. The reduced currents can result from many factors such as channel blockade by a drug, decreased expression of the pore forming subunit and/or auxiliary proteins, altered regulatory pathways, and altered cellular metabolism, among others. 33, 34 Repolarizing potassium channel downregulation can lead to action potential prolongation, which in turn can promote the formation of cellular arrhythmias such as delayed after depolarization

(DADs) and/or early after depolarization (EADs) which can trigger lethal ventricular arrhythmias.35

Many potassium channels are known to be modulated by reactive oxygen species

(ROS) and reactive nitrogen species (RNS) commonly known as oxidant species. 13

Oxidant species are normally occurring products of cellular metabolism through multiple sources. Cellular anti-oxidant defenses are used to buffer the ROS and RNS and prevent oxidative damage. However, when the balance between oxidant production and anti- oxidant capacity is disrupted, as occurs in many cardiac diseases, oxidative stress occurs.

Some sources of oxidant species are the mitochondria, xanthine oxidase, induction of nitric oxide synthase, uncoupling of nitric oxide synthase and NADPH oxidase, while the main anti-oxidants are superoxide dismutase and glutathione. Oxidants are known to modulate calcium handling proteins therefore, affecting calcium cycling in cardiac myocytes though post-translational modification. Additionally several studies have

36-46 demonstrated that oxidants also modulate IK1, INa, IKr, Ito, IKur and ICaL, therefore, contributing to arrhythmia formation. Major forms of altered potassium channel function are discussed in more detail below.

Acquired K+ channel remodeling

Some forms of cardiac diseases alter cardiac function, which may be part of an adaptive response to the primary abnormality, but may lead to maladaptive secondary cardiac dysfunction.47

K+ current blockade and drug induced Torsade de Pointes (TdP)

Drug-induced modulation of potassium channels can interfere with cardiac repolarization and induce triggered activity. Drug-induced cardiac arrhythmias can arise from abnormalities of inward currents or outward currents. For the purposes of this review the focus will be on the blockade of outward K+ currents. Inhibition of outward

14 repolarizing current(s) usually gives rise to a prolongation of the action potential duration which in turn increases the chances of EADs due to a decrease in repolarizing K+ current.

Prolongation of the ventricular action potential can translate to prolongation on the ECG of the QT interval (ventricular repolarization). Since drug-induced prolongation of the action potential is usually not homogeneous throughout the thickness of the ventricular wall (attributed in part to inhomogeneities in K+ channel distribution), the result may be increased transmural dispersion of repolarization which provides a substrate for the development of dangerous ventricular arrhythmias such as torsade de pointes (TdP).48

Drug-induced APD prolongation can occur as a result of a drug effect on one or multiple ion channels. However, the most common effect of drug induced APD and QT

49 prolongation is the blockade of the rapid component of the delayed rectifier current IKr.

Of particular interest is the association of IKr blockade and the genesis of life-threatening ventricular arrhythmias such as TdP. The first class of drugs linked to the formation of this lethal arrhythmias were the anti-arrhythmic drugs with class III action, which may block various K+ channels to prolong the AP.

However, the biggest concern is the growing number of non-cardiovascular drugs that also have been shown to cause TdP mainly attributed to IKr blockade, a phenomena often referred to as cardotoxicity.50 Around 27% of recent drug withdrawals from the market have been due to cardiotoxicity, which is often a barrier in the development of

51 new drugs. Thus, drug companies focus on assessments of IKr blockade in new drugs early in the development process, mostly relying on in vitro assessment of IKr blockade and in vivo models to evaluate QT prolongation on the surface ECG.52 However, the

15 increasing number of drug withdrawals from the market due to ventricular arrhythmias suggest the need for more sensitive ways to identify proarrhythmic risk. Recently developed techniques such as beat to beat variability in the ventricular action potential and QT interval variability have shown a higher sensitivity to identify proarrhythmic drugs and suggest the potential for better ways to identify and prevent drug induced cardiotoxicity.53

K+ channel remodeling in ischemic heart disease and its contribution to SCD

Ischemic heart disease occurs primarily when there is death of cardiac tissue, commonly known as myocardial infarction (MI), due to a critical decrease in coronary blood flow. MI is a risk factor for sudden cardiac death due to ventricular arrhythmias.54

Ion channel remodeling is responsible for changes in the electrical activity of ventricular myocytes, and MI is known to cause changes in ion channel expression and function. The

MI border zone has been described as a site of ionic remodeling in animal models, and the border zone differs from other “normal” zones of the ventricle following MI.55, 56

Here, the discussion will be focused on changes that occur in the normal zones of the heart (non-ischemic) after a cardiac insult rather than the infarct border zone.

Two of the most common post-MI remodeling features in the non-ischemic area of the ventricle from several animal models such as rat, rabbit and canine are: action potential duration prolongation and cellular arrhythmias.57-59 Common manifestations of post-MI ion channel remodeling are decreases in Ito, IK1, and in total delayed rectifying

57-59 currents (IKs and/ or IKr). However, the channel protein changes responsible for the 16 observed ionic current changes are less uniform between animal models making it challenging to pinpoint a specific unifying mechanism for the observed electrophysiological remodeling. For example Ito decreases in post MI rats and correlates to a decrease in the pore forming protein sub unit Kv4.2.58 In contrast, in canines reduced

59 Ito is attributed to a decrease in the auxiliary protein KChiP2. In addition to potassium channel remodeling it is also important to note that changes in calcium handling, calcium current and sodium current are also known to contribute to the electrophysiological changes that occur in both animal and human post MI ventricular myocytes.58, 60-62

Additional physiological changes that occur post-MI and which can contribute to arrhythmogenesis are: increased oxidative stress, fibrosis and altered cell to cell coupling.63, 64

Pharmacological approaches for K potassium channel manipulation

Due to the high risk of ventricular arrhythmia in patients with MI the most common treatment includes defibrillator implantation. Unfortunately, the number of drugs available to reduce arrhythmias in post MI patients is limited due to an increased risk of pro-arrhythmias in these patients. Other forms of treatment which may indirectly modulate K+ currents include drugs with antioxidant properties (e.g. , beta- blockers such as carvedilol)65, 66 or exercise.67 Additional discussion about the effects of exercise in K+ channels in a post-MI canine model will covered in chapter 2.

17

K+ channels remodeling in HF

Heart failure (HF) refers to the impaired ability of the heart to pump sufficient blood to meet the metabolic needs of the body. About 50% of among HF patients occur due to ventricular arrhythmias and sudden cardiac death.68 In addition to ventricular arrhythmias, patients with heart failure have an increased risk of developing atrial arrhythmias. There are different forms of HF however in this review I will focus on dilated , which is discussed further in chapters 5 and 6. In , the heart cannot contract properly (systolic dysfunction) and in order to compensate, several physiological and neurohormonal changes occur. However, with time those changes become detrimental. As the name says the hallmark of dilated cardiomyopathy is the development of a thin ventricular wall and dilated ventricular chamber. In addition to the physiological changes that occur in dilated cardiomyopathy, electrophysiological remodeling also occurs in the cardiac myocytes. Differential changes between atria and ventricle occur in the setting of dilated cardiomyopathy, and described further below and in chapters 4, 5 and 6.

Atria

Structural and electrophysiological remodeling in the atria in HF are key factors in the development of atrial fibrillation. Some of the structural changes include: increased left atrial diameter, thinning of the atrial wall and increased fibrosis.69, 70 Several studies in tachypacing induced HF canine models are contradictory in terms of the electrophysiological remodeling; these differences may be attributed to the differences in the duration of the HF (5 weeks vs. 16 weeks). In five week tachypacing induced HF 18

71, 72 there is a prolongation of the atrial action potential, and decreases in Ito and IKs.

Meanwhile, the changes observed in the 4 month HF canine model are a shortening of the atrial action potential accompanied by an increase in Ito and decreases in IK1, IKur, and

73 IKs. In both durations of HF, the rapid component of the delayed rectifying current IKr, is unchanged. Human studies from failing atrial myocytes have demonstrated that Ito increases with worsening ventricular function which is consistent with the chronic (4 month) HF canine model.74

Ventricle

Electrophysiological and structural remodeling during the progression of HF contributes to the formation of deadly ventricular arrhythmias. Dilated cardiomyopathy is characterized by left ventricular enlargement, thinning of the wall and an increase in interstitial fibrosis. Of special interest are the electrophysiological changes that occur during HF since most pharmacological antiarrhythmic approaches target these changes.

The most common finding in HF ventricular cells from humans and animal disease models is a decrease in Ito, IK1, and IKs accompanied by a prolongation in the action potential.75-77 Several changes in calcium and sodium currents have also been described78,

79 and are outside the scope of this review. IKr changes are variably reported in the literature, with some studies finding a decrease and others finding no change.75-77

Understanding the electrophysiological changes that occurs in the atria versus the ventricle might provide “ventricle specific or atrial specific” pharmacologic targets thus,

19 avoiding nonspecific or undesirable effects in the atria while preventing or treating ventricular arrhythmias.

K+ channel remodeling in atrial fibrillation

Atrial fibrillation causes irregular, disorganized excitation of the atria. It is the most common sustained arrhythmia in the developed world and is a source of cardiovascular morbidity and mortality.80 The high atrial rate during AF itself contributes to further remodeling and maintenance of AF.81 As with the cardiac diseases discussed above, AF also causes structural and electrophysiological remodeling. Some of the structural remodeling includes increased oxidative stress and fibrosis accompanied by .82, 83 The most common change is a reduction in the transient outward

+ 84, 85 K current density, Ito, in human AF and animal AF models. Less clear are the changes that occur in IKur, IKr and IKs since some studies report increases in current density meanwhile others report no change.84, 86, 87Important changes in inward rectifier

+ K currents IK1 and IKach have also been reported, with atrial myocytes from AF patients

28, 84, 87-89 and animal models showing significant increases in IK1 and IKach current density.

Anti-arrhythmic drugs

Because of the major contribution of potassium channels to repolarization it is not surprising that many antiarrhythmic drug therapies available today target K+ channels.

However even though these drugs have been found to be beneficial for arrhythmia treatment they are among the most dangerous pharmacological agents due to their potential for serious adverse effects such as proarrhythmias.

20

Antiarrhythmic drugs are commonly classified using the Harrison’s revised version of Vaughan Williams & Singh classification90-93 into four different groups based on their mechanism of action. Briefly, Class I drugs block sodium channels (further subdivided into IA,IB and IC denoting the degree of channel block as moderate, mild and marked), Class II drugs are β- blockers, Class III drugs prolong the action potential mainly through K+ channel blockade, and Class IV drugs that block L- type calcium channels. As K+ channels are the major focus of this dissertation, I will focus here on class III antiarrhythmic drugs and their mechanism(s) of action. Drugs classified as Class III all prolong the action potential thus prolonging the refractory period and decreasing reentrant arrhythmias; effects that are independent of conduction velocity. Due to the powerful modulation of ventricular repolarization by these drugs, and their use in patients with acquired forms of electrical remodeling, they have to be carefully monitored due to the risk of proarrhythmia. Therefore, it is imperative to find new targets and drug therapies with decreased risks of proarrhythmia.

During the course of my thesis I will expand on mechanisms of arrhythmia formation which may elucidate new drug targets, discuss alternative ways to decrease arrhythmias whether ventricular and/or atrial in origin, and talk about the sensitivity and importance of detecting proarrhythmic drug risk, with consideration of diseases such as

HF, AF and post- MI arrhythmias.

21

Figures

Figure 1.1: ECG tracing. Letters represent their respective waves.

22

Figure 1.2: Representative atrial and ventricular action potential tracings

Schematic representation of the main currents that contribute to the action potential in atrial (left panel) and ventricular (right panel) canine myocytes. For clarity inward and outward currents are depicted in green and purple, respectively. The phases of the action potential (zero to four) are labeled on the representative action potential waveforms. Modified version from: Nerbonne JM, Kass RS Molecular physiology of cardiac repolarization. (2005) Physiol Rev 85:1205-1253.

23

Chapter 2 : Endurance Exercise Training Normalizes

Repolarization and Calcium Handling Abnormalities

Preventing Ventricular Fibrillation in a Model of

Sudden Cardiac Death

Ingrid M. Bonilla1, Andriy Belevych2, Arun Sridhar1, Yoshinori Nishijima1, Hsiang-Ting

Ho2, Quanhua He1, Monica Kukielka2, Dmitry Terentyev2, Radmila Terentyeva2, Bin

Lu2, Victor P. Long1, Sandor Györke2, Cynthia A. Carnes1,2 and George E. Billman2

1College of Pharmacy, 2Department of Physiology and Cell , and the Dorothy M. Davis Heart and Research Institute, The Ohio State University, Columbus, Ohio 43210 Corresponding Author: Cynthia A. Carnes College of Pharmacy 500 W. 12th Avenue Columbus, OH43210 PH: 614-292-1715 FAX: 614-292-1335 Email: [email protected]

Previously Published: J Appl Physiol. 2012 Dec;113(11):1772-83

24

Abstract:

The risk of sudden cardiac death is increased following myocardial infarction.

Exercise training reduces arrhythmia susceptibility, but the mechanism is unknown. We used a canine model of sudden cardiac death (healed infarction, with ventricular tachyarrhythmias induced by an exercise plus ischemia test, VF+); we previously reported that endurance exercise training was antiarrhythmic in this model (Billman GE.

Am J Physiol Heart Circ Physiol 297: H1171–H1193, 2009). A total of 41 VF+ animals were studied, after random assignment to 10 wk of endurance exercise training (EET; n =

21) or a matched sedentary period (n = 20). Following (>1 wk) the final attempted arrhythmia induction, isolated myocytes were used to test the hypotheses that the endurance exercise-induced antiarrhythmic effects resulted from normalization of cellular electrophysiology and/or normalization of calcium handling. EET prevented VF and shortened in vivo repolarization (p < 0.05). EET normalized action potential duration and variability compared with the sedentary group. EET resulted in a further decrement in transient outward current compared with the sedentary VF+ group (p< 0.05). Sedentary

VF+ dogs had a significant reduction in repolarizing K+ current, which was restored by exercise training (p< 0.05). Compared with controls, myocytes from the sedentary VF+ group displayed calcium alternans, increased calcium spark frequency, and increased phosphorylation of S2814 on ryanodine receptor 2. These abnormalities in intracellular calcium handling were attenuated by exercise training (p < 0.05). Exercise training

25 prevented ischemically induced VF, in association with a combination of beneficial effects on cellular electrophysiology and calcium handling.

Keywords: arrhythmia, myocardial infarction, ion channels, sarcoplasmic reticulum

26

Introduction:

Sudden cardiac death (SCD) remains the leading cause of death in industrialized countries. The risk of SCD is increased in patients who have survived a myocardial infarction (MI).54, 94 Multiple lines of evidence suggest that abnormalities in balance are associated with an increased risk of sudden death.95,

96Specifically, reduced parasympathetic or increased sympathetic activity increases the risk of ventricular tachyarrhythmias.97, 98

Endurance exercise training is well known to alter autonomic nervous system activity favorably,99, 100 by increasing parasympathetic regulation and by reducing sympathetic nervous system activity, actions that could be antiarrhythmic.99 Endurance exercise training has previously been reported to improve cardiac autonomic balance and to prevent malignant arrhythmias in a canine model of post-MI sudden death.99, 101

Similar results have been reported in other animal models.102, 103 Importantly, exercise training has also shown to improve autonomic balance and reduce the incidence of sudden death in patients with myocardial infarction or . 104 The mechanisms that mediate the possible antiarrhythmic effects of endurance exercise training remain to be elucidated fully.

Abnormal repolarization (due to altered potassium currents) or intracellular calcium (Ca2+) dysregulation may occur individually or in concert to decrease cardiac electrical stability and increase the propensity for SCD. It is well established that MI

27 induces repolarization abnormalities that can provide an electrophysiological substrate for ventricular arrythmias.59, 96 In a similar manner, abnormal Ca2+ regulation during acute myocardial ischemia or following myocardial infarction elicits elevations in cytosolic Ca2+ (Ca2+ overload) that can provoke oscillations in membrane potential

(delayed afterdepolarizations) that, if of sufficient magnitude, can trigger extrasystoles.105

The effects of endurance exercise training on repolarization abnormalities or ventricular myocyte Ca2+ regulation have not been extensively studied. There are only a few studies providing indirect evidence that exercise training can alter repolarization.

Aerobic conditioning has been shown to reduce QT interval in young patients with long

QT syndrome (LQTS).106Exercise training has also been shown to reduce regional differences in repolarization in heart failure patients 107 and prolong the effective refractory period in rabbits.108 There is a similar paucity of information concerning the effects of exercise training on cardiac Ca2+ regulation. Further, the few published studies have often yielded conflicting results. For example, similar exercise training protocols have both increased and decreased sodium/calcium exchanger (NCX) activity.109 Thus, the anti- arrhythmic effects of endurance exercise training on repolarization abnormalities or ventricular myocyte Ca2+ regulation remain to be elucidated.

Our laboratory previously reported that ventricular myocytes obtained from post

MI animals susceptible to ventricular fibrillation (VF+) exhibited both repolarization abnormalities 59 and intracellular Ca2+ alternans.36 It is possible that exercise training could reverse these abnormalities and, as a consequence, protect against VF. Therefore, it was the purpose of the present study to test the hypothesis that exercise training would

28 improve cardiac electrophysiological properties and Ca2+ handling and thereby protect against VF in a well-characterized canine model of sudden death.

Materials and Methods

In Vivo Studies

All animal procedures were approved by the Ohio State University Institutional

Animal Care and Use Committee and conformed with the Guide for the Care and Use of

Laboratory Animals published by the US National Institute of Health (NIH publication

No. 85-23, revised 1996).

Surgical Preparation

Mixed breed dogs were anesthetized and instrumented as has been previously described 110-112. Briefly, a 20 MHz pulsed Doppler flow transducer and a hydraulic occluder was placed around the left circumflex coronary artery. Two bipolar insulated stainless steel pacing electrodes were also sutured on the epicardial surface of the heart and used to obtain a ventricular electrogram. One bipolar electrode was placed in the potentially ischemic area (posterolateral surface of the left ventricle, an area supplied by the left circumflex artery) and a non-ischemic region (anterior left ventricle proximal to the occluder). A two-stage occlusion of the left anterior descending artery was then performed distal to the first large diagonal branch in order to produce an anterior wall myocardial infarction. This vessel was partially occluded for 20 minutes and then tied off.

29

Exercise plus Ischemia Test: selection for susceptibility to malignant arrhythmia

The studies began 3-4 weeks after the production of the myocardial infarction. The susceptibility to ventricular fibrillation was tested as previously described110-112. Briefly, the animals ran on a motor- driven treadmill while workload progressively increased until a heart rate of 70% of maximum (approximately 210 beats/min) had been achieved.

During the last minute (on average during the 18th minute) of exercise, the left circumflex coronary artery was occluded, the treadmill stopped and the occlusion maintained for an additional minute (total occlusion time = 2 min.). The exercise plus ischemia test reliably induced that rapidly deteriorated into ventricular fibrillation in susceptible animals. Therefore, large defibrillation electrodes (Stat-padz,

Zoll Medical, Burlington, MA) were placed across the animal’s chest so that electrical defibrillation (Zoll M series defibrillator, Zoll Medical, Burlington, MA) could be achieved with a minimal delay but only after the animal was unconscious (~10-20 s after the onset of ventricular fibrillation). The occlusion was immediately released if ventricular fibrillation occurred. This exercise plus ischemia test was repeated, using the same exercise intensity, after the completion of the 10-week exercise training or 10-week sedentary time period (see below).

Exercise Training Protocol

Forty-one mixed breed dogs (male/female 2-3 years old) weighing 19.8± 0.3 kg developed VF during the exercise plus ischemia test. These susceptible (VF+) dogs were then randomly assigned to either a 10-week exercise training period (VF+ exercise, n =

21) or an equivalent sedentary period (VF+ sedentary n = 20). The dogs in the VF+

30 exercise group ran on a motor driven treadmill for 10-weeks, 5 days/week at approximately 70-80% of maximum heart rate.113 The exercise intensity and duration progressively increased as follows: 1st week, 20 min at 4.8 kph/0% grade; 2nd week, 40 min at 5.6 kph/10% grade; 3rd week, 40 min at 6.4 kph/10% grade; 4th week, 60 min at

6.4 kph/10% grade; 5th week, 60 min at 6.4 kph/12% grade; 6th week, 75 min at 6.4 kph/12% grade, 7th week, 90 min at 6.4 kph/12% grade; 8th – 10th weeks, 90 min at 6.4 kph/14% grade. Each exercise session included 5-min warm-up and 5-min cool-down periods (running at a low intensity, 0% grade and speed, 4.8 kph). The dogs in the VF+ sedentary group were placed in transport cage for equivalent time periods but without exercise.

The ECG data were digitized (1 kHz) and recorded using a Biopac MP-100 data acquisition system (Biopac Systems, Inc. Goleta, CA). The exercise data were averaged over the last 30 seconds of each exercise level. The coronary occlusion data were averaged over the last 5 seconds before, and at the 60 sec time point (or ventricular fibrillation onset) after occlusion onset. The ECG variables assessed were HR, heart rate variability, PR interval, QRS duration, QTc (Bazett’s correction factor),descending portions of the T wave (Tpeak-Tend) an index of the dispersion of repolarization.114, 115

Heart rate variability (0.24 to 1.04 Hz component of R-R interval variability; an index of cardiac index) was obtained using a Delta-Biometrics vagal tone monitor triggering off the electrocardiogram R-R interval (Urbana-Champaign, IL). This device employs the time-series signal processing techniques as developed by Porgesto estimate

31 the amplitude of respiratory sinus arrhythmia.116 Details of this analysis have been described previously.117

In vitro Studies

Myocyte Isolation

At the end of the 10 week study period, the VF+ sedentary (n = 14) and VF+ exercise (n = 19) dogs where anesthetized with pentobarbital sodium (50 mg/kg iv;

Nembutal, Abbott Laboratories). The heart was rapidly removed and perfused with cold cardioplegia (containing the following (in mM) 110 NaCl, 1.2 CaCl2, 10 mM NaHCO3,

16 KCl and 16 of MgCl2) injected into the coronary ostia. The left circumflex coronary artery was cannulated for myocyte isolation as previously described.23, 59 After washout of blood from the heart, collagenase (Worthington type 2; 1.15mg/ml) was added to the perfusate. After 30-45 min of enzyme , the digested midmyocardial section of the left ventricle was separated from the epicardial and endocardial sections, avoiding the infarcted area and the “border zone”. The typical yield of this procedure is 30-70% rod shaped myocytes with sharp margins and staircase ends. The myocytes were suspended and stored at room temperature in standard IB containing (in mM): 118 NaCl, 4.8 KCl,

1.2 KH2PO4, 0.68 Glutamine, 10 Glucose, 5 Pyruvate, 1 CaCl2, 1µM/L Insulin and 1%

BSA, pH adjusted to pH 7.35. A group of five mixed breed dogs (male/female, 2-3 years old) were used as a source of control myocytes. All myocyte experiments were conducted within 10 h of isolation.

32

Protein Expression

Potassium channel subunits and calcium handling proteins were assessed by immunoblot analysis. 20 mg protein from tissue homogenates were subjected to 4~15%

SDS-PAGE (Bio-Rad labs, CA, USA), blotted onto nitrocellulose membranes (Bio-Rad labs). Kv channel interacting protein 2 (KChip2), Anti-DPP6 and KV4.3 antibodies were from Santa Cruz Biotech (Santa Cruz, CA, USA). Anti-Kir2.1, Kir2.2 and Kir2.3 antibodies were from Alomone Lab (Jerusalem, Israel). Anti- RyR2 antibody was from

ThermoFisherScientific; Anti-NCX1 antibody from Milipore (CA); Anti- SERCA and

Anti-PLN were a gift from Dr. Mark Ziolo (The Ohio State University, Columbus OH),

Anti GAPDH antibody was used to measure GAPDH as a loading control (Abcam,

Cambridge, MA USA). Phosphorylation status of RyR2 was examined using phosphor- specific antibodies to S2808 and S2814 and normalized to total RyR2. Blots were developed with super Signal West Pico (Pierce) and quantified using ImageJ (NIH, USA) and Origin 8 (OriginLAb, Nothampton, MA, USA) software.

Electrophysiological studies

Amphotericin-B perforated patch clamp technique was used with a bath temperature of 36 ± 0.5 ºC as previously described.23, 59Myocytes were placed in a laminin coated cell chamber (Cell Microcontrols, Norfolk, VA) and superfused with bath solution containing (in mM): 135 NaCl, 5 MgCl2, 5 KCl, 10 glucose, 1.8 CaCl2, and 5

HEPES with pH adjusted to 7.40 with NaOH at temp of 36 ± 0.5 ºC. Borosilicate glass micropipettes with tip resistance of 1.5-3 MΩ, were filled with pipette solution containing the following (in mM): 100 K-aspartate, 40 KCl, 5 MgCl, 5 EGTA, 5 HEPES, 33 pH adjusted to 7.2 with KOH. Action potential duration (APD) data were obtained as the average of the last 10 traces (steady state) from 25 elicited at each stimulation rate. The standard deviation of the APD90 for the last ten traces (i.e. from trace 15-25) was used to evaluate repolarization variability.59

For potassium current measurements, nifedipine (2 µM) was added to the superfusate to block the L-Type calcium current. For voltage clamp experiments, only cells with access resistance less than 20 MΩ were included in the analysis. IK1 was elicited with a holding potential of -40 mV and 100 ms voltage steps from -140 mV to

+40mV and measured as the steady-state current. Inward slope conductance was calculated from the inward I-V relationship between -140 and -100 mV.(19; 48) Ito was elicited from a holding potential of -60 mV by a series of 100-ms test potentials from -20

23 to +50 mV, as previously described. IKr and IKs were elicited by voltage steps from -20 to +50 mV from a holding potential of -50 mV; IKr was separated pharmacologically by superfusion with d- (100 μM).23, 59

Ca2+ Handling measurements

Calcium currents were recorded using conventional whole-cell configuration of voltage clamp technique. For the calcium current recordings the external solution contained in mmol/L: 140 NaCl, 5.4 CsCl, 2.0 CaCl2, 0.5 MgCl2, 10 HEPES, and 5.6 glucose (pH 7.4). Patch pipettes were filled with a solution that contained in mmol/L: 123

CsCl, 20 TEA Cl, 5 MgATP, 5 NaCl, 1 MgCl2, 0.1 Tris GTP, 10 HEPES, and 0.1 Fluo-3

K-salt (pH 7.2). Ca2+ currents were evoked by 300 ms depolarizing steps from a holding potential of -50 mV to 0 mV at 1 Hz to measure the time-dependent profile of Ca2+- 34 activated cytosolic Ca2+ transients. Intracellular Ca2+ imaging was performed using an

Olympus Fluoview 1000 confocal microscope in line-scan mode. Ca2+alternans amplitude was defined as 100 – (A2/A1)*100 (%), where A1 and A2 are the amplitudes of two consecutive cytosolic Ca2+ transients, evoked as described above, with the addition of 100 nM isoproterenol.

Calcium sparks were studied in myocytes permeabilized with saponin using an intracellular solution of (mmol/L): 120 potassium aspartate, 20 KCl, 3 MgATP, 10 phosphocreatine, 5 U/ml creatine phosphokinase, 0.5 EGTA (pCa 7) and 20 HEPES (pH

7.2).

Statistical analysis

All data are presented as means ± SE. The ECG data were analyzed using a three factor [Group (2 levels sedentary vs. exercise trained); Pre-post, (2 levels, i.e., before and after either exercise or sedentary time period), & time, (3 levels, pre-exercise onset, exercise, and occlusion)] mixed designed ANOVA with repeated measures on two factors

(pre-post and time). In a similar fashion, the heart rate and heart rate variability responses to submaximal exercise were analyzed using a three factor ANOVA [pre-post (2 levels) x group (sedentary vs. exercise trained) x time (exercise 7 levels) with repeated measures on two factors (pre-post and time)]. Since repeated measures ANOVA depends on homogeneity of covariance, this sphericity assumption (i.e., the assumption that the variance of the difference scores in a within-subject design are equal across the groups) was tested using Mauchley’s test. If the sphericity assumption was violated, then the F- ratio was corrected using Huynh-Feldt correction (NCSS, Kaysville, UT). If the F-ratio

35 was found to exceed the critical value (p<0.05) then the difference between the means was determined using Scheffe’s test (NCSS, Kaysville, UT). The effect of exercise training on susceptibility to VF was evaluated using Fisher’s Exact test.

Electrophysiological data were analyzed using Clampfit 10.2 (Axon Instruments) and Origin 8.0 (OriginLab). Currents were normalized to the cell capacitance and expressed as (pA/pF). APD and current densities and Ca2+ spark frequency were analyzed by ANOVA with post hoc least significant difference testing as appropriate (SAS for

Windows v9.1, Cary NC). Based on non-normality of IKr and calcium transient alternans amplitude data distribution, the Kruskall-Wallis test was used to test for differences between the groups.

Results

Confirmation of Exercise Training

The effects of the 10-wk exercise training on the heart rate and heart rate variability responses to submaximal exercise training are displayed in Figure. 2.1. In agreement with previous studies99 the exercise training program elicited significant

(p<0.01) decreases in heart rate in response to submaximal exercise after the 10 week treatment period that were accompanied by significant increases in heart rate variability, an indicator of vagal tone index.99, 118 In contrast, these variables did not change in the sedentary group (data not shown). These data are consistent with an exercise training induced , a well-established hallmark of an effective training program.99, 119,

Effect of Exercise Training on Susceptibility to VF

36

Of the 41 dogs randomized to either exercise training or a matched duration sedentary period, 2 dogs in the VF+ exercise group died suddenly (2-3 weeks after training began), while 6 dogs in the VF+ sedentary group died suddenly during the 10 week sedentary period (p=0.130). Six dogs (2 in the exercise group; 4 in the sedentary group) did not complete the protocol due to occluder failure. Exercise training prevented exercise plus ischemia-induced tachyarrhythmias in 16 of 17 dogs tested, while in the sedentary group, 10 of the 10 dogs tested continued to have VF after the observation period (p<0.00001).

ECG analysis

The effects of exercise training or the 10-week sedentary period on ECG variables in response to the exercise plus ischemia test are displayed in Table 2-1. The coronary artery occlusion provoked similar changes in heart rate in the sedentary and exercise trained groups before and at the end of the 10-week study period.

The baseline QTc interval significantly (p<0.02) increased during ischemia in both the VF+ sedentary and VF+ exercise trained groups before the onset of the 10 week study period. However, in marked contrast, QTc still increased during ischemia in the

VF+ sedentary group but not in the VF+ exercise dogs at the end of the 10 wk training period (Fig 2-2A). In a similar manner Tpeak –Tend, a measure of the ventricular repolarization dispersion,114, 115 was also increased during ischemia before the onset of the study period in both the sedentary and the exercise training groups, but only in the sedentary group at the end of the 10 wk training period a (p<0.0002) (Figure 2.2 C)

37

Neither QRS duration nor PR intervals were altered by the coronary occlusion in either the sedentary or the exercise trained groups (Table 2-1)

Electrophysiological recordings

In agreement with our laboratory’s previous study,59 ventricular myocytes from

VF+ sedentary animals had significant prolongations of APD50 and APD90 at both 0.5 and 1 Hz, and increased beat to beat variability in APD compared with control myocytes

(p<0.05). Exercise training reduced APD50 and APD90 (p<0.05, Figure 2.3) compared to the sedentary VF+ group, to values no different from control myocytes. In a similar manner, exercise training normalized the beat to beat variability (measured by standard deviation) in APD90 (p<0.05, Figure 2.3B).

The resting membrane potential in the myocytes isolated from the VF+ exercise group had a more negative resting potential when compared with VF+ sedentary group

(-82.3 ± 0.8 mV vs. -79.6 ± 0.8 mV, p<0.05), returning to values no different from those seen in controls (i.e., no MI, -81.7 ± 0.3 mV) dogs. Inward IK1 conductance in both VF+ groups was reduced to a similar extent relative to controls (p<0.05), while peak outward

IK1 did not differ between the three groups (Figure 2.4). Transient outward current (Ito) was significantly reduced in myocytes from VF+ sedentary dogs when compared with the control group and was further reduced in the VF+ exercise group (p<0.05, Figure 2.4).

Finally, IKr was significantly reduced in myocytes from the VF+ sedentary group compared to controls. Exercise training significantly increased IKr (p<0.05 vs VF+ sedentary group) to levels no different from control, although it is notable that this

38 increase was highly heterogeneous (Figure 2.4). No change in IKs was found between the three groups (Figure 2.4).

Ventricular cell capacitance was significantly increased in myocytes from VF+ sedentary group compared to the myocytes from the control group (204.7± 9.7 vs. 151.7±

11.4 pF, p<0.05) while myocytes from the VF+ exercise group did not differ from either controls or the VF+ sedentary group (183.5 ± 13.1pF, p=NS).

Protein Expression

Potassium channel subunit expression was measured to assess possible contributions to current alterations. Kv4.3, which contributes to Ito, and the Kv 4.3 associated auxiliary proteins DPP6 and KCHIP2 were measured;18KChIP2 was significantly reduced (p<0.05) in VF+ sedentary samples compared to either control or

VF+ exercise groups, while Kv4.3 and DPP6 expression did not differ between groups.

Kir 2.1, Kir 2.2 and Kir 2.3 analyses showed that Kir 2.1 was significantly (p<0.05) reduced in both sedentary and exercised trained VF+ groups, while the expression of Kir

2.2 and Kir 2.3 proteins did not differ among the three groups (p=NS) (Figure 2.5).

Calcium Handling

We previously reported that ventricular myocytes isolated from VF+ dogs displayed increased susceptibility to Ca2+ transient alternans.36 This effect was attributed, at least partially, to increased activity of the ryanodine receptors (RyRs).36 In the present study the effect of exercise on RyR activity was evaluated by measuring Ca2+alternans and spark frequency in permeabilized myocytes. As shown in Fig. 2-6, myocytes from the VF+ sedentary group exhibited a significant (p <0.05) increase in Ca2+ spark

39 frequency when compared to controls, while exercise treatment almost completely restored Ca2+ spark frequency to control levels. Furthermore, endurance exercise training significantly reduced the amplitude of Ca2+alternans (p<0.05), while the incidence of

Ca2+alternans was returned toward control values (Fig. 2-7). Steady-state peak Ca2+ current density (1 Hz) was not different between the groups (Control -6.2 ± 0.7 pA/pF n=11, VF+ sedentary -5.0± 0.3 pA/pF n=11, VF+ exercise -5.1± 0.6 pA/pF n=14).

Western blot analyses showed that in the sedentary VF+ group, protein levels of

RyR, (PLN) and SERCA2a were significantly (p<0.05) reduced when compared to control (Fig. 2-8). Exercise training increased PLN and SERCA2a to values no different than control, while RyR2 density remained reduced in the VF+ exercise group relative to control (p<0.05). As shown in Fig. 2-8, the PLN to SERCA2a ratio was not altered in the exercise VF+ group consistent with the previously reported conserved functional activity of SERCA in sedentary VF+.36 NCX1 levels were not different between groups in the present study, in contrast to a previous report were we found that VF+ exercise group tissues had reduced NCX1 expression relative to a VF+ sedentary group.120 The phosphorylation status of the PKA site on RyR2 (S2808) was unchanged among the three groups. There was hyperphosphorylation of S2814, the

CAMKII-dependent site in the VF+ sedentary group; exercise training reduced this to levels no different from control (Figure 2.9). We have previously reported that this modification contributes to cellular arrhythmias, and suggest that much of the improved arrhythmia phenotype after exercise training is attributable to reduced CAMKII- dependent phosphorylation.

40

Discussion

The present study investigated the effects of endurance exercise training on myocyte electrophysiology and Ca2+ handling in a Post-MI ischemia induced model of

VF. The primary findings of our study are as follows: 1) In agreement with previous studies,109, 112, 113, 121endurance exercise training prevented malignant arrhythmias in a canine post MI model of SCD; 2) in post-MI VF+ animals, there were significant increases in QTc interval and the descending potion of the T wave (Tpeak-Tend), which were significantly attenuated by a 10-wk endurance exercise training program; 3) in agreement with previous studies,59 ventricular myocytes from dogs susceptible to VF exhibited prolonged action potential duration (APD) and increased beat to beat variability in APD, changes which were attenuated by a 10 wk endurance exercise program; 4) myocytes from the animals susceptible to VF also exhibited abnormalities in ryanodine receptor (RyR) activity (increased Ca2+ leak from the sarcoplasmic reticulum, and

Ca2+alternans, attributable to CAMKII-dependent phosphorylation of RyR2),36 changes which were abrogated by exercise training. Notably, these exercise-induced alterations were not evident in the sedentary group, suggesting that exercise training rather than time-dependent healing caused the observed effects. Thus, endurance exercise normalizes both in vivo and in vitro repolarization abnormalities, improved myocyte Ca2+ handling, and prevented malignant arrhythmias in post MI animals shown to be susceptible to VF induced by ischemia.

41

Effect of exercise training on cellular electrophysiology: normalization of repolarization abnormalities

In agreement with the present study, most animal studies report that exercise training can prevent arrhythmias particularly those following ischemia or MI.99, 122, 123

Epidemiological studies indicate that high levels of physical activity may protect against coronary artery disease and reduce cardiac mortality.124Endurance exercise training favorable alters autonomic balance, enhancing cardiac parasympathetic regulation 125,

126,and decreasing sympathetic activity.127, 128 Exercise training has been shown to improve both cardiac parasympathetic function and to reduce cardiac mortality in patients recovering from myocardial infarction.129 Furthermore, our laboratory recently demonstrated that exercise training normalized abnormal beta-adrenergic receptor activity 113 and enhanced cardiac parasympathetic regulation in dog that were susceptible to VF prior to exercise training.130 However, this protection did not appear to result solely form improved cardiac vagal regulation, as treatment with the cholinergic antagonist atropine did not re-introduce ventricular arrhythmias.130 Thus, additional factors most likely contribute to the protection observed following exercise training. Consistent with that, we observed normalization of cellular electrophysiology and calcium handling in cardiac myocytes, in the absence of vagal modulation.

Our laboratory previously reported that myocytes from dogs susceptible to VF

59 exhibited abnormal repolarization. Specifically, APD90, APD alternans and APD variability (beat to beat variability) were increased in VF+ myocytes compared to VF- or control myocytes.36, 59 Similarly, in the present study, exercise training reversed these

42 cellular abnormalities in repolarization. In the present study, exercise training abolished ischemically induced changes in QTc and Tpeak-Tend data consistent with improved ventricular repolarization. Furthermore, as QRS duration was not altered by either ischemia or exercise training, the improvement in QTc interval did not result as a consequence of changes in ventricular electrical conduction. One potential mechanistic explanation for this finding is increased expression and/or activity of sarcolemmal KATP channels, which could modulate the ischemia-induced changes observed in vivo; previous studies have demonstrated a role for these channels in exercise-induced cardioprotection 131, 132. Similarly, exercise training decreased the repolarization abnormalities observed in LQTS patients, heart failure patients and prolonged the refractory period of rabbit hearts106-108.

In contrast to our findings, Bito et al. found that APD was similar in myocytes from either exercise-trained or sedentary post- MI mice 133. Furthermore, Such and colleagues reported that physical training in normal (non-infarcted) rabbits prolonged ex vivo ventricular refractory period accompanied by a decrease in ventricular fibrillation susceptibility in rabbit 108. Species differences (rabbit or mouse vs. dog), absence of MI (rabbit study) or differences in the intensity of the exercise training (uncontrolled voluntary wheel running vs. supervised progressively increasing exercise) most probably accounts for differences noted between these studies and the present investigation.

In contrast to our results of normalization of the APD90 by endurance exercise training, it has been reported that exercise training causes a prolongation in the APD90 in

134 epicardial monophasic action potential in female rats . Reductions in Ito in response to

43 exercise have also been reported in rats,134, 135 and may explain exercise-induced prolongation of APD90 in rats, given the strong modulatory effect of Ito on overall APD in rat ventricle. These findings in rats are congruent with the further reduction in Ito which we observed following endurance exercise training, relative to the sedentary group. Although Kv4.3 was unchanged by exercise, the Kv4.3 modulatory protein

KChIP2 was normalized, which is unexpected considering the further reduction in Ito we observed. Thus, we suggest that post-translational modifications or altered channel trafficking, or alterations of other channel subunits (e.g. Kv1.4) contributing to Ito expression may explain the observed exercise training induced decrement in Ito, possibilities that require further investigation. Additionally, this result demonstrates that

Ito reduction in isolation is not a critical determinant of APD90 in canine ventricular myocytes, as our laboratory has previously observed in VF+ post-MI canine ventricular

59 myocytes . This is congruent with the studies examining the role of the Ito-dependent notch on calcium current using AP clamp, where shallower APD notches resulted in less reactivation of ICaL (an effect which reduces height of the plateau of the AP and the

136 duration of the AP) . Thus, the modulatory effect of Ito in canine ventricular myocytes is highly complex due to the interplay with other ion currents.

Our laboratory previously reported that IKr is decreased in the VF+ canine model,

59 In the present study, this was confirmed in the VF+ sedentary group, and we found that endurance exercise training increased IKr, although there was significant variability in response to the exercise training. Notably, exercise training did not increase outward IK1,

IKs or Ito; thus, the shortening of the action potential we observed is at least partially

44 attributable to alterations in IKr and possibly to other currents (INa-Late) not measured in the present study. It is notable that we observed a small shift to a more negative resting potential after exercise training, a change which would increase bathmotropy and improve diastolic electrical stability. The basis of this observation is not attributable to

137 IK1 (which was unaltered by exercise training); while the mechanism is unclear, it is possible that some other change (e.g. Na-K-ATPase) may have resulted in the exercise- induced improvement in resting potential

Effect of exercise training on intracellular calcium handling

Abnormalities in intracellular Ca2+ handling in disease states may lead to contractile dysfunction and/or triggered tachyarrythmias.95, 96 Several disease states such as heart failure and MI are well known to result in elevated cytosolic Ca2+

(Ca2+overload), which can trigger malignant arrhythmias especially when accompanied by prolongation of the action potential, effects that act in concert to increase the formation of after-depolarizations.95, 96

Our laboratory previously reported in this canine model, that susceptibility to VF is associated with abnormal Ca2+ handling, manifested as increased frequency and amplitude of Ca2+ alternans compared to controls.36 This is notable as Ca2+ alternans is associated with an increased vulnerability to cardiac arrhythmias.138 In the present study, exercise training reduced the amplitude and incidence of Ca2+ alternans, which may contribute to the attenuation of VF observed after exercise training. Several factors are known to contribute to Ca2+ alternans including slowed SERCA-mediated SR Ca2+ uptake 138; however, in the VF model used in the present study, myocyte Ca2+ alternans 45 are associated with dysregulated RyR function manifested as increased frequency of Ca2+ sparks in the absence of notable changes in SR Ca2+ uptake 36. In the present study, a normalization of spark frequency was observed, suggesting normalization of RyR function (Figure 2.6). Altered RyR function in various disease states, including this model of VF, has been attributed to posttranslational modification of the channel protein by CAMKII phosphorylation 139. Notably, we found that exercise training reduced

CAMKII-dependent hyperphosphorylation of RyR at S2814 (but not PKA-dependent phosphorylation of S2808, Figure 2.9). Normalization of CAMKII-dependent phosphorylation of RyRs could contribute to normalization of calcium-dependent arrhythmogenic alterations in VF myocytes and to the beneficial effects of exercise in VF animals. This finding is consistent with work by Stolen et al.140, who demonstrated that normalization of Ca2+ handling by exercise training in a mouse model of involves a reduction of CaMKII-dependent phosphorylation of RyR.

Limitations

In summary, endurance exercise training in a post-MI SCD canine model decreases APD, beat-to-beat variability, and the incidence and amplitude of Ca2+

2+ alternans. In addition, it normalizes the Ca spark frequency, IKr, QTc interval, and the descending portion of the T wave (Tpeak-Tend). At the molecular level, normalization of calcium handling can be attributed to modulation of CAMKII-dependent phosphorylation of the RyR. Collectively, our data suggest that the antiarrhythmic effects of endurance exercise training in the post-MI setting can be attributed to normalization of in vivo and in vitro ventricular repolarization, as well as improved Ca2+ handling. In addition to the

46 improved autonomic balance observed after exercise training, alterations in repolarization and calcium regulation may be antiarrhythmic following MI.

Conclusion

In summary, endurance exercise training in a post-MI SCD canine model, decreases: APD, beat to beat variability, and the incidence and amplitude of

2+ 2+ Ca alternans. In addition it normalizes: the Ca spark frequency, IKr, QTc interval and the descending portion of the T wave (Tpeak- Tend). At the molecular level, normalization of calcium handling can be attributed to CAMKII-dependent phosphorylation of the RyR. Collectively our data suggests that the antiarrhythmic effects of endurance exercise training in the post-MI setting can be attributed to normalization of in vivo and in vitro ventricular repolarization, as well as improved Ca2+ handling. In addition to the improved autonomic balance observed after exercise training, alterations in repolarization and calcium regulation may be antiarrhythmic following myocardial infarction.

Funding Sources:

This work was supported by grants from the NIH: HL086700 (G.E.B), HL063043 (S.G),

HL089836 (C.A.C.)

47

Figures

Figure 2.1 The effect of 10-week exercise training (n=19) on the heart rate and heart rate variability responses to submaximal exercise in animals susceptible to VF.

Exercise elicited significantly smaller increases in heart rate A, and smaller reductions in the heart rate variability (cardiac vagal tone index 0.24 to1.04 Hz frequency component of R-R interval variability, B) after exercise training. Tshe variables were not altered in the sedentary group (data not shown). Pre = before beginning the 10-wk exercise training or sedentary period, Post = after the completion of the 10 wk study period. * p<0.01 pre- vs. post, Exercise levels: 1 = 0 kph/0% grade, 2 = 4.8 kph/0% grade, 3 = 6.4 kph/0% grade, 4 = 6.4 kph/4% grade, 5 = 6.4 kph/8% grade, 6 = 6.4 kph 12% grade, 7 = 6.4 kph/16% grade.

48

Figure 2.2 Exercise training normalized repolarization abnormalities in dogs susceptible to VF. Note that ischemia provoked increases in QTc in both the exercise trained A and sedentary B group at the beginning (pre) of the 10 week study period. These increases were abolished by exercise training A, but unchanged in the sedentary group B. Exercise training had a significant effect ischemia-induced increase of the descending portion of the T wave (Tpeak – Tend) an index of ventricular repolarization evaluated in response to the exercise plus ischemia test in susceptible animals before (pre) and after (post) either a 10-wk exercise training (n = 17, Panel C). Conversely, in the sedentary group there was no difference in the ischemia-induced increase of the descending portion of the T wave (Tpeak-Tend) after the 10-wk sedentary period (n = 10). * p<0.01 occlusion vs. exercise; # p<0.01 pre vs. post treatment period.

49

Table 2.1: Effect of exercise training on ECG parameters at baseline, during exercise and during coronary artery occlusion Pre- Post- Control Exercise Occlusion Control Exercise Occlusion

Heart Rate (beats/min) Ex Train 123.5 ± 6.1 207.9 ± 5.6 229.8 ± 9.8* 119.4 ± 6.9 187.9 ± 5.6# 211.4 ± 6.6*# Sed 128.8 ± 6.2 198.1 ± 7.4 218.5 ± 9.9* 120.9 ± 5.6 202.0 ± 4.3 227.9 ± 6.5*

PR interval (ms) Ex Train 104.3 ± 3.1 84.7 ± 2.6 82.6 ± 3.4 102.6 ± 2.9 82.9 ± 3.0 84.0 ± 4.3 Sed 99.5 ± 3.5 82.1 ± 2.8 79.7 ± 6.9 92.8 ± 3.5 78.4 ± 4.0 75.8 ± 5.6

5 QRS duration (ms)

0

Ex Train 80.6 ± 2.1 83.5 ± 2.9 83.9 ± 2.9 83.9 ± 2.2 81.8 ± 2.2 82.3 ± 2.8 Sed 85.1 ±1.7 82.0 ± 2.4 86.8 ± 4.1 85.6 ± 2.1 82.4 ± 2.2 87.0 ± 6.0

QTc interval (ms) Ex Train 299.3 ± 4.9 306.7 ± 8.8 328.2 ± 5.8* 301.5 ± 6.6 309.3 ± 7.5 316.3 ± 8.4# Sed 309.4 ± 9.1 304.6 ± 10.2 340.0 ± 9.8* 307.4 ± 8.1 317.9 ± 8.5 347.6 ± 11.7*

Tpeak – Tend (corrected) (ms) Ex Train 63.2 ± 5.4 78.2 ± 5.6 114.8 ± 8.9* 63.0 ± 4.6 67.4 ± 5.4 72.5 ± 5.2# Sed 67.1 ± 8.1 81.5 ± 9.9 107.6 ± 8.3* 72.2 ± 6.1 84.2 ± 8.6 118.2 ± 13.3*

* p<0.01 exercise vs. coronary artery occlusion (60 s after occlusion onset or last 5 s before VF onset); # = p<0.01 Pre- vs. Post- Pre = before exercise training or sedentary period, Post = after 10 weeks of exercise training or equivalent sedentary period. 50

Figure 2.3: Exercise normalizes repolarization abnormalities in myocytes from animals susceptible VF

A. Representative action potentials recorded at 1Hz in control, VF+ sedentary and VF+ exercise groups. B. Beat to beat variability of APD90 is significantly increased in the VF+ sedentary group, relative to controls. Endurance exercise training restores beat to beat variability of repolarization to values no different from controls (measured as the standard deviation of the APD90 between beats). C and D. Action potential duration at 50% (APD50) and 90% (APD90) repolarization is prolonged in the sedentary VF+ group at both 0.5 (Panel C) and 1 (Panel D) Hz. Exercise training has a normalizing effect on APD50 and APD90 relative to the VF+ Sedentary group. (*p<0.05 vs. Control, #p<0.05 vs. VF+ Sedentary). N = 11-13 cells and 2-4 dogs per group.

51

Figure 2.4: Effects of endurance exercise training on potassium currents.

A. rapidly activating delayed rectifier K+ current (IKr) was significantly reduced in the VF+ sedentary group relative to control. Exercise training increases IKr relative to the VF+ sedentary group, to values no different from control. IKr data were obtained at a test potential of +50 mv. Dotted line shows the median value for each group. B. slowly activating delayed rectifier K+ current (IKs) (D- sotalol insensitive current) at +50 mV did not differ among the three groups. C. inward rectifier K+ current (IK1) density-voltage relationships. D. inward IK1 conductance is significantly reduced in the VF+ sedentary group and similar in the VF+ exercise group. (both p<0.05 vs. control). E. peak outward IK1 did not differ between groups. F. transient outward K+ current (Ito) density was reduced in the sedentary VF+ group relative to control (p<0.05). Exercise training results in a further decrement in Ito (p<0.05). Values are means ± SE; N=7-28 cells per group (* p<0.05 vs. Control, # p<0.05 vs. VF+ sedentary).

52

Figure 2.4

53

Figure 2.5: Western blot analyses of potassium channel subunit expression.

A: representative Western blots of left ventricular potassium channel expression. B: + the Ito auxiliary protein voltage-gated K (KV) channel interacting protein 2 (KChip2) was reduced in sedentary VF+ group and normalized after exercise training. C: no change in the Ito pore-forming subunit KV4.3 was observed between groups. D: + protein subunits inward rectifier K (Kir) channel 2.1 (Kir2.1) were similarly reduced in the sedentary and exercise-trained VF+ groups, relative to control. Kir2.2 and Kir2.3 did not differ among groups. Values are means ± SE; n = 4 dogs per group. *P < 0.05 vs. control. #p < 0.05 vs. VF+ sedentary.

54

Figure 2.6: Effect of exercise training on calcium spark frequency.

A: representative examples of calcium sparks from the three groups. B: calcium spark frequency was increased in the sedentary VF+ group and restored toward control by exercise training. F/F0, fluorescence ratio. Values are means ± SE. p < 0.05 vs. *control and #sedentary VF+.

55

Figure 2.7: Effect of endurance exercise training on the incidence and amplitude of calcium alternans.

A: representative recordings of L-type Ca2+ current (top), line scan (middle), and calcium transients (bottom) from each group. B: the amplitude of calcium alternans was significantly increased in the sedentary VF+ group and normalized by exercise training. C: the incidence of alternans was increased in the sedentary VF+ group and reduced by exercise training. Values are means ± SE. *p < 0.05 vs. control.

56

Figure 2.8: Western blot analyses of calcium handling proteins.

A: representative gels of the calcium-handling proteins analyzed: ryanodine receptor 2 (RyR2), sarco(endo)plasmic reticulum calcium ATPase (SERCA2a), sodium/calcium exchanger (NCX1), and phospholamban (PLN). B: pooled data of the RyR2 protein normalized to GAPDH. RyR2 protein was decreased in the VF+ sedentary group and unchanged by exercise training. C: SERCA2a was downregulated in the VF+ sedentary group and restored by exercise training. D: in addition to SERCA normalization, PLN was normalized after exercise as well. E: the SERCA-to-PLN ratio was unchanged as these proteins change in parallel. F: NCX shows no change between groups. Values are means ± SE. *p < 0.05 vs. control.

57

Figure 2.9 : Phosphorylation of ryanodine receptor protein. A: serine 2814, the CAMKII-dependent phosphorylation site, was significantly hyperphosphorylated in the VF+ sedentary group (*p < 0.05 vs. control). Exercise training normalizes the phosphorylation status of serine 2814 (#p < 0.05 vs. VF+ sedentary) to values no different than control (p = not significant vs. control). B: serine 2808, the PKA-dependent phosphorylation site, did not differ among the three groups. Values are means ± SE.

58

Chapter 3 : Ibandronate and Ventricular Arrhythmia

Risk

Ingrid M Bonilla1*, Pedro Vargas-Pinto DVM, Ph.D.2,a*, Yoshinori Nishijima DVM,Ph.D.1,b, Adriana Pedraza-Toscano DVM, Ph.D.2, Hsiang-Ting Ho, MS3,4, Victor P. Long III PharmD.1, Andriy E. Belevych Ph.D.3,4 , Patric Glynn 4, Mahmoud Houmsse M.D.3,4, Troy Rhodes M.D.3,4, Raul Weiss M.D3,4 , Thomas J. Hund, Ph.D.3,4, Robert L Hamlin DVM, PhD.2,4,5, Sandor Györke Ph.D.3,4, Cynthia A Carnes PharmD, Ph.D.1-4 1 College of Pharmacy, The Ohio State University, Columbus, OH, USA; 2College of Veterinary Medicine, The Ohio State University, Columbus, OH, USA; 3 College of Medicine, The Ohio State University, Columbus, OH, USA; 4Dorothy M. Davis Heart and Lung Research Institute, The Ohio State University, Columbus, OH, USA; 5QTest Labs, Columbus, OH,USA

*These authors contributed equally to the work Running Title: Ibandronate-induced proarrhythmia Corresponding Author: Cynthia A. Carnes College of Pharmacy 500 W. 12th Avenue Columbus, OH43210 PH: 614-292-1715 FAX: 614-292-1335 Email: [email protected] aCurrent address: Universidad de La Salle, Bogotá, Colombia bCurrent address: Dept. of Medicine, Cardiovascular center, Medical college of Wisconsin, Milwaukee, WI 53226 Conflicts of interest: none

Previously Published: J Cardiovasc Electrophysiol. 2013 Nov 20. doi: 10.1111/jce.12327

59

Abstract

Introduction: Bisphosphonates, including ibandronate, are used in the prevention and treatment of osteoporosis.

Methods and Results: We report a case of suspected ibandronate-associated proarrhythmia, following a first dose of ibandronate in a 55 year old female. ECG at presentation revealed frequent ectopy and QT/QTc interval prolongation; at follow-up nine months later the QT/QTc intervals were normalized. Proarrhythmic potential of ibandronate were assessed with a combination of in vivo and in vitro approaches in canines and canine ventricular myocytes. We observed late onset in vivo repolarization instability after ibandronate treatment. Myocytes superfused with ibandronate exhibited action potential duration (APD) prolongation and variability, increased early afterdepolarizations (EADs) and reduced Ito (p<0.05), with no change in IKr.

Ibandronate-induced APD changes and EADs were prevented by inhibition of intracellular calcium cycling. Ibandronate increased sarcoplasmic reticulum calcium load; during washout there was an increase in calcium spark frequency and spontaneous calcium waves. Computational modeling was used to examine the observed effects of ibandronate. While reductions in Ito alone had modest effects on APD, when combined with altered RyR inactivation kinetics, the model predicted effects on APD and SR Ca2+ load consistent with observed experimental results.

Conclusion: Ibandronate may increase the susceptibility to ventricular ectopy and arrhythmias. Collectively these data suggest that reduced Ito combined with abnormal

RyR calcium handling may result in a previously unrecognized form of drug-induced

60 proarrhythmia.

Key words: arrhythmia; calcium; ion channels; potassium; pharmacology

61

Introduction

Bisphosphonates, such as ibandronate, are used in the prevention and treatment of osteoporosis. Osteoporosis is prevalent in post-menopausal women, and bisphosphonates are often used as first-line preventative therapy. Here we report a suspected case of ibandronate-associated arrhythmia and a subsequent integrated investigation of the proarrhythmic potential of ibandronate.

Cardiotoxicity has been estimated to cause 27% of drug withdrawals from the market, and limits the development of new drug entities.51 One form of drug-induced proarrhythmia is (TdP), a potentially fatal cardiac arrhythmia which occurs in the setting of prolonged ventricular repolarization. In vitro assessments of arrhythmogenic potential for new drug entities are primarily focused on prolongation of ventricular repolarization, assessed by blockade of IKr, the fast component of the delayed rectifier potassium rectifier current in cardiac cells, or blockade of hERG, the human

141 channel protein encoding IKr, in heterologous expression systems. Arrhythmogenic potential may be assessed in vitro or in vivo in mammals to evaluate prolongation of the

QT/QTc on the ECG.52 Beat-to beat variability of the QT interval or cellular action potential duration (APD) has been suggested as a more reliable predictor of drug induced

TdP 53, 142.

While calcium handling dysregulation in cardiac myocytes is known to contribute to cardiac arrhythmias, as has long been recognized for glycosides,143, 144 this form of proarrhythmia is not systematically assessed in the drug development process.

We examined the effects of ibandronate in a pre-clinical model, evaluated

62 electrophysiologic and calcium handling effects in isolated myocytes which was followed by computational modeling to examine the proarrhythmic effects of combined altered repolarization and dysregulated myocyte calcium handling.

Methods

Case Report:

A 55 year old woman presented to an outside hospital following a witnessed syncopal episode. Fifteen days prior to the event, she received her first monthly dose of ibandronate; the patient was not taking any other medications. Admission laboratory values were normal except for (3.1 mEq/L). After correction of hypokalemia, her presenting electrocardiogram (ECG) showed a prolonged QTc of

575ms (Figure 3.1A). After transfer to our institution, ventricular ectopy and non- sustained polymorphic ventricular (PMVT) were seen on telemetry. Given her syncopal episode and persistently prolonged QT interval in the absence of a recognized cause for QT prolongation or arrhythmias, an implantable cardioverter- defibrillator (ICD) was placed (8). Ibandronate therapy was discontinued after the initial dose. Nine months after initial presentation, her QT and QTc were 400 and 426 ms, respectively (Figure 3.1B); no ICD therapies had been administered. Given the normalization of ventricular repolarization following discontinuation of ibandronate, we used preclinical testing to evaluate the proarrhythmic potential of ibandronate as a contributor to sudden .

63

Myocyte Isolation

All animal procedures were approved by the Institutional Animal Care and Use

Committee of the Ohio State University. Thirty adult mixed breed dogs (male/female, age

9 months to 5 years) weighing between 8 and 20 kg with normal cardiac function were used for the experiments. On the day of the experiment, the dogs where anesthetized with pentobarbital sodium (50 mg/kg IV; Nembutal, Abbott Laboratories). The heart was rapidly removed and perfused with cold cardioplegic solution containing the following in mM: NaCl 110, CaCl2 1.2, KCl 16, MgCl2 16 and NaHCO3 10.

Cannulation of the left circumflex was used for the enzymatic myocyte isolation, as previously described.23 Cells were stored at room temperature until use. 23

This isolation procedure typically yields 70-90% rod shaped ventricular myocytes with clear striations and margins. All myocyte experiments were conducted within 10 hours of isolation. Amphotericin-B perforated patch clamp techniques were used with a bath temperature of 36 ± 0.5 ºC as previously described.67 Myocytes were placed in a laminin coated cell chamber (Cell Microcontrols, Norfolk, VA) and superfused with bath solution containing (in mM): 135 NaCl, 5 MgCl2, 5 KCl, 10 glucose, 1.8 CaCl2, and 5 HEPES with pH adjusted to 7.40 with NaOH. Borosilicate glass micropipettes with tip resistance of 1.5-3 MΩ, were filled with pipette solution containing the following (in mM): 100 K- aspartate, 40 KCl, 5 MgCl, 5 EGTA, 5 HEPES, pH adjusted to 7.2 with KOH.

APDs were measured from the average of the last 10 traces (steady state) from a train of 25 action potentials elicited at each stimulation rate. The standard deviation of the

APD90 for the last ten traces (i.e. from trace 15-25) was used to evaluate repolarization

64 variability.145 To further evaluate beat to beat variability, Poincaré plots of the last ten consecutive beats were drawn by plotting each APD90 (APD90 n+1) against the APD90 of

142 the previous beat (APD90 n) as previously reported . Short term variability of APD90 , expressed in ms, was calculated by using the following formula:∑

√ -1 as previously reported.142 Data was collected at baseline, after superfusion with ibandronate (0.01-10 µg/L) and after washout of the drug. Cells demonstrating multiple early afterdepolarizations after drug treatment (as seen in Figure

3.2, panel B) were not included in the APD and beat to beat variability measurements.

EAD propensity was assessed as the percentage of cells exhibiting EADs during ibandronate superfusion and/or washout.

To assess the potential contribution of intracellular calcium handling to the effects of ibandronate, myocytes were incubated in buffer supplemented with BAPTA-

AM (10 mM) or ryanodine (100 nM) for at least half an hour to either buffer or deplete intracellular calcium, respectively. After incubation, action potentials were recorded at described above, before and after ibandronate (10 µg/L) superfusion.

For potassium current measurements, nifedipine (2 μM) was added to the bath solution and calcium concentration reduced to 1 mM.23, 145 Only cells with access resistance less than 20 MΩ were included in the analysis. Data was collected at baseline and after superfusion with 10 µg/L of ibandronate. Voltage protocols are shown in the insets of Figure 3.2. Data acquisition was performed with a low noise data acquisition system Digidata 1440A (Molecular devices, Sunnyvale, CA), Clampex 10.2 software and an Axopatch 200A amplifier (Axon Instruments, Sunnyvale, CA).

65

Calcium handling

Calcium transients were measured by Rhod-2 (10 µM) using an Olympus

Fluoview 1000 confocal microscope in line-scan mode. Myocytes were loaded with dye at room temperature and Rhod-2 was excited at 543 nm and fluorescence was collected at wavelength >590 nm. Myocytes were paced by extracellular platinum electrodes at 0.5

Hz for 20 seconds, and recordings made at 0.3 Hz in the presence of 100 nM isoproterenol with external solution containing (in mM): 140 NaCl, 5.4 KCl, 2 CaCl2, 0.5

2+ MgCl2, 10 HEPES and 5.6 glucose (pH 7.3). To assess SR Ca load, 20mM was applied at the end of the experiments.

Calcium sparks were studied in myocytes permeabilized with saponin (0.01% for

20–30 s) using an internal solution containing the following (in mM): 120 potassium

2+ aspartate, 20 KCl, 0.81 MgCl2, 1 KH2PO4, 0.5 EGTA (free [Ca ] ∼100 nM), 3 MgATP,

10 phosphocreatine, 0.03 Fluo-3 pento potassium salt, 20 HEPES (pH 7.2) and 5 Uml−1 creatine phosphokinase.

Ibandronate concentrations:

The range of in vitro ibandronate concentrations were selected to bracket the expected peak unbound plasma concentrations following a single dose.146 Concentrations were selected with consideration of bioavailability, volume of distribution, and plasma protein binding.146, 147

Computational modeling:

A well-validated model of the human ventricular cardiomyocyte was used to

66 simulate ion channel kinetics, action potential, and calcium cycling.148 Sensitivity analysis was performed, as described,149 to determine likely candidates for observed changes in AP and SR Ca2+ load. Briefly, parameters related to SR Ca2+ release, uptake and intracellular Ca2+ were perturbed one parameter at a time +30% and -

30%. Action potential duration at 90% repolarization (APD90) and maximal diastolic

Ca2+ concentration in the junctional sarcoplasmic reticulum ([Ca2+]JSR) were determined following steady-state (change in APD < 0.1%) pacing at a cycle length of 1000 ms. For each property (X) and parameter (p), sensitivity was calculated according to equation [1] and expressed relative to maximal value for all parameters.

XXpp, 30% , 30% SXp,  [1] 0.6X con

Data analysis:

Canine ECG data was analyzed using an IOX EMKA station. Differences in QTc short term variability between baseline and 1 month after treatment were tested by

Fisher’s exact test. (Origin 9.0). Cellular electrophysiology data were analyzed using

Clampfit 10.3 software (Axon Instruments) and Origin 9.0 software (OriginLab,

Northampton, MA, USA). Currents were normalized to cell capacitance in picofarads

(pF) and are expressed as pA/pF. Comparisons between baseline and drug exposed cells were made by the appropriate t-test (Origin Pro 9.0, OriginLab). Comparisons between

67

APDs and APD variability at baseline, during ibandronate exposure and during washout were made using one way ANOVA with post hoc least significant difference testing.

(Origin 9.0, Origin Lab). Differences in EAD occurrence were tested with Pearson’s

Chi-Square test. Differences in Ca2+ handling were tested by ANOVA. A level of p

<0.05 was accepted as statistically significant. All data are presented as mean ± SE and p<0.05 was the criterion for statistical significance for all comparisons.

Results

Ibandronate effects on canine QT

Monitoring of canines treated with ibandronate revealed no prolongation of the

QT or QTc interval up to six weeks post-dosing compared to baseline. However, delayed onset of increased short term variability of the QTc was observed in three of the four dogs (Figure 3.6 Supplemental) consistent with late-onset drug-induced repolarization instability.

Cellular electrophysiology:

Treatment of isolated canine ventricular myocytes with ibandronate revealed prolongation of repolarization over a concentration range of 0.01 µg/L to 10 µg/L (Figure

3.2), with rate-independent prolongation of APD50. The effects of ibandronate were not mitigated by washout, with washout of up to 15 minutes not relieving (p<0.05 vs baseline) the ibandronate-induced APD50 prolongation. The effects of ibandronate on

APD90 were similar to those observed in the APD50. As observed with the APD50, washout of ibandronate did not relieve ibandronate-dependent APD90 prolongation

68

(p<0.05 vs baseline).

In addition to the observed ibandronate-induced prolongation of repolarization, a significant (p<0.001 vs baseline) number of cells exhibited EADs either during drug superfusion (~35%) or washout (~30%, Figure 3.2). Notably, no cells exhibited EADs at baseline.

Potassium currents

To investigate potential mechanism(s) for ibandronate-induced APD prolongation, we evaluated the effects of drug treatment on major repolarizing K+ currents. Ibandronate significantly (p<0.05 vs baseline) reduced Ito slope conductance and the maximal current density (at +50mV) as shown in Figure 3.2.

Neither IK1 nor IKr was affected by ibandronate superfusion (Figure 3.2).

Calcium depletion and buffering experiments

Myocytes were pre- treated with ryanodine and then superfused with ibandronate to assess the role of SR calcium release via the ryanodine receptor in ibandronate- dependent cellular pro-arrhythmia. Repolarization instability was evaluated at an ibandronate concentration of 10 µg/L ibandronate, which is within the expected unbound plasma concentration range.146 Ryanodine pretreatment prevented ibandronate-dependent increases in APD50, APD90, beat to beat repolarization variability and short term repolarization variability.(p=NS vs baseline, Figure 3.3) . In addition, experiments were conducted with the calcium chelator, BAPTA. BAPTA pre-treatment prevented ibandronate-induced prolongation of APD50 or APD90. (Figure 3.7 Supplemental)

69

Furthermore, EADs were not observed in either the ryanodine or BAPTA pre-treatment experiments (Figures 3.3 and 3.7), suggesting that the arrhythmogenic effects observed after ibandronate treatment are calcium- and RyR-dependent.

Calcium Handling

Superfusion of myocytes with ibandronate did not affect the amplitude of Ca2+ transients or change the rate of occurrence of spontaneous Ca2+ waves. (Figure 3.4 B-C).

However, ibandronate treatment significantly (p<0.05 vs baseline and washout) increased the sarcoplasmic reticulum calcium load, assessed from the amplitude of caffeine- induced Ca2+ transients. Interestingly, ibandronate washout significantly increased the frequency of spontaneous calcium waves accompanied by a significant (p<0.05 vs baseline and ibandronate superfusion) decrease in Ca2+ transient amplitude (Figure 3.4).

Additionally, washout of ibandronate resulted in normalization of the sarcoplasmic reticulum Ca2+ load.

Calcium sparks were measured with ibandronate (100 µg/L) in saponin- permeabilized myocytes (Figure 3.5). A significant decrease in spark amplitude was observed during ibandronate perfusion (p<0.05 vs baseline). Ibandronate washout caused a significant increase in the spark frequency (normalized to the SR calcium load) compared to either baseline or ibandronate superfusion (p<0.05).

Computational modeling

Computational modeling was used to predict the likely mechanism(s) responsible for the pro-arrhythmia phenotypes of APD prolongation and increased SR Ca2+ load

70 observed with ibandronate. Incorporation of the measured defect in Ito into a human cardiac myocyte model 148 produced only a slight prolongation of APD. Sensitivity analysis was then performed to determine model parameters that had the greatest

2+ influence on APD and SR Ca load. The forward RyR inactivation rate (kiCa) was found to have the greatest effect on APD while also changing SR Ca2+ load. Combining decreased Ito with an increase in kiCa produced an increase in SR calcium load and prolongation in the action potential similar to the experimentally observed effects (Figure

3.8 Supplemental).

Discussion

Ibandronate is a bisphosphonate used to prevent and treat osteoporosis by increasing bone density and decreasing bone metabolism.149 Bisphosphonates have been suggested to increase incident atrial fibrillation150, although this is not a consistent finding.151 However, to our knowledge the and associated QTc prolongation experienced by our patient have not been previously attributed to ibandronate. Together, the normalization of the QTc after ibandronate discontinuation (Figure 3.1B), and the lack of arrhythmias requiring ICD therapy during the follow-up period, suggests that the syncopal episode was caused by the drug. While the mechanism responsible for the observed ibandronate induced cardiac arrest experienced by the patient is unknown, as there is no ECG from the time of the event, ventricular arrhythmias were considered as the likely cause based on the patient’s presentation.59 The purpose of this study was to evaluate the possible mechanism(s) behind this rare, but possibly fatal side effect.

Bisphosphonate elimination following intravenous administration is 71 multiphasic.146 The drug is rapidly distributed into bones and circulating drug is eliminated in the urine. The terminal elimination rate is much slower and occurs as the drug is redistributed from the bone into the blood; some long term adverse effects of bisphosphonates have been linked to drug redistribution occurring years after discontinuation.150 Thus, long term drug redistribution could conceivably contribute to a delayed onset of arrhythmia in the patient case.

Drug-induced arrhythmias are increasingly recognized to occur during the use of non-cardiac drugs. The increased number of recent drug withdrawals due to ventricular arrhythmias suggests that current risk stratification approaches (valuation of the fast component of the delayed rectifier potassium current (IKr) encoded by the ether a go-go

(HERG) channel in isolated cardiac myocytes or heterologous expression systems, and evaluation of ECG recordings with QT prolongation in conscious or anesthetized animals.152) are not sufficiently sensitive to predict arrhythmic risk. It can years to decades to detect this rare, but potentially fatal adverse drug effect.153

To further investigate the effects of ibandronate on cardiac repolarization and elucidate potential arrhythmia-inducing mechanisms, we used canine ventricular cardiomyocytes. Action potential prolongation is known to be arrhythmogenic by increasing the propensity to develop EADs, which are generally accepted as the most common proximate cause of TdP.154 The observed ibandronate- induced increase in beat to beat and short term variability indicates that ibandronate increases repolarization instability not only in the action potential but also in intact myocardium as the QT interval in 3 out 4 dogs treated demonstrated instability (Figure 3.6 Supplemental). The in

72 vivo observations are consistent with an increased susceptibility to develop arrhythmias such as TdP.53 Since, ibandronate caused EADs without affecting phase 3 repolarizing currents, another contributor to the EADs was sought.

EAD formation has been also attributed to abnormal calcium release from the SR which in turn may lead to inward current through the sodium calcium exchanger

(NCX).155, 156 Thus, since not only repolarizing current abnormalities, but calcium handling abnormalities can also lead to arrhythmia formation,142, 157 we investigated ibandronate effects on calcium handling. Inhibition of calcium cycling by RyR blockade or calcium buffering, with ryanodine or BAPTA, respectively protected the cell from the detrimental effects of ibandronate (i.e. EADs and APD variability). Ibandronate treatment caused a significant increase in calcium load and a decrease in calcium spark amplitude.

Drug washout increased the incidence of spontaneous calcium waves, decreased the calcium transient amplitude, and increased calcium sparks. The relationship between intra-SR Ca2+ concentration (Ca2+ load) and Ca2+ release through RyR gating is complicated as recently reviewed by Radwanski et. al.158 At elevated SR Ca2+ loads the amount of Ca2+ available for release (fractional SR calcium release) increases. In addition, an increase intra-SR Ca2+ load enhances the frequency of spontaneous calcium release (Ca2+ sparks).136 Considered collectively, our data suggests that ibandronate may block the RyR, causing a Ca2+ load increase. When the drug is washed out, the RyR blockade is relieved, and coupled with the transiently increased SR Ca2+ load leads to an increase in spontaneous calcium waves and spark frequency.

In silico experiments using a physiological model of a human ventricular

73 cardiomyocyte to simulate the effects of ibandronate revealed that the drug likely affects multiple targets in the myocyte, including Ito and RyR calcium release. While the model predicted that either defect alone was insufficient to produce significant APD prolongation and/or increased SR Ca2+ load, the simulations demonstrate that they have an additive effect that gives rise to the overall phenotype. Thus, pro-arrhythmia associated with ibandronate may be due to “multiple hits” that act synergistically to alter cell function. The “multiple hit” nature of ibandronate also highlights unique challenges related to screening of drugs for cardiac safety.

Limitations

This report was initiated by a single case report and it is possible that there were unique patient characteristics which increased susceptibility to late-onset arrhythmias following ibandronate initiation. The myocyte studies were conducted in canine myocytes, while the in silico experiments were conducted in a human ventricular cardiomyocyte model. Ion channels and currents vary by species, although there is substantial similarity between humans and canines. The lack of QT prolongation in vivo, which contrasts with the myocyte APD prolongation, may have resulted from differences in coupled vs. uncoupled myocytes.

L-type calcium current was not directly measured in these studies, nor was the potential contribution of Na+-Ca2+ exchanger. The steady-state nature of the in silico experiments may have underestimated dynamic effects of calcium dysregulation on INCX, which is accepted as a known contributor to afterdepolarizations.

74

Conclusion

Ibandronate, and possibly other bisphosphonates, may be associated with an unrecognized risk for significant cardiac arrhythmias. Furthermore, the presentation may be insidious with a delayed onset, reducing the likelihood of associating such events with drug exposure. There may be a reduced sensitivity to detect proarrhythmia for drugs which are primarily administered to an older population (i.e. prevention of post- menopausal osteoporosis). The potential for this adverse side effect of ibandronate should be considered, particularly when prescribing the drug to patients with heart abnormalities which may reduce repolarization reserve.

Current testing paradigms to screen for proarrhythmic potential could be strengthened by evaluating beat to beat instability and repolarization reserve in repolarization as well as effects on myocyte calcium handling. Integrated computational modeling of experimental data may also assist in risk stratification for possible proarrhythmic liability. Considered collectively, our results suggest that evaluation of proarrhythmic risk could be improved.

Acknowledgments:

The authors thank Ms. Jeanne Green for expert technical assistance and Ms. Jessica

Smith for assistance with data analysis.

Funding Sources

IMB: Supplement to NIH award HL089836; TJH: NIH HL096805, HL114893

75

Figures

Figure 3.1: Patient ECG after syncopal episode, and nine months after discontinuation of ibandronate; ibandronate discontinuation normalized the QT interval. A. Electrocardiogram showing ectopic beats, with a prolonged QT interval of 484 ms and Bazett corrected QT interval of 575 ms. ECG was obtained following correction of presenting hypokalemia. B. Repeat electrocardiogram (9 months after initial presentation) showing occasional ectopic beats. The QT interval is 400 ms and the Bazett corrected QT is 426 ms.

76

Figure 3.2: Ibandronate prolongs the action potential, causes EADs and decreases Ito current density and conductance in left ventricular cardiomyocytes. A.Representative action potential tracings recorded at 0.5 Hz show ibandronate- induced prolongation, which was unrelieved by washout. B. Ibandronate caused a significant increase in the % of myocytes with EADs compared to baseline. (p<0.001 vs Baseline). The inset shows a representative EAD. Washout of ibandronate did not reduce EADs (p<0.05 vs baseline). C and D. Ibandronate prolonged the action potential duration at 50 (C) and 90 (D) percent repolarization (closed symbols). Data shown as percent change from baseline. Open symbols are data obtained during washout. E. No effect of ibandronate on IK1 current density is observed. (p =NS from baseline) F. Ibandronate treatment significantly decreases Ito current density and conductance compared to baseline. (p <0.05 vs baseline) G. Ibandronate did not alter IKr (p =NS vs baseline). Insets show voltage protocols.

77

Figure 3.2

78

Figure 3.3: Inhibition of calcium cycling prevents ibandronate-induced action potential prolongation and action potential variability. A and B. Ryanodine pretreatment prevented ibandronate-induced prolongation of the action potential and 50 and 90 percent of repolarization (p =NS vs baseline). Data presented as percentage change from baseline. C. Representative Poincaré plots of APD90 during baseline, ibandronate treated, ibandronate washout and rynaodine pre- treated + ibandronate treatment groups. D. Ibandronate washout significantly increases (p <0.05) short term variability at 0.5 Hz compared to baseline. Ryanodine pretreatment prevented ibandronate-induced increases in short term variability of APD90 (p =NS vs baseline)

79

Figure 3.4 Ibandronate treatment increases SR calcium load and washout induces spontaneous calcium waves. A. Line scan and tracing representation of calcium transients at baseline, during treatment and during washout. B. Ibandronate washout significantly (p<0.05) increases the spontaneous calcium waves frequency compared to treatment and baseline. C. Ibandronate washout significantly (p<0.05) decreases calcium transient amplitude compared to baseline and ibandronate treatment. D. Ibandronate treatment significantly (p<0.05) increases SR calcium load defined as caffeine-induced calcium transient compared to baseline.

80

Figure 3.5: Ibandronate washout increases calcium spark frequency. A. Line scan representation of the calcium spark measured at baseline, during ibandronate treatment and during washout. B. Ibandronate treatment and washout significantly (p <0.05) decreases calcium spark amplitude compared to baseline. C. Ibandronate treatment and washout did not alter calcium spark frequency. D. Calcium spark frequency, normalized to SR calcium load, was significantly increased during ibandronate washout (p <0.05) compared to baseline and ibandronate treatment.

81

Figure 3.6 (Supplemental): Ibandronate increases repolarization instability in vivo. A. Ibandronate treatment did not prolong QT interval. B. One month after initial treatment ibandronate increased QT short term variability in 3 out of the 4 dogs studied. C. Representative Poincaré plot of baseline and 1 month after ibandronate treatment QT intervals.

82

Figure 3.7 (Supplemental): Calcium chelation prior to ibandronate treatment prevents the ibandronate-induced increase in short term variability of repolarization. A. Representative Poincaré plot of baseline, ibandronate treated and BAPTA pre- treated + ibandronate action potentials at 90% repolarization. B. Ibandronate washout significantly (p<0.05) increases APD90 short-term variability at 0.5 Hz compared to baseline. BAPTA pre-treatment results in post-ibandronate short term variability in APD90 which is no different than baseline. (p=NS)

83

Figure 3.8 (Supplemental): Mathematical modeling of electrophysiological changes with ibandronate-mimicking conditions. A. Model of ryanodine receptor SR Ca2+ release channel (RyR) showing open (O), inactive (I), resting (R), and resting inactive (RI) states with associated transition rates. B. Ratio of APD sensitivity to SR Ca2+ sensitivity for key model parameters related to intracellular Ca2+ cycling. A positive sensitivity ratio signifies that 2+ 2+ parameter affects APD90 and SR Ca load ([Ca ]JSR,dia) in the same way (e.g. both properties increase as parameter changes, designated by black bars). The baseline had the greatest influence on APD90, in a manner consistent with observed defects 2+ (e.g. drug causes increase in both APD90 and SR Ca load). C-D. Parametric analysis was performed by simultaneously varying maximal conductance of the transient + 2+ outward K current (Ito) and kiCa. APD90 and [Ca ]JSR,dia were plotted for each simulation (color bar on right indicates value). Reducing Ito alone increased APD but not to the extent seen experimentally. However, combining a reduction in Ito with effects on kiCa provided a better fit to experimentally observed changes in APD and Ca2+. E. Simulated action potentials and F-G. comparison between experimentally 2+ measured and simulated APD and [Ca ]JSR for the combination of model parameters that provides best fit to experimental data (54% reduction in Ito and three-fold increase in kiCa). Experimental data recorded at 0.5 Hz. Abbreviations are as follows: kiCa, koCa – baseline, non-SR-dependent transition rate constants for RyR; kim, kom – reverse transition rates for RyR inactivation and activation, respectively; ks – RyR SR 2+ 2+ Ca release rate constant; EC50SR – half-maximal concentration for [Ca ]SR- dependent RyR activation; 푵푪푿̅̅̅̅̅̅ – Maximal Na+/Ca2+ exchanger current; 푵푲푨̅̅̅̅̅̅̅ – + + 2+ Maximal Na /K ATPase current; 푽̅ – Maximal velocity of SR Ca ATPase.

84

Figure 3.8

85

Chapter 4 : Differential effects of the peroxynitrite

donor, SIN-1, on atrial and ventricular myocyte

electrophysiology

Ingrid M. Bonilla MS1,2, Arun Sridhar PhD3, Yoshinori Nishijima DVM , PhD1, Sandor

Györke Ph.D 2,4, Arturo J. Cardounel Ph.D 2,4, and Cynthia A. Carnes Pharm D, PhD1,2*

1 College of Pharmacy, The Ohio State University, Columbus, OH, USA 2 Dorothy M. Davis Heart and Lung Research Institute, The Ohio State University, Columbus, OH, USA 3 Glaxo Smith Kline, Ware, UK 4 College of Medicine, The Ohio State University, Columbus, OH, USA *Correspondence: Cynthia A. Carnes, PharmD, PhD, College of Pharmacy, The Ohio State University, 500 West 12th Avenue, Columbus, OH 43210, USA. email: [email protected]

Funding: This project was supported by NIH/NHLBI grant HL089836. Conflict of interest: None of the authors in the present manuscript have any conflict of interest to declare.

Previously Published: Front Physiol. 2012 Apr 23;3:105.

86

Abstract

Oxidative stress has been implicated in the pathogenesis of heart failure and atrial fibrillation and can result in increased peroxynitrite production in the myocardium. Atrial and ventricular canine cardiac myocytes were superfused with SIN-1 (3- morpholinosydnonimine-N-ethylcarbamide), a peroxynitrite donor, to evaluate the acute electrophysiologic effects of peroxynitrite. Perforated whole cell patch clamp techniques were used to record action potentials. SIN-1 (200 µM) increased the action potential duration (APD) in atrial and ventricular myocytes; however, in the atria, APD prolongation was rate-independent, while in the ventricle APD prolongation was rate dependent. In addition to prolongation of the action potential, beat to beat variability of repolarization was significantly increased in ventricular, but not atrial myocytes. We examined the contribution of intracellular calcium cycling to the effects of SIN-1 by treating myocytes with the sarcoendoplasmic reticulum calcium ATPase (SERCA) blocker, thapsigargin (5-10 µM). Inhibition of calcium cycling prevented APD prolongation in the atrial and ventricular myocytes, and prevented the SIN-1 induced increase in ventricular beat to beat APD variability. Collectively, these data demonstrate that peroxynitrite affects atrial and ventricular electrophysiology differentially. A detailed understanding of oxidative modulation of electrophysiology in specific chambers is critical to optimize therapeutic approaches for cardiac diseases.

Keywords: electrophysiology; cardiac myocyte; peroxynitrite

87

Introduction

Reactive nitrogen species (RNS) and reactive oxygen species (ROS) are normally produced by cellular metabolism where they serve as cell signaling moieties. Nitrosative and/or oxidative stress occur when there is an imbalance between ROS/RNS production and their reduction by antioxidant agents or enzymes. Oxidative and nitrosative stress have been suggested to participate in the pathophysiology of numerous cardiovascular diseases including: progression of heart failure (HF), myocardial infarction (MI) and atrial fibrillation (AF) 159-162.

ROS and RNS are produced by multiple sources, among them uncoupling of mitochondrial electron transport, cytokine activity, NADH/NADPH oxidase

(nicotinamide adenine dinucleotide in its reduced form/ nicotinamide dinucleotide phosphatase in its reduced form) and uncoupled nitric oxide synthases. Endogenous antioxidant defenses include superoxide dismutase, catalase, and glutathione, among others 163. When nitric oxide (NO) and superoxide anion are produced in close proximity they interact to form a very powerful oxidant called peroxynitrite (OONO-) 164.

Peroxynitrite rapidly causes nitration of proteins, including tryptophan and tyrosine residues.165, As a result of tyrosine nitration, 3-nitrotyrosine (3-NT) is formed, a modification which is either irreversible or slowly reversible 166. Increased 3-NT is therefore a biomarker for the presence of OONO-. Increased peroxynitrite has been associated with the pathology of a number of cardiac diseases, as recently reviewed 167.

Specifically, increased 3-NT has been found in the atria of patients with persistent atrial

88 fibrillation, and is thought to contribute to the impaired contractility associated with atrial fibrillation 168. In human ventricle, increased SERCA 3-NT content has been reported during heart failure 169. Thus, 3-NT is increased in both atrial fibrillation and heart failure. In addition, peroxynitrite has been implicated in the modulation of excitation- contraction coupling 38, 169-171. Notably, there may be differential chamber-specific effects on action potential duration, the integrated cellular response to sarcolemmal ion currents, induced by cardiac diseases. Specifically, in chronic heart failure, the ventricular action potential is prolonged while the atrial action potential is shortened 172-174,73. The potential for peroxynitrite to elicit chamber-specific changes in action potential are undefined. In the present study we tested the hypothesis that there are differential effects of peroxynitrite on atrial and ventricular myocyte electrophysiology. SIN-1, a molsidomine metabolite, releases both superoxide and NO which react in a diffusion limited manner to produce ONOO-; thus, SIN-1 was used to test the acute electrophysiologic effects of

OONO-.

Methods

Myocyte Isolation

All animal procedures were approved by the Institutional Animal Care and Use

Committee of the Ohio State University. Twenty adult mixed breed dogs (male/female, age 9 months to 5 years) weighing between 8 and 20 kg with normal cardiac function were used for the experiments. On the day of the experiment, dogs where anesthetized with pentobarbital sodium (50 mg/kg IV; Nembutal, Abbott Laboratories). The heart was rapidly removed and perfused with cold cardioplegia solution (containing: 5% glucose,

89

0.1% mannitol, 22.4 mM NaHCO3 and 30 mM of KCl) injected into the coronary ostia.

The left circumflex artery was cannulated for myocyte isolation as previously described 175. Collagenase (Worthington, type 2; 0.65 mg/ml) was added to the perfusate for enzymatic dispersion of the myocytes. After 30-45 minutes of enzyme perfusion, the left atrial appendage was separated from the left ventricle. The left atrial appendage was minced and placed into 5 mL of perfusate in a shaking water bath at 37ºC for 5-15 minutes as previously described 73. Simultaneously, the digested left ventricular mid- myocardial section of the left ventricular free wall was separated from the epicardial and endocardial section; digested tissue was shaken in a water bath at 37º C for an additional

5 minutes. After secondary digestion the myocytes were washed and stored at room temperature, as previously described 82. Storage solution contained (in mM): 118 NaCl,

4.8 MgCl2, 1.2 KH2PO4, 0.68 glutamine, 10 glucose, 5 pyruvate, 1 CaCl2; 1 µM insulin and 1% BSA. This isolation procedure typically yielded 40-60% and 70-90% of rod shaped atrial and ventricular myocytes, respectively, with sharp margins and clear striations.

Electrophysiological Studies

Data acquisition was performed with Clampex version 9.0 software (Axon

Instruments Union City, CA, USA) and Axopatch 200A patch clamp amplifiers (Axon

Instruments Inc.) Perforated whole cell patch clamp (using amphotericin B) was used to minimize alterations in intracellular milieu during action potential (AP) and K+ current recordings. Myocytes were placed in a laminin coated cell chamber (Cell Microcontrols,

Norfolk, VA, USA) and superfused (1 mL/min) with bath solution containing the 90 following (in mM): 135 NaCl, 5 KCl, 5 MgCl2, 10 D-glucose, 1.8 CaCl2, 5 4-(2- hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pH adjusted at 7.40 with

NaOH, was used. After obtaining baseline data, the cells were superfused with bath solution supplemented with SIN-1 (200 µM), a concentration previously reported to inhibit phospholamban function in isolated cardiac myocytes 176 as well as NOS activity177 which is relevant as we previously demonstrated that NOS activity is an important mediator of atrial electrophysiology 82 Specificity of SIN-1 effects through peroxynitrite formation was evaluated by co-superfusion with Uric Acid (500 µM), a peroxynitrite scavenger.

To assess the contribution of calcium cycling to the observed SIN-1 effects, the

SERCA inhibitor, thapsigargin was used. Thapsigargin (5-10 µM) was added to the bath solution and recordings were taken at baseline, after thapsigargin, and after subsequent

SIN-1 superfusion. Recordings were obtained a minimum of five minutes of drug.

Action potentials were recorded at 0.5 Hz, 1 Hz and 2 Hz applied in a random order. The average of the last 10 (steady state) action potentials, recorded during a train of 25 action potentials at each stimulation rate were used to calculate the action potential duration at 50% repolarization (APD50) and 90% of repolarization (APD90). To analyze beat to beat variability in the action potential recordings we measured the standard deviation (SD) of the action potential duration at 90% of repolarization (APD90) of the last ten traces (i.e. from trace 15-25) as previously described 23.

+ For K current recordings, the bath solution CaCl2 was reduced to 1 mM, and 2

µM of nifedepine was added to the bath solution to block the L-type calcium current. K+

91 currents were recorded at baseline and after SIN-1 perfusion as previously described 73.

Action potentials and current recordings were made at 35 ± 0.5 ºC.

Statistical Analysis

Acquired data were analyzed using Clampfit 8.0 software (Axon Instruments) and

Origin 8.6 software (OriginLab, Northampton, MA, USA). Currents were normalized to cell capacitance in picofarads (pF) and are expressed as pA/pF. Comparisons between baseline and drug exposed cells were made by the appropriate t-test (Origin Pro 8.6,

OriginLab). Comparison between baseline, thapsigargin and SIN-1 post thapsigargin were made using one way ANOVA with post hoc least significant difference testing.

(Origin 8.6, Origin Lab). Rate- dependence of drug effects was analyzed by two way

ANOVA with post hoc least significant difference testing (Originpro 8.6, OriginLab). All data are presented as mean ± SE and p<0.05 was the criterion for statistical significance for all comparisons.

Chemicals

All chemicals used were purchased from Sigma-Aldrich (St. Louis, MO, USA) and Fisher Scientific (Pittsburg, PA, USA). All buffers and solutions were prepared daily.

Results

Specificity of effects of SIN-1 on action potential duration

Pilot experiments were conducted to assure that SIN-1 effects observed in the cardiac myocytes result from the production of ONOO-. For this myocytes were co- 92 superfused with the peroxynitrite scavenger, uric acid (500 µM) and SIN-1 200µm. Atrial action potentials were recorded at baseline, with uric acid and with uric acid + SIN-1

(Figure 2.1A). No statistical change in APD90 was observed between the baseline, uric acid and uric acid + SIN-1 treatments (p>0.05). Thus, any effects observed with SIN-1 exposure are consistent with peroxynitrite-dependent effects. SIN-1 treatment caused a significant (p<0.05 vs baseline) prolongation in the action potential (APD90) in both atrial and ventricular myocytes (Figure 3.1 Panels B and C, Figure 3.2 Panels A and B). The action potential prolongation observed in the atria was rate independent (p = 0.98 at rates of 0.5 Hz, 1 Hz and 2 Hz) while the observed prolongation in the ventricle was rate dependent (p<0.05, Figure 3.2 panels A and B, respectively). A significant increase in

APD50 was only observed at 2 Hz in the atria (p<0.05 vs baseline Figure 3.2 panel A).

Ventricular APD50 was unaffected by SIN-1 perfusion (p=0.73 Figure 3.2 panel B). The resting membrane potential was not affected by SIN-1 in either atrial (Baseline -78.82 ±

2.20 vs. SIN-1 -75.5 ± 3.30 p= 0.43) or ventricular myocytes (Baseline -80.40 ± 0.96 vs.

SIN-1 -79.78 ± 0.76 p=0.41).

Beat to beat variability

Instability of repolarization is associated with an increased propensity to arrhythmias. We assessed beat to beat variability of APD90, and found it was increased only in ventricular myocytes after SIN-1 treatment at 1 and 2 Hz (p <0.05 vs baseline

Figure 3.2 panels C and D).

93

Potassium currents

Potassium currents measured in the atria were: inward rectifier (IK1) at -80 mV, transient outward (Ito) and IKsus (sustained potassium current which contains IKur, IKs and

84 IKr). In the ventricle: IK1, Ito and the fast component of the delayed rectifier potassium current (IKr) were measured. No significant differences (p>0.05) were found in any of the currents measured in the atria (Table 3-1). In the ventricle the only current was affected by SIN-1 was Ito which was significantly decreased after SIN-1 treatment (p<0.05 vs baseline, Table 3-1).

Thapsigargin experiments

To assess the specificity of the effects observed after SIN-1 exposure, a comparison between baseline and thapsigargin were made. No significant difference between the groups in either cell type was found (p>0.05 data not shown). Thus, thapsigargin itself did not have direct electrophysiologic effects on the APD.

After thapsigargin pre-treatment, the cells were superfused with SIN-1.

Thapsigargin pre-treatment prevented the SIN-1 dependent prolongation of atrial APD50 previously observed at 2 Hz (p=0.09 with thapsigargin + SIN-1 vs baseline, data not shown). Thapsigargin pretreatment abolished the previously observed APD90 prolongation in both atrial and ventricular myocytes (P=0.85 and P=0.68, respectively).

In summary, no statistical differences between baseline and SIN-1 + thapsigargin were found in atrial or ventricular cells. (Figure 3.3 Panels A and B, respectively).

The SIN-1 dependent increase in beat to beat variability (SD APD90) previously observed in ventricular myocytes was prevented by thapsigargin pretreatment. Therefore 94

SERCA inhibition prevented ONOO- dependent ventricular repolarization instability

(p>0.05 vs baseline Figure 3.3 Panels D).

Discussion

The main findings of this study are that peroxynitrite causes a significant prolongation of atrial and ventricular myocyte repolarization, increases repolarization instability in ventricular myocytes, and decreases Ito current density only in ventricular myocytes. Surprisingly, we found no inhibition of delayed rectifier repolarizing potassium currents in either atrial or ventricular myocytes with SIN-1. However, SERCA blockade and the resulting depletion of calcium in the sarcoplasmic reticulum with thapsigargin prevented the peroxynitrite-dependent effects on repolarization.

In the present study SIN-1 was used as a peroxynitrite donor as it releases both superoxide anion and NO, which at physiological pH combine to form peroxynitrite164.

We ascertained that the effects observed after SIN-1 were due to peroxynitrite action and not due a non-specific effect, by conducting co-superperfusion experiments with uric acid

(a natural peroxynitrite scavenger). Thus, the effects we observed are attributable to peroxynitrite.

The role of increased production of reactive species in the pathogenesis of several cardiovascular diseases is a well-established phenomenon. Increased peroxynitrite production has been shown to occur in diseases such as chronic heart failure, myocardial infarction and atrial fibrillation. Indeed 3-NT (commonly used as a biomarker for the production of ONOO-) has been found to be increased in patients with persistent atrial fibrillation and patients with inflammatory heart disease 38, 168, 178. It is formed when 95

ONOO- modifies the tyrosine residues of proteins via nitration. When proteins are nitrated and 3-NT is formed it can cause a loss or gain of function or increased degradation of nitrated protein 166. Interestingly, another common finding in heart disease, uncoupling of inducible nitric oxide synthase (iNOS) has also been shown to result in an increase in peroxynitrite production in cardiovascular diseases 38, 82. In addition to tyrosine it is notable that methionine, cysteine and tryptophan are also susceptible to nitration, therefore, it is possible that part of the differential effects of

ONOO- between chambers observed in the present study are due to differential nitration of key protein residues, 165, 179 as recently reviewed.180

Differential responses to SIN-1 between atria and ventricle were observed, and atrial specific changes in APD90 manifested as rate independent prolongation. The prolongation of APD in the atrial cells suggests that acute or transient increases in peroxynitrite are unlikely to participate in the shortening of the APD observed in chronic heart failure 73 or in atrial fibrillation172,181,81. In contrast to the atria, ventricular effects of SIN-1 on APD90 prolongation were rate dependent, accompanied by a significant increase in the beat to beat variability and a significant decrease in Ito current density.

The differential response between atrial and ventricular myocytes is not unexpected as the chambers differ in many ways. The action potential morphology and duration, gene and protein expression and calcium handling mechanisms are some of the differences between atrial and ventricular myocytes 182. In addition it has been reported that electrical remodeling during HF (tachypacing induced congestive HF in canine model) also differs between the chambers 183, 184. Chronic HF is well known to prolong

96 ventricular repolarization, 174, 185 while in the atria, chronic HF shortens repolarization 73,

82, 172. Because of the differences between disease-induced electrical remodeling between chambers, it is not surprising to observe variable effects with SIN-1. The prolongation in the action potential and the increased beat to beat variability observed in this study are very common findings observed in the pathogenesis of heart failure and ventricular fibrillation canine models and have been linked to increased arrhythmogenesis 59, 73.

SERCA inhibition and subsequent calcium depletion of the sarcoplasmic reticulum with thapsigargin before SIN-1 perfusion in ventricular myocytes prevented the

SIN-1 dependent APD90 prolongation and increase in beat to beat variability, suggesting that calcium cycling might be responsible for the observed SIN-1 dependent effects in the ventricle. The effects of SIN-1 (ONOO-) on calcium handling have been previously described as affecting phospholamban (PLB) phosphorylation in murine ventricular myocytes 171, 186 SERCA pump is regulated by PLB phosphorylation to increase the SR

Ca2+ load and decrease cytosolic Ca2+ concentration under normal conditions (for a review see 187). In our experiments with thapsigargin we are not only inhibiting SERCA, but by depletion of calcium stores also inhibiting calcium handling in the myocyte. Thus, we suggest that SIN-1 is exerting electrophysiologic effects via a component of the Ca2+ handling system: SERCA, ryanodine receptor (RyR), PLB, etc.

Cardiac myocyte calcium handling has been reported to be redox sensitive in canine ventricular myocytes; reactive species cause an increase in spontaneous calcium release and calcium alternans, and are associated with heart failure and ventricular tachyarrhythmias 188, 189. Specifically, ROS are known to activate serine/ threonine

97 kinases such as CAMKII, which in turn which can phosphorylate the RyR, PLB, sodium

Channels and L-type calcium channels, as recently reviewed190 Under normal conditions

CAMKII activation and subsequent phosphorylation of proteins are calcium-dependent, but in the presence of ROS, CAMKII may be oxidized resulting in calcium-independent activation191, possibly leading to abnormal phosphorylation of key components of the calcium handling system. Another mechanism for calcium-handling dependent effects of peroxynitrite may be through oxidation of specific components of the calcium-handling apparatus, as L-type calcium channels, sodium channels, RyR, Sodium calcium exchanger (NCX) and SERCA190. Thus, inhibition of calcium-cycling by thapsigargin may inhibit the acute effects of peroxynitrite through multiple mechanisms.

The basis of the atrio-ventricular differential effects of thapsigargin in mediating

SIN-1 effects are not clear. Possible explanations might be that peroxynitrite targets proteins that are differentially expressed in the ventricle and the atria: e.g. PLB is the primary regulator of SERCA pump function in the ventricle, while sarcolipin is an atrial- specific regulator of SERCA pump function 192. Another possible explanation for the chamber specific responses to SIN-1 exposure is the potential for differential expression of the intracellular antioxidant defenses such as superoxide dismutase, catalase and glutathione in cardiac myocytes, or differential expression of proteins that detoxify peroxynitrite such as hemoglobin or peroxiredoxins,179, 193 any of which could give rise to differential reactions during exposure to reactive species.

Reactive oxygen species and peroxynitrite reportedly decrease Ito density and prolong the action potential in rat ventricular myocytes 194. Whether the observed

98 decrease in Ito current density would cause prolongation in the canine action potential is a matter of debate. Some reports suggest that this is only true in rodents, suggesting that in larger mammals (canines, humans) the effect of decreased Ito current density alters the notch observed in phase 1 of the action potential 195-197. Another possible explanation for the chamber specific differences we observed after thapsigargin may be that Ito is reduced in the ventricular, but not the atrial myocytes by peroxynitrite. Ito is known to modulate

12, 17, 136 calcium current in very complex ways and the effects of SIN-1 on ventricular Ito may be contributing to the differential modulation of beat to beat variability. Further investigation will be needed to determine the specific target(s) and mechanism(s) by which SIN-1 modulates calcium handling with secondary effects on repolarization.

Limitation

Prolonged (2-24h) SIN-1 exposure in cardiac myocytes has been reported to induce caspases and subsequent apoptosis.198, 199 The time course from caspase activation to apoptotic signaling in adult cardiac myocytes appears to be at least 2 hours,200 while our maximum duration of exposure to SIN-1 was 15 minutes. Thus, apoptotic signaling is unlikely to contribute to the altered electrophysiology during SIN-1 exposure. The contribution of the late sodium, L-type calcium, and NCX currents to the effects of peroxynitrite were not taken into consideration in the present study. It is well known that late sodium current and L-type calcium current contribute to rate adaptation of the action potential 201. However, our goals were to focus primarily on the action potential duration, the integrated response to all ion currents, and action potential duration instability, a measure of arrhythmic potential, in the present study. 99

Myocytes were isolated from specific regions of the two chambers, and do not fully reflect the spatial differences occurring within the heart. All studies were conducted in myocytes from normal controls, and thus, the effects of peroxynitrite we observed may vary in heart diseases such as heart failure and atrial fibrillation.

Conclusions

Cardiac remodeling is comprised of multiple complex processes, and an increase in reactive species is only one component of the remodeling. Examining the changes that occur due to the increase in peroxynitrite provides potential insights into the pathophysiology of heart disease. Our data suggests a potential antiarrhythmic role of acute increases in peroxynitrite in the atria; however, it has proarrhythmic potential in the ventricle due to enhanced action potential instability 202. A complete understanding of alterations due to peroxynitrite requires further study in the context of overall redox modulation of electrophysiology.

Acknowledgments

The authors thank Ms. Jeanne Green for excellent technical support.

100

Figures

Figure 4.1: SIN-1 prolongs the atrial and ventricular action potential. A. The peroxynitrite scavenger uric acid was used to evaluate specificity of SIN-1. Atrial action potentials were recorded at baseline, with uric acid and with uric acid + SIN-1. No statistical change in atrial APD90 was observed after co-superfusion. B. SIN-1 prolongs the atrial action potential and C. SIN-1 prolongs the action potential and decreases the phase 1 notch in ventricular myocytes.

101

Figure 4.2: SIN-1 affects atrial and ventricular repolarization

A. SIN-1 causes a significant (p<0.05 vs baseline) prolongation in the atrial APD50 only at 2 Hz while the APD90 is significantly increased (p<0.05 vs baseline) after SIN-1 in a rate independent manner. B. Contrary to the atria, no change was observed in the ventricular APD50 after SIN-1. At baseline ventricular APD90 was significantly shorter at 2 Hz than 0.5 Hz. A significant increase (p<0.05 vs baseline) was observed in the ventricular APD90 at all rates during SIN-1 superfusion. While SIN-1 prolonged the ventricular APD90, in contrast to the atria, APD90 retains rate-dependence, with APD90 being significantly shorter at 2 Hz, compared to 0.5 or 1 Hz (P<0.05). C. SIN-1 does not affect atrial beat to beat variability of repolarization (SD APD90). D. A significant (p<0.05 vs baseline) increase in beat to beat variability of repolarization (SD APD90) occurs during SIN-1 superperfusion in ventricular myocytes at 0.5 and 1 Hz, but not at 2 Hz. (n=5-8 cells in each group, N= 3-4 animals per group).

102

Table 4.1: Electrophysiologic effects of SIN-1 in cardiac myocytes Ventricle (n=6, N=3) Control SIN-1 P value

IK1 at -100mV(pA/pF) -6.552 ±1.911 -6.584 ± 0.548 0.185 IK1 Peak Outward (pA/pF) 1.774 ± 0.127 1.688 ± 0.214 0.393 Ito at +50 (pA/pF) 7.143 ±1.664 6.803 ±1.585 0.003

IKr Peak (pA/pF) 1.533 ± 0.147 1.502 ± 0.115 0.365 Atria (n=8, N=3) Control SIN-1 P value

IK1 at -80mV (pA/pF) -0.328 ± 0.264 -0.239 ± 0.306 0.348 Ito at +50mV (pA/pF) 5.842±0.654 5.206± 0.775 0.597 IKsus (pA/pF) 6.271 ± 0.926 5.834 ± 1.088 0.326

IK1 (inward rectifier potassium current), Ito (transient outward potassium current), IKr (fast component of the delayed rectifier potassium current), IKsus (sustained outward potassium current which contains IKur, IKs and IKr), n (number of cells per group), N (number of animal per group)

103

Figure 4.3 : Peroxynitrite-induced changes in atrial and ventricular repolarization and variability of repolarization are prevented with thapsigargin pre-treatment.

A. and B. Thapsigargin pre-treatment (to deplete SR calcium stores and prevent calcium cycling) prevented both atrial and ventricular SIN-1 dependent APD90 prolongation (p>0.05 from baseline). C. Atrial beat to beat variability in repolarization is unchanged by peroxynitrite (p=NS vs baseline). D. Peroxynitrite-induced increase in ventricular beat to beat variability in repolarization is prevented by thapsigargin (p=NS vs baseline). (n=5-8 cells in each group and N= 3-5 animals per group)

104

Chapter 5 : Calcium-activated Potassium Current

Modulates Ventricular Repolarization in Chronic Heart

Failure

Ingrid M. Bonilla1, Victor P. Long III1, Pedro Vargas-Pinto3, Patrick Wright2, Andriy

Belevych2, Qing Lou2, Kent Mowrey4, Jae Yoo1, Philip F. Binkley2, Vadim V. Fedorov2,

Sandor Györke2, Paulus M.L. Janssen2, Ahmet Kilic3, Peter J. Mohler2, Cynthia A.

Carnes1,2

1 College of Pharmacy, The Ohio State University, Columbus, OH, USA; 2Dorothy M. Davis Heart and Lung Research Institute, The Ohio State University Wexner Medical Center, Columbus, OH, USA;3 College of Veterinary Medicine, The Ohio State University, Columbus, OH, USA; 4St Jude Medical,Sylmar, CA, USA

Corresponding Author: Cynthia A. Carnes College of Pharmacy 500 W 12th Avenue Columbus OH, 43210 PH: 614-292-1715 FAX: 614-292-1335 Email: [email protected]

105

Abstract

Background: IKCa has been suggested as a novel therapeutic target for the treatment of cardiac arrhythmias. Objective: We examined the cellular electrophysiologic effects of

IKCa blockade in controls, chronic heart failure (HF) and HF with sustained atrial fibrillation. Methods: Using action potential recordings to maintain calcium cycling, the effects of IKCa block (apamin 100 nM) were examined in atrial and ventricular myocytes.

A canine tachypacing induced model of HF (1 and 4 months, n=5 per group) was used, and compared to a group of 4 month HF with 6 weeks of superimposed atrial fibrillation

(n=7). A group of age-matched controls were used (n=5). Human atrial and ventricular myocytes isolated from explanted hearts obtained from transplant recipients were studied in parallel. Results: Atrial myocyte action potentials were unchanged by IKCa blockade in any of the groups studied. IKCa blockade did not affect ventricular myocyte repolarization in controls. HF caused prolongation of ventricular myocyte action potential repolarization. IKCa blockade caused further prolongation of ventricular repolarization in

HF and also caused repolarization instability and early afterdepolarizations. SK2 and SK3 expression in the atria and SK3 in the ventricle were increased in canine heart failure.

Conclusion: During HF, IKCa blockade in ventricular myocytes results in cellular arrhythmias. Our data suggest an important role for IKCa in the maintenance of ventricular repolarization stability during heart failure. Novel antiarrhythmic therapies should be assessed in both the atria and ventricle to evaluate safety in addition to efficacy. Key Words: calcium-activated potassium current; repolarization; heart failure; atrial fibrillation; proarrhythmia

106

Introduction

Heart Failure (HF) is a chronic disease that develops over months to years, and is defined by insufficient to meet the physiologic and metabolic needs of the body. Atrial fibrillation (AF) and HF are common coexisting disease states, and HF results in a 4.5 to 5.9 fold increase in the risk of developing AF.203 Moreover, in patients with HF, the development of AF significantly increases the risk of death.80 Thus, identifying and elucidating pharmacological targets to treat AF may significantly reduce mortality and morbidity in HF.

Small-conductance Ca2+- activated K+ (SK) channels are expressed in multiple tissues such as skeletal and smooth muscle, the central and peripheral nervous system and the heart.29, 204, 205 Cardiac myocytes express SK1, SK2 and SK3 gene products.206 SK- encoded current is voltage-independent and activated by intracellular calcium.32 All three members of the SK family have similar calcium sensitivity for activation (0.6 - 0.7

µM)207. SK-encoded current is blocked by the bee venom apamin.31, 32

IKCa, the potassium current conducted by SK channels, contributes to

29, 208 repolarization, but the effect of IKCa on repolarization remains poorly elucidated. For example, ventricular IKCa shortens repolarization and promotes peri-infarct arrhythmias in

209 rats. Conversely, blockade of IKCa promotes ventricular arrhythmias in human HF and a

210, 211 non- ischemic rabbit HF model, suggesting a protective role for IKCa. The contribution of IKCa to atrial repolarization is also unclear as some reports demonstrate

212, 213 that IKCa is proarrhythmic while in other circumstances it is considered protective.

107

We measured the impact of IKCa block on action potentials in intact myocytes to maintain intrinsic Ca2+cycling. We utilized a well-validated canine model that emulates all key features of human HF including chamber dilatation, impaired contractility, impaired functional capacity, repolarization abnormalities, dysregulated myocyte calcium handling, increased predisposition to AF and increased myocardial fibrosis.188, 214

Complementary experiments were conducted in end-stage human HF. The role of IKCa in

AF was evaluated in atrial myocytes from a canine model of chronic HF with sustained

AF.

Methods

Heart failure model

Atrial fibrillation was induced in dogs with HF using a customized pacemaker (St Jude

Medical, Sylmar, CA). One pacing lead was implanted in the right atria (RA) and the second lead was implanted in the RV. The standard protocol described was used to induce HF. After 10 weeks of RV tachypacing, RA tachypacing was initiated, with the

RA stimulated at 10 Hz for 60 seconds, followed by a 10 second pause for automated interrogation of atrial rhythm. This cycle of RA tachypacing was repeated every 70 seconds until AF was detected. Subsequent detection of normal atrial rhythm resulted in resumption of the atrial tachypacing. The total HF duration in the HF + AF group was 4 months. Ventricular pacing was stopped during atrial stimulation, ventricular rate was

150-200 BPM during atrial pacing or AF.

108

In vivo

Serial echocardiograms and ECGs were done as previously reported.174, 214 Serial pacemaker interrogations were used to monitor cardiac rhythm.

Myocyte Isolation and Tissue Collection

On the day of the terminal procedure, the dogs were anesthetized with pentobarbital sodium (50 mg/kg intravenously; Nembutal, Abbott Laboratories). The heart was rapidly removed and perfused with cold cardioplegia solution containing the following in mM: NaCl 110, CaCl2 1.2, KCl 16, MgCl2 16 and NaHCO3 10. Cannulation of the left circumflex artery was used to perfuse left atria and ventricle following removal of the right atrium and right ventricle, as previously described.215 Adjacent tissue samples were collected and snap frozen for protein analyses. Tyrode’s solution (mM) containing

NaCl 130, KCl 5.4, MgCl2 3.5, NaH2PO4 0.5, Glucose 10, HEPES 5 and taurine 20, was used as the initial perfusate. During the cell isolation process the heart was perfused with three different solutions (36ºC). First the heart was perfused for 10 minutes with Tyrode’s solution with 0.1 mM EGTA; followed by perfusion with Tyrode’s solution containing

0.3 mM Calcium, 0.12 mg/ml of Trypsin Inhibitor (NIBCO) and 1.33mg/ml of collagenase (Type II, Worthington), for a maximum of 45 minutes. Following enzymatic digestion, the heart was perfused with normal Tyrode’s solution for five minutes to remove residual enzyme. Subsequently, left ventricular mid-myocardial and left atrial myocytes were obtained through secondary digestion, as previously described.215 After secondary digestion the cells were re-suspended in incubation buffer.38 This isolation

109 procedure typically yields 70-90% and 40-60% rod shaped ventricular and atrial myocytes, respectively. All myocyte electrophysiology experiments were conducted within 10 hours of isolation.

In parallel experiments, left ventricular mid-myocardial and left atrial appendage myocytes were isolated and adjacent tissues were collected from explanted end-stage failing hearts (n=6; obtained from the Ohio State University Wexner Medical Center transplant program)Tissues from left ventricular mid myocardium and left atrial appendege were collected from non-failing heart for western blotting purposes (n=8 obtained from Lifeline of Ohio). Procedures were approved by the Ohio State University

Institutional Review Board in accordance with the 1964 Declaration of Helsinki and its later amendments. Non-HF status was confirmed in these tissues by lack of CaMKII pS286 hyperphosphorylation. Methods for myocyte isolation and tissue collection were as described for canine samples above.

Action Potential Measurements

Amphotericin-B perforated patch clamp techniques with a bath temperature of 36

± 0.5ºC were used to assess myocyte electrophysiology. The myocytes were placed in a laminin coated cell chamber (Cell Microcontrols, Norfolk, VA) and superfused (~1 mL/min) with bath solution containing (in mM): 135 NaCl, 5 MgCl2, 5 KCl, 10 glucose,

1.8 CaCl2, and 5 HEPES with pH adjusted to 7.40 with NaOH. Borosilicate glass micropipettes with tip resistance of 1.5-3 MΩ, were filled with pipette solution containing the following (in mM): 100 K-aspartate, 40 KCl, 5 MgCl, 5 EGTA, 5 HEPES, pH adjusted to 7.2 with KOH. 110

Action potentials (AP) were recorded in a train of 25 traces at 0.5, 1 and 2 Hz at baseline and after drug perfusion. The average of the last 10 traces (i.e. from trace 15-25) was used to calculate the action potential duration (APD). APD was calculated at 50 and

90 percent of repolarization (APD50 and APD90).

To evaluate repolarization instability, beat to beat variability (BTBV) of APD90 was measured as previously described. 216, 217 Early afterdepolarization (EAD) propensity was assessed as the percentage of cells exhibiting EADs. Recordings exhibiting EADs were excluded from APD and BTBV measurements.

Data collection was done at baseline and after superfusion with the IKCa blocker apamin (100 nM), a concentration known to block SK1, SK2 and SK3 encoded- current.218-220 A low noise system (Digidata 1440A , Molecular devices, Sunnyvale, CA),

Axopatch 200A amplifier (Axon Instruments, Sunnyvale, CA) and Clampex 10.2 software was used for data acquisition.

Calcium transient Measurements

Calcium transients were recorded using Ca2+ sensitive dye Fluo-4AM (10 µM) and an Olympus Fluoview 1000 confocal microscope in line scan mode. Myocytes were loaded with dye for 25 minutes at room temperature. Fluo-4 was excited with a 488 nm argon laser and fluorescence collected at wavelength 500-600 nm. Myocytes were paced by extracellular platinum electrodes at 1 Hz stimulation. External solution contained (in mM): 140 NaCl, 5.4 KCl, 2 CaCl2, 0.5 MgCl2, 10 HEPES and 5.6 glucose (pH 7.3).

111

Immunoblots

Following protein quantification, tissue lysates were analyzed on Mini-

PROTEAN tetra cell (BioRad) on a 4-15% precast TGX gel (BioRad) in

Tris/Glycine/SDS Buffer (BioRad). Gels were transferred to a nitrocellulose membrane using the Mini-PROTEAN tetra cell (BioRad) in Tris/Glycine buffer with 20% methanol

(v/v, BioRad). Membranes were blocked for 1 hour at room temperature using a 3% BSA solution and incubated with primary antibody overnight at 4°C. Antibodies included:

SK2 (Alomone, Santa Cruz), SK3 (Alomone, Santa Cruz), GAPDH (Fitzgerald), and actin (Sigma). Secondary antibodies included donkey anti-mouse-HRP and donkey-anti- rabbit-HRP (Jackson Laboratories). Densitometry was performed using Image lab software and all data was normalized to GAPDH or actin levels present in each sample.

Data Analysis

Cellular electrophysiology and Ca2+ imaging data were analyzed using Clampfit

10.3 software (Axon Instruments) and Origin 9.0 software (OriginLab, Northampton,

MA, USA). All APD paired data were compared by paired student t-test. Unpaired data and comparisons between groups were analyzed by one-way ANOVA with post hoc least significant difference testing (Originpro 8.6, OriginLab). Differences in EADs incidence were tested with Pearson’s Chi-Square test. All data are presented as mean ± SE and p<0.05 was the criterion for statistical significance for all comparisons. For protein experiments, data are presented as mean ± SEM. P values were assessed with a paired

Student’s t test (2-tailed) or ANOVA, as appropriate, for continuous data. The Bonferroni test was used for post-hoc testing. The null hypothesis was rejected for p < 0.05. 112

Chemicals

All chemicals used were purchased from Sigma-Aldrich (St. Louis, MO, USA) and

Fisher Scientific (Pittsburg, PA, USA), unless otherwise noted. All buffers and solutions were prepared daily.

Results

In vivo cardiac remodeling

Left ventricular fractional shortening (LVFS) was similarly reduced in the 1 month HF, 4 month HF and 4 month HF + AF groups (Figure 5.1B), consistent with HF.

Electrocardiograms (ECGs) in all canines assigned to the HF+ AF group demonstrated sustained atrial tachyarrhythmias, evident as the absence of P waves and the irregularly irregular ventricular rate characteristic of AF (Figure 5.1A). Additionally, atrial contractility, measured as left atrial fractional area change (FAC), was significantly reduced in both the 4 month HF and 4 month HF + AF groups compared to baseline

(p<0.05) as shown in Figure 5.1C. Notably, the presence of sustained AF did not cause a further decrement in atrial contractility compared to HF alone.

IKCa inhibition in control ventricular myocytes

Action potentials before and after apamin treatment were recorded from control canine ventricular myocytes. Varying apamin concentrations (0.5-100 nM) were tested in order to generate a concentration response curve. Apamin did not alter APD50 or APD90 in control ventricular myocytes (Figure 5.2A).

113

IKCa inhibition and SK expression in failing ventricle

Canine

We observed no HF-induced difference in APD50 or APD90 in the 1 month HF myocytes compared with control myocytes. However, apamin (100 nM) caused a significant prolongation of the APD90 in one month HF (p<0.05; Figure 5.3 A-D). In contrast to 1 month HF, 4 month HF significantly increased APD90 relative to control ventricular myocytes (p<0.05 vs control). Furthermore, when IKCa was blocked in 4 month HF ventricular myocytes (100 nM apamin), there was a significant prolongation of the APD50 at lower rates (i.e. p<0.05 vs control at 0.5 and 1 Hz) and a further prolongation in the APD90 (p<0.05 vs control and baseline 4 month HF).

In order to assess repolarization instability induced by block of IKCa, we measured the beat to beat variability (BTBV) of APD90 ± apamin. HF alone did not increase the

BTBV in either HF group compared to controls. Block of IKCa significantly increased

BTBV in the 4 month, but not one month, HF group. (Figure 5.4A, B)

Canine cardiac IKCa encoding proteins SK2 and SK3 were measured in control, 1 month HF and 4 month HF ventricular tissues. No significant change in SK2 protein expression in either HF group (p>0.05 vs control) was found. An ~4-fold increase in SK3 expression was found in both 1 month and 4 month HF groups. (p<0.05 vs control)

(Figure 5.4 C, D)

End-stage human heart failure

In end-stage human HF, inhibition of IKCa (100 nM apamin) caused a significant prolongation of both APD50 and APD90 compared to baseline (p<0.05, Figure 5.5 A, B).

114

In addition to AP prolongation, BTBV was significantly increased at 2 Hz in end-stage human HF ventricular cells compared to baseline (Figure 5.5 C). IKCa blockade induced late phase 3 early afterdepolarizations (EADs) in ~40% of myocytes; while no EADs were observed at baseline (Figure 5.5D, E).

Human cardiac IKCa encoding proteins SK2 and SK3 were measured in control and end-stage human HF ventricular tissue lysate (Figure 5.5 F). While there was a trend toward increased SK2 expression in HF, this did not achieve statistical significance. SK3 protein also showed a tendency to increase in human end-stage HF compared to control

(p=0.14)

IKCa inhibition and SK expression in atrial myocytes

Canine

Neither HF nor AF caused any change in APD50 compared to control. HF with superimposed AF caused significant APD90 shortening compared to control and 4 month

HF (p<0.05 vs control and 4 Mo HF). IKCa blockade (100 nM apamin) in control atrial myocytes did not change APD50 but caused an unexpected shortening of the APD90 at 0.5 and 1 Hz (p<0.05 vs baseline) as shown in Figure 5.6. IKCa blockade in the 4 month HF and 4 month HF + AF atrial cells did not cause any change in the APD50 or the APD90

(Figure 5.6). In further contrast to what we observed in the ventricle, no change in the

BTBV of repolarization was observed in either the control, 4 month HF or the 4 month

HF+AF groups after IKCa blockade (data not shown). Thus, contrary to what we observed in the ventricle, IKCa in the atria does not modulate repolarization in our chronic HF model. 115

The cardiac IKCa encoding proteins SK2 and SK3 were measured in left atrial appendage tissue from the three groups (i.e. Control, 4 month HF and 4 month HF + AF).

A 3-fold and 2-fold increase in SK2 and SK3, respectively in the 4 month HF group was observed compared to both control and 4 month HF + AF groups (p<0.05, Figure 5.7A,

B); while the 4 month HF + AF group did not differ from control. HF, with or without sustained AF, caused a similar significant decrease in calcium transient amplitude compared to controls (p< 0.05 vs control, Figure 5.7C, D).

End-stage human heart failure

Human end-stage HF atrial myocytes showed no significant change in either

APD50 or APD90 when treated with apamin (100 nM). Contrary to what we observed in the ventricular cells no difference was observed in BTBV or afterdepolarizations after apamin treatment in human HF atrial cells (Figure 5.8). SK2 and SK3 were significantly increased in atrial human HF samples compared with atrial samples from non-failing individuals (p<0.05, Figure 5.8).

Discussion

It is well known that HF is a substrate for AF and these are common co-existing disease states.73, 203 HF patients are at an increased risk for both atrial and ventricular arrhythmias, which contribute to morbidity and mortality.80 Our main findings were two- fold: first, we did not find any modulation of atrial myocyte repolarization by IKCa in the settings of normal, failing or sustained AF hearts. Secondly, IKCa is activated during HF contributing to stability of ventricular repolarization. Thus, block of IKCa in chronic HF ventricular myocytes prolonged repolarization and increased repolarization instability; 116 these effects have been shown to predict proarrhythmia.53 Consistent with our findings,

IKCa has been previously suggested to play a protective role in the human ventricle during

HF.211

One interesting question is how IKCa becomes an important modulator of ventricular repolarization during heart failure. Potential explanations for this finding include 1) increased channel expression; 2) altered channel sensitivity to calcium; 3) increased calcium concentrations; or 4) loss of other repolarizing current, thereby unmasking the role of IKCa,

In considering these possibilities, we observed an increase in SK3, but not SK2 in our canine HF model. However, we did not observe a statistically significant increase in either SK2 or SK3 in human HF, although there was a trend toward an increase in SK3

(p=0.14); our findings are in contrast to a previous report where SK2 expression was increased in human HF.211 While the expression was not significantly increased in human

HF, the inter-species differences we observed may be explained by the intrinsic enhanced variability in explanted human end-stage heart failure samples resulting from inhomogeneities in etiology, comorbidities and drug treatments.

Other possible explanations for our findings are altered channel sensitivity to calcium, altered calcium cycling, or altered repolarization. Recently it was reported in human end stage HF that SK channel sensitivity to calcium was increased in ventricular myocytes,211 which could contribute in part to our findings. Of note, other proteins such as: protein kinase, calmodulin and protein phosphatase A,32, 221 are also known to contribute to the regulation of SK channels, and thus may modulate IKCa during HF.

117

Since IKCa is a calcium-activated potassium current, HF-induced changes in ventricular calcium handling should directly affect the current. We have previously reported that in our HF model, there is a significant reduction in SR calcium release and

222 calcium transient amplitude, which would reduce rather than augment IKCa. In support of this interpretation, a recent report indicates that SR release is necessary and sufficient

223 for IKCa activation. Considering the HF-induced reduction in calcium cycling, and the lack of apamin effect in control cells where calcium cycling is robust, this suggests that altered calcium cycling is not responsible for the protective role of IKCa in heart failure.

Reduced ventricular repolarization reserve may unmask the role of small currents

224 such as IKCa . Decreased repolarization reserve is well-described in the ventricle during

HF, and attributed to reductions in repolarizing currents such as IK1, IKr and IKs, changes

196, 224, 225 which predispose to repolarization instability and/or arrhythmias. Since IKCa blockade prolonged the AP only during HF and not in controls, we suggest that the contribution of IKCa becomes evident only in settings of decreased repolarization reserve.

Thus we suggest that increased channel expression, altered calcium sensitivity of SK channels, or altered repolarization reserve may contribute to the stabilizing role of IKCa .

226 IKCa has been suggested as a therapeutic target for AF. One limitation of previous studies evaluating IKCa blockade has been a focus on primarily one cardiac chamber; this is limiting since electrical remodeling during HF is chamber-dependent.

Specifically during chronic HF, the atrial action potential is shorter,73, 74, 172 while in the ventricle the action potential is prolonged.227

118

Interest in IKCa as a therapeutic target for atrial arrhythmias followed reports of a genetic predisposition to lone AF attributed to a single nucleotide mutation in the gene

KCNN3, which encodes for SK3.228, 229 The exact mechanism(s) by which a single mutation affects SK channel function remains unclear. Data supporting both loss of function and gain of function as possible mechanisms for AF have been reported in multiple models.212, 213 Additionally, SK2 and SK3 down regulation have been associated with human AF.230

One goal of this study was to elucidate the role of IKCa in atrial electrophysiology during HF, and HF with superimposed AF. Considering HF alone, contrary to previous

212, 213 reports, IKCa blockade failed to prolong the atrial action potential in either control or

HF atrial myocytes, at physiologic rates. Our findings are different from a recent report where IKCa blockade prolonged the atrial action potential in a whole atrial preparation in a reverse rate-dependent fashion; however this only occurred at rates slower than those used in the present study.213

Despite an increase in both SK2 and SK3 expression in HF atrial tissue, the atrial action potential was not prolonged with IKCa block. Possible explanations include altered protein trafficking, altered channel calcium sensitivity or altered myocyte calcium handling. We previously reported that HF causes a decrease in calcium current in our 4 months HF tachypacing induced canine model,73, 175 and in the present study we report reduced calcium transient amplitude. Surprisingly, even in control myocytes where the calcium transient and current are normal, apamin failed to prolong the action potential.

119

208, 226 Since many studies have suggested a role for IKCa blockers for AF treatment we also evaluated a HF model with superimposed AF. In a recent report in a canine atrial tachypacing AF model, with preserved LV function, IKCa reduction via a drug which reduced calcium sensitivity of the channel caused a significant prolongation of left atrial action potentials.212 This contrasts with our AF results, in the setting of chronic HF, where IKCa blockade failed to prolong the action potential. Notably, we observed that atrial HF myocytes had similar calcium transient amplitudes whether or not AF was superimposed, suggesting that calcium cycling in HF may be insufficient to activate the current.

In agreement with a previous study of patients with chronic AF who had decreased expression of SK proteins, we found that AF superimposed on HF caused a decrease in the SK2 and SK3 protein expression relative to HF alone.230 Thus, the lack of apamin effect in the 4 months HF + AF atrial cells may be explained by a decrease in protein expression and/or a decrease in the calcium available for current activation. Since

231 IKCa is a very small current (~14 pS) , and repolarization is accelerated in AF (AP is

173 shortened ) it may be less likely that a change in IKCa would affect the overall AP duration. The same logic might apply in chronic HF, where atrial repolarization is also accelerated.73, 172

While we did not find a role for IKCa block in HF or AF, IKCa blockade might have utility in disease states where atrial repolarization is prolonged, or if there is spatial dispersion of atrial repolarization. Additionally, a recent study shows that IKCa blockade

120 in pulmonary terminates AF suggesting a potential role for IKCa blockers in paroxysmal AF.212

Limitations

Several studies have shown a gradient of SK channel expression and IKCa current density across the human ventricular wall.211 However our experiments used only the mid myocardium as our myocyte source. A similar limitation occurred in the atria, where we only studied cells from the left appendage, and there may be a difference in IKCa between free wall and appendage.213 Additionally we used only single cells which may differ in response compared to coupled cells or intact tissue.

One confounding variable in studying IKCa is that in physiologic settings, the activity varies during the in a calcium concentration-dependent manner.223

To assess the role of IKCa in an integrated system, we used perforated patch action potential recordings to permit maintenance of intrinsic calcium cycling, rather than conducting voltage clamp studies to assess the current.

Conclusions

These experiments highlight the need to evaluate novel therapeutic targets for arrhythmias in both atrial and ventricular chambers. In chronic HF, IKCa plays a protective role in the ventricle and currrent block is proarrhythmic. Notably, in early HF

(1 month canine HF), IKCa blockade is not proarrhythmic, perhaps reflecting a relatively preserved repolarization reserve with a diminished dependence on IKCa for repolarization stability, compared to chronic HF.

121

In the atria we found that IKCa does not play a role in repolarization as current block does not prolong the action potential in either human or canine HF, or during sustained AF with concurrent chronic HF, despite increased protein expression.

Collectively, our data do not support a role for IKCa blockers for the treatment of atrial arrhythmias. Rather, we suggest that therapeutic strategies to reduce IKCa may be unsafe in the setting of concurrent HF.

Conflict of Interest Statement

On behalf of all authors, the corresponding author states that there is no conflict of interest.

Acknowledgments

Technical support provided by Jeanne Green and Destiny Allen.

122

Figures

Figure 5.1 : In Vivo data from 1 month (1 Mo), 4 months (4 Mo), and heart failure with sustained AF (4 Mo HF + AF) canine groups. A. Representative ECG recording from a 4 month HF + AF dog showing the absence of P waves and irregularly irregular QRS complexes characteristic of AF. B. LVFS was similarly decreased in the 1 month HF, 4 month HF and the 4 month HF + AF groups compared to baseline. (p<0.05 vs baseline). C. Atrial contractility was decreased in 4 month HF and 4 month HF + AF groups compared to baseline. (p<0.05, N=5-7 per group).

123

Figure 5.2 : IKCa inhibition does not alter repolarization in control ventricular cells. A. Representative action potential tracing before and after 100 nM apamin recorded at 1 Hz. B. APD and C. APD dose response data (0-100nM apamin) recorded at 50 90 stimulation rates of 0.5, 1 and 2 Hz. (p=NS, n=5-11 cells per group).

124

Figure 5.3: Apamin modulates ventricular repolarization during HF. Representative action potential of a 1 month (A) HF and 4 month HF(B) ventricular cell before and after apamin superfusion. C. Summary data of APD50 in control,1 month and 4 month HF before and after apamin treatment. No difference between 1 month HF, 4 month HF and control is observed. Apamin treatment of 1 month HF cells causes a prolongation only at 2 Hz (p<0.05), likewise apamin treatment of 4 month HF cells causes a prolongation at 0.5 Hz (p<0.05) and 1 Hz (p < 0.05). D. Apamin prolongs APD90 in both 1 and 4 month HF (p<0.05).

125

Figure 5.4: IKCa contributes to ventricular repolarization stability in canine HF, and HF increases SK3 expression.

A. Beat to beat variability of APD90 (BTBV) is unchanged in either 1 month or 4 month HF vs. controls. IKCa block increases the BTBV only in the 4 month, but not the 1 month HF group (p<0.05 vs control, 1 month HF and 4 month HF). B. Representative AP tracings of control, 1 month HF and 4 month HF during superfusion with 100nm apamin. C. Representative blots of SK2 and SK3. D. SK3 in the 1 and 4 month HF groups is increased in SK3 (p<0.05 vs control) while SK2 is unchanged (N=4-5).

126

Figure 5.5 : Apamin modulates ventricular repolarization in end stage human HF. A. Representative action potential recorded at 1 Hz from an end-stage human HF ventricular myocyte before and after apamin. B. Apamin superfusion prolongs APD50 and APD90 in end-stage human HF ventricular myocytes at all rates (p<0.05 vs baseline, n=7). C. Apamin superfusion increases (p<0.05 vs baseline) BTBV at 2 Hz. D. Representative action potential showing late phase 3 EADs after apamin superfusion. E. Apamin treatment increases EAD incidence in failing human ventricular myocytes. (p<0.05 vs baseline) F. Representative blots of control and end stage human HF SK2 and SK3 proteins (SK2 p=0.556 and SK3 p=0.141 vs. control). HF: N=7 (4 male/3 female); age=52 ± 13 years and LV of 14.5 ± 5.2%; non-failing controls: n=5 (2 male/ 3 female); age = 47 ± 12 years.

127

Figure 5.5

128

Figure 5.6: IKCa block does not affect atrial repolarization in normal or diseased myocytes.

A. 100nM apamin does not affect atrial APD50 in any of the studied groups (i.e. control, 4 month HF and 4 month HF + AF, n=7-9 cells per group) B. 100nM apamin shortened the APD90 in controls at 0.5 and 1 Hz. (p<0.05). HF + AF had a shorter baseline APD90 compared to control and HF (p<0.05), however no change in APD90 was observed after apamin treatment in either 4 month HF or 4 month HF + AF groups. (n= 7-9 cells per groups) C and D. Atrial action potential tracings before and after apamin treatment from the 4 month HF group and the 4 month HF+ AF group.

129

Figure 5.7 : SK expression and calcium transients in chronic HF with and without AF. A. Representative Western blots of SK2 and SK3. B. SK2 and SK3 are increased 3- and 2-fold, respectively in the 4 month HF atria. (p<0.05 vs control and 4 month HF + AF) No differences between control and 4 month HF + AF were found in any of the subunits. (N=3) C. Representative calcium transient line scans. D. Calcium transient amplitude was decreased in the 4 month HF and 4 month HF + AF groups compared to control (p<0.05 vs control).

130

Figure 5.8 : Apamin does not modulate repolarization in end-stage human HF atrial myocytes. A. Representative atrial action potential tracing recorded at 1 Hz. B and C. Apamin did not change APD50 or APD90 in human atrial myocytes. (n=3) D. Apamin superfusion did not increase BTBV. E. Representative SK2 and SK3 Western blots. F. HF increased SK2 and SK3 expression in left atrial tissue (p<0.05 vs non-failing). HF: N=4 (2 male/2 female); age = 56 ± 4 years and LV ejection fraction of 14.5 ± 1.1%; non-failing controls: n=4 (2 male/ 2 female); age = 46 ± 14 years.

131

Figure 5.8

132

Chapter 6 Atrial electrical remodeling in heart failure:

modulation by sustained atrial fibrillation

Abstract

Background: Atrial fibrillation (AF) is the most common sustained arrhythmia in the

US, and heart failure (HF) is a common antecedent to AF. When AF and HF occur together, there is an increased risk of mortality. AF and HF both promote the progression and/or occurrence of AF, however it is unclear whether HF- and AF-induced AF are mechanistically similar.

Methods and Results: To assess atrial remodeling during the progression of HF we used an RV tachypacing induced heart failure canine model. Two HF durations were studied

(one month n=8 and four months n=7). To assess the effects of AF in the setting of HF, a group of four month HF with AF superimposed for the last six weeks (n=5) was studied.

Age matched controls (n=10) were used for comparison. Atrial repolarization was assessed via action potentials, IK1, Ito and IKsus. Structural remodeling was assessed by measures of atrial dilation and fibrosis. We found that electrophysiological remodeling translates to a change in the action potential duration late in HF. The addition of AF to

HF causes a decrease in Ito, and increased IK1 with no change in the action potential duration compared to time-matched HF. Fibrosis increases early in HF (one month) and

133 is unchanged at four months of HF. The addition of AF to four month HF further increases fibrosis. Conclusions: Even though IKsus decreases early in HF, the integrated response (the action potential) remains unchanged. While the APD is reduced in four month HF irrespective of AF, the ion currents contributing to APD shortening differ.

This suggests that etiology-specific therapeutic intervention may be needed to normalize atrial electrophysiology.

Introduction

Heart failure (HF) is an irreversible and chronic syndrome that is increasingly prevalent, and results in substantial mortality and morbidity.232 Atrial fibrillation (AF) is a common co-morbidity during HF, and HF increases the chances of developing AF by approximately 4-6 fold.80 This is significant because the development of AF in patients with advanced HF significantly increases the risk of death.80

Electrophysiologic and structural remodeling of the atria have been implicated in the pathogenesis of AF, and these forms of remodeling have been shown to progress during clinical AF and in atrial tachypacing models of AF.83, 84, 173, 233 Similarly, chronic

HF results in structural (i.e.fibrosis and atrial dilation) and electrophysiological remodeling contributing to a substrate for AF.73 Pharmacologic approaches to treat and prevent AF are often ineffective; therapeutic failures may be attributed to variable, potentially progressive remodeling of arrhythmic substrates.

To evaluate the time-dependent evolution of the atrial substrate for AF during HF, we used a canine one (acute) and four (chronic) month model of HF. The potential

134 differential effects of concurrent HF and AF were evaluated using the four month HF model with sustained AF superimposed for the final six weeks of HF (HF+AF). Here we report that while both chronic HF, and HF+AF, shorten the atrial action potential, there are differential effects on ion currents, suggesting the possibility of etiology-specific therapeutic targets for AF.

Methods

Heart failure and heart failure + atrial fibrillation canine models

All animal procedures were approved by the Institutional Animal Care and Use

Committee of the Ohio State University. A total of 30 adult mixed breed dogs of either sex (2- 5 years of age) weighing between 8 and 20 kg with normal cardiac function were used. Fifteen dogs had heart failure (HF) induced by right ventricular (RV) tachypacing as previously described.73 The duration of the RV tachypacing induced HF was one (n=8) or four months (n=7). Another group of five dogs had heart failure with superimposed sustained atrial fibrillation (HF+AF) induced by a customized pacemaker (donated by St

Jude Medical, Sylmar, CA). In the HF+AF group, two leads were implanted, one in the right atria (RA) and the second in the RV. The same protocol described above was initially used to induce HF. After 10 weeks of RV tachypacing, RA tachypacing was initiated, with the RA stimulated at 10 Hz for 60 seconds, followed by a 10 second pause for automated interrogation of atrial rhythm. This cycle of RA tachypacing was repeated every 70 seconds until atrial fibrillation (AF) was detected. Subsequent detection of normal atrial rhythm resulted in resumption of the atrial tachypacing. The duration of

135 atrial stimulation/atrial fibrillation was six weeks. Thus, the HF + AF group had a total of four months of HF prior to termination of the in vivo study, to match the HF duration of the four month HF group.

Serial echocardiograms and ECGs were done as previously reported. 174, 214 Serial pacemaker interrogations were used to monitor cardiac rhythm. A group of ten normal controls were studied in parallel. Left atrial fractional area change (FAC) and left atrial areas (diastolic and systolic) were measured as previously described.73 Each animal served as its own control with data reported as percent change from baseline.

Myocyte Isolation

On the day of the terminal procedure, the dogs were anesthetized with pentobarbital sodium (50 mg/kg IV; Nembutal, Abbott Laboratories). The heart was rapidly removed and perfused with cold cardioplegia solution containing the following in mM: NaCl 110, CaCl2 1.2, KCl 16, MgCl2 16 and NaHCO3 10. Cannulation of the left circumflex artery was used to perfuse the left atria and ventricle, as previously described.73, 215 Adjacent tissue samples were collected and snap frozen or formalin fixed for subsequent tissue analyses. Tyrode’s solution (mM) containing NaCl 130, KCl 5.4,

MgCl2 3.5, NaH2PO4 0.5, Glucose 10, HEPES 5 and taurine 20, was used as the initial perfusate. During the cell isolation process the heart was perfused with three different solutions (36ºC). First the heart was perfused for 10 minutes with Tyrode’s solution with

0.1 mM EGTA; followed by perfusion with Tyrode’s solution containing 0.3 mM calcium, 0.12 mg/ml of trypsin Inhibitor (NIBCO) and 1.33mg/ml of collagenase (Type

II, Worthington), for a maximum of 45 minutes. Following enzymatic digestion, the heart 136 was perfused with normal Tyrode’s solution for five minutes to remove residual enzyme.

Subsequently left atrial myocytes were obtained through secondary digestion, as previously reported. 73, 82, 215 After secondary digestion the cells were re-suspended in incubation buffer. This isolation procedure typically yields 40-60% rod shaped atrial myocytes. All myocyte electrophysiology experiments were conducted within 10 hours of myocyte isolation.

Electrophysiological recordings

To assess myocyte electrophysiology amphotericin-B perforated patch clamp techniques with a bath temperature of 36 ± 0.5ºC were used. The myocytes were placed in a laminin coated cell chamber (Cell Microcontrols, Norfolk, VA) and superfused with bath solution containing (in mM): 135 NaCl, 5 MgCl2, 5 KCl, 10 glucose, 1.8 CaCl2, and

5 HEPES with pH adjusted to 7.40 with NaOH. Borosilicate glass micropipettes with tip resistance of 1.5-3 MΩ, were filled with pipette solution containing the following (in mM): 100 K-aspartate, 40 KCl, 5 MgCl, 5 EGTA, 5 HEPES, pH adjusted to 7.2 with

KOH.

Action potentials (AP) were recorded in a train of 25 traces at 0.5, 1 and 2 Hz.

The average of the last 10 traces (i.e. from trace 16-25) was used to calculate the action potential duration (APD). APD was measured at 50 and 90 percent of repolarization

(APD50 and APD90).

For current recordings, only recordings with an access resistance <20 MΩ were included in the analyses. Calcium in the bath solution was reduced to 1.0 mM and 2 µM nifidepine was added to block L-type Ca2+ current. Transient outward potassium current 137

(Ito) was elicited from a holding potential of -60 mV by a series of 100 ms test potentials from -30 to +50 mV and measured as peak current minus steady state current. The steady state current was used to define the sustained outward potassium current (IKsus). The

59 potassium inward rectifying current (IK1) was elicited as previously described. Data was collected with a low noise data acquisition system Digidata 1440A (Molecular devices, Sunnyvale, CA), Clampex 10.2 software and an Axopatch 200A amplifier (Axon

Instruments, Sunnyvale, CA) was used for data acquisition.

Interstitial fibrosis

Left atrial appendage tissue was formalin fixed and embedded in paraffin and prepared in 5 µm sections, using standard procedures. Tissue sections were then stained with Masson’s trichrome which was used to define the percentage area of fibrosis, as previously described.174

Data Analysis

Cellular electrophysiology data were analyzed using Clampfit 10.3 software

(Axon Instruments) and Origin 9.0 software (OriginLab, Northampton, MA, USA).

Comparisons between three or more groups was analyzed by one-way ANOVA with post hoc least significant difference testing (Originpro 8.6, OriginLab). Paired data was analyzed using paired students t-test. All data are presented as mean ± SE and p<0.05 was the criterion for statistical significance for all comparisons.

138

Chemicals

All chemicals used were purchased from Sigma-Aldrich (St. Louis, MO, USA) and Fisher Scientific (Pittsburg, PA, USA), unless otherwise noted. All buffers and solutions were prepared daily.

Results

Echocardiographic measurements

Left ventricular fractional shortening (FS) measurement was used to assess ventricular function. In all the experimental groups (i.e. one month HF, four month HF and four month HF + AF), the left ventricular FS was significantly lower than baseline

(p<0.05 vs baseline), consistent with impaired ventricular function (Figure 6.1). Left atrial area at end-ventricular systole (LA-ES) was used to assess atrial dilation and presented as percentage change from baseline. No significant change in atrial dimension

(12.9 ± 7.5%; p = 0.31 vs baseline) occurred in the one Mo HF group. In contrast, a significant % change from baseline was found in both the four month HF and the HF +

AF groups (77.0 ± 16.7% and 20.5 ± 8.2%; both p<0.05 vs baseline) respectively, consistent with chronic but not acute HF causing atrial dilation.

To assess left atrial contractile function the fractional area change of the atria

(FAC) was measured in each of the aforementioned groups. In the one month HF group the FAC was not changed compared to baseline, while both the four month HF and the

HF + AF groups had significantly decreased atrial contractility vs. baseline (p<0.05,

Figure 6.1).

139

Electrophysiological data

Left atrial appendage action potentials from each group were measured. No change in APD50 was found between the groups at any rate (0.5, 1 and 2 Hz, Figure 6.2).

While no change in APD90 was seen in the one month HF group compared to controls, we found that the four month HF group had a significantly shorter APD90 at 0.5 and 1 Hz

(p<0.05 vs control) similar to what we previously reported.7 Similarly, we found that the four month HF + AF group had a significantly shortened APD90 (p<0.05 vs control and one month HF at 0.5,1 and 2 Hz) as seen in Figure 6.2. Furthermore, the HF+AF group had a significant shortening of the APD90 at 2 Hz (p<0.05) vs four month HF. No significant differences in the resting membrane potential occurred between groups

(Figure 6.2). Taken together, this data suggests that AP remodeling occurs during chronic

HF but not in acute (one month) HF. Since superimposing AF on the four month HF did not cause a further change in the action potential length, this suggests that the AP shortening observed in the HF+AF group mainly resulted from chronic HF, rather than the superimposed AF.

Atrial cell capacitance was used as a measure of myocyte hypertrophy. Atrial cell capacitance (pF) was 91.7 ± 6.1 for controls, 93.3 ± 11.3 in one month HF, 140.7 ± 12.2 in the four month HF and 113.2 ± 8.6 in the HF + AF group. Significant myocyte hypertrophy relative to controls was only found in the four month HF (p= 0.00026 vs control) and HF + AF (p= 0.02 vs control) groups, suggesting that hypertrophy occurs in chronic but not acute HF.

140

Inward rectifier current (IK1) was significantly decreased only in the four month

HF group as seen in Figure 6.3 (p<0.05 vs control, one month HF and the HF + AF groups). Thus, superimposed AF mitigated the effects of chronic HF on IK1.

IKsus density, which is largely IKur, decreased in one month HF and doesn’t further change with four months HF relative to control. (Figure 6.3) Superimposed AF did not further alter IKsus density beyond the effects of HF, and remained reduced relative to control.

We found that one month HF did not cause any change in the transient outward

73 current (Ito) density. However, as we previously reported, chronic HF caused a significant increase in Ito density (p<0.05 vs control and one month HF). On the other hand, the addition of AF to four month HF caused a significant decrease in Ito density

(p<0.05 vs control, one month HF and four month HF) as seen in Figure 6.4. Taken together this data suggests that increased Ito contributes to the observed APD shortening during chronic HF. However, the lack of Ito in the AF group, while previously reported in

AF84 is not consistent with the characteristic APD shortening observed with AF, supporting the role of another current or currents, such as the known AF-induced reduction in L-type Ca2+ current to the observed APD changes.173, 233 Interestingly, the addition of AF to chronic HF supersedes the effects of chronic HF alone on Ito. To more fully evaluate the contribution of Ito to AP length we analyzed Ito time constants and voltage-dependent steady state inactivation.

141

The recovery from inactivation did not differ between groups. Inactivation of Ito was best fit as the sum of two exponentials, based on the Akaike information criterion and the sum of squares. At +50 mV, the mean time constants(were 21.8 ± 3.6 ms and 112.1 ± 29.0 ms, 17.3 ± 2.3 ms and 103.6 ± 49.4 ms, and 28.9 ± 4.6 ms and 119.4 ±

19.6 ms for control (n=16), one month HF (n=13), and four month HF (n=14), respectively. The rapidly inactivating component ( contributed 54.1 ± 6.1% of the peak current in controls, 49.3 ± 5.1% in one month HF (p=NS vs control), but was significantly increased to 69.4 ± 4.5% in four month HF (p<0.05 vs control). To assess steady state inactivation of Ito, a series of 500 ms prepulses to voltages between -80 mV and +10 mV (from a holding potential of -60 mV) were followed by a 300 ms step to +50 mV (peak current). Ito elicited during each prepulse was normalized to the peak current and plotted against the respective prepulse potential. A plot of steady state activation, obtained by normalizing Ito at a given voltage to peak current (+50 mV) plotted as a function of test potential. The superimposed activation and inactivation graphs identify

234 the presence of a larger “window” current in chronic HF vs. controls (Figure 6.5). The low amplitude of Ito in HF+AF cells prevented a kinetic analysis. Taken together, the increased Ito density and window current suggests a possible role for Ito activation at plateau potentials in chronic HF.

Fibrosis

Interstitial fibrosis is known to play a role in atrial fibrillation and facilitate re- entry.83, 235 Interestingly we found that fibrosis is significantly increased in all the experimental groups (p<0.05 vs control) as shown in Figure 6.6. However, we also 142 found that fibrosis in the HF + AF group is increased relative to four month HF (p<0.05), suggesting that while fibrosis formation begins at early stages of HF, AF increases the fibrosis more than seen with HF alone.

Discussion

HF and AF commonly co-exist and the addition of one to the other significantly increases the risk of death beyond that of either disease alone.80 Clinical HF is a chronic syndrome and possible mechanisms through which HF can promote AF include structural remodeling (dilation and fibrosis), and electrophysiological changes. AF itself also promotes the progression and/or maintenance of AF through electrophysiologic changes

84, 173 and increased fibrosis.70, 236 In this study, we found that the HF- and AF-induced remodeling which promote a substrate for AF have some similarities, and some key differences. We found that chronic, but not short-term, HF causes atrial dilation, increased interstitial fibrosis, and shortens the APD, which is attributable to an increase in Ito density. Interestingly, AF superimposed on HF results in continued shortening of the atrial APD, but decreases Ito, and results in further increases in interstitial fibrosis compared to HF alone.

Interestingly, we only found a decrease in IKsus at one month HF. However this didn’t translate to a change in the action potential, as there were no differences in the action potential between our control and acute (one month HF) groups, despite the development of HF (evident as impaired ventricular function). This is interesting as the electrophysiologic atrial remodeling due to short-term canine tachypacing-induced HF

143 has been extensively studied by other groups,86, 237 where increased APD and reduced

86 atrial Ito has been reported. The differences between previous reports (4 to 5 weeks of

HF) and our one month HF group may be attributable to variations in tachypacing protocols (our rates were 180-200 BPM vs. the 240-250 BPM used in other studies). In the present study, similar to what we previously reported, chronic HF (four months HF)

73 caused shortening of the atrial APD and an increase in Ito density. This is consistent with a previous report where increased human atrial Ito density was inversely related to

74 decreased ventricular function. Thus, atrial Ito appears to be modulated by duration and/or severity of ventricular dysfunction.

In order to further investigate the possible involvement of Ito in the modulation of the AP, we examined the activation and inactivation of the current. While Ito has been previously reported to have monoexponetial inactivation in rabbit atrial cells,238 we found biexponential inactivation, with chronic HF increasing the contribution of the fast inactivating component similar to that found in ventricular and Purkinje fiber cells.239

Potential explanations for these results could be altered expression of the auxiliary proteins, such as KChIP2 or DPP6, which can alter inactivation kinetics,240, 241 a hypothesis supported by our previous finding of increased KChIP2 expression in four month HF.73 Using an existing method234 to examine “window current” by plotting the inactivation curve with the activation curve of Ito (Figure 6.5), we observed that chronic

HF increases window current. This suggests that in chronic HF Ito is still active at membrane potentials occurring during phase 2 of the action potential, which would contribute to shortening of the AP. This concept is supported by dynamic clamp studies 144 in rabbit and human atrial myocytes where increased Ito in fact shortens the action

242 potential. Collectively, this suggests that the increased Ito in chronic HF does

73 contribute to the observed and previously reported APD shortening. In addition to Ito, we and others have previously reported chronic HF changes in atrial myocytes of

73 73, 233 73 233 233 decreased IKur, IKs, and IK1, increased NCX, as well as decreased ICaL.

As previously reported with AF-induced electrical remodeling,84, 87 we found that

AF superimposed on chronic HF decreases Ito. This has two implications: first that the

APD shortening observed in AF is not attributable to altered Ito, as reduced Ito would be expected to lengthen not shorten APD; second, the modulation of Ito by AF overrides HF- induced changes in Ito. The reasons for this AF-induced change are not clear, and will require further examination to determine the mechanism(s).

84, 85 It has previously been reported that AF induces increased IK1, and we found that the addition of AF to chronic HF restores the HF-induced decreases in IK1. As with

Ito, this suggests that rapid atrial rate itself can induce effects on atrial currents which supersede the effects of chronic HF. Notably, AF does not change sustained outward currents beyond the changes caused by chronic HF. Thus, while both HF and HF+AF have shortened APs, the mechanisms appear to vary, and require further investigation.

For example, the shortening of the AP could result from changes in calcium current, which was not evaluated in this study, but would be consistent with previous reports in atrial myocytes from animal AF models and human AF85, 87, 173, 233 of decreased in L-

Type Ca+2 current,. L-Type Ca2+ current blockade in normal atrial myocytes results in

145 abbreviation of the AP similar to that seen in in our chronic HF and AF,86, 87 and thus may provide a plausible explanation for the shorter AP following AF. Additionally, it has also been reported that patients with chronic AF have a constitutively active IKach, which could contribute to the shortening of the AP in AF.28

Atrial fibrosis increases the chances of reentrant arrhythmias due to altered conduction velocity and regional inhomogeneities of conduction.70, 83 Patients with AF or

HF are known to have increased atrial fibrosis.243 In our study, we found that increased fibrosis occurs early in response to HF, while the electrophysiological changes take longer to develop. However, AF superimposed on chronic HF causes a further increase in fibrotic tissue furthersupporting a role for fibrosis is one of the factors contributing to

AF-induced AF progression.83, 236

Another structural change that has been linked with fibrosis and subsequent conduction heterogeneity is atrial dilation.235 Atrial dilation has been implicated in the pathogenesis and perpetuation of atrial fibrillation as atrial size is an important predictor of AF susceptibility and recurrence in patients.236, 244 However, whether atrial size predicts AF, or AF causes the dilation is not clear due to the observational nature of clinical studies looking at this issue. Based on our findings we suggest that atrial dilation occurs due to chronic HF and that chronic AF itself does not further increase the atrial diameter.

In addition to structural remodeling, the characteristic short atrial APD during AF contributes to a decreased refractory period facilitating the development of faster

146 reentrant circuits.245 Prolongation of the AP (and refractory period) is a preferred therapeutic target for the termination of AF. However, it is not surprising that considering the multiple changes and factors in AF formation and perpetuation, AF treatment remains challenging. Since most AF occurs in the presence of a preexisting condition such as HF, it might be more effective to try to mitigate the changes favoring the development and progression of AF such as electrical and structural remodeling.

Limitations

We only used atrial cells from the left atrial appendage and it is known that cells from different parts of the atria such as the atrial wall have different electrophysiological

246 properties. Additionally, L-type calcium current and IKAch are important modulators of the atrial action potential and were not measured in this study.28, 245

Summary and Conclusions

In summary, atrial structural remodeling appears to occur relatively early during

HF, with action potential remodeling occurring later in the disease process. AF occurring in the setting of HF results in altered electrophysiologic remodeling beyond that of HF alone, although the integrated cellular response – the action potential remains shortened.

These findings will require further investigation to determine the causes of differential

HF- and AF-mediated atrial changes, which could be used to develop etiology-specific therapeutic interventions.

147

Figures

Figure 6.1: In Vivo data from one and four month HF and HF + AF canine groups

A.Representative baseline and atrial fibrillation ECGs with the characteristic irregularly irregular QRS waves and absence of P waves. B. Left ventricular fractional shortening shows a significant impairment in ventricular function in all HF groups compared to baseline (p<0.05 vs. baseline). C. Left atrial fractional area change was significantly decreased in both the four month HF (p<0.05 vs baseline) and the HF + AF groups (p<0.05 vs baseline and one month HF).

148

Figure 6.2: Chronic but not acute HF causes a shortening of the atrial action potential, while AF superimposed on chronic HF doesn't cause any further change.

A. Representative action potential from each group at 1 Hz. B. No difference between groups was found in the APD . C. APD was significantly shorter in the four month HF 50 90 group compared to control (p<0.05 vs Baseline at 0.5 and 1 Hz) and in the HF + AF group compared to control and one month HF (p<0.05 vs baseline and one month HF at 0.5, 1 and 2 Hz). The HF + AF APD was significantly shorter vs four month HF only at 90 2 Hz. D. No difference was found between groups in the resting membrane potential.

149

Figure 6.3: Four month HF and HF+AF alter I and I . K1 Ksus

A. I at -80 mV is significantly reduced only in the four month HF group (p<0.05 vs K1 control, one month HF and four month HF + AF); superimposed AF restores I to values K1 no different from control. B. I is significantly decreased in all groups compared to Ksus control (p<0.05 vs control).

150

Figure 6.4: Chronic HF causes a significant increase in I density while to superimposed AF significantly reduces I vs. control. to

A-B. I-V curve and summarized I density at +50 mV in control, one month HF, four to month HF and four month HF + AF canine atrial myocytes. I density at +50 mV is to significantly increased in the four month HF group compared to control and one month HF (p<0.05 vs control and one month HF). AF superimposed on four month HF caused a significant decrease in the I current density at +50 mV compared to control, one month to HF and four month HF(p<0.05). C. I slope conductance was significantly decreased in to one month HF compared to control (p<0.05). In four month HF, I slope conductance to was significantly increased compared to one month HF (p<0.05 vs 1 Mo HF), while in the HF + AF group it was significantly decreased compared to all other groups (p<0.05).

151

Figure 6.5: Chronic heart failure alters I inactivation kinetics. to

A-C. Plot of steady state inactivation and activation of Ito in controls (A) and chronic HF(B) fitted to a Boltzmann Function. Superimposed fits of (A) and (B) reveals the presence of an Ito “window current” that is greater in four month HF than control (C). D-

E. Heart failure did not alter the time constants of Ito inactivation: Ito,fast (D) and Ito,slow

(E). F. Contribution of Ito,fast to total inactivation is significantly increased in four month HF compared to control and one month HF.

152

Figure 6.6: Atrial fibrosis occurs early in HF, and superimposed AF causes more fibrosis than chronic HF alone.

A. Representative Masson’s trichrome stain of left atrial appendage tissue for each group. B. Fibrosis is significantly increased from control in one month HF (p<0.05 vs control) and four month HF (p<0.05 vs control). A significant increase in fibrosis was observed in the 4 Mo HF + AF group (p<0.05 vs control and 4 Mo HF).

153

Chapter 7 : Summary and Conclusions

The studies presented in this dissertation describe different forms of acquired electrophysiological remodeling, mainly focused on repolarizing K+ currents and action potential abnormalities. One common feature of many cardiac diseases is altered repolarization, which in turn can predispose to the development of malignant arrhythmias. The studies presented in this dissertation address different mechanisms and settings where electrophysiological remodeling can occur. In some cases these studies provide insights into novel therapeutic targets and in other cases they provide insights into mechanisms behind proarrhythmic or antiarrhythmic effects.

Sudden cardiac death (SCD) is the leading cause of death in industrialized countries and its risk is increased in patients who have survived a myocardial infarction

(MI). 54, 94Even though it has been previously reported that endurance exercise training prevents malignant arrhythmias, the electrophysiologic mechanisms behind exercise- induced protection are not entirely known. In chapter 2 we used a well-established SCD

(susceptible to ventricular arrhythmias) canine model developed by our collaborator Dr.

Billman. In this study the dogs that were susceptible to ventricular arrhythmias and were subjected to endurance exercise training, were protected against ventricular arrhythmias.

Our results indicate that some of that protection conferred by exercise training can be 154 attributed to an exercise induced normalization of the action potential/QTc duration, a reduction in beat to beat variability of repolarization, normalization of IKr current, resting membrane potential, and improved Ca2+ handling. Therefore we can conclude that endurance exercise protection in post-MI setting can be attributed in part to normalization of ventricular repolarization measured both in vivo and in vitro.

Another way of developing ventricular arrhythmias is by the action of drugs on cardiac repolarization in the absence of an underlying cardiac disease. This is called cardiotoxicity and it accounts for 27% of drug withdrawals from the market. 51The most common form of malignant drug induced proarrhythmia is Torsade de Pointes (TdP).

Drug companies assess a drug’s torsadogenic potential primarily by analyzing its IKr blockade potential. However in chapter 3, based on a case report of an otherwise healthy woman that developed TdP after initiation of ibandronate therapy, we found mechanisms other than IKr blockade that contribute to the ibandronate arrhythmogenic effect.

Specifically, we found that in normal canine ventricular myocytes, ibandronate treatment increased the action potential duration, beat to beat variability and caused cellular arrhythmias, at therapeutic concentrations. Because of the lack of ibandronate effect on

IKr current, we decided to study other potassium currents and calcium handling mechanisms. We found that ibandronate causes a reduction in Ito, inhibits the ryanodine receptor (RyR) and washout of the drug releases the RyR inhibition leading to an increase in spontaneous Ca2+ waves and spark frequency. In addition to in vitro studies we also used in silico studies, in collaboration with Dr. Thomas Hund, where we found that a

2+ synergistic effect of Ito blockade and RyR Ca release is sufficient to cause altered

155 electrical function. This study demonstrates that drug cardiotoxicity can occur by means of other mechanisms in addition to the known IKr blockade and highlights the challenges and limitations in the actual safety screening used by drug companies.

In chapter 4 we analyzed the differential effects of peroxynitrite on atrial and ventricular myocyte electrophysiology. Increased reactive oxygen (ROS) and reactive nitrogen species (RNS) have been implicated in the pathophysiology of numerous cardiovascular diseases such as MI, atrial fibrillation (AF), and heart failure (HF). 159-

162One of the major problems of previous studies assessing the effects of ROS and RNS is their focus on one single chamber (i.e. atria or ventricle). Due to the inherent electrophysiological differences between atria and ventricle we studied the effect of peroxynitrite in myocytes from both chambers in order to elucidate chamber specific effects. Control atrial and ventricular cells were treated with the peroxynitrite donor SIN-

1. SIN-1 caused a prolongation of the action potential in both chambers; however, the atrial APD prolongation was rate independent while in the ventricle it was rate dependent. Additionally, an increase in beat to beat variability of repolarization was observed only in ventricular cells. Since atrial K+ currents were unaffected by SIN-1 and only Ito was decreased in the ventricle, to further understand the cause of the observed

APD prolongation we used the SERCA blocker thapsigargin to assess the involvement of

Ca2+ handling in the observed SIN-1 effects. Thapsigargin pre-treatment abolished the

SIN-1 effects on repolarization in both cell types. Collectively, this data suggests that acute increases in peroxynitrite in the atria may have an antiarrhythmic effects while in the ventricle it may have proarrhythmic effect. However, in order to determine the

156 specific molecular mechanisms by which peroxynitrite modulated Ca2+ handling further investigation will be required.

Chamber specific effects or remodeling are important to take into consideration when talking about heart failure (HF) since it induces chamber specific remodeling. For example we previously reported that during chronic HF the atrial AP gets shorter73 , while it is well known that the ventricular AP prolongs. Contrary to our observations of atrial APs during HF, other labs have reported that the atrial AP gets longer during HF; however, they use a different canine right ventricular tachypacing model of relatively short (up to 5 weeks) duration.233 In agreement with our canine model, human atrial data also suggests a decreased in the APD in the atria of HF patients. 172These chamber specific differences have to be taken into consideration when talking about drug treatments since drugs that act by shortening the action potential could be beneficial for the ventricle but promote arrhythmias in the atria. Thus, the importance of chamber specific drug targets as therapeutic strategies for the treatment of HF is clear, while safety must be considered in both chambers.

In chapter 5 we evaluated the contribution of the calcium activated K+ channels to the atrial and ventricular repolarization in health and disease by using two durations of canine HF, canine HF with superimposed AF and human heart failure cardiac myocytes.

Contrary to previous reports where IKCa blockade with apamin was atrial specific and was antiarrhythmic,247 in our studies apamin had no effect on atrial AP in any of the studied groups (control, HF, HF+AF and human HF). No effect of apamin in ventricular cells from control animals was observed. However it is known that during HF there is a

157 decrease in repolarization reserve 224 in the ventricle, and that repolarization is prolonged.

To test the effects of apamin in the settings of decreased repolarization reserve we used two time points of HF (one and four months) in addition to using ventricular myocytes from end-stage human heart failure. No effect of apamin was observed on the ventricular

AP in the one month HF myocytes. However in the four month canine HF myocytes, apamin caused an increase in the AP in addition to increasing the beat to beat variability of repolarization. The same happened in the human heart failure ventricular myocytes: increased APD, increased beat to beat variability of repolarization, and induction of cellular arrhythmias. This data agrees with previous studies where apamin was also found to be proarrhythmic in human heart failure ventricular cells. Previous studies have

73, 203 suggested a potential role for IKCa blockade for the treatment of atrial arrhythmias.

However our data do not support the use of IKCa blockers for atrial arrhythmias, and this approach during HF may be unsafe. Additionally this study highlights the importance of evaluating both chambers when screening for novel therapeutic approaches.

Atrial fibrillation (AF) is the most common sustained arrhythmia in the US. HF increases the chances of developing AF and as a consequence increases the risk of death.

80, 203 The possible mechanisms through which HF can promote AF include increased fibrosis, reactive oxygen species and electrophysiological changes. It is known that AF begets AF, thus the need to identify factors that promote AF as well as factors that facilitate AF perpetuation. Chapter 6 examines atrial electrophysiological and structural remodeling during the progression of HF and that during HF with superimposed AF. As in the previous chapter here we used two different time points of HF (one and four

158 months), in order to assess atrial remodeling during the progression of HF, and in four month HF with superimposed AF, to examine the AF-induced remodeling. We found that electrophysiological remodeling does not occur after only one month of HF, therefore we can speculate that a more prolonged or worsened disease state is needed for electrophysiological remodeling to occur. As we previously reported four month HF caused a shortening of the atrial AP, increased Ito, decreased IK1 and decreased IKsus. AF superimposed on the HF reversed the HF-induced increase in Ito and caused a further decrease in Ito without an accompanying change the HF-induced APD shortening. It has been reported that Ito density increases in concert with increased ventricular dysfunction in human atria.74 The APD shortening in HF is consistent with altered Ito in the four month HF group but not the HF + AF group. Most interestingly, the effects of sustained high atrial rate (AF) can reverse or supersede the effects of HF on atrial ion currents. In addition to electrophysiological remodeling, structural remodeling of the atria also occurs; therefore we measured atrial dilation and fibrosis. We found that the atria was dilated at four months of HF but not at one month of HF; the addition of AF to the 4 months HF did not caused any further change in atrial dimension. Fibrosis is known to promote reentrant arrhythmias 70, 83 and patients with HF and AF have increased atria fibrosis.243 In our study we found that fibrosis appears early in HF (1 month) and the progression of HF (4 months) does not increase it, but the addition of AF to four month

HF further increased the amount of fibrosis suggesting that it might play a role in AF perpetuation, and that AF itself may drive fibrosis independent of HF. These results suggest that targeting fibrosis might be a plausible therapeutic strategy to avoid AF

159 perpetuation. However, more in depth investigation into the mechanisms of increased fibrosis formation in HF and AF will be needed before it can be used as a therapeutic target.

In summary, acquired electrophysiological remodeling caused by either disease or drug actions can be harmful. Understanding these changes can help avoid detrimental outcomes and develop new safer treatments. However when interpreting patch clamp data from isolated cardiac myocytes we have to take into consideration that transmural heterogeneity of the AP exists in the ventricle and that regional variability of the AP exists in the atria. One limitation of the use of isolated cardiomyocytes in our studies is that they come from only one part of the ventricular wall (in our case midmyocardium), while the atrial myocytes were isolated from the left atrial appendage. Despite the aforementioned limitations, the data presented in this work offers insights in the causes of repolarization abnormalities that occur in a diverse set of cardiovascular conditions that can predispose to cardiac arrhythmias.

160

Reference List

Reference List

1. Berndt Lüderitz MD. History of the Disorders of Cardiac Rhythm. Futura Publishing Company, Inc.; 1998.

2. Barold SS (2003) Willem Einthoven and the birth of clinical electrocardiography a hundred years ago. Card Electrophysiol Rev: 7:99-104.

3. Schuetze SM (1983) The discovery of the action potential. Trends Neurosci: 6:164-168.

4. Schwiening CJ (2012) A brief historical perspective: Hodgkin and Huxley. J Physiol: 590:2571-2575.

5. Hodgkin AL, Huxley AF, KATZ B (1952) Measurement of current-voltage relations in the membrane of the giant axon of Loligo. J Physiol: 116:424-448.

6. Kleber AG, Niggli E, McGuigan JA, Weingart R (2006) The early years of cellular cardiac electrophysiology and Silvio Weidmann (1921-2005). Heart Rhythm: 3:353-359.

7. DRAPER MH, WEIDMANN S (1951) Cardiac resting and action potentials recorded with an intracellular electrode. J Physiol: 115:74-94.

8. WEIDMANN S (1951) Effect of current flow on the membrane potential of . J Physiol: 115:227-236.

9. Brown H, DiFrancesco D (1980) Voltage-clamp investigations of membrane currents underlying pace-maker activity in rabbit sino-atrial node. J Physiol: 308:331-351.

161

10. Niggli E, Kleber A, Weingart R (2006) Founder of cardiac cellular electrophysiology: honouring Silvio Weidmann, 7 April. J Physiol: 570:431-432.

11. Fozzard HA (2002) Cardiac sodium and calcium channels: a history of excitatory currents. Cardiovasc Res: 55:1-8.

12. Fulop L, Banyasz T, Magyar J et al. (2004) Reopening of L-type calcium channels in human ventricular myocytes during applied epicardial action potentials. Acta Physiol Scand: 180:39-47.

13. Barry DM, Nerbonne JM (1996) Myocardial potassium channels: electrophysiological and molecular diversity. Annu Rev Physiol: 58:363-394.

14. Oudit GY, Kassiri Z, Sah R et al. (2001) The molecular physiology of the cardiac transient outward potassium current (I(to)) in normal and diseased myocardium. J Mol Cell Cardiol: 33:851-872.

15. Guo W, Xu H, London B, Nerbonne JM (1999) Molecular basis of transient outward K+ current diversity in mouse ventricular myocytes. J Physiol: 521 Pt 3:587-599.

16. Nabauer M, Beuckelmann DJ, Uberfuhr P, Steinbeck G (1996) Regional differences in current density and rate-dependent properties of the transient outward current in subepicardial and subendocardial myocytes of human left ventricle. Circulation: 93:168-177.

17. Sah R, Ramirez RJ, Oudit GY et al. (2003) Regulation of cardiac excitation- contraction coupling by action potential repolarization: role of the transient outward potassium current (I(to)). J Physiol: 546:5-18.

18. Radicke S, Cotella D, Graf EM, Ravens U, Wettwer E (2005) Expression and function of dipeptidyl-aminopeptidase-like protein 6 as a putative beta-subunit of human cardiac transient outward current encoded by Kv4.3. J Physiol: 565:751- 756.

19. Ward CA, Giles WR (1997) Ionic mechanism of the effects of hydrogen peroxide in rat ventricular myocytes. J Physiol: 500 ( Pt 3):631-642.

20. Sanguinetti MC, Jurkiewicz NK (1990) Two components of cardiac delayed rectifier K+ current. Differential sensitivity to block by class III antiarrhythmic agents. J Gen Physiol: 96:195-215.

21. Nerbonne JM, Guo W (2002) Heterogeneous expression of voltage-gated potassium channels in the heart: roles in normal excitation and arrhythmias. J Cardiovasc Electrophysiol: 13:406-409.

162

22. Schram G, Pourrier M, Melnyk P, Nattel S (2002) Differential distribution of cardiac ion channel expression as a basis for regional specialization in electrical function. Circ Res: 90:939-950.

23. Sridhar A, da Cunha DN, Lacombe VA et al. (2007) The plateau outward current in canine ventricle, sensitive to 4-aminopyridine, is a constitutive contributor to ventricular repolarization. Br J Pharmacol: 152:870-879.

24. Postema PG, Ritsema van Eck HJ, Opthof T et al. (2009) IK1 modulates the U- wave: insights in a 100-year-old enigma. Heart Rhythm: 6:393-400.

25. Lopatin AN, Nichols CG (2001) Inward rectifiers in the heart: an update on I(K1). J Mol Cell Cardiol: 33:625-638.

26. Rials SJ, Xu X, Wu Y et al. (2001) Restoration of normal ventricular electrophysiology in renovascular hypertensive rabbits after treatment with losartan. J Cardiovasc Pharmacol: 37:317-323.

27. Koumi S, Backer CL, Arentzen CE, Sato R (1995) beta-Adrenergic modulation of the inwardly rectifying potassium channel in isolated human ventricular myocytes. Alteration in channel response to beta-adrenergic stimulation in failing human hearts. J Clin Invest: 96:2870-2881.

28. Dobrev D, Friedrich A, Voigt N et al. (2005) The G protein-gated potassium current I(K,ACh) is constitutively active in patients with chronic atrial fibrillation. Circulation: 112:3697-3706.

29. Xu Y, Tuteja D, Zhang Z et al. (2003) Molecular identification and functional roles of a Ca(2+)-activated K+ channel in human and mouse hearts. J Biol Chem: 278:49085-49094.

30. Kohler M, Hirschberg B, Bond CT et al. (1996) Small-conductance, calcium- activated potassium channels from mammalian brain. Science: 273:1709-1714.

31. Grunnet M, Jensen BS, Olesen SP, Klaerke DA (2001) Apamin interacts with all subtypes of cloned small-conductance Ca2+-activated K+ channels. Pflugers Arch: 441:544-550.

32. Xia XM, Fakler B, Rivard A et al. (1998) Mechanism of calcium gating in small- conductance calcium-activated potassium channels. Nature: 395:503-507.

33. Kuo HC, Cheng CF, Clark RB et al. (2001) A defect in the Kv channel-interacting protein 2 (KChIP2) gene leads to a complete loss of I(to) and confers susceptibility to . Cell: 107:801-813.

163

34. Nabauer M, Kaab S (1998) Potassium channel down-regulation in heart failure. Cardiovasc Res: 37:324-334.

35. Keating MT, Sanguinetti MC (2001) Molecular and cellular mechanisms of cardiac arrhythmias. Cell: 104:569-580.

36. Belevych AE, Terentyev D, Viatchenko-Karpinski S et al. (2009) Redox modification of ryanodine receptors underlies calcium alternans in a canine model of sudden cardiac death. Cardiovasc Res: 84:387-395.

37. Ho HT, Stevens SC, Terentyeva R et al. (2011) Arrhythmogenic adverse effects of cardiac glycosides are mediated by redox modification of ryanodine receptors. J Physiol: 589:4697-4708.

38. Bonilla IM, Sridhar A, Gyorke S, Cardounel AJ, Carnes CA (2012) Nitric oxide synthases and atrial fibrillation. Front Physiol: 3:105.

39. Ahmmed GU, Xu Y, Hong DP et al. (2001) Nitric oxide modulates cardiac Na(+) channel via protein kinase A and protein kinase G. Circ Res: 89:1005-1013.

40. Ahern GP, Hsu SF, Klyachko VA, Jackson MB (2000) Induction of persistent sodium current by exogenous and endogenous nitric oxide. J Biol Chem: 275:28810-28815.

41. Campbell DL, Stamler JS, Strauss HC (1996) Redox modulation of L-type calcium channels in ferret ventricular myocytes. Dual mechanism regulation by nitric oxide and S-nitrosothiols. J Gen Physiol: 108:277-293.

42. Nunez L, Vaquero M, Gomez R et al. (2006) Nitric oxide blocks hKv1.5 channels by S-nitrosylation and by a cyclic GMP-dependent mechanism. Cardiovasc Res: 72:80-89.

43. Gomez R, Nunez L, Vaquero M et al. (2008) Nitric oxide inhibits Kv4.3 and human cardiac transient outward potassium current (Ito1). Cardiovasc Res: 80:375-384.

44. Taglialatela M, Pannaccione A, Iossa S, Castaldo P, Annunziato L (1999) Modulation of the K(+) channels encoded by the human ether-a-gogo-related gene-1 (hERG1) by nitric oxide. Mol Pharmacol: 56:1298-1308.

45. Taglialatela M, Castaldo P, Iossa S et al. (1997) Regulation of the human ether-a- gogo related gene (HERG) K+ channels by reactive oxygen species. Proc Natl Acad Sci U S A: 94:11698-11703.

46. Asada K, Kurokawa J, Furukawa T (2009) Redox- and calmodulin-dependent S- nitrosylation of the KCNQ1 channel. J Biol Chem: 284:6014-6020. 164

47. Nattel S, Khairy P, Schram G (2001) Arrhythmogenic ionic remodeling: adaptive responses with maladaptive consequences. Trends Cardiovasc Med: 11:295-301.

48. Antzelevitch C (2005) Role of transmural dispersion of repolarization in the genesis of drug-induced torsades de pointes. Heart Rhythm: 2:S9-15.

49. Bednar MM, Harrigan EP, Anziano RJ, Camm AJ, Ruskin JN (2001) The QT interval. Prog Cardiovasc Dis: 43:1-45.

50. Haverkamp W, Breithardt G, Camm AJ et al. (2000) The potential for QT prolongation and proarrhythmia by non-antiarrhythmic drugs: clinical and regulatory implications. Report on a policy conference of the European Society of . Eur Heart J: 21:1216-1231.

51. Casartelli A, Lanzoni A, Comelli R et al. (2011) A novel and integrated approach for the identification and characterization of drug-induced cardiac toxicity in the dog. Toxicol Pathol: 39:361-371.

52. Guo L, Dong Z, Guthrie H (2009) Validation of a guinea pig Langendorff heart model for assessing potential cardiovascular liability of drug candidates. J Pharmacol Toxicol Methods: 60:130-151.

53. Thomsen MB, Oros A, Schoenmakers M et al. (2007) Proarrhythmic electrical remodelling is associated with increased beat-to-beat variability of repolarisation. Cardiovasc Res: 73:521-530.

54. Kannel WB, Sorlie P, McNamara PM (1979) Prognosis after initial myocardial infarction: the Framingham study. Am J Cardiol: 44:53-59.

55. Dun W, Baba S, Yagi T, Boyden PA (2004) Dynamic remodeling of K+ and Ca2+ currents in cells that survived in the epicardial border zone of canine healed infarcted heart. Am J Physiol Heart Circ Physiol: 287:H1046-H1054.

56. Dun W, Boyden PA (2005) Diverse phenotypes of outward currents in cells that have survived in the 5-day-infarcted heart. Am J Physiol Heart Circ Physiol: 289:H667-H673.

57. Liu N, Niu H, Li Y et al. (2004) The changes of potassium currents in rabbit ventricle with healed myocardial infarction. J Huazhong Univ Sci Technolog Med Sci: 24:128-131.

58. Kaprielian R, Sah R, Nguyen T, Wickenden AD, Backx PH (2002) Myocardial infarction in rat eliminates regional heterogeneity of AP profiles, I(to) K(+) currents, and [Ca(2+)](i) transients. Am J Physiol Heart Circ Physiol: 283:H1157- H1168.

165

59. Sridhar A, Nishijima Y, Terentyev D et al. (2008) Repolarization abnormalities and afterdepolarizations in a canine model of sudden cardiac death. Am J Physiol Regul Integr Comp Physiol: 295:R1463-R1472.

60. Spear JF, Horowitz LN, Hodess AB, MacVaugh H, III, Moore EN (1979) Cellular electrophysiology of human myocardial infarction. 1. Abnormalities of cellular activation. Circulation: 59:247-256.

61. Litwin SE, Zhang D, Bridge JH (2000) Dyssynchronous Ca(2+) sparks in myocytes from infarcted hearts. Circ Res: 87:1040-1047.

62. Kim YK, Kim SJ, Kramer CM et al. (2002) Altered excitation-contraction coupling in myocytes from remodeled myocardium after chronic myocardial infarction. J Mol Cell Cardiol: 34:63-73.

63. Fukuda K, Davies SS, Nakajima T et al. (2005) Oxidative mediated lipid peroxidation recapitulates proarrhythmic effects on cardiac sodium channels. Circ Res: 97:1262-1269.

64. Gardner PI, Ursell PC, Fenoglio JJ, Jr., Wit AL (1985) Electrophysiologic and anatomic basis for fractionated electrograms recorded from healed myocardial infarcts. Circulation: 72:596-611.

65. Halici Z, Dengiz GO, Odabasoglu F et al. (2007) Amiodarone has anti- inflammatory and anti-oxidative properties: an experimental study in rats with carrageenan-induced paw edema. Eur J Pharmacol: 566:215-221.

66. Chen-Scarabelli C, Saravolatz L, Jr., Murad Y et al. (2012) A critical review of the use of carvedilol in ischemic heart disease. Am J Cardiovasc Drugs: 12:391- 401.

67. Bonilla IM, Belevych A, Sridhar A et al. (2012) Endurance Exercise Training Normalizes Repolarization and Calcium Handling Abnormalities Preventing Ventricular Fibrillation in a Model of Sudden Cardiac Death. J Appl Physiol: 113:1772-1783.

68. Kjekshus J (1990) Arrhythmias and mortality in congestive heart failure. Am J Cardiol: 65:42I-48I.

69. Karayannis G, Kitsios G, Kotidis H, Triposkiadis F (2008) Left atrial remodelling contributes to the progression of asymptomatic left ventricular systolic dysfunction to chronic symptomatic heart failure. Heart Fail Rev: 13:91-98.

166

70. Aldhoon B, Kucera T, Smorodinova N et al. (2013) Associations between and permanent atrial fibrillation in advanced heart failure. Physiol Res: 62:247-255.

71. Li D, Fareh S, Leung TK, Nattel S (1999) Promotion of atrial fibrillation by heart failure in dogs: atrial remodeling of a different sort. Circulation: 100:87-95.

72. Li D, Melnyk P, Feng J et al. (2000) Effects of experimental heart failure on atrial cellular and ionic electrophysiology. Circulation: 101:2631-2638.

73. Sridhar A, Nishijima Y, Terentyev D et al. (2009) Chronic heart failure and the substrate for atrial fibrillation. Cardiovasc Res: 84:227-236.

74. Schreieck J, Wang Y, Overbeck M, Schomig A, Schmitt C (2000) Altered transient outward current in human atrial myocytes of patients with reduced left ventricular function. J Cardiovasc Electrophysiol: 11:180-192.

75. Tsuji Y, Opthof T, Kamiya K et al. (2000) Pacing-induced heart failure causes a reduction of delayed rectifier potassium currents along with decreases in calcium and transient outward currents in rabbit ventricle. Cardiovasc Res: 48:300-309.

76. Li GR, Lau CP, Ducharme A, Tardif JC, Nattel S (2002) Transmural action potential and ionic current remodeling in ventricles of failing canine hearts. Am J Physiol Heart Circ Physiol: 283:H1031-H1041.

77. Li GR, Lau CP, Leung TK, Nattel S (2004) Ionic current abnormalities associated with prolonged action potentials in cardiomyocytes from diseased human right ventricles. Heart Rhythm: 1:460-468.

78. Ouadid H, Albat B, Nargeot J (1995) Calcium currents in diseased human cardiac cells. J Cardiovasc Pharmacol: 25:282-291.

79. Zicha S, Maltsev VA, Nattel S, Sabbah HN, Undrovinas AI (2004) Post- transcriptional alterations in the expression of cardiac Na+ channel subunits in chronic heart failure. J Mol Cell Cardiol: 37:91-100.

80. Wang TJ, Larson MG, Levy D et al. (2003) Temporal relations of atrial fibrillation and congestive heart failure and their joint influence on mortality: the Framingham Heart Study. Circulation: 107:2920-2925.

81. Wijffels MC, Kirchhof CJ, Dorland R, Allessie MA (1995) Atrial fibrillation begets atrial fibrillation. A study in awake chronically instrumented goats. Circulation: 92:1954-1968.

167

82. Nishijima Y, Sridhar A, Bonilla I et al. (2011) Tetrahydrobiopterin depletion and NOS2 uncoupling contribute to heart failure-induced alterations in atrial electrophysiology. Cardiovasc Res: 91:71-79.

83. Allessie M, Ausma J, Schotten U (2002) Electrical, contractile and structural remodeling during atrial fibrillation. Cardiovasc Res: 54:230-246.

84. Van Wagoner DR, Pond AL, McCarthy PM, Trimmer JS, Nerbonne JM (1997) Outward K+ current densities and Kv1.5 expression are reduced in chronic human atrial fibrillation. Circ Res: 80:772-781.

85. Bosch RF, Scherer CR, Rub N et al. (2003) Molecular mechanisms of early electrical remodeling: transcriptional downregulation of ion channel subunits reduces I(Ca,L) and I(to) in rapid atrial pacing in rabbits. J Am Coll Cardiol: 41:858-869.

86. Yue L, Feng J, Gaspo R et al. (1997) Ionic remodeling underlying action potential changes in a canine model of atrial fibrillation. Circ Res: 81:512-525.

87. Bosch RF, Zeng X, Grammer JB et al. (1999) Ionic mechanisms of electrical remodeling in human atrial fibrillation. Cardiovasc Res: 44:121-131.

88. Dobrev D, Wettwer E, Kortner A et al. (2002) Human inward rectifier potassium channels in chronic and postoperative atrial fibrillation. Cardiovasc Res: 54:397- 404.

89. Dobrev D, Graf E, Wettwer E et al. (2001) Molecular basis of downregulation of G-protein-coupled inward rectifying K(+) current (I(K,ACh) in chronic human atrial fibrillation: decrease in GIRK4 mRNA correlates with reduced I(K,ACh) and muscarinic receptor-mediated shortening of action potentials. Circulation: 104:2551-2557.

90. Harrison DC (1985) Antiarrhythmic drug classification: new science and practical applications. Am J Cardiol: 56:185-187.

91. Harrison DC (1992) The Sicilian Gambit: reasons for maintaining the present antiarrhythmic drug classification. Cardiovasc Res: 26:566-567.

92. Vaughan Williams EM (1984) A classification of antiarrhythmic actions reassessed after a decade of new drugs. J Clin Pharmacol: 24:129-147.

93. Nattel S, Singh BN (1999) Evolution, mechanisms, and classification of antiarrhythmic drugs: focus on class III actions. Am J Cardiol: 84:11R-19R.

168

94. Gottlieb SH, Ouyang P, Gottlieb SO (1988) Death after acute myocardial infarction: interrelation between left ventricular dysfunction, arrhythmias and ischemia. Am J Cardiol: 61:7B-12B.

95. Rubart M, Zipes DP (2005) Mechanisms of sudden cardiac death. J Clin Invest: 115:2305-2315.

96. Zipes DP, Rubart M (2006) Neural modulation of cardiac arrhythmias and sudden cardiac death. Heart Rhythm: 3:108-113.

97. Zipes DP (2008) Heart-brain interactions in cardiac arrhythmias: role of the autonomic nervous system. Cleve Clin J Med: 75 Suppl 2:S94-S96.

98. Zipes DP (1990) Influence of myocardial ischemia and infarction on autonomic innervation of heart. Circulation: 82:1095-1105.

99. Billman GE (2009) Cardiac autonomic neural remodeling and susceptibility to sudden cardiac death: effect of endurance exercise training. Am J Physiol Heart Circ Physiol: 297:H1171-H1193.

100. Raczak G, nilowicz-Szymanowicz L, Kobuszewska-Chwirot M et al. (2006) Long-term exercise training improves autonomic nervous system profile in professional runners. Kardiol Pol: 64:135-140.

101. Billman GE, Castillo LC, Hensley J, Hohl CM, Altschuld RA (1997) Beta2- adrenergic receptor antagonists protect against ventricular fibrillation: in vivo and in vitro evidence for enhanced sensitivity to beta2-adrenergic stimulation in animals susceptible to sudden death. Circulation: 96:1914-1922.

102. De AK, Wichi RB, Jesus WR et al. (2004) Exercise training changes autonomic cardiovascular balance in mice. J Appl Physiol: 96:2174-2178.

103. Negrao CE, Moreira ED, Brum PC, Denadai ML, Krieger EM (1992) Vagal and sympathetic control of heart rate during exercise by sedentary and exercise- trained rats. Braz J Med Biol Res: 25:1045-1052.

104. Iellamo F, Legramante JM, Massaro M, Raimondi G, Galante A (2000) Effects of a residential exercise training on sensitivity and heart rate variability in patients with coronary artery disease: A randomized, controlled study. Circulation: 102:2588-2592.

105. Clusin WT (2003) Calcium and cardiac arrhythmias: DADs, EADs, and alternans. Crit Rev Clin Lab Sci: 40:337-375.

169

106. Perhonen MA, Haapalahti P, Kivisto S et al. (2006) Effect of physical training on ventricular repolarization in type 1 long QT syndrome: a pilot study in asymptomatic carriers of the G589D KCNQ1 mutation. Europace: 8:894-898.

107. Ali A, Mehra MR, Malik FS et al. (1999) Effects of training on indices of ventricular repolarization in patients with chronic heart failure. Chest: 116:83-87.

108. Such L, Alberola AM, Such-Miquel L et al. (2008) Effects of chronic exercise on myocardial refractoriness: a study on isolated rabbit heart. Acta Physiol (Oxf): 193:331-339.

109. Cheung JY, Song J, Rothblum LI, Zhang XQ (2004) Exercise training improves cardiac function postinfarction: special emphasis on recent controversies on na+/ca2+ exchanger. Exerc Sport Sci Rev: 32:83-89.

110. Billman GE (2006) A comprehensive review and analysis of 25 years of data from an in vivo canine model of sudden cardiac death: implications for future anti-arrhythmic drug development. Pharmacol Ther: 111:808-835.

111. Billman GE, Schwartz PJ, Stone HL (1982) control of heart rate: a predictor of sudden cardiac death. Circulation: 66:874-880.

112. Billman GE, Schwartz PJ, Stone HL (1984) The effects of daily exercise on susceptibility to sudden cardiac death. Circulation: 69:1182-1189.

113. Billman GE, Kukielka M, Kelley R, Moustafa-Bayoumi M, Altschuld RA (2006) Endurance exercise training attenuates cardiac beta2-adrenoceptor responsiveness and prevents ventricular fibrillation in animals susceptible to sudden death. Am J Physiol Heart Circ Physiol: 290:H2590-H2599.

114. Opthof T, Coronel R, Wilms-Schopman FJ et al. (2007) Dispersion of repolarization in canine ventricle and the electrocardiographic T wave: Tp-e interval does not reflect transmural dispersion. Heart Rhythm: 4:341-348.

115. Yan GX, Antzelevitch C (1998) Cellular basis for the normal T wave and the electrocardiographic manifestations of the long-QT syndrome. Circulation: 98:1928-1936.

116. Porges SW, Byrne EA (1992) Research methods for measurement of heart rate and respiration. Biol Psychol: 34:93-130.

117. Billman GE, Dujardin JP (1990) Dynamic changes in cardiac vagal tone as measured by time-series analysis. Am J Physiol: 258:H896-H902.

170

118. Berntson GG, Bigger JT, Jr., Eckberg DL et al. (1997) Heart rate variability: origins, methods, and interpretive caveats. Psychophysiology: 34:623-648.

119. Goldsmith RL, Bigger JT, Jr., Steinman RC, Fleiss JL (1992) Comparison of 24- hour parasympathetic activity in endurance-trained and untrained young men. J Am Coll Cardiol: 20:552-558.

120. Kukielka M, Holycross BJ, Billman GE (2011) Endurance exercise training reduces cardiac sodium/calcium exchanger expression in animals susceptible to ventricular fibrillation. Front Physiol: 2:3.

121. Hull SS, Jr., Vanoli E, Adamson PB et al. (1994) Exercise training confers anticipatory protection from sudden death during acute myocardial ischemia. Circulation: 89:548-552.

122. Frederico MJ, Justo SL, da LG et al. (2009) Exercise training provides cardioprotection via a reduction in reactive oxygen species in rats submitted to myocardial infarction induced by isoproterenol. Free Radic Res: 43:957-964.

123. Mazzuero G, Lanfranchi P, Colombo R, Giannuzzi P, Giordano A (1992) Long- term adaptation of 24-h heart rate variability after myocardial infarction. The EAMI Study Group. Exercise Training in Anterior Myocardial Infarction. Chest: 101:304S-308S.

124. Sundquist K, Qvist J, Johansson SE, Sundquist J (2005) The long-term effect of physical activity on incidence of coronary heart disease: a 12-year follow-up study. Prev Med: 41:219-225.

125. Billman GE, Kukielka M (2007) Effect of endurance exercise training on heart rate onset and heart rate recovery responses to submaximal exercise in animals susceptible to ventricular fibrillation. J Appl Physiol: 102:231-240.

126. Blomqvist CG, Saltin B (1983) Cardiovascular adaptations to physical training. Annu Rev Physiol: 45:169-189.

127. Smith ML, Hudson DL, Graitzer HM, Raven PB (1989) Exercise training bradycardia: the role of autonomic balance. Med Sci Sports Exerc: 21:40-44.

128. Wyatt HL, Mitchell JH (1974) Influences of physical training on the heart of dogs. Circ Res: 35:883-889.

129. Mimura J, Yuasa F, Yuyama R et al. (2005) The effect of residential exercise training on baroreflex control of heart rate and sympathetic nerve activity in patients with acute myocardial infarction. Chest: 127:1108-1115.

171

130. Billman GE, Kukielka M (2006) Effects of endurance exercise training on heart rate variability and susceptibility to sudden cardiac death: protection is not due to enhanced cardiac vagal regulation. J Appl Physiol: 100:896-906.

131. Brown DA, Chicco AJ, Jew KN et al. (2005) Cardioprotection afforded by chronic exercise is mediated by the sarcolemmal, and not the mitochondrial, isoform of the KATP channel in the rat. J Physiol: 569:913-924.

132. Zingman LV, Zhu Z, Sierra A et al. (2011) Exercise-induced expression of cardiac ATP-sensitive potassium channels promotes action potential shortening and energy conservation. J Mol Cell Cardiol: 51:72-81.

133. Bito V, de Waard MC, Biesmans L et al. (2010) Early exercise training after myocardial infarction prevents contractile but not electrical remodelling or hypertrophy. Cardiovasc Res: 86:72-81.

134. Stones R, Billeter R, Zhang H, Harrison S, White E (2009) The role of transient outward K+ current in electrical remodelling induced by voluntary exercise in female rat hearts. Basic Res Cardiol: 104:643-652.

135. Natali AJ, Wilson LA, Peckham M et al. (2002) Different regional effects of voluntary exercise on the mechanical and electrical properties of rat ventricular myocytes. J Physiol: 541:863-875.

136. Banyasz T, Fulop L, Magyar J et al. (2003) Endocardial versus epicardial differences in L-type calcium current in canine ventricular myocytes studied by action potential voltage clamp. Cardiovasc Res: 58:66-75.

137. Sung RJ, Wu SN, Wu JS, Chang HD, Luo CH (2006) Electrophysiological mechanisms of ventricular arrhythmias in relation to Andersen-Tawil syndrome under conditions of reduced IK1: a simulation study. Am J Physiol Heart Circ Physiol: 291:H2597-H2605.

138. Weiss JN, Nivala M, Garfinkel A, Qu Z (2011) Alternans and arrhythmias: from cell to heart. Circ Res: 108:98-112.

139. Gyorke S, Carnes C (2008) Dysregulated sarcoplasmic reticulum calcium release: potential pharmacological target in cardiac disease. Pharmacol Ther: 119:340- 354.

140. Stolen TO, Hoydal MA, Kemi OJ et al. (2009) Interval training normalizes cardiomyocyte function, diastolic Ca2+ control, and SR Ca2+ release synchronicity in a mouse model of diabetic cardiomyopathy. Circ Res: 105:527- 536.

172

141. Bowes J, Brown AJ, Hamon J et al. (2012) Reducing safety-related drug attrition: the use of in vitro pharmacological profiling. Nat Rev Drug Discov: 11:909-922.

142. Varkevisser R, Wijers SC, van der Heyden MA et al. (2012) Beat-to-beat variability of repolarization as a new biomarker for proarrhythmia in vivo. Heart Rhythm: 9:1718-1726.

143. Wier WG, Hess P (1984) Excitation-contraction coupling in cardiac . Effects of cardiotonic steroids on the intracellular [Ca2+] transient, membrane potential, and contraction. J Gen Physiol: 83:395-415.

144. Ho HT, Stevens SC, Terentyeva R et al. (2011) Arrhythmogenic adverse effects of cardiac glycosides are mediated by redox modification of ryanodine receptors. J Physiol: 589:4697-4708.

145. Barrett J, Worth E, Bauss F, Epstein S (2004) Ibandronate: a clinical pharmacological and pharmacokinetic update. J Clin Pharmacol: 44:951-965.

146. Kimmel DB (2007) Mechanism of action, pharmacokinetic and pharmacodynamic profile, and clinical applications of nitrogen-containing bisphosphonates. J Dent Res: 86:1022-1033.

147. Grandi E, Pasqualini FS, Bers DM (2010) A novel computational model of the human ventricular action potential and Ca transient. J Mol Cell Cardiol: 48:112- 121.

148. Romero L, Pueyo E, Fink M, Rodriguez B (2009) Impact of ionic current variability on human ventricular cellular electrophysiology. Am J Physiol Heart Circ Physiol: 297:H1436-H1445.

149. Salari P, Abdollahi M (2012) Long term bisphosphonate use in osteoporotic patients; a step forward, two steps back. J Pharm Pharm Sci: 15:305-317.

150. John CA (2010) Review of the cardiovascular safety of zoledronic acid and other bisphosphonates for the treatment of osteoporosis. Clin Ther: 32:426-436.

151. Dumotier BM, Georgieva AV (2007) Preclinical cardio-safety assessment of torsadogenic risk and alternative methods to animal experimentation: the inseparable twins. Cell Biol Toxicol: 23:293-302.

152. Cavero I, Crumb W (2005) ICH S7B draft guideline on the non-clinical strategy for testing delayed cardiac repolarisation risk of drugs: a critical analysis. Expert Opin Drug Saf: 4:509-530.

173

153. Friedrichs GS, Patmore L, Bass A (2005) Non-clinical evaluation of ventricular repolarization (ICH S7B): results of an interim survey of international pharmaceutical companies. J Pharmacol Toxicol Methods: 52:6-11.

154. Redfern WS, Carlsson L, Davis AS et al. (2003) Relationships between preclinical cardiac electrophysiology, clinical QT interval prolongation and torsade de pointes for a broad range of drugs: evidence for a provisional safety margin in drug development. Cardiovasc Res: 58:32-45.

155. Hoffmann P, Warner B (2006) Are hERG channel inhibition and QT interval prolongation all there is in drug-induced torsadogenesis? A review of emerging trends. J Pharmacol Toxicol Methods: 53:87-105.

156. Pugsley MK, Towart R, Authier S, Gallacher DJ, Curtis MJ (2011) Innovation in safety pharmacology testing. J Pharmacol Toxicol Methods: 64:1-6.

157. Laverty H, Benson C, Cartwright E et al. (2011) How can we improve our understanding of cardiovascular safety liabilities to develop safer medicines? Br J Pharmacol: 163:675-693.

158. Qian JY, Guo L (2010) Altered cytosolic Ca2+ dynamics in cultured Guinea pig cardiomyocytes as an in vitro model to identify potential cardiotoxicants. Toxicol In Vitro: 24:960-972.

159. Sugamura K, Keaney JF, Jr. (2011) Reactive oxygen species in . Free Radic Biol Med: 51:978-992.

160. Han W, Fu S, Wei N et al. (2008) Nitric oxide overproduction derived from inducible nitric oxide synthase increases cardiomyocyte apoptosis in human atrial fibrillation. Int J Cardiol: 130:165-173.

161. Sun Y (2007) Oxidative stress and cardiac repair/remodeling following infarction. Am J Med Sci: 334:197-205.

162. Kameda K, Matsunaga T, Abe N et al. (2003) Correlation of oxidative stress with activity of matrix metalloproteinase in patients with coronary artery disease. Possible role for left ventricular remodelling. Eur Heart J: 24:2180-2185.

163. Wattanapitayakul SK, Bauer JA (2001) Oxidative pathways in cardiovascular disease: roles, mechanisms, and therapeutic implications. Pharmacol Ther: 89:187-206.

164. Reiter CD, Teng RJ, Beckman JS (2000) Superoxide reacts with nitric oxide to nitrate tyrosine at physiological pH via peroxynitrite. J Biol Chem: 275:32460- 32466.

174

165. Nuriel T, Hansler A, Gross SS (2011) Protein nitrotryptophan: formation, significance and identification. J Proteomics: 74:2300-2312.

166. Souza JM, Peluffo G, Radi R (2008) Protein tyrosine nitration--functional alteration or just a biomarker? Free Radic Biol Med: 45:357-366.

167. Pacher P, Beckman JS, Liaudet L (2007) Nitric oxide and peroxynitrite in health and disease. Physiol Rev: 87:315-424.

168. Mihm MJ, Yu F, Carnes CA et al. (2001) Impaired myofibrillar energetics and oxidative injury during human atrial fibrillation. Circulation: 104:174-180.

169. Lokuta AJ, Maertz NA, Meethal SV et al. (2005) Increased nitration of sarcoplasmic reticulum Ca2+-ATPase in human heart failure. Circulation: 111:988-995.

170. Chesnais JM, Fischmeister R, Mery PF (1999) Positive and negative inotropic effects of NO donors in atrial and ventricular fibres of the frog heart. J Physiol: 518 ( Pt 2):449-461.

171. Kohr MJ, Davis JP, Ziolo MT (2009) Peroxynitrite Increases Protein Phosphatase Activity and Promotes the Interaction of Phospholamban with Protein Phosphatase 2a in the Myocardium. Nitric Oxide: 20:217-221.

172. Workman AJ, Pau D, Redpath CJ et al. (2009) Atrial cellular electrophysiological changes in patients with ventricular dysfunction may predispose to AF. Heart Rhythm: 6:445-451.

173. Van Wagoner DR, Pond AL, Lamorgese M et al. (1999) Atrial L-type Ca2+ currents and human atrial fibrillation. Circ Res: 85:428-436.

174. Nishijima Y, Sridhar A, Viatchenko-Karpinski S et al. (2007) Chronic cardiac resynchronization therapy and reverse in a model of nonischemic cardiomyopathy. Life Sci: 81:1152-1159.

175. Kubalova Z, Terentyev D, Viatchenko-Karpinski S et al. (2005) Abnormal intrastore calcium signaling in chronic heart failure. Proc Natl Acad Sci U S A: 102:14104-14109.

176. Kohr MJ, Wang H, Wheeler DG et al. (2008) Targeting of phospholamban by peroxynitrite decreases beta-adrenergic stimulation in cardiomyocytes. Cardiovasc Res: 77:353-361.

177. Chen W, Druhan LJ, Chen CA et al. (2010) Peroxynitrite induces destruction of the tetrahydrobiopterin and heme in endothelial nitric oxide synthase: transition from reversible to irreversible enzyme inhibition. Biochemistry: 49:3129-3137. 175

178. Kooy NW, Lewis SJ, Royall JA et al. (1997) Extensive tyrosine nitration in human myocardial inflammation: evidence for the presence of peroxynitrite. Crit Care Med: 25:812-819.

179. Alvarez B, Radi R (2003) Peroxynitrite reactivity with amino acids and proteins. Amino Acids: 25:295-311.

180. Jones LH (2012) Chemistry and biology of biomolecule nitration. Chem Biol: 19:1086-1092.

181. Franz MR, Karasik PL, Li C, Moubarak J, Chavez M (1997) Electrical remodeling of the human atrium: similar effects in patients with chronic atrial fibrillation and . J Am Coll Cardiol: 30:1785-1792.

182. Kaab S, Barth AS, Margerie D et al. (2004) Global gene expression in human myocardium-oligonucleotide microarray analysis of regional diversity and transcriptional regulation in heart failure. J Mol Med (Berl): 82:308-316.

183. Hanna N, Cardin S, Leung TK, Nattel S (2004) Differences in atrial versus ventricular remodeling in dogs with ventricular tachypacing-induced congestive heart failure. Cardiovasc Res: 63:236-244.

184. Cardin S, Pelletier P, Libby E et al. (2008) Marked differences between atrial and ventricular gene-expression remodeling in dogs with experimental heart failure. J Mol Cell Cardiol: 45:821-831.

185. Beuckelmann DJ, Nabauer M, Erdmann E (1993) Alterations of K+ currents in isolated human ventricular myocytes from patients with terminal heart failure. Circ Res: 73:379-385.

186. Snook JH, Li J, Helmke BP, Guilford WH (2008) Peroxynitrite inhibits myofibrillar protein function in an in vitro assay of motility. Free Radic Biol Med: 44:14-23.

187. Frank KF, Bolck B, Erdmann E, Schwinger RH (2003) Sarcoplasmic reticulum Ca2+-ATPase modulates cardiac contraction and relaxation. Cardiovasc Res: 57:20-27.

188. Terentyev D, Gyorke I, Belevych AE et al. (2008) Redox modification of ryanodine receptors contributes to sarcoplasmic reticulum Ca2+ leak in chronic heart failure. Circ Res: 103:1466-1472.

189. Belevych AE, Terentyev D, Viatchenko-Karpinski S et al. (2009) Redox modification of ryanodine receptors underlies calcium alternans in a canine model of sudden cardiac death. Cardiovasc Res: 84:387-395.

176

190. Wagner S, Rokita AG, Anderson ME, Maier LS (2012) Redox Regulation of Sodium and Calcium Handling. Antioxid Redox Signal:

191. Erickson JR, Joiner ML, Guan X et al. (2008) A dynamic pathway for calcium- independent activation of CaMKII by methionine oxidation. Cell: 133:462-474.

192. Babu GJ, Bhupathy P, Carnes CA, Billman GE, Periasamy M (2007) Differential expression of sarcolipin protein during muscle development and cardiac pathophysiology. J Mol Cell Cardiol: 43:215-222.

193. Dhalla NS, Temsah RM, Netticadan T (2000) Role of oxidative stress in cardiovascular diseases. J Hypertens: 18:655-673.

194. Fernandez-Velasco M, Ruiz-Hurtado G, Hurtado O, Moro MA, Delgado C (2007) TNF-alpha downregulates transient outward potassium current in rat ventricular myocytes through iNOS overexpression and oxidant species generation. Am J Physiol Heart Circ Physiol: 293:H238-H245.

195. Priebe L, Beuckelmann DJ (1998) Simulation study of cellular electric properties in heart failure. Circ Res: 82:1206-1223.

196. Winslow RL, Rice J, Jafri S, Marban E, O'Rourke B (1999) Mechanisms of altered excitation-contraction coupling in canine tachycardia-induced heart failure, II: model studies. Circ Res: 84:571-586.

197. Sun X, Wang HS (2005) Role of the transient outward current (Ito) in shaping canine ventricular action potential--a dynamic clamp study. J Physiol: 564:411- 419.

198. Rabkin SW, Klassen SS (2007) Nitric oxide differentially regulates the gene expression of caspase genes but not some autophagic genes. Nitric Oxide: 16:339- 347.

199. Tao L, Gao E, Hu A et al. (2006) Thioredoxin reduces post-ischemic myocardial apoptosis by reducing oxidative/nitrative stress. Br J Pharmacol: 149:311-318.

200. Suzuki K, Kostin S, Person V, Elsasser A, Schaper J (2001) Time course of the apoptotic cascade and effects of caspase inhibitors in adult rat ventricular cardiomyocytes. J Mol Cell Cardiol: 33:983-994.

201. Guo D, Lian J, Liu T et al. (2011) Contribution of late sodium current (I(Na-L)) to rate adaptation of ventricular repolarization and reverse use-dependence of QT- prolonging agents. Heart Rhythm: 8:762-769.

177

202. Varkevisser R, Wijers SC, van der Heyden MA et al. (2012) Beat-to-beat variability of repolarization as a new biomarker for proarrhythmia in vivo. Heart Rhythm: 9:1718-1726.

203. Kannel WB, Wolf PA, Benjamin EJ, Levy D (1998) Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates. Am J Cardiol: 82:2N-9N.

204. Vergara C, Latorre R, Marrion NV, Adelman JP (1998) Calcium-activated potassium channels. Curr Opin Neurobiol: 8:321-329.

205. Ro S, Hatton WJ, Koh SD, Horowitz B (2001) Molecular properties of small- conductance Ca2+-activated K+ channels expressed in murine colonic smooth muscle. Am J Physiol Gastrointest Liver Physiol: 281:G964-G973.

206. Tuteja D, Xu D, Timofeyev V et al. (2005) Differential expression of small- conductance Ca2+-activated K+ channels SK1, SK2, and SK3 in mouse atrial and ventricular myocytes. Am J Physiol Heart Circ Physiol: 289:H2714-H2723.

207. Wei AD, Gutman GA, Aldrich R et al. (2005) International Union of Pharmacology. LII. Nomenclature and molecular relationships of calcium- activated potassium channels. Pharmacol Rev: 57:463-472.

208. Li N, Timofeyev V, Tuteja D et al. (2009) Ablation of a Ca2+-activated K+ channel (SK2 channel) results in action potential prolongation in atrial myocytes and atrial fibrillation. J Physiol: 587:1087-1100.

209. Gui L, Bao Z, Jia Y et al. (2013) Ventricular tachyarrhythmias in rats with acute myocardial infarction involves activation of small-conductance Ca2+-activated K+ channels. Am J Physiol Heart Circ Physiol: 304:H118-H130.

210. Chua SK, Chang PC, Maruyama M et al. (2011) Small-conductance calcium- activated potassium channel and recurrent ventricular fibrillation in failing rabbit ventricles. Circ Res: 108:971-979.

211. Chang PC, Turker I, Lopshire JC et al. (2013) Heterogeneous upregulation of apamin-sensitive potassium currents in failing human ventricles. J Am Heart Assoc: 2:e004713.

212. Qi XY, Diness JG, Brundel B et al. (2013) Role of Small Conductance Calcium- Activated Potassium Channels in Atrial Electrophysiology and Fibrillation in the Dog. Circulation: 430-440.

178

213. Hsueh CH, Chang PC, Hsieh YC et al. (2013) Proarrhythmic effect of blocking the small conductance calcium activated potassium channel in isolated canine left atrium. Heart Rhythm: 10:891-898.

214. Nishijima Y, Feldman DS, Bonagura JD et al. (2005) Canine nonischemic left ventricular dysfunction: a model of chronic human cardiomyopathy. J Card Fail: 11:638-644.

215. Bonilla IM, Sridhar A, Nishijima Y et al. (2013) Differential effects of the peroxynitrite donor, SIN-1, on atrial and ventricular myocyte electrophysiology. J Cardiovasc Pharmacol: 61:401-407.

216. Thomsen MB, Volders PG, Beekman JD, Matz J, Vos MA (2006) Beat-to-Beat variability of repolarization determines proarrhythmic outcome in dogs susceptible to drug-induced torsades de pointes. J Am Coll Cardiol: 48:1268- 1276.

217. Bonilla IM, Vargas-Pinto P, Nishijima Y et al. (2013) Ibandronate and Ventricular Arrhythmia Risk. J Cardiovasc Electrophysiol:

218. Weatherall KL, Seutin V, Liegeois JF, Marrion NV (2011) Crucial role of a shared extracellular loop in apamin sensitivity and maintenance of pore shape of small-conductance calcium-activated potassium (SK) channels. Proc Natl Acad Sci U S A: 108:18494-18499.

219. Weatherall KL, Goodchild SJ, Jane DE, Marrion NV (2010) Small conductance calcium-activated potassium channels: from structure to function. Prog Neurobiol: 91:242-255.

220. Grunnet M, Jespersen T, Angelo K et al. (2001) Pharmacological modulation of SK3 channels. Neuropharmacology: 40:879-887.

221. Bildl W, Strassmaier T, Thurm H et al. (2004) Protein kinase CK2 is coassembled with small conductance Ca(2+)-activated K+ channels and regulates channel gating. : 43:847-858.

222. Belevych AE, Terentyev D, Terentyeva R et al. (2011) The relationship between arrhythmogenesis and impaired contractility in heart failure: role of altered ryanodine receptor function. Cardiovasc Res: 90:493-502.

223. Terentyev D, Rochira JA, Terentyeva R et al. (2013) Sarcoplasmic reticulum Ca2+ release is both necessary and sufficient for SK channel activation in ventricular myocytes. Am J Physiol Heart Circ Physiol: In Press.

179

224. Roden DM (1998) Taking the "idio" out of "idiosyncratic": predicting torsades de pointes. Pacing Clin Electrophysiol: 21:1029-1034.

225. Kaab S, Nuss HB, Chiamvimonvat N et al. (1996) Ionic mechanism of action potential prolongation in ventricular myocytes from dogs with pacing-induced heart failure. Circ Res: 78:262-273.

226. Nattel S (2009) Calcium-activated potassium current: a novel ion channel candidate in atrial fibrillation. J Physiol: 587:1385-1386.

227. Glukhov AV, Fedorov VV, Kalish PW et al. (2012) Conduction remodeling in human end-stage nonischemic left ventricular cardiomyopathy. Circulation: 125:1835-1847.

228. Chang SH, Chang SN, Hwang JJ et al. (2012) Significant association of rs13376333 in KCNN3 on chromosome 1q21 with atrial fibrillation in a Taiwanese population. Circ J: 76:184-188.

229. Ellinor PT, Lunetta KL, Glazer NL et al. (2010) Common variants in KCNN3 are associated with lone atrial fibrillation. Nat Genet: 42:240-244.

230. Yu T, Deng C, Wu R et al. (2012) Decreased expression of small-conductance Ca2+-activated K+ channels SK1 and SK2 in human chronic atrial fibrillation. Life Sci: 90:219-227.

231. Park YB (1994) Ion selectivity and gating of small conductance Ca(2+)-activated K+ channels in cultured rat adrenal chromaffin cells. J Physiol: 481 ( Pt 3):555- 570.

232. Kannel WB (2000) Incidence and epidemiology of heart failure. Heart Fail Rev: 5:167-173.

233. Cha TJ, Ehrlich JR, Zhang L, Nattel S (2004) Atrial Ionic Remodeling Induced by in the Presence of Congestive Heart Failure. Circulation: 110:1520-1526.

234. Virag L, Jost N, Papp R et al. (2011) Analysis of the contribution of I(to) to repolarization in canine ventricular myocardium. Br J Pharmacol: 164:93-105.

235. Goudis CA, Kallergis EM, Vardas PE (2012) Extracellular matrix alterations in the atria: insights into the mechanisms and perpetuation of atrial fibrillation. Europace: 14:623-630.

236. Eckstein J, Verheule S, de Groot NM, Allessie M, Schotten U (2008) Mechanisms of perpetuation of atrial fibrillation in chronically dilated atria. Prog Biophys Mol Biol: 97:435-451. 180

237. Yeh YH, Wakili R, Qi XY et al. (2008) Calcium-handling abnormalities underlying atrial arrhythmogenesis and contractile dysfunction in dogs with congestive heart failure. Circ Arrhythm Electrophysiol: 1:93-102.

238. Ogbaghebriel A, Shrier A (1994) Inhibition of metabolism abolishes transient outward current in rabbit atrial myocytes. Am J Physiol: 266:H182-H190.

239. Xiao L, Koopmann TT, Ordog B et al. (2013) Unique cardiac Purkinje fiber transient outward current beta-subunit composition: a potential molecular link to idiopathic ventricular fibrillation. Circ Res: 112:1310-1322.

240. Radicke S, Cotella D, Sblattero D et al. (2009) The transmembrane +¦-subunits KCNE1, KCNE2, and DPP6 modify pharmacological effects of the on the transient outward current I to. Naunyn- Schmied Arch Pharmacol: 379:617-626.

241. Deschenes I, Disilvestre D, Juang GJ et al. (2002) Regulation of Kv4.3 current by KChIP2 splice variants: a component of native cardiac I(to)? Circulation: 106:423-429.

242. Workman AJ, Marshall GE, Rankin AC, Smith GL, Dempster J (2012) Transient outward K+ current reduction prolongs action potentials and promotes afterdepolarisations: a dynamic-clamp study in human and rabbit cardiac atrial myocytes. J Physiol: 590:4289-4305.

243. van Brakel TJ, van der Krieken T, Westra SW et al. (2013) Fibrosis and electrophysiological characteristics of the atrial appendage in patients with atrial fibrillation and structural heart disease. J Interv Card Electrophysiol: 38:85-93.

244. Tian T, Wang Y, Sun K et al. (2013) Clinical profile and prognostic significance of atrial fibrillation in hypertrophic cardiomyopathy. Cardiology: 126:258-264.

245. Colman MA, Aslanidi OV, Kharche S et al. (2013) Pro-arrhythmogenic effects of atrial fibrillation-induced electrical remodelling: insights from the three- dimensional virtual human atria. The Journal of Physiology: 591:4249-4272.

246. Feng J, Yue L, Wang Z, Nattel S (1998) Ionic mechanisms of regional action potential heterogeneity in the canine right atrium. Circ Res: 83:541-551.

247. Xu Y, Zhang Z, Timofeyev V et al. (2005) The effects of intracellular Ca2+ on cardiac K+ channel expression and activity: novel insights from genetically altered mice. J Physiol: 562:745-758.

181