A SERIES OF AMINO-2-CYCLOHEXYL ESTERS, THEIR ELECTROPHYSIOLOGICAL AND ANTIARRHYTHMIC EFFECTS AS RELATED TO ACTIONS ON ISCHEMIA-INDUCED ARRHYTHMIAS by Sandro Luis Yong B.Sc. The University of British Columbia, Vancouver, B.C., Canada, 1992 M.Sc. The University of British Columbia, Vancouver, B.C., Canada, 1995 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF PHARMACOLOGY & THERAPEUTICS FACULTY OF MEDICINE We accept this thesis as conforming -To-the-xecjuired standard THE UNIVERSITY OF BRITISH COLUMBIA May 2000 © SANDRO L U I S YONG, 2000 U D ^ opeciai ^ u i i c c u u i i s - n i c s i s /-Yuuiuusauuii r u i u i r a g e i u i i In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y a v a i l a b l e for reference and study. I further agree that permission for extensive copying of th i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It i s understood that copying or pu b l i c a t i o n of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of The U n i v e r s i t y of B r i t i s h Columbia Vancouver, Canada http://\vww.library.ubc.ca/spcoll/thesauth.html 5/25/00 Abstract Indiscriminate blockade of cardiac ion channels, in addition to affects on extra-cardiac tissues, is primarily responsible for cardiac side-effects associated with conventional antiarrhythmic drug use. The inactive R,R-enantiomers of cyclohexylbenzenacetamide analgesics (e.g., U -50,488H and RSD920, Parke Davis compounds) exhibited antiarrhythmic activity independent of central or peripheral opioid activity. This pharmacophore offered a novel approach for rational drug design for the purpose of limiting drug toxicity. The cyclohexylbenzenacetamide structure was soon replaced with less cardiac-depressant ester derivatives. The present thesis examined a series of 16 structurally-related amino-2-cyclohexyl ester compounds to identify the chemical components within this series for cardiac ion channel blockade and, ultimately, antiarrhythmic actions against ischemia-induced arrhythmias. The compounds were tested against arrhythmias produced by myocardial ischemia and by electrical stimulation. Additional studies were performed in isolated rat ventricular cells to characterize their actions on the major depolarizing and repolarizing currents (Kja and ITo, respectively). Cloned sodium channels expressed in Xenopus oocytes served as a separate cell line in which to directly evaluate effects on lN a- The major conclusions from this study were as follows: 1. In this series, antiarrhythmic activity against ischemia-induced arrhythmias was dependent on chemical structure, in particular, with the nature of the aromatic group (Ri) and amine heterocyclic group (R 2) held in a trans position by the cyclohexyl backbone. i i i 2. Suppression of ischemia-induced arrhythmias with minimal actions on blood pressure and heart rate was best when Ri was 1-naphthalene and R 2 , N-morpholino (RSD1000). 3. Naphthalene analogues of RSD1000 with different ionizable nitrogen groups were shown to inhibit I N a and/or ITo currents in rat ventricular myocytes. 4. Inhibition of IN 2 , but not ITo, was potentiated by acid pH. 5. The blocking potency for Kja was proportional to the pKa of the compound, i.e., the higher the pKa, the lower the IC50. 6. Antiarrhythmic effectiveness and selectivity for ischemia-induced arrhythmias was inversely proportional to pKa. The overall finding was that suppression of ischemia-induced arrhythmias by RSD1000 was a function of mixed I N 3 and ITo blockade and, unlike predominantly charged antiarrhythmic agents (e.g., quinidine), the former blocking component was localized to myocardial tissue with external acid pH. iv T A B L E OF C O N T E N T S Chapter Page no. Abstract i Table of Contents I V List of Figures vii List of Tables «x List of Abbreviations x l 1 I N T R O D U C T I O N 1 1.1 Background 1 1.1.1 Etiology of ischemia-induced cardiac arrhythmias: 4 Reentry mechanism 1.1.2 Antiarrhythmic drug therapy against ischemia- 5 induced arrhythmias: preface 1.1.2.1 Therapy using blockers of the cardiac inward 6 sodium current 1.1.2.2 " Therapy using blockers of the cardiac outward 8 potassium current 1.1.2.3 Therapy using mixed channel blockade 10 1.2 Drug development and the ion channel targets for 12 modulating cardiac action potential 1.2.1 Cardiac ion channels: preface 12 1.2.1.1 Fast inward sodium current (IN a) 13 1.2.1.2 Transient outward potassium current (ITO) 16 1.3 Ion channels (TWITO) at a molecular level 19 1.3.1 Molecular basis of Na + channel function 20 1.3.2 Molecular basis of K + channel function 28 1.4 Molecular basis of class I and III antiarrhythmic binding 33 interaction 1.4.1 Class I Antiarrhythmics 33 1.4.2 Class III Antiarrhythmics 38 1.5 Acute myocardial ischemia and cardiac ion channels 39 1.5.1 Metabolic changes associated with acute myocardial 39 ischemia 1.5.2 Influence of intra- and extra-cellular pH on cardiac 42 ion channels V 1.5.2.1 Inward sodium current (IN3) 43 1.5.2.2 Transient outward potassium current (ITO) 44 1.5.3 Effects associated with raised extracellular K + 46 1.5.4 Raised extracellular K + and ischemia-induced 48 arrhythmias 1.6 Selectivity for myocardial ischemia 51 1.6.1 Rationale 51 1.6.2 Role of molecular pKa and selectivity for myocardial 54 ischemia 1.7 Objectives 57 2 M E T H O D S A N D M A T E R I A L S 62 2.1 In vivo studies 62 2.1.1 Ischemia-induced arrhythmias 62 2.1.1.1 Exclusion criteria 64 2.1.1.2 Classification of ischemia-induced arrhythmias and 65 the arrhythmia score (AS) 2.1.1.3 Antiarrhythmic analysis 66 2.1.2 Ventricular arrhythmias induced by electrical 67 stimulation 2.2 In vitro studies 70 2.2.1 Isolated rat cardiac ventricular myocytes 70 2.2.1.1 Isolation 70 2.2.1.2 Recording solutions 71 2.2.1.3 Data recording 72 2.2.1.4 Current measurements 72 2.2.2 Cloned sodium channels 74 2.2.2.1 Expression of cardiac sodium channels in Xenopus 74 oocytes 2.2.2.1.1 Transcription of R N A and expression in 74 Xenopus oocytes 2.2.2.2 Solutions and drugs 74 2.2.2.3 Data recording and analysis 75 2.3 Data and statistical analysis 76 2.4 RSD Compounds 77 2.4.1 pKa determination 77 2.4.2 HPLC measurement of hydrophobicity for RSD 79 compounds V 3 RESULTS 83 3.1 Effects on haemodynamic and E C G parameters in rats 83 subjected to myocardial ischemia 3.2 Effects of compounds on ischemia-induced arrhythmias 94 3.2.1 Effects of increasing aromatic chain length 94 3.2.2 Effects of substitutions on the phenyl ring 103 3.2.2.1 3,4-phenyl substitutions 103 3.2.2.2 Para-substitutions: electron-withdrawing 105 effect of the phenyl group 3.2.3 Naphthalene position: 1-versus 2-naphthalene 108 3.3 Effects of RSD1000 on isolated rat ventricular myocytes 108 3.4 Influence of pKa and effects on ischemia-induced 114 arrhythmias 3.5 Electrically-induced arrhythmias 118 3.6 Effects of RSD 1000, RSD 1046, RSD 1049, and RSD 1025 127 on cloned sodium channels 3.7 Effects ofRSD1000,RSD1009,RSD1015, and RSD1025 128 on ITo currents in isolated rat ventricular myocytes 3.8 Effects df RSD 1009 IN 3 currents in isolated rat ventricular 132 myocytes 3.9 Correlation matrices 133 4 DISCUSSION 139 4.1 Biological and chemical activity of conventional 139 antiarrhythmics 4.2 Limitations 145 4.3 RSD compounds: Biological activity 152 4.3.1 Antiarrhythmic activity against ischemia- and electrically- 152 induced arrhythmias 4.3.2 Cardiac and cardiovascular responses 155 4.3.3 Effects in cloned sodium channels and isolated rat 157 ventricular cells vi 4.3.4 Summary of biological activity of RSD compounds 160 4.4 RSD compounds: Chemical activity 160 4.4.1 RSD compounds: Comparison of biological and chemical 162 activity 4.4.1.1 Quantitative versus qualitative structure-activity analysis 162 4.4.2 RSD compounds: Qualitative comparison of antiarrhythmic 165 selectivity and chemical substituents 4.4.2.1 Analysis of the aromatic side group 165 4.4.2.2 Ionizable amine group: pKa and its role in antiarrhythmic 169 activity 4.4.2.3 p H 0 effects on the I N a blocking component of RSD 171 compounds 4.4.2.4 p H 0 effects on the I T O blocking component of RSD 180 compounds 4.4.2.5 The role of the ITo blocking component in antiarrhythmic 182 activity 4.5 Summary 184 4.6 Metabolism of RSD compounds: possible chemical sites 185 5. REFERENCES 188 viii List of Figures Figure Page 1 Proposed molecular structure of the voltage-gated cardiac sodium 21 channel 2 Proposed molecular structure of the voltage-gated cardiac potassium 22 channel 3 An overview of the chemical changes to Ri and R 2 of the general 60 RSD pharmacophore. 4 Effects of RSD 1000, RSD 1025 and RSD 1049 on lowering blood 90-91 pressure and increasing P-R interval in rats prior to coronary occlusion as representative graphs to illustrate the determination of D25% values. 5 Original polygraph traces from the coronary artery ligation model in 95-96 which representative vehicle- and RSDIOOO-treated animals are shown. 6 RSD1053, RSD1010, RSD1050 and RSD1050 tested against 99-100 ischemia-induced arrhythmias to investigate the extension of the benzene aromatic group. 7 RSD 1012 and RSD 1072 tested against ischemia-induced 104 arrhythmias to investigate 3,4-phenyl substitutions. 8 RSD1019 and RSD1014 tested against ischemia-induced 106 arrhythmias to investigate 4-phenyl substitutions. 9 RSD 1000 and RSD 1009 tested against ischemia-induced 107 arrhythmias to investigate positional arrangement of the naphthalene group. 10 Inhibition of I N a currents in isolated rat ventricular myocytes by 109 RSD1000atpH7.3 and 6.4. 11 Use-dependent inhibition of i N a by RSD 1000 in isolated rat 110 ventricular myocytes at pH 7.3 and 6.4. 12 Inhibition of ITo currents in isolated rat ventricular myocytes by 111 RSD 1000 at pH 7.3 and 6.4. 13 Effects of RSD 1000 on Ic a current in isolated rat ventricular 112 ix myocytes. 14 RSD 1025 and RSD 1015 tested against ischemia-induced 115 arrhythmias to investigate the amine heterocyclic group and the role of pKa. 15 RSD 1046 and RSD 1049 tested against ischemia-induced 116 arrhythmias to investigate the amine heterocyclic group and the role of pKa. 16 Effects of RSD 1000, RSD 1046, RSD 1049 and RSD 1025 in the 119-120 electrically-induced arrhythmia rat model. 17 The inhibitory actions of RSD 1000, RSD 1025, RSD 1046, and 123 RSD 1049 at pH 7.3 and 6.4 in cardiac sodium channel isoforms (SkM2) expressed in Xenopus oocytes. 18 Inhibition of rat ventricular I T 0 currents by RSD 1009, RSD 1025 and 129 RSD1015. 19 Inhibition of I_Na currents in isolated rat ventricular myocytes by 134 RSD1009atpH7.3 and 6.4. 20 Infusion models with "small" and "large" V d 149 X List of Tables Table Page 1 A list of chemical structures for all 16 RSD compounds used in the 58 study. 2 Arrhythmia scoring table. 65 3 The chemical names and values of pKa and hydrophobicity (log Q) 81 for all 16 RSD compounds. 4 Effects on blood pressure, heart rate, and E C G parameters by 85-86 RSD1000, RSD1046, RSD1049, RSD1025, RSD1009 and RSD1015 in rats prior to coronary occlusion. 5 A n overall summary of D 25% values for 14 RSD compounds and 88 their effects on blood pressure (BP), heart rate (HR), E C G parameters and electrical stimulation variables (iT, ERP, VFt). 6 Antiarrhythmic activity against ischemia-induced arrhythmias for the 102 formate and acetate analogues in the coronary-occluded rats. 7 List of I C 5 0 values for RSD1000, RSD1025, RSD1046 and RSD1049 125 on the cardiac sodium channel isoform (in Xenopus oocytes) at pH 7.3 and 6.4. 8 List of IC50 values for inhibition of ITO currents by RSD 1000, 131 RSD1009, RSD1025 and RSD1015 in rat ventricular myocytes. 9 Correlation matrices between biological and chemical properties for 136-137 14 RSD compounds. 10 List of references of animal models for the testing of conventional 141 antiarrhythmic agents against ischemia-induced arrhythmias in the early phase. 11 Literature references of class I and III drugs and their concentrations 143 for producing 50% inhibition of IN 3 and ITO currents in rat ventricular cells performed in whole-cell studies. 12 Therapeutic index comparison table - a qualitative comparison of biological activity and chemical substituents. 166 List of Abbreviations AS arrhythmia score 4-AP 4-aminopyridine AP action potential APD action potential duration ATP adenosine 5'-triphosphate b (or H) Hi l l coefficient BP blood pressure bpm beats per minute C a 2 + calcium ion C m membrane capacitance COF cardiac output failure C T X charybdotoxin D25% dose producing a 25% change from control D T X alpha-dendrotoxin E A D early afterdepolarizations E C G electrocardiogram EC50 50% effective concentration ED50 50%) effective dose E G T A ethylene glycol-bis(b-aminoethyl ether)-N,N,N',N'-tetraacetic acid ERP effective refractory period gm gram g gravity HEPES (N-[2-Hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid]) HPLC high performance liquid chromatography HR heart rate IC50 50%) inhibitory concentration I K "repolarizing potassium current" IKI inward rectifying potassium current I N a fast inward sodium current iT current threshold ITo transient outward potassium current K + potassium ion Kv voltage-gated potassium channel isoform log Q (or P) hydrophobicity M Q megaohm uA microamp(s) uL microliter(s) umol/kg/min micromoles/kilogram/minute mV millivolt(s) mL milliliter(s) m M millimoles min minute(s) n - sample size Na + sodium ion N d distal nitrogen N E potency Not Estimated N P proximal nitrogen OZ occluded zone P probability value pF picofarad pH hydrogen ion concentration PIPES 1,4-piperazine bissulphonic acid pKa dissociation (ionization) constant PVC premature ventricular contraction QSAR quantitative/qualitative structure-activity relationship r correlation index Ra access resistance Rm membrane resistance Rn R-group RSD Rhythm Search Developments SAR structure-activity relationship(s) SD standard deviation sec second(s) S E M standard error of the mean X time constant required to charge C m t retention time TES (N-tris[hydroxymethyl]methyl-2-aminoethanesulfonic acid) Tris (Tris [hydroxy ethyl] aminomethane) tT duration threshold TTX tetrodotoxin VT ventricular tachycardia V F ventricular fibrillation VFt ventricular fibrillo-flutter threshold V H holding potential v m voltage command Vi„f steady state potential 2° secondary 3° tertiary []i intracellular []o extracellular 1 1. I N T R O D U C T I O N 1.1. Background The heart is a muscular organ that is in a constant state of mechanical and metabolic flux. The actions and interactions of active and passive cellular properties determine each phase of cardiac excitability. Cardiac cells undergo individual and sequential regenerative depolarization and repolarization that allows them to communicate with each other and propagate a normal cardiac rhythm. A n arrhythmia is an abnormality of rate, regularity, or site of origin of the cardiac impulse, or a disturbance in conduction that causes an alteration in the normal sequence of activation of the atria and ventricles. Cardiac arrhythmias can be classified into three main categories on the basis of rate: premature contraction - depolarizations independent of the pacemaker, tachycardia - a series of five or more consecutive premature contractions at a regular rate, fibrillation - a series or groups of premature contractions at an irregular rate. The latter class of arrhythmia is the most life-threatening since the level of asynchronous depolarizations prohibits adequate contraction of heart chambers and as a consequence there is little or no cardiac output. The etiology of cardiac arrhythmias is multifaceted and their severity can range from asymptomatic to life-threatening abnormalities. The genesis of cardiac arrhythmias is associated with many precursors, both pathophysiological (e.g., myocardial ischemia, ionic imbalance) and, most recently identified, genetic (e.g., long QT syndrome). Many of the mechanisms underlying cardiac arrhythmias have been identified in cellular and animal experiments. Precise mechanisms are known and treatment targeted against those mechanisms can be used. In some cases, mechanisms can be inferred, and the choice of treatment is based largely on results of prior experience. The application of this knowledge for clinical practice has been, for the most part, by a trial-and-error approach with the intention to alleviate symptoms of rhythm disorders and prevent sudden death due to heart failure. Clinically, cardiac arrhythmias can be divided as having two main locations, those of atrial and ventricular origin. Current medical interventions for these arrhythmias include surgery (e.g., ablation), electric shock (i.e., electric defibrillation), and drug therapy, or a combination of all three. Just as there are many types and causes of arrhythmias, there are many types of antiarrhythmic drug therapy and these continue to evolve and be refined. It is not the intention of this thesis to discuss the overall spectrum of cardiac arrhythmias and their treatments. Aspects of past research landmarks, and recent developments, will be highlighted to help develop and lend support to the perspectives of this thesis. Much of the present work will centre on ventricular arrhythmias produced after occlusion of a coronary artery. Two distinct time-dependent early (ischemia-) and late (necrosis-induced) arrhythmia phases have been identified and are caused by different mechanisms. The etiology of the early arrhythmia phase along with past and present antiarrhythmic therapies will be the focus on the following introduction. Over the last two decades, antiarrhythmic drug development has undergone extensive evolution. Presently, there is reduced emphasis on drug therapy as a means of controlling cardiac arrhythmias as a result of the inadequacies of many drug classes in the clinical setting, despite some positive and encouraging experimental results. The finding that is common for all of these drugs is their narrow margin between antiarrhythmic and 3 proarrhythmic actions, to which is added extra-cardiac toxicity. The role and value of sodium channel blockers continues to be questioned, and data from clinical trials indicate that the use of this class of drugs should be limited to control symptoms in patients who have arrhythmias and either no, or minimal, heart disease. The decline in the use of sodium channel blockers has led to greater use of (3-blockers and complex Class III agents, such as sotalol and amiodarone, as both primary therapy and adjunctive therapy with or without implantable defibrillators in patients with cardiac disease. However, their multiple mechanisms of actions and non-selectivity for substrates of arrhythmogenesis, i.e., non-diseased myocardium is also targeted, adds a level of complexity and warrants careful dose titration and patient monitoring to reduce drug-induced side-effects, especially during long-term therapy. One of the underlying aims of this thesis is to approach antiarrhythmic drug development from the standpoint that antiarrhythmic drugs need to be limited in their actions on non-diseased myocardium. The proposed objective is to investigate non-selective ion channel blockers with an appropriate pKa that may provide a pathophysiological approach for the suppression of arrhythmias due to myocardial ischemia. The rationale is to minimize drug-induced toxicity and increase drug selectivity for cardiac tissue with abnormal activity by virtue of localizing or having high concentrations of the "active" drug form in the cardiac zone-at-risk. Such an agent may act to preferentially suppress electrical activity in ischemic relative to normal myocardial tissue and reduce the degree of inhomogeneity and increase the threshold for reentrant mechanisms between zones. 4 1.1.1. Etiology of ischemia-induced cardiac arrhythmias: Reentry mechanism A physiological "barrier" may exist in an area(s) of the heart to impair normal cardiac signal conduction. Most notably, this "barrier" (for example, as a result of ischemia) may exist as an interface or gradient of normal to diseased myocardial cells that separates the normal, conducting and excitable from the diseased, non-conducting and inexcitable areas of the heart (Janse et al., 1979). Cardiac arrhythmias due to signal inhomogeneity between these adjacent zones in the heart are often explained by the reentry mechanism, first demonstrated by Wit et al. (1972) based on the original model of reentry proposed by Schmitt and Erlanger (1929). The reentry mechanism depends critically on the relations between refractoriness and conduction velocity and requires the presence of a unidirectional block(s) along the conduction pathways (El-Sherif et al., 1977a, 1977b; Janse et al., 1980; Janse and Capelle, 1982). In general, reentry begins when a propagating wavefront enters a "barrier" of cells that has a higher threshold for excitability. Conduction of the propagating impulse is slowed sufficiently that when it encounters an area of normal cells that was previously excited in the same cardiac cycle but is no longer refractory, a recirculating and self-perpetuating rhythm is established. Factors controlling refractoriness and conduction include action potential duration (APD), sodium and calcium currents, membrane passive properties, and cell-to-cell coupling; current concepts of the. genesis of the cardiac action potential have been reviewed by Noble (1975; 1984). The first experiment that characterized the incidence of cardiac arrhythmias following coronary artery occlusion and evaluated the protective action of a few antiarrhythmic drugs was that of Stephenson and his group (1960). Antiarrhythmic drugs can stop propagation of the reentrant impulse, or prevent reentry by 5 either improving or depressing conduction. Improved conduction can either eliminate the unidirectional block or accelerate the returning wavefront, which will extinguish reentry by encroaching on refractory fibers (Arnsdorf and Bigger, 1972). Drugs that depress conduction can terminate or prevent reentry by transforming a unidirectional block into a bidirectional block (Giardina and Bigger, 1973; Kupersmith, et al., 1975; El-Sherif et al., 1977). Agents that prolong repolarization and refractoriness may prevent or terminate reentry i f the reentering impulse finds the fiber inexcitable (i.e., bidirectional block) (Steinberg etal., 1981). 1.1.2. Antiarrhythmic drug therapy against ischemia-induced arrhythmias: preface The cardiac action potential is a consequence of many transmembrane ionic currents flowing as the result of the opening and closing of selective, pore-forming protein channels. Antiarrhythmic drugs are able to modify ionic currents in cardiac tissues by affecting ion channels, pumps, receptors, or second messenger systems. These drugs have been generally classified on the basis of their electrophysiological actions. Thus, the Vaughan Williams (Vaughan Williams, 1975) or modified schemes (Task Force, 1991) include four main classes that exhibit characteristic effects on the morphology of cardiac action potentials (AP). Agents with class I activity block the fast sodium channels and thereby decrease V m a x ; agents with class II activity are p-adrenoceptor blockers; drugs with class III activity prolong APD; and the class IV activity is due to L-type calcium channel blockade. However, it is now evident that most existing drugs should be categorized into more than one class due to their complex 6 molecular mechanism. Therefore, it is better to speak about activity classes, rather than drug classes, within the context of these classification systems. Much of the research described herein is based on experiments performed on rat cardiac tissue. Therefore, the present study focuses on actions on the fast inward sodium current (INa) and transient outward potassium current (I x 0) since these are the two major ionic currents responsible for the generation of the rat cardiac action potential. Effects on other ionic currents, such as the inward calcium (ICa), will also be considered but in a limited manner. 1.1.2.1. Therapy using blockers of the cardiac inward sodium current The first antiarrhythmic drug was identified by Frey (1918) and it was later shown to be a blocker of the sodium channel (Weidmann, 1955; Johnson, 1957). Since then, drugs of this class were shown to have antiarrhythmic actions and the study of sodium current modulation as an antiarrhythmic mechanism and the development of novel blockers for this channel ensued. Until 1989, sodium channel blockers (class I agents) represented the most favored group of antiarrhythmic drugs for use in the clinic. This class of drugs, at therapeutic concentrations, blocks cardiac sodium channels and thereby reduces V m a x of the cardiac action potential and, as a result, reduces impulse conduction. Many of these agents produce more blockade at fast rather than slow heart rates ("use-dependence"; Courtney, 1975; Hondeghem and Katzung, 1977). The modulated receptor model (Hondeghem and Katzung, 1977; Hille, 1977) has been the most enduring model to address the issues of state-specific blockade by a drug. Although the model has been most extensively studied for sodium and calcium channels in cardiac tissues, it may apply 7 equally well to other membrane channels. According to this hypothesis, affinity of drug binding, and subsequent sodium channel block, depends on the functional state of the sodium channel. During the action potential (open and inactivated states) drug association with the channel occurs, while during diastole (resting state) drug dissociation, and therefore relief from block, is preferred. Class la and Ic drugs dissociate relatively slowly (compared to physiological cycle lengths) from the sodium channels during diastole (Varro et al., 1985, 1988, 1990; Campbell and Vaughan Williams, 1983; Langenfeld et al., 1990; Lee and Rosen, 1991; Winslow et al., 1989), such that the action can be characterized by a recovery time constant longer than 2 sec. As a consequence, class la and Ic drugs such as quinidine, disopyramide, flecainide, encainide, propafenone, etc. exert strong sodium channel depression at a normal heart rate, whereas class lb drugs such as lidocaine, mexilitine and tocainide inhibit impulse conduction only at heart rates faster than resting physiological rates. The use-dependent behavior (onset and recovery kinetics) of the sodium channel block by a given drug may relate to their physiochemical properties, such as lipid solubility and molecular weight (Courtney, 1980). Serious side effects have been identified that considerably limit the use of sodium channel blockers. This was particularly realized in the Cardiac Arrhythmia Suppression Trial (The CAST Investigators, 1989), in which the placebo group had better survival in ischemic heart disease, or postinfarction than patients treated with flecainide or encainide (CAST I; The CAST Investigators, 1989) or moricizine (CAST II; The CAST II Investigators, 1992). Therefore, agents that depress sodium channels at normal heart rates (with the exception of class lb agents) do not represent a promising direction for future drug development. 8 In our laboratory, experimental evidence in rats subjected to coronary artery occlusion lends further support to the toxicity of these drugs and their lack of cardiac selectivity (Igwemezie et al., 1992; Barrett et al., 1995). In the clinic, the use of other class I antiarrhythmics (Velebit et al., 1982; Abbott, 1981; Pentecost et al., 1981; Rademaker et al., 1986; Denaro and Benowitz, 1989; Jaffe, 1993) has also suffered a similar fate to class Ic class (e.g., flecainide), in that drug-induced proarrhythmia and toxicity outweigh the antiarrhythmic usefulness of these agents. 1.1.2.2. Therapy using blockers of the cardiac outward potassium current Prolongation of the action potential is considered an important antiarrhythmic intervention based on a number of clinical observations. For example, it is known that in cases of extreme hypocalcemia, there is marked prolongation of cardiac repolarization but the occurrence of arrhythmias is rare (Oshita et a l , 1980). Similarly, in hypothyroidism, a situation in which there is bradycardia and uniform prolongation of repolarization (Freedberg et al., 1970; Johnson et al., 1973), arrhythmias of all kinds were reported to be uncommon. Reducing potassium (K +) outward currents can bring about effective prolongation of the action potential and, for example, in reentry type arrhythmias with a short excitable gap, increasing refractoriness could constitute an antifibrillatory mechanism. Following the CAST trials, this area of cardiac current modulation gained more attention from the standpoint of drug development. There are a number of K + currents that contribute to the cardiac repolarization process: 1) the transient outward (ITO) responsible for the rapid initial repolarization process from the crest of the action potential to the plateau level (Coraboeuf and 9 Carmeliet, 1982), 2) the delayed K + current (IK), involved in the overall repolarization process during the plateau (Noble, 1975) and 3) the I K 1 or inward rectifier responsible for final repolarization and the maintenance of the resting potential (Nobel, 1965; 1975). Block of the delayed K + current has been extensively studied and much effort has been expanded in the development of specific I K blockers, particularly for the rapid component, IK r. Clinical results with such agents have not been encouraging in view of the fact that the clinical evidence of reduced mortality of patients suffering from infarction or ischemia-related arrhythmias paralleled the outcomes from the CAST trials (see SWORD Trial, Waldo et a l , 1993). Several of the findings suggest that I K blockers preferentially prolong the action potential at low heart rates ("reverse use-dependence"; Hondeghem and Snyders, 1990). A recent experiment (Vanoli et al., 1995) following the SWORD trial analyzed the role of heart rate modulation associated with blockade and offered some explanation for the clinical outcomes with d-sotalol. In canine models with pre-existing (30 day) M i ' s subjected to transient myocardial ischemia, the reduction in reflex heart rate increase in response to exercise by d,l-sotalol and propranolol, but not d-sotalol, was associated with a lower incidence of VF-induced mortality (Vanoli et al., 1995). Simply, selective I K r blockade (d-sotalol), in the absence of adequate P-adrenergic blockade (d,l-sotalol) may not be sufficient to prevent sudden death when the mechanism is a lethal ischemia-related arrhythmia associated with sympathetic hyperactivity. The clinical relevance of heart rate associated with myocardial ischemia was implicated to be due to the contribution of autonomic activation in ischemia-related sudden cardiac death (Vanoli et al., 1995). However, excessive slowing of heart rate, or a pause and a concomitant 10 increase in QT interval (i.e., due to I K blockade) predisposes the initiation of early afterdepolarizations (EAD)'s and EAD-related cardiac arrhythmias (see review by Roden et al., 1996). The most common arrhythmia associated with EAD-related triggered activity is a series of rapid repetitive polymorphic QRS complexes with characteristic undulating peaks (i.e., torsades des pointes; Roden et al., 1986; Surawicz, 1989; Cranefield and Aronson, 1988;). Therefore, reduction of I K current as a target for achieving APD prolongation remains to be clearly established. If slowing conduction or increasing refractoriness are the desired antiarrhythmic actions, the focus should, perhaps, be diverted to unselective ion channel blockade, i.e., neither sodium nor potassium channel blockade alone, but rather, a mixed blockade of both sodium and potassium channels. 1.1.2.3. Therapy using mixed channel blockade There is evidence to suggest that the combination of class I and III antiarrhythmic drug therapies enhances efficacy in suppressing cardiac arrhythmias as compared with high dose monotherapy (reviewed in Marcus, 1992). Amiodarone is one agent that has multiple ion channel blocking properties and has been convincingly shown to reduce the incidence of cardiac death in postmyocardial infarction patients by reducing the incidence of sudden cardiac death (Ceremuzynski et al., 1992). First developed by Labaz Laboratory in Belgium as an antianginal compound, it was later found to possess antiarrhythmic and electrophysiological actions (Singh et al., 1989). In addition to its class III properties, amiodarone also exhibits sodium and calcium channel blockade (Nattel et al., 1987; Singh et al., 1989; Sheldon, et al., 1989), nonspecific sympathetic 11 blockade (Polster and Broekhuysen, 1976; Nokin et al., 1983), inhibition of thyroid metabolism (Nademanee et al., 1989) and acts as a phospholipase inhibitor (Shaikh et al., 1987). Many of its actions have a delayed onset that, coupled with the drug's long elimination half-life (Mason, 1987; Singh et al., 1989; Nattel et al., 1994), makes it difficult to optimize doses. Given these multiple and complex pharmacological characteristics, it has been difficult to pinpoint amiodarone's mechanism of antiarrhythmic action. It is not surprising that there has been quite a disparity in the reports of the rates of efficacy and toxicity for amiodarone (Mason, 1987; Nademanee et al., 1983; Heger et al., 1981; Singh et al., 1984; Herre et al., 1989). Amiodarone's poorly understood pharmacology, and its reputed toxicity, have made it an agent of last resort. Attention has therefore focussed on newer compounds that have the propensity to block more than one ion channel. This strategy has not been widely accepted since there is limited evidence to support an increase in antiarrhythmic efficacy for individual agents with mixed sodium and potassium channel blockade (Duff et al., 1990; Morganroth, 1987). Quinidine is one such agent, a moderately frequency-dependent sodium channel blocker with blocking actions on I T 0 (Imaizumi and Giles, 1987), whose limited efficacy (i.e., a poor dose-related antiarrhythmic activity with < 100% maximum protection) and non-cardiac selective actions have been confirmed in rats with ischemia-induced arrhythmias (Barrett et. al., 1995). Much of this drug's limitations as an antiarrhythmic may be associated with its class la properties and extra-cardiac toxicities. Multiple ion channel blockade has the potential to achieve optimal modification of the cardiac AP. Based on work in the last two decades, future antiarrhythmic drug 12 development must ask the following questions: "Which cardiac ion channels should be targeted to formulate a combination best suited for optimal AP modification in the setting of myocardial ischemia?" and "What is that nature of this combination or the degree to which these targets are modulated that affords selectivity for cardiac (ischemia-related) arrhythmias and limited drug-induced toxicity?". 1.2. Drug development and the ion channel targets for modulating cardiac action potential 1.2.1. Cardiac ion channels: preface Voltage-gated sodium, calcium and potassium channels are responsible for the generation of propagated electrical signals in neurons and other excitable cells (Hille, 1992). These ion channels respond to changes in transmembrane potential, often produced by the binding of neurotransmitters to ligand-gated ion channels, and when activated generate action potentials and propagate the membrane potential change over the surface of the cell, along a nerve axon or muscle fiber. Activation of the ion channels involves opening of selective transmembrane pores through which specific ions can diffuse down their electrochemical gradient into or out of the cell. Thus, in most cases, activation of sodium or calcium channels leads to inward movement of these ions and depolarization of the cell whereas activation of potassium channels leads to outward movement of potassium ions and repolarization or hyperpolarization of the cell. The permeability increase resulting from activation of voltage-gated ion channels is biphasic. Upon depolarization, permeability to sodium, calcium or potassium increases dramatically over a period of 0.5 to hundreds of milliseconds and then decreases to the baseline level over a period of 2 milliseconds to seconds. This biphasic behavior results 13 from two experimentally separable gating processes that control ion channel function: activation, which controls the rate and voltage dependence of the permeability increase following depolarization, and inactivation, which controls the rate and voltage dependence of the subsequent return of the ion permeability to the resting level during a maintained depolarization. The voltage-gated ion channels can therefore exist in at least three functionally distinct states or groups of states: resting, active, and inactive. Both resting and inactivated states are non-conducting, whereas the ion conductance of the activated ion channels is highly selective. Since much of the present work has been performed on rat cardiac tissue, the discussion of the main cardiac ion channels for this thesis will comprise the fast inward sodium (INa) and transient outward potassium current (I x o), major currents for this species. 1.2.1.1. Fast inward sodium current (IN a) The sodium channel is critical to the genesis of action potentials in the working myocardium. This is because the initial upstroke of the action potential of nerves and most cardiac myocytes is due to a transient inward current caused by an influx of sodium ions (INa) through the channel (Noble, 1975; Hille, 1984 and for reviews see Coraboeuf 1978; Noble, 1984; Fozzard and Arnsdorf, 1986). Much of the current and functional understanding of the sodium channel is based on the work of Hodgkin and Huxley (1952) who first unequivocally demonstrated that a transient inward current of sodium ions was responsible for action potential propagation in squid giant-axons. Although voltage-gated sodium channels demonstrate a range of sensitivities to various toxins amongst cell types 14 (e.g. brain, skeletal, and cardiac), they do not show the extreme diversity of function between species that is characteristic of calcium and potassium channels. The results of the CAST trial (the CAST Investigators, 1989) have highlighted the risks inherent in the oversimplistic interpretation of the antiarrhythmic actions of sodium channel blockers. No longer is ventricular ectopy believed to be a high risk or a primary pre-requisite associated with sudden cardiac death in patients with myocardial infarction, but rather, asynchronous activation and signal propagation patterns between adjacent areas of the heart that lead to reentrant arrhythmias. The aim of reducing I N a as an antiarrhythmic mechanism in the treatment of reentrant-related arrhythmias is to slow or block conduction sufficiently in those areas of the heart where there is already reduction of impulse conduction (unidirectional block) so as to terminate the impulse's reemergence (bidirectional block). For I N a to be an effective antiarrhythmic target, block should develop during the action potential and decrease during diastole. The narrow margin between the extent of blockade that is antiarrhythmic and that which is proarrhythmic is due to the time- and voltage-dependent block of sodium channels that favors net binding or unbinding. Time-dependent blockade is best explained as development of drug-associated blockade increases with increasing rate of channel activation. By definition, this defines "use-dependency" and this pattern of channel block is central to antiarrhythmic action since the dynamic nature of the blocking process facilitates the occurrence of antagonism (antiarrhythmia) or agonism (proarrhythmia or over-expression) depending on the drug's onset and offset kinetics. For example, sodium channel blockade with relatively fast recovery from block (e.g., class Ib; lidocaine = 180 ms; Weld and Bigger, 1975; Grant et 15 al., 1980; Clarkson and Hondeghem, 1985) shows marked direct use-dependency, with effects appearing only at fast heart rates (e.g., during tachycardia). However, with very slow recovery from block (e.g., class Ic; flecainide > 2 sec), most of the use-dependent effects may be already manifested at physiological heart rates, a condition associated with proarrhythmic risk (Hondeghem, 1987; Brugada et al., 1991). Voltage-dependence is another aspect of use-dependent action that must be considered since drug affinities to different sodium channel states, as a function of membrane potential, can significantly affect drug binding and/or unbinding. The "ideal" sodium channel blocker is one that exhibits use-dependent blockade with fast diastolic recovery from block that is a function of membrane potential. Lidocaine is one such agent whose fast diastolic recovery allows it to be well tolerated in normal myocardium (at physiological heart rates, e.g., 70 bpm) and be more depressant (as an inactivated-state blocker) in depolarized myocardium (e.g., ischemic tissue). The description for lidocaine, thus far, suggests that it acts selectively on myocardial tissue under arrhythmogenic conditions (i.e., depolarized tissue at risk of high-frequency automaticity). As mentioned in the outset (see Section 1.1.2.1.), lidocaine has extra-cardiac side effects (e.g., convulsions) and this lack of cardiac selectivity limits its antiarrhythmic usefulness (Igwemezie et al., 1992; Barrett et al., 1995). Except for its convulsiogenic potential, one circumstance that can be extrapolated from lidocaine or agents with "ischemia-selective" properties is that differences in the blocking action between diseased and normal myocardium determines a margin of therapeutic safety with respect to antiarrhythmia and proarrhythmia. Therefore, the issue of antiarrhythmic 16 selectivity between normal and diseased myocardium as part of the antiarrhythmic profile will be discussed later in detail. 1.2.1.2. Transient outward current (Ixo) Potassium currents are fundamentally important in initiating and modulating repolarization. In the rat heart, the main repolarizing potassium current is I T O (Josephson et al., 1984) and its prevalence in this species explains the short repolarization time in rat ventricular tissue. As an initial repolarizing current, it has a significant effect on the height of the early plateau, thus influencing activation of other plateau currents that control repolarization (Gintant et al., 1992). Studies from sheep and rabbit cardiac tissues have determined that I T O is composed of two components; these have been termed I x o l and I x o 2 by Tseng and Hoffman (1989): I x o l , is a Ca2+-independent and 4-aminopyridine (4-AP)-sensitive outward current and I x o 2, is a Ca2+-dependent and 4-AP-resistant outward current, the latter being dependent on the Ca 2 + inward current (ICa) and the resulting release of Ca 2 + from the sarcoplasmic reticulum (Kenyon and Gibbons, 1979; Coraboeuf and Carmeliet, 1982; Escande et al., 1987; Hiraoka and Kawano, 1989). The charge carrier of I x 0 2 has been identified as being Cl" rather than K + ions (Zygmunt et al., 1991, 1992). In single cells from human atrium (Escande et al., 1987; Shibata et al., 1989; Heidbuchel et al., 1990) and ventricle (Corabeouf and Nargeot, 1993), I x o l has been identified and the general properties of this current and other K + currents in human cardiac cells have been summarized by Corabeouf and Nargeot, (1993). During the cardiac action potential, I x o shows rapid activation followed immediately by a much slower recovery from inactivation (several seconds), hence the 17 name, "transient outward". A reduction of I x o conductance slows the recovery from inactivation prolonging the initial repolarization phase and, ultimately prolonging the ventricular action potential. The processes of inactivation (Kukushkin et al., 1983; Josephson et al., 1984; Clark et al., 1988; Shibata et al., 1989; Fermini et al., 1993) and recovery from inactivation (Shibata et al., 1989; Fermini et al., 1993) are strongly voltage-dependent such that the kinetics of these two processes significantly influences the modulation of the size of I x o current in mammalian heart. For example, removing inactivation by hyperpolarization results in shortening of the action potential, whereas increasing the time spent at resting potential, or at depolarized membrane potentials, results in less inactivation and a shortened action potential upon subsequent initiation. In fact, the time-course of reactivation appears to be a key physiological variable that controls the functional features of I T O (Shibata et al., 1989; Fermini et al., 1993). For this reason, I T 0 also undergoes significant changes as a function of frequency, such that action potential duration and amplitude are decreased with short-coupled activation intervals. Since I x o in mammalian heart has the appropriate kinetics and magnitude to be significant in early repolarization, it inherently has a strong influences upon subsequent plateau currents (e.g., I C a, IK). In the rat ventricle, blockade of I T O has profound prolonging effects on APD. However, the exact role of I x 0 for action potential duration is more complex in cardiac tissues where it may be one of many plateau currents that control repolarization but this remains to be clarified. Tedisamil, a non-selective but potent I T O blocker, blockade of I T 0 in rat ventricle that is deficient in other functional repolarizing currents, e.g., IK, against electrically- (Beatch et al., 1991) and ischemia-induced arrhythmias (Adaikan et al., 1992), was shown to only occur at near-toxic doses. 18 At these high doses, the antiarrhythmic mechanism may be explained by increases in cardiac refractoriness such that minimum path lengths for reentry are reduced to a sufficient degree that multiple reentry circuits responsible for ischemia-related arrhythmias are reduced. The bradycardic actions of tedisamil also increased the cycle length and thereby extended the time "window" (i.e., non-refractory or diastolic period) for potential reentry. This may explain the incomplete abolition of cardiac arrhythmias (i.e., low-frequency tachycardias) by tedisamil reported in that study (Adaikan et al., 1992). Moreover, there is evidence to indicate that the regional population of I T 0 channel density is variable in cardiac tissue (i.e., epicardium, midmyocardium, endocardium), helps give rise to the heterogeneity of the action potential within the ventricular wall (Antzelevitch et al., 1991). Therefore, in cardiac tissue with other prominent repolarizing currents, the resultant dispersion of refractoriness that may be amplified by I T 0 blockade is potentially proarrhythmic. These considerations further emphasize the need to investigate antiarrhythmic actions with selective rather than indiscriminant blockade of cardiac ion channels. In this section, the main conducting ion channels, I N a and I x o , in the rat have been described. Their antiarrhythmic potential, or lack thereof, has been described above with agents that selectively reduce their conductance. Work in our laboratory to profile the antiarrhythmic activity of these drugs in our rat models has provided further evidence in support of the limited antiarrhythmic protection afforded by selective ion channel blockade (Igwemezie et al., 1992; Barrett et al., 1995). Before continuing with the discussion of possible alternative(s) to antiarrhythmic drug development, the following sections will provide background material for drug interactions at a molecular level. 19 1.3. Ion channels ( I N a / I T O ) at a m o l e c u l a r level This preface is a summary of the molecular research that has been performed to elucidate the structure and function of cardiac ion channels. A detailed discussion of the molecular evidence for sodium and potassium currents will be presented in the following sections. The structures of the voltage-gated sodium, calcium, and potassium channels are similar at a molecular level, yet are functionally diverse in the core ion channel motif which forms the basis for ion selectivity and, possibly, drug binding. Essentially all models for the structure of voltage-gated ion channels include a transmembrane pore in the center of an array of four homologous transmembrane domains. Identification of the segments that line the transmembrane pore and define the conductance and ion selectivity of the channels is of great interest in relation to binding interactions between compound and channel. Although each channel contains associated auxilliary subunits, the principal subunits can carry out the basic functions of the voltage-gated ion channels by themselves and, therefore, must contain the necessary structural elements for ion channel function within them. Various toxins, drugs, and inorganic cations are blockers of the voltage-gated ion channels and, in several biophysical analyses, their mechanism of action indicates that these molecules enter and bind within the transmembrane pores of the channels and compete with permeant ions for occupancy of the pore (reviewed in Hille, 1992). Such studies have also identified amino acid residues that form the extra-cellular and intra-cellular mouths of the transmembrane pores. Molecules binding at an extra-cellular 20 region near the mouth of the channel (from the cytoplasm or extra-cellular side) or along the transmembrane domains from within the lipid phase may serve as alternative sites to channel blockade from within the pore. This may explain the inherent diversity (and selectivity) of the different molecules/ion channel blockers that exists for each cardiac channel that cannot be accounted for by direct interaction with the channel pore itself. It should be noted that the different channel states (i.e., resting, activated, or inactivated) that are present prior to binding by a molecule or blocker also adds a kinetic or conformational factor that could affect their binding interactions. A brief overview of the molecular make-up of I N a and molecular channel isoforms of I T 0 with respect to drugs and toxins that act as specific blockers of these channels, as well as mutagenic interventions, may identify certain ion channel similarities and/or differences that may offer possible binding interaction and selectivity. 1.3.1. Molecular basis of Na + channel function a-Subunit. Since the first cloned ion-channel gene from the eel electroplax was isolated (Noda et al., 1984), a great deal of information has been obtained over the past decades concerning the molecular basis of sodium channel structure and function. Two main experimental approaches have proven valuable in this molecular pursuit: antibodies against short peptide segments of the principal a-subunits and oligonucleotide-directed mutagenesis on cDNAs encoding the principal a-subunits. Much of the molecular work has identified the a-subunit as a highly conserved unit of sodium channels (Noda et al., 1984) with the following evidence to indicate that it is a transmembrane polypeptide. The a-subunit is glycosylated (Cohen and Barchi, 1981; 21 Figure 1 Proposed structure and functional domains of the voltage-gated sodium channel. Top: The transmembrane topology with four homologous domains (I-IV), each consisting of six membrane-spanning segments (S1-S6). A l l four S4 segments consist of a charged residue every third position (see also lower right). Amino acid sequences involved in tetrodotoxin resistance, local anesthetic binding, and inactivation (I (He), F (Phe), M (Met)) are also indicated. Mutations of amino acids indicated by (*) have been identified in the congenital long Q-T syndrome. Lower left: Schematic representation of the folding conformation of the 24 membrane spanning segments that form the sodium channel with the 5 t h (white) and 6 t h (uniform gray) segments lining the pore. The figure was reproduced from Roden and George, 1997. 23 Figure 2 Proposed structure for the voltage-gated potassium channel. A) Membrane topology showing hydrophobic membrane-spanning domains (S1-S6) of one tetrameric subunit (top and middle) and the entire tetrameric complex (bottom; the nearest subunit has been removed to expose the interior of the pore and the farthest subunit has been colored white for clarity). The thick cylinders represent a-helices, the medium width cylinders represents p-strands of the "P" loop, and the thin cylinders represent connecting linkers. This figure was reproduced from Durrell and Guy, 1992. Ribbon (B) and cartoon (C) representation of the KcsA K + channel as modelled by Doyle et al. (1998). In the latter model, a large aqueous inner cavity is shown stabilizing an (K + ) ion. 24 Miller et al., 1983; Barchi, 1983; Messner and Catterall, 1985) and is covalently labeled by a- and p-scorpion toxins (Beneski and Catterall, 1980; Barhanin et al., 1983; Darbon et al., 1983; Sharkey et al., 1984) and by tetrodotoxin (TTX) derivatives (Lombet et al., 1983; Noda et al., 1989), which act only on the extra-cellular side of the channel. The a-subunit is also exposed on the intra-cellular side of the membrane since it is phosphorylated on four sites by cAMP-dependent protein kinase within intact synaptosomes (Costa and Catterall, 1984). Channel pore. Based on the above evidence and the predicted amino acid sequence, the a-subunit of the sodium channel consists of four homologous domains, termed I-IV, along with five hydrophobic (Sl-3, S5, and S6) segments at equivalent positions in the four domains (Noda et al., 1984). Evidence initially developed in studies with potassium channels (see below) and now further confirmed in the sodium channel (Backx et al., 1992; Heinemann et al., 1992) indicates that the pore through which sodium ions permeate is lined by the P loop between S5 and S6. Mutations in this region can completely disrupt ion-channel function (i.e., ion selectivity and/or permeability) or the binding interaction of pore blockers (see below). Neutralization of the negatively charged residues in the first and second repeat (Terlau et al., 1991; Heinemann et al., 1992) as well as amino acid mutations in the third and fouth domains (Heinemann et al., 1992) has resulted in reduced flux and selectivity of Na + ions, respectively. "Voltage"-sensor. In addition to the five transmembrane segments, a fourth segment, S4, in each of the four domains contains five to eight positively charged residues spaced at three-amino acid intervals, whereby every third residue is an arginine or lysine (Noda et al., 1984). This led to the prediction that the S4 region might serve as 25 the "voltage sensor", moving in response to a change in membrane potential (Catterall, 1988; Numa, 1989; Guy and Conti, 1990). Studies with site-directed mutagenesis (i.e., charge neutralization mutations) and expression of mutant channels in heterologous systems have supported this concept (Summer et al., 1989; Yang et al., 1996). The sequence encoding the S4 segment is well conserved amongst the sequences of all the cloned voltage-gated ion channels (Tanabe et al., 1987; Pongs et al., 1988) but the mechanism by which the S4 segment "senses" voltage changes is not known. Models whereby the S4 segment moves outward during the gating process and conformational changes of the sodium channel have been proposed (Catterall, 1988; Guy and Conti, 1990) and, perhaps, substantiated by recent mutagenic evidence (Chahine et al., 1994). Accessory (3-subunits may or may not accompany ot-subunits nor is it clear whether (3,-subunits are required to recapitulate native I N a or associates with the cardiac sodium channel (Cohen et al., 1993). Thus, the exact roles and functions of these accessory units in many channel types and species remains unexplained. Inactivation. The ion conductance of sodium channels is terminated within a few milliseconds after opening by the process of inactivation. Inactivation can be specifically prevented by treatment of the infra-cellular surface of the sodium channel with proteolytic enzymes (Armstrong, 1981). Such observations led to the proposal of an autoinhibitory, "ball-and-chain" model for sodium channel inactivation in which an inactivation particle tethered on the surface of the sodium channel (the ball) diffuses to a receptor site in the intra-cellular mouth of the pore, binds, and blocks the pore during the process of inactivation (Armstrong, 1981). The sodium channel segments that are 26 required for fast inactivation has been identified as a short intra-cellular segment of 15 -20 amino acid residues connecting homologous domains III and IV. This proposed inactivation-gating loop contains highly conserved clusters of positively charged and hydrophobic amino acid residues. Further analysis of this loop revealed a highly conserved three-residue, hydrophobic cluster, IFM, with the single phenylalanine (F) in the center of this cluster at position 1489 (phel489) in sodium channel domain II as the critical residue for sodium channel inactivation. The interaction of phel489 with the receptor of the inactivation gating particle is likely to be hydrophobic because there is a close correlation between the hydrophobicity of the residue at that position and the extent of fast sodium channel inactivation (West et al., 1992). In addition, a candidate for the inactivation gate receptor region has been identified to be three adjacent hydrophobic residues at the intra-cellular end of the S6 segment in domain IV (McPhee et al., 1994). Sodium channels can also be inactivated (slow inactivation) by processes that occur over a much longer time scale, from 100 msec to minutes (Simoncini and Stuhmer, 1987; Ruff et al., 1987, Ruben et al., 1992). This phenomenon of slow inactivation has been reported previously following long depolarizing pulses (Narahashi, 1974; Aimers et al., 1983) but has only recently been linked to several human neurological disorders (reviewed in Barchi, 1995; Cannon, 1996) and long QT syndrome (Wang et al., 1995a, 1995b; Wang et al., 1996). Proteolytic removal of fast inactivation by intra-cellular application of enzymes does not prevent slow inactivation (Valenzuela and Bennett, 1994). However, the molecular mechanism underlying this slow inactivation phenomenon for voltage-gated sodium channels remains unknown. 27 Toxin and local anesthetic block. Sodium channels expressed in cardiac muscle cells are distinguished by the fact that they are resistant to nanomolar concentrations of tetrodotoxin (Brown et al., 1981), but are more sensitive than neuronal channels to inhibition by lidocaine (Bean et al., 1983). Cardiac sodium channels are generally referred to as tetrodotoxin-resistant, although they are blocked by micromolar concentrations of the toxin. The hydrophilic and permanently charged guanidinium toxins, tetrodotoxin and saxitoxin, block sodium channels by binding with high affinity to the extra-cellular mouth of the pore (Noda et al., 1989). Block of their binding by protonation or covalent modification of carboxyl residues led to the model that these cationic toxins bind to a ring of carboxyl residues at the extra-cellular mouth of the pore (Noda et al., 1989; Terlau et al., 1991; Hille, 1992). Biophysical studies suggest that local anesthetics and related antiarrhythmic drugs bind at a site on the sodium channel which is accessible only from the infra-cellular side of the membrane and is more accessible when the sodium channel is open (Frazier et al., 1970; Strichartz, 1973). In addition, such drugs also interact with amino acid residues in S6 segments. Analysis of the S6 segment in domain IV of the sodium channel a-subunit reveals two aromatic amino acid residues in positions 1764 and 1771 which are required for high-affinity binding of local anesthetics (Ragsdale et al., 1994). These two residues are located approximately 11 A apart on the same face of the proposed IVS6 a helix and may interact with the aromatic and positively charged amino groups of the drug molecule which are spaced 10 to 15 A apart. It is likely that these pore-blocking drugs also interact with amino acid residues in the S6 segment of other domains of the sodium channel as they bind within the pore surrounded by all four domains. Recent site-directed mutagenic 28 studies have identified that phenylalanine in position 1710 in the transmembrane segment IVS6 may be involved in stabilizing local anesthetics binding interactions to open and inactivated channel states via cationic-rc or aromatic-aromatic interactions between the charged or aromatic moieties on the drug molecule and the aromaticity of the residue (Li etal., 1998). 1.3.2. Molecular basis of K + channel function Amongst the cardiac voltage-gated ion channels, K + channels are more numerous with respect to their diversity as ion channels than any of the other types of channel that is expressed in cardiac cells (reviewed in Hille, 1984; Christie, 1995, Deal et al., 1996, and Catterall, 1996). This diversity has a physiological significance in the heart, and more importantly in different regions of the heart, such that the various types of K + channels play specific roles in controlling resting membrane potentials and shaping action potential waveforms, as well as influencing refractoriness and automaticity (Nobel, 1975). For this reason, there is interest in defining the molecular correlates of functional cardiac K + channels. Unfortunately, the low abundance of K + channels (on average, one channel per square micrometer) has made attempts to purify K + channels difficult (Rehm and Lazdunski, 1988; Schmidt and Betz, 1988). It should be noted that no cardiac-specific K + channels have been cloned. An alternative approach, making use of Drosophila genetics, has allowed the cloning of a K + channel gene. The first voltage-gated K + channel (Kv) pore-forming (oc)-subunit was cloned from the Shaker locus in Drosophila (Papazian et al., 1987; Kamb et al., 1988; Pongs et al., 1988). Three highly homologous subfamilies of K v a-subunits, Shab, Shaw, and 29 Shal, were subsequently cloned from Drosophila (Butler et al., 1989; Wei et a l , 1990). Classification of the K + channel a-subunit genes of the Shaker, Shab, Shaw, and Shal subfamilies are presently referred to as K v l . x , Kv2.x, Kv3.x, and Kv4.x, respectively (Chandy, 1991). When copy cDNAs from each of these subunits is injected into Xenopus oocytes, depolarization-activated K + channels are expressed but their biophysical and pharmacological properties do not normally match those of the native K + channels (reviewed in Christie, 1995, Deal et al., 1996, and Catterall, 1996). Of the voltage-gated K + channel oc-subunits cloned from heart, three, K v l . 4 (Tseng-Crank et al., 1990; Roberds and Tamkun, 1991; Tamkun et al., 1991; Po et al., 1992; Comer et al., 1994), Kv4.2 (Baldwin et al., 1991; Blair et al., 1991; Roberds and Tamkun, 1991), and Kv4.3 (Dixon et al., 1996) produce 4-AP-sensitive rapidly activating and inactivating voltage-gated K + currents that qualitatively resemble I x o . Additional biochemical (Dixon and MacKinnon, 1994), immunohistochemical .(Barry et al., 1995; Xu et al., 1996) and electrophysiological (Fiset et al., 1997; Nakamura et al., 1997; Yeola and Snyders, 1997) findings have concluded that Kv4.2 is the closest correlate to rat ventricular I T 0 . The accepted membrane topology of K v channels and other homologous sequences has been most thoroughly modelled by Durell and Guy (1992). The main core structure that is contained within the Shaker sequence is highly conserved among all Kv sequences determined to date. Sequence analysis revealed a protein with six transmembrane domains and a region (P or H5 loop) between the fifth and sixth membrane-spanning domains that contributes to the K+-selectivity pore (Jan and Jan, 1992; Pongs, 1992a, 1992b). The fourth transmembrane domain is homologous to the corresponding region in voltage-gated Na + and Ca + channels, placing Shaker in the S4 30 superfamily of voltage-gated channels (Jan and Jan, 1992). Heterologous expression of Shaker produces voltage-gated K + channels (Timpe et al., 1988) composed of four subunits (Schwarz, 1988; MacKinnon, 1991). Whereas the central core is highly conserved, differences in amino acid sequence, that reside primarily in the N H 2 - and COOH-terminal regions (Deal et al., 1996), define the different subfamilies/isoforms. The following structural and functional predictions have been strongly supported by site-directed mutagenesis. Voltage sensor. The S4 segment which carries a number of positively charged amino acid residues is involved in channel gating and acts as a voltage sensor (Durell and Guy, 1992; Jan and Jan, 1992). In most examples, every third amino acid is positively charged along the full length of the predicted a-helix. These charges are suitably located to translate changes within the membrane electric field into channel gating. Papazian et al. (1991) systematically mutated each of these arginine or lysine residues in Shaker, either to an uncharged residue (glutamine) or a different, positively charged residue. Unlike the results of similar experiments performed on voltage-dependent Na + channels (Stuhrner et al., 1989), the effects of charge removal were not equivalent to elimination of gating charges from the voltage sensor, which should decrease the slope of the activaiton curve in a predictable manner. In contrast, a good correlation was found among native Kv channels and charge density of the S4 region (Logothetis et al., 1993). The gating charge of the Shaker K + channel has been estimated to be about four in voltage-clamp studies of the whole cell current (i.e., channel activation is associated with the translocation of the equivalent of about four positive charges from the cytoplasmic to the extra-cellular side of the membrane) (Timpe et al., 1988). Single channel analysis, which 31 allows better separation of activation from the inactivation process, gives a higher estimate of eight for the gating charge oi Shaker K + channel (Koren et al., 1990; Zagotta etal., 1990). The channel pore. Expression of the resultant Shaker oc-subunits in oocytes and monitoring current modulation by various toxins has identified the P or H5 region as comprising the ion channel pore (Hartmann, et al., 1991; Yellen et al., 1991; Yool and Schwartz, 1991). This short stretch of uncharged and largely hydrophobic (-19) amino acids between S5 and S6 is highly conserved, critical for K+-selectivity (Yool and Schwarz, 1991; Slesinger et al., 1993) and permeation (Jan and Jan, 1992; Sather et a l , 1994). The domain linking segments S4 and S5, as well as the carboxy terminal end of segment S6, also contribute to the channel pore (Pongs, 1992a, 1992b; Hoshi and Zagota, 1993). The narrow inner mouth (~3A; Hille, 1984) of the pore is formed by a P-barrel structure composed of four P-hairpins, one from each subunit (Durell and Guy, 1992). According to the crystal structure of the KcsA K + channel (from Streptomyces lividan) a more detailed description of the K + channel pore and selectivity filter is depicted (described in Doyle et al., 1998). The overall length of the pore has been estimated to be 45A. From inside the cell, the pore begins as a tunnel 18A in length (inner pore) and then opens into a water-filled central cavity (~10A across). The inner pore and cavity are presumed to lower electrostatic energy barriers required by a cation entering the low dielectric bilayer by providing a "pseudo-aqueous" inner channel environment that dose not disrupt the electrostatic properties of the cation. In essence, the K + channel achieves a high ion throughtput by providing a low resistance pathway from the cytoplasm to the selectivity filter (Dolye et al., 1998). To further ensure a high 32 ionic throughput, it has also been proposed (Doyle et al., 1998) that there is stabilization offered by the surrounding amino acids of the selectivity filter that is specific for K+ ions to counterbalance the energy loss due to water-shell dehydration following passage from the inner cavity. One appealing hypothesis involving the K+-selectivity of the pore proposes a model involving a ring of aromatic residues (a ring of tyrosine from the G Y G sequence; Durell and Guy, 1992) located at a constricted region of the pore for coordinating the passage of K + ions via cationic-7t interactions (Kump and Dougherty, 1993). The inner and outer limits of the channel lining in the membrane have been defined by blockade of K + conductance with T E A (MacKinnon and Yellen, 1990; Yellen et al., 1991) and toxins such as C T X (MacKinnon and Miller, 1989; MacKinnon et al., 1990), and a-DTX (Stocker et al., 1991) acting at extra- or intra-cellular sites. Generally, the mutations involve negatively charged (e.g., D or aspartic acid) or hydroxyl-containing (e.g., Y or tyrosine) amino acids to nonpolar (K or valine) or charged (L or lysine) amino groups. For example, the external binding site for T E A ions is well characterized and it involves a single amino acid substitution from tyrosine to valine. In fact, further evidence suggests that T E A interacts simultaneous with four tyrosine residues contributed by each of the channel subunits (Christie et al., 1990). Inactivation. Inactivating domains have been characterized within the amino terminus, the pore itself and segment S6 (Pongs, 1992; Hoshi and Zagota, 1993). Several modes of inactivation have been identified in K v channels: N-type (for N-terminal region; Hoshi et a l , 1990), C- (for C-terminal region; Hoshi et al., 1991), and P- (for "P" region) inactivation (Jan and Jan, 1992; Pongs, 1992; Aldrich, 1994). N-type inactivation has 33 been the most studied simply from the observation that Shaker channels that differ greatly in their rates of inactivation differ structurally only near the N-terminus. Deletions of the first 20-40 amino acids nearest the N-terminus eliminated inactivation and these observations were consistent with the "ball and chain" model proposed for the Na + channel by Bezanilla and Armstrong, (1977). The native complex of voltage-gated K + channels is also composed of a hydrophilic P-subunit, which probably binds to and plugs the internal mouth of the a-subunit pore structure (Pongs, 1992; Isom et al., 1994; Adelman, 1995). In response to positive potential changes, the currents produced by the cardiac cell are carried by ions moving down their respective electrochemical gradients via their respective conducting channels. The selectivity for certain ions is reflected in differences of the channel protein and/or the channel pore at a molecular level. Recognizing these differences may assist in the structure-activity relationship analysis profiled in this work for novel ion channel blockers. 1.4. Molecular basis of class I and III antiarrhythmic binding interaction In the field of antiarrhythmic agents, classical two-dimensional quantitative structure-activity relationship (QSAR) analyses and traditional structure-activity relationship (SAR) studies have been mainly described so far. Class I and III antiarrhythmic agents will be discussed together with a brief overview of their modes of actions at a molecular level. 34 1.4.1. Class I antiarrhythmics The structural requirements of class I antiarrhythmics and their interactions with cardiac sodium channels are not well defined. Early comparisons of antiarrhythmic drug binding interactions with the cardiac sodium channel involved in vivo and in vitro kinetic measurements of drug binding and unbinding and led to the formulation of models such as the modulated receptor model (Hondeghem and Katzung, 1977; Hille, 1977). Models such as this were an early attempt to describe the molecular mechanism of antiarrhythmic-channel interaction. The modulated receptor model implies that class I antiarrhythmics bind to a specific common receptor site in cardiac sodium channels and that affinity to the receptor site(s) is modulated by the channel state. Clarkson and Hondeghem (1985) showed competitive displacement of one drug by another using lidocaine, bupivacaine or quidinine, thus providing experimental evidence for a specific receptor site. Biochemical evidence in support of a specific binding site for class I antiarrhythmics proposed by such models as the modulated receptor hypothesis, which was based on electrophysiological data, was provided by Sheldon and his group (1987). Using radioligand binding assay, a [ 3 H]BTXB binding site in rat cardiac myocytes was identified and that binding of antiarrhythmic drugs to this site is saturable, reversible, and sterospecific (Sheldon et al., 1987). The structural determinants (see below) of local anesthetic-class I antiarrhythmics consists of an aromatic lipophilic moiety connected via a linking alkyl chain to a hydrophilic (2° or 3°) amine group (Adamson and Bush, 1976; Strother et a l , 1977; Bokesch et al., 1986; Sheldon et al., 1991). Studies using stereoisomers of these antiarrhythmics revealed that the orientations of both the aromatic 35 and amine groups to the rest of the molecule are important determinants of drug binding (Hill et al., 1988; Valenzuela et al., 1995). Furthermore, the binding affinity of local anesthestic-type blockers remained unchanged or increased with chemical substitutions (see below) to the aromatic ring (X) or the linking chain (R 3 ' , up to six carbons), respectively (Sheldon et al., 1994). The precise requirement for an ether, ester or amide moiety between the alkyl link and the aryl ring is uncertain (Courtney, 1987). However, using closely related homologues of lidocaine, it was determined that the optimal link length was two carbons (R,') (Sheldon et al., 1991). Recent molecular studies by L i and his group (1999) have proposed critical binding interactions between local anesthetics and phenylalanine amino acid at position 1710 (F1710) in the transmembrane segment IVS6 of the sodium a-subunit. The proposed model describes stabilization of local anesthetic drug binding to open or inactivated channel states by either cationic-Tc or aromatic-aromatic interactions between the aromatic side chain or the amino acid and charged or aromatic moieties on the drug molecule (Li etal., 1999). Structural differences among class Ia-c agents on a molecular level have also been investigated. Conformational analysis by Marrer (1989) of quinidine and lidocaine (in protonated forms) as representative class la and lb agents, respectively, and modeling their binding interactions led to results consistent with the modulated receptor hypothesis. LIDOCAINE * CHIRAL AMINE GROUP GROUP LINKING CHAIN 36 In the open/activated state, the class la agent binds to the receptor with two hydrogen bonds involving the ammonium hydrogen and an acceptor group (in the case of quinidine, the hydroxyl group) of the ligand. In the case of lidocaine, such interactions are only possible with an energetically less favorable conformer; in the minimum energy conformer, lidocaine is able to form only one hydrogen bond. In another study (Glowka et al., 1991), crystal structures and conformational spaces of pyrrolidinylmethylphenols with antiarrhythmic activity were analyzed in comparison with representatives of all three class I subclasses. These results indicated three possible recognition sites for the ligand - two polar groups (a nitrogen and a carbonyl oxygen) and a hydrophobic group interact with the receptor. Distances between these sites are characteristic for each subclass. In particular, the N. . .0 distance has been found to be significantly shorter in class Ib agents than in class la and Ic agents (summarized in Glowka et al., 1991). Moreover, in the same molecules, distances from the amino nitrogen to the hydrophobic center (aryl-carbonyl groups) follow the same trend (i.e., shorter for class Ib versus la and Ic). The separation between the recognition points (aryl-carbonyl groups and the amino nitrogen) has been proposed to affect the time constant for recovery from sodium channel blockade (Glowka et al., 1991). Molecular size, as measured by the distance between the recognition points, provides support for Courtney's size-limiting hypothesis (1988, 1990) that the rate of recovery from sodium channel block is proportional to the end-on dimensions of the agent. A receptor that binds all of these agents must undergo structural changes so that the separation between the recognition points can vary. The simultaneous binding of both points would require that the receptor have a series of intermediate states and that in 37 one particular state the separation between recognition points matches the geometry of the drug. In a more recent theoretical study of lidocaine, tocainide and mexiletine, principally taken as the conformations of their salts found in the solid state, a two-center "zipper" model has been proposed (Remko et al., 1994) based on the original proposal by Burgen et al. (1975) for flexible ligands binding to macromolecules. In the first recognition step the protonated amine associates with a carboxyl group of the membrane protein through a strong hydrogen bond. In the next step, eventually with a rearrangement, the second polar group of the drug molecule (e.g., an ether or carbonyl oxygen) forms another hydrogen bond with an ionized amine group of the membrane. Interactions of polar groups of the drug with the receptor sites were calculated by quantum chemical calculations (Remko et al., 1994). In summary, the main feature of local anesthetic-type class I antiarrhythmics that has been identified as being essential for their activity as sodium channel blockers includes the two recognition points of the molecule: the hydrophilic amino nitrogen and the hydrophobic chain. It should be noted that the binding interaction of conventional class I antiarrhythmics as described above takes into account a common and definable receptor site. Novel sodium channel blockers with chemical structures that are unconventional or do not visibly conform to the generic local anesthetic pharmacophore may encounter binding interactions to an unrelated receptor site(s) that may be influenced by molecular factors, such as molecular conformation, molecular size, lipophilicity and pH. 38 1.4.2. Class III antiarrhythmics The mode of action of agents with class III properties is poorly understood at the single channel level. Two major problems are associated with investigation of drug-potassium channel interactions: the diversity of cardiac potassium channels and the presence of many overlapping currents in cardiac tissue. Recently many potassium channels have been cloned and stably expressed in various systems, including mammalian cell lines (reviewed in Cook, 1990; Catterall, 1996). Although this makes it possible to analyze drug-protein interactions directly, effective incorporation of the results into the drug development process also requires identification of the physiological relevance and full pharmacological characterization of the expressed channel. A convenient classification has been made on a molecular structural basis, and potassium channels have been divided into three main families. These are: 1) the Shaker voltage-gated archetype family (Kvl-4; Papazian et al., 1987), 2) inwardly rectifying K + channels (K I R; Ho et al , 1993; Kubo et al., 1993; Perier et al., 1994) and 3) channels not related to the Kv superfamily (IsK; Takumi et al., 1988). The Shaker-related delayed rectifier and inwardly rectifying channel proteins form tetramers, each containing six and two transmembrane segments, respectively, with a common K+-selective filter region. The third family is represented by the slowly activating channels that contain a single transmembrane segment. Few SAR studies have been performed on cloned channels but the structural requirements for blockade of the human cardiac hKvl.5 channel have been investigated. Properties of this channel closely resemble a rapidly activating delayed rectifier in the heart. Quinidine, quinine (a diastereomer of quinidine), clofilium and 39 tetrapentylammonium have all been shown to bind to this channel with high affinity, but no block has been observed with methanesulfonamide-type antiarrhythmics at therapeutic doses. Analysis of drug action indicated that the former type of compounds act as cationic channel blockers and bind to a common receptor site. Hydrophobic interactions have been found primarily responsible for the stabilization of the drug-receptor complex (Snyders and Yeola, 1995). 1.5. Acute myocardial ischemia and cardiac ion channels Functional characteristics of ion channels are known to be highly dependent on the extra- and infra-cellular ionic milieu. Alterations of membrane currents as a consequence of these ionic changes underlie alterations in excitability, abnormal automaticity, refractoriness and conduction in cardiac tissue. These characteristic changes of electrophysiologic properties culminate in loss of excitability and failure of impulse propagation and, as such, form the precursors necessary for cardiac arrhythmias. The cellular changes that occur during acute myocardial ischemia will be described in the following sections with emphasis on IN a and IT O current modulation as a function of changes in extra-cellular pH and [K+]. 1.5.1. Metabolic changes associated with acute myocardial ischemia Myocardial ischemia is defined as a condition in which arterial perfusion is inadequate to meet energy needs of the cells, leading to biochemical mechanisms that alter ionic homeostasis and ultimately leads to cell death. The acute phase of myocardial ischemia refers to events occurring within the first hour after the sudden onset of partial 40 or complete reduction of blood flow through a coronary artery. "No-flow" versus "low-flow" conditions results in "anoxia" versus "hypoxia" both of which influence the time-course and magnitude of metabolic and ionic changes and, ultimately, the extent of alterations in electrophysiological properties and impulse propagation. Regardless, any deprivation of blood flow causes alterations in cation homeostasis, an increase in extra-cellular K + and H + concentrations, induction of hypoxia, production of high P C 0 2 , depletion of substrates, and accumulation of substances such as inorganic phosphate, phospholipid metabolites (e.g., lysophosphoglycerides) and catecholamines (reviewed in Cascio etal , 1995). The following brief discussion is confined to conditions of no-flow ischemia. Within the early minutes of ischemia, cardiac aerobic metabolism ceases and shifts to anaerobic metabolism resulting in an increase in intra-cellular lactate levels (Braasch et al., 1968; Opie, 1976; Cross et al., 1995). This is accompanied by a decrease in intra-cellular pH (pHj) by ~ 0.5-0.8 units (Garlick et al., 1979). A n intra-cellular pH of 6.8 after 10 minutes of ischemia has been reported (Nattel et al., 1981; Kingsley et al., 1991). As a result of diffusion of lactate from myocardial cells into the intercellular space and activation of the Na+-proton pump antiporter, a decrease in extra-cellular pH (pH0) also develops. As the time in ischemia increases, production of high-energy phosphates decreases and their supply is exceeded by demand. Creatine phosphate donates its high-energy group to ADP and helps maintain ATP levels. When cellular ATP levels have dropped below normal, the energy-dependent maintenance of transmembrane ionic gradients becomes impaired (Steenbergen et al., 1985). This is due in part to a decrease in the activity of the ATP-dependent Na + , K + 41 pump and in part to an increase in Na + uptake. The increased Na + uptake results from the intra-cellular FT load described above being exchanged for extra-cellular Na + by the Na +-Ff antiporter (Tani and Neely, 1989). Inhibition of the Na +, K + pump also results in reduction of intra-cellular K + and, from an electrophysiologic point of view, the most important consequence of this is an accumulation of K + in extra-cellular spaces and the consequent attendant depolarization of nearby cells. Differences in K + concentrations at the interface between normal and ischemic myocardial tissue may produce local differences in the resting membrane potential (Kleber, 1983) and, potentially, affect AP shape and duration. Other ionic changes that occur include an increase in intra-cellular Ca 2 + (Steenbergen et al , 1985) which would predispose the development of afterpotentials and oscillations in resting potential. Also changes in phospholipid metabolism, due to ischemia-induced activation of phospholipase A 2 (Sobel et al., 1978), lead to an increase in tissue levels of lysophospatidylcholine (LPC), the result of which has been shown to have deleterious electrophysiological effects leading to arrhythmogenesis (Corr et al., 1981). Finally, acute myocardial ischemia both facilitates and causes the release of noradrenaline from adrenergic nerves. Evidence to support the role of the sympathetic nervous system has included use of drugs to modify the system or surgical denervation (see review by Schomig A, 1995). Work from our laboratory has shown that the sympathetic system, at least in rats, does not play a direct role in the genesis of acute ischemia-induced arrhythmias (Guo et al., 1999; Palleta et al., 1989; Curtis et al., 1985). 42 The above plethora of ionic changes leads to the following cardiac electrical changes: decreases in amplitude and upstroke velocity, reduction in membrane excitability, shortening of APD, and prolongation of recovery of excitability following an action potential (Gettes et al., 1992a, 1992b; Wit and Janse, 1992; Cascio et al., 1995). Of the ionic changes, the major pathophysiological components of acute myocardial ischemia are elevated extra-cellular potassium ([K+]0) and [H+] 0 and these will be described in detail in relation to cardiac channel function, namely IN a and I T O . 1.5.2. Influences of intra- and extra-cellular pH on cardiac ion channels. Extra- and intra-cellular acidosis is an early consequence of ischemia. In multicellular preparations and isolated myocytes studied under experimental conditions of reduced pH s, action potential duration has been shown to be increased (Boyett et al., 1988; Rozanski and Witt, 1991) or decreased (Kurachi, 1982), whereas resting membrane potential has been shown to be unchanged or reduced (Bielen et al., 1990). Changes in pH may induce complex electrical changes on the different ion channel proteins that in turn change the magnitude and kinetic properties of the ionic currents that control the shape of the action potential. The mechanisms of proton modulation of cardiac ion channels are not fully known and are likely to vary from one class of channel protein to another. In general, however, it has been proposed that ion channel regulation occurs through protonation of specific amino acid residues on the channel protein or fixed negative surface changes near the channel (Irisawa and Sato, 1986; Ito et al., 1992). 43 1.5.2.1. Inward sodium current (INa) The effects of protons on IN a in cardiac muscle have been investigated primarily by lowering the pH of the medium, externally (Chesnais, et al., 1975; Kohlhardt, et al., 1976; Vogel and Sperelakis, 1977; Brown, et al., 1978; Yatani, et al., 1984), internally (Kurachi, 1982), or both (Watson and Gold, 1995). Intracellular acidification is an early and primary component of the response to ischemia with extra-cellular acidosis occurring as a secondary condition due to p l i regulatory mechanisms transporting FT into the extra-cellular space (see Section 1.5.1.). In accord with this, the work of Watson and Gold (1995) was the first systematic investigation that looked at the combined effects of intra-and extra-cellular acidosis on IN a in neonatal ventricular myocytes. Lowering pH 0 independent of pH; was found to slow the activation kinetics of IN a and shift its activation in the depolarizing direction (Watson and Gold, 1995). The reduction of peak IN a with extra-cellular, but not intra-cellular, acidosis at more depolarized test potentials is consistent with the hypothesis that the action of H + ions on IN a amplitude is due to voltage-dependent block of open Na + channels (Woodhull, 1973; Watson and Gold, 1994). In contrast, intra-cellular acidosis alone produced a hyperpolarizing shift in steady-state inactivation presumably via intra-cellular Na + loading through activation of the Na7H + exchanger. Changes in the kinetics of inactivation with acidosis suggests that a prolonged IN a current under ischemic conditions would have important implications for open-channel blockers of IN a whose effect may be accentuated by slowed inactivation. These effects of ischemia-induced acidosis on IN a are, in part, expected to result in a decrease conduction velocity and thereby predispose cardiac tissue to reentrant arrhythmias. In agreement with such findings, upstroke velocity of the cardiac action 44 potential is reduced in the presence of raised extra-cellular [FT] (Grant et al., 1980; Nattel et al., 1981; Moorman et al., 1986). A common observation for sodium channel blockers in acidotic media is a slowing of their dissociation from sodium channels (Betancourt and Dresel, 1979, Grant et al., 1980; Moorman et al., 1986; Wendt et al., 1993), and this is observed as potentiation of I N a inhibition. Although some effect(s) of pH 0 on drug action may be due to protonation (ionization) of the drug, protonation of specific channel sites involved in drug binding may also make an important contribution (Yeh and Tanguy, 1991; Backx and Tue, 1991; Zhang and Siegelbaum, 1991). The significance of extra-cellular [H+] and its influence on modulating sodium channels or the blockers involved wil l be discussed in later sections. 1.5.2.2. Transient outward potassium current (Ixo) Information available in the literature regarding effects of intra- and extra-cellular pH on K + currents is inconsistent and quite variable. In rat myocytes, exposure to acidic extra-cellular solution (pH 6.0) either increases, decreases, or has no effect on the Ca 2 + -independent component of I x o , depending on both the holding and test potentials (Stengl et al., 1998). When the holding potential was set to -80 mV, acidosis suppressed the current at most test potentials except very positive ones (> +60 mV), where acidosis did not elicit any effect. In contrast, when the holding potential was less negative (-30 mV), the I T 0 current increased. From these experiments, it was concluded that acid-induced rightward shifts of steady-state activation and inactivation of I T O were responsible for the observed effects (Stengl M et al., 1998). However, preconditioning of cells with external 45 acidic solutions results in the internal pH changing accordingly, i.e., being reduced from control, prior to establishing whole cell recording (Wallert and Frohlich, 1989). Studies by X u and Rozanski (1997) have shown in rat myocytes that I T 0 channels are modulated predominantly by intra-cellular protons, a property that has also been shown for L-type Ca 2 + and IKi channels (Irisawa and Sato, 1986; Ito et al., 1992). Despite the marked reduction of I x o produced by decreased intra-cellular pH (with V H = -80 mV to +60 mV), steady-state activation and inactivation kinetics of I x o was altered but this was not statistically significant (Xu and Rozanski, 1997). Under physiological conditions, when the resting membrane potential is negative to -70 mV in non-ischemic myocardium and the action potential overshoot is less positive than +60 mV, extra-cellular protons from the border zone between normal and ischemic myocardium, should inhibit I x o . However, in ischemic tissue where the resting membrane potential is reduced (< -70 mV) acidosis will stimulate (Stengl et al., 1998) or (if resting membrane potential remains unchanged) reduce I T O (Xu and Rozanski, 1997). In summary, differences in I T 0 modulation by acidosis appear to be dependent on the proton-buffering capacity of the cell (intra- versus extra-cellular [H+]) and the resting membrane potential. Clearly, these influences of protons on I T 0 can alter action potential duration and, thus, lead to dispersion of refractoriness and arrhythmogenesis. Current K + channel blockers have been shown to have a reduced blocking activity under ischemic conditions and it is unclear whether this loss of activity is a direct function of acidosis, or other consequences of ischemia (ionic and/or cellular). 46 1.5.3. Effects associated with raised extra-cellular K + Myocardial K + balance is extremely sensitive to ischemia (Kleber, 1984; 1987). The mechanism of cellular K + loss during the initial stages of ischemia may involve the following mechanisms (Aksnes, 1992): partial inhibition of the Na7K + pump, reduction of the extra-cellular space by osmotic changes, cotransport with anions, and activation of the ATP sensitive K + channel. During acute myocardial ischemia, K + efflux from the cells at risk may occur heterogeneously (Hill and Gettes, 1980; Coronel et al., 1988). At the cellular level, the resting membrane potential is primarily determined by the ratio of the intra-cellular and extra-cellular K + concentrations, since the permeability to K + is higher than to other ions (Goldman, 1943; Hodgkin and Katz, 1949). The steady-state activation and inactivation of I N a (Hodkin and Huxley, 1952) and I x o (Josephson et al., 1984; Clark et a l , 1988; Shibata et al., 1989; Apkon and Nerbonne, 1991) and the kinetics of both these processes have been studied. By changing the K + equilibrium potential in the absence of ischemia (i.e., increasing or decreasing [K +] 0), the availability of the fast sodium channels is increased (hyperpolarization) or decreased (depolarization) (Aldeman and Palti, 1969). In other words, a reduced membrane potential shifts the activation and inactivation for I N a to more positive potentials and, as such, reduces the threshold for myocardial excitability and conduction velocity. With respect to drug-channel interactions involving sodium channel or class I-type blockers, modulation of sodium channel states, as a function membrane depolarization may increase channel availability for drug binding (e.g., increase blockade of inactivated-state channel blockers in depolarized tissue) or simply produce the opposite effect (e.g., reduce blockade of open-state channel blockers in depolarized tissue). 47 The influence of extra-cellular K + on I T O is discussed with respect to the time-course for inactivation since this parameter modulates the size and magnitude of this current and, as such, the initial phase of repolarization in mammalian hearts (Surawicz, 1992; Carmeliet et al., 1993). The time course of I T O reactivation is strongly voltage-dependent such that recovery is prolonged at less polarized potentials (e.g., < -80 mV; Shibata et al., 1989; Apkon and Nerbonne, 1991). However, Pardo and Stuhmer (1994) have demonstrated that when extra-cellular K + ([K+] 0) is reduced, the Ca2+-independent transient outward component in rat atrium was decreased (i.e., recovery from inactivation was slowed). This was unexpected given the circumstances that reducing extra-cellular K + should produce an increase in the electrochemical gradient and thereby produce any increase in current carried by K + . This phenomenon was also demonstrated in K v l . 4 channels (transient outward K + channel clone; Pardo and Stuhmer, 1994). This apparent paradox for I x o or ITO-type channels, however, was not unique amongst the K + currents. Early studies looking at the effects of external K + on repolarization found a shortening of cardiac Purkinje APD by increasing [K + ] 0 (Brooks et al., 1955; Hoffman and Cranefield, 1960). In high [K + ] 0 , there was depolarization of the resting potential accompanied by a reduced AP duration, whereas at low [K + ] 0 there was hyperpolarization of the membrane and a prolongation of the APD. Similar effects were recorded in ventricular muscle (Surawicz and Gettes, 1963). This is also consistent with the brevity of APD associated with activation of K + A T P channels (Imanishi et al., 1983, 1984; Steinberg et al., 1988; Bri l et al., 1990; Lathrop et al., 1990) or its attentuation with K + A T P blockers, such as glibenclamide (Kantor, 1990). 48 A plausible explanation for the increase in net repolarizing current following an increase in [K + ] 0 (as well as for the observations made by Pardo and Stuhmer, 1994) may involve a causal relationship between inward flow of ions and an acceleration of recovery from inactivation. Much of the early formulations (Noble, 1965) and experiments (McAllister and Noble, 1966; Cleeman, 1981) aimed at explaining this paradox of [K + ] 0 -induced increase in net repolarizing current were based on models that investigated the inwardly rectifying K + current since it was known that the K+-conducting system in Purkinje fibers exhibited inward-rectification (Hutter and Noble, 1960; Carmeliet, 1961; Trautwein and Kassebaum, 1961; Hall, 1963; Deck and Trautwein, 1964) and many of the important effects of extra-cellular K + were adequately explained by this current. Subsequent experiments showed that [K + ] 0 is required to open inward rectifier channels (Cleemann and Morad, 1979; Cleemann, 1981) and that the same current is suppressed during cellular K + loss (Cleemann, 1981). Thus, a reduction in the outward electrochemical gradient on K + ions by raised [K + ] 0 followed by a reduction of potassium ion permeability and, perhaps, an increase in potassium ion permeability by [K+]0-induced opening of inward K + currents (Cleemann and Morad, 1979) may be sufficient to increase net repolarizing current despite a reduction in the gradient. 1.5.4. Raised extra-cellular K + and ischemia-induced arrhythmias The rise in extra-cellular K + has been implicated to be a major factor contributing to the development of ischemia-related arrhythmias since the early coronary-ligation studies of Harris et al. (1950; 1954) in dogs. In addition, regional elevation of extra-cellular [K +] in the absence of ischemia has been shown to produce electrophysiological 49 changes which are similar to those produced by ischemia (Gettes and Surawicz, 1968; Morena, 1980; Curtis, 1991), and are liable to precipitate arrhythmias. There are, however, certain inconsistencies with regard to the etiologic role of extra-cellular [K +] alone in the development of ischemia-related arrhythmias. Downar and his group (1977) presented evidence for factors other than, or in addition to, the ischemia-induced K + rise that are associated with electrophysiological changes in relation to arrhythmogenesis. Nordrehaug and Von der Lippe (1983; 1986) have reported that, in the clinical setting of myocardial infarction, the occurrence of ventricular arrhythmias is inversely related to serum [K +]. Podrid (1990) reviews the clinical observation that cardiac arrhythmias are aggravated by hypokalemia (serum [K +] < 4 mM). Experiments from our laboratory have also shown that the frequency and severity of arrhythmias induced by coronary artery occlusion and subsequent ischemia is inversely related to serum [K +] (Curtis et al., 1985a; 1985b; 1986; Saint et a l , 1992). Curtis and Hearse demonstrated a similar relationship in isolated rat hearts perfused with varying [K +] and subjected to the same model of ischemia. It should be noted that in the above clinical and experimental studies, the range of K + concentration was 2-8 m M and there is an upper limit of 15 m M for the proposed inverse relationship. Monophasic complexes accompanied by ventricular fibrillation were produced as extra-cellular [K +] was raised to 15 m M in the perfusate of isolated porcine and rabbit hearts (Morena et al., 1980; Curtis, 1991). The relationship between serum [K +] and ischemia-induced arrhythmias is best explained on the basis of signal activity and phase variations thereof between ischemic and non-ischemic tissue. Action potentials recorded during ischemia or in ischemic 50 tissue have a much lower amplitude, lower upstroke velocity, and show alternation in amplitude and duration (Gettes and Surawicz, 1968; E l Sherif et al., 1975; Downar et al., 1977; Morena et al., 1980; Janse et al., 1980) than normal. Provided the cells are well coupled and there is minimal collateral blood supply to the ischemic site (to minimize inhomogeneous and graded ischemic:normal border zones), the signal patterns in normal and ischemic zones are out of phase and potentially arrhythmogenic (via dispersion of refractoriness and facilitation of reentry). Raising serum [K +] decreases phase disparity in signals between zones by modulating AP signals in the normal zone to closely approximate those in the ischemic zone (i.e., reduced AP shape and duration). In the upper (15 mM) and lower (< 4 mM) serum [K +] extremes, action potential signals in the adjacent zones reestablish phase disparity due to progressive conduction slowing and onset of inexcitability at high [K + ] D (Gettes et al., 1968) and increased conduction and faster rate of diastolic depolarization at low [K + ] 0 (Gettes et al., 1968). When acidosis and hypoxia are taken into account, the effects of systemic and regional hyperkalemia on excitation and conduction can not be explained entirely by their effects on resting membrane potential. Acute myocardial ischemia offers potential approaches for developing channel blockers that are selective for conditions of ischemia. One approach is to target the channel state or channels that are present in the early stages of myocardial ischemia. The other is to modulate the drug or compound species as a function of ischemia-induced increases in extra-cellular ionic components. This would localize the action of the blocker to where it is needed and by this action, selectivity for the pathological tissue is 5 1 established. A more detailed description of selectivity for ischemic myocardial tissue will be discussed in the following chapters. 1.6. Selectivity for myocardial ischemia 1.6.1. Rationale The following is a summary of the premises that have been alluded to in the earlier sections with regard to the "ineffectiveness" of currently available antiarrhythmic agents and why there should be a renewed and unconventional approach to future antiarrhythmic drug design. Selectivity for cardiac tissue and, in particular, ischemic myocardial tissue, has been inadequately explored as an avenue for antiarrhythmic drug development. Presently, there are a myriad of arrhythmia models that induce arrhythmias in atrial or ventricular tissue. Since these types of cardiac tissues vary in terms of their current type and density, activation and repolarization characteristics, the present discussion on the rationale for the selection of ischemic myocardium wil l focus on ventricular myocardial tissue. In summary, current antiarrhythmic drugs are unsatisfactory against ischemia-induced arrhythmias because of the following reasons. They 1) act on normal (non-pathological) cardiac tissue, 2) have the potential for causing arrhythmias by overexpression of the same mechanisms by which they produce their antiarrhythmic actions, 3) produce extra-cardiac effects (e.g., neurotoxicity), and 4) have been shown to be ineffective in reducing human mortality in survival studies. As previously mentioned, sodium channel blockade alone is inadequate as an antiarrhythmic mechanism due to adverse side-effects. Clinical evidence (e.g., CAST, CAST II, SWORD; see sections 52 1.1.1. and 1.1.2.) has unequivocally shown that monotherapy involving sodium channel blockers against ischemia-related arrhythmias fails to improve survival rate compared with combinational antiarrhythmic therapy, either with two agents of the same class or with one possessing multiple actions (e.g., quinidine) (Duff et al., 1991; Marcus, 1992). Current agents, e.g., d,l-sotalol, amiodarone, that prolong APD via "class III" actions have additional blocking properties (e.g., p-blockade; see section 1.1.2.) and their superiority to their specific ion channel blocker predecessors is bolstered by recent controlled clinical trials (Mason et al., 1993; C A S C A D E Investigators, 1993). This recent clinical evidence supports a lack of antiarrhythmic activity and a prevention of mortality due to sudden cardiac death by virtue of reductions in repolarizing currents as a sole antiarrhythmic mechanism (Singh et al., 1990; Vanoli et al., 1995; see review by Singh, 1998). The focus in examining drug actions has also been on the normal cardiac action potential in the hope that its modulation should establish higher thresholds against aberrant signal generation. Although cardiac action potential modulation is an important antiarrhythmic determinant, this action must be localized i f it is to reduce inhomogeneous signal dispersion between normal and diseased myocardial zones. The extra-cellular milieu that precipitates, and is associated with acute myocardial ischemia, offers targets for drug action that may be separate from normal myocardium. The ionic extra-cellular milieus of acutely ischemic myocardium, such as raised H + and K + ions, offer a pathophysiological approach, such that drug selectivity for ischemic myocardium may be described as a function of the extra-cellular ionic conditions of myocardial ischemia rather than some affinity for ischemia-induced channel 53 states. Localizing drug action to ischemic tissue may reduce APD phase disparity between normal and ischemic zones. As previously mentioned, lidocaine is a less commonly used antiarrhythmic that has been experimentally shown to selectively depresses cardiac excitability in ischemic tissue or in conditions that mimic myocardial ischemia (for a review, see Carson et al., 1986). This selective action is primarily due to preferential block of sodium channels in their inactivated state (due to [K+]0-induced reduction in membrane potential). The resultant electrophysiological effect of lidocaine, due to a reduction in sodium channel availability in ischemic tissue, is a reduction in conduction velocity with minimal effects in normal tissue. Effects on A P D may also be shortened by lidocaine (Davis and Temte, 1969; Bigger and Mandel, 1970) due to blockade of the TTX-sensitive inward sodium channel during the plateau phase (Colatsky, 1982; Carmeliet and Saikawa, 1982), the result of which may add to the limited antiarrhythmic activity of lidocaine (see below). In the absence of sodium channel blockers, the conditions of ischemia depress conduction and excitability (see above). However, in the presence of agents such as lidocaine, the setting can be described as a drug-induced progression of ischemia, whereby the "window" of critical depression that leads to the onset of arrhythmias is shifted earlier in time after the onset of ischemia (Carson et al., 1986; Hondeghem, 1987). In order to render the ischemic tissue inexcitable, higher concentrations of lidocaine would be required, but these may be associated with extracardiac side-effects. Clearly, sodium channel block and prolongation of action potential duration (e.g., reduced g ^ by themselves are inadequate antiarrhythmic interventions. It is also difficult to confirm or refute the effectiveness of mixed sodium and potassium channel blockers (e.g., quinidine) 54 against ischemia-related arrhythmias when their actions in normal and ischemic zones decrease homogeneous signal transmission in the heart. Thus, the rationale for developing a blocker that is selective for ischemic myocardium may be more desirable, particularly one that exhibited sodium and potassium channel blocking properties since conduction velocity and effective refractory periods determine reentrant arrhythmias. Minimizing the effective drug concentration in ischemic tissue will also reduce extra-cardiac side-effects. The potential of such selectivity for ischemic myocardium is a higher therapeutic margin between doses for effective antiarrhythmic protection and doses for drug-induced toxicity. Therefore, a focus on the properties of a compound/agent that are specific for conditions of ischemic myocardium should accomplish the required antiarrhythmic selectivity independent of channel types and their states but rather on the environment of the ischemic myocardium. 1.6.2. Role of molecular pKa and selectivity for myocardial ischemia The two major ischemia-related ionic constituents, raised [K + ] 0 and [H +] 0, have different influences on antiarrhythmic drug actions. As mentioned above, [K+]0-induced depolarization is the result of a reduction of the transmembrane K + gradient and this has indirect actions on drug-channel interactions. For example, the reduced membrane potential affects voltage-sensitive sodium channels by shifting activation and inactivation. Open-state sodium channel blockers, for example, would be less effective since fewer channels would be available for activation and the majority of them would be in the inactivated forms. The depolarized myocardium, however, favors lidocaine-like agents or inactivated-state blockers. The literature and the existence of state-dependent 55 K + channel blockers appears limited. Attempts to modify the chemical structure(s) or components of a compound for the purpose of achieving blockade of I N a or I K in a preferred channel state while minimizing the compound's activity in normal myocardium appear complex and it may be accompanied by a multitude of variables. In contrast, the influence of extra-cellular H + has a direct action on the blocking drug or compound. Drug ionization, as determined by a molecule's pKa, serves to regulate the amount of charged versus uncharged drug forms. Differences in drug action between charged and uncharged drug forms have been well studied, particularly for sodium channel blockers (Frazier et al., 1970; Strichartz, 1973; Hille, 1977). Since the 1970's, it has been well known that there is a potency difference between charged and uncharged sodium channel blockers. Those who studied the local anaesthetic actions of QX (e.g., QX514) or quaternary (3° or 4° amines) compounds (Frazier et al., 1970; Strichartz, 1973; Hille, 1977) observed that the charged form of these compounds reduced sodium current much more than the uncharged form. The apparent potentiation of I N a block by these agents in acidosis is explained by a rate-limiting step of drug dissociation from blocked channels (Narahashi and Frazier, 1971; Hille, 1977; Schwarz et al., 1977). Since these agents bind to an intra-cellular site, the route(s) of egress of the charged drug form is via a hydrophilic (open channel) or hydrophobic membrane route (only after deprotonation of tertiary nitrogen-containing agents) (Narahashi and Frazier, 1971; Hille, 1977; Schwarz et al., 1977). Membrane diffusion of hydrophobic and quaternized compounds also occurs but at much slower rates (e.g., the rate of action for QX-572 is > 200 sec versus 1 sec for benzocaine; Frazier et a l , 1970; Hille, 1977). Thus, the rate-limiting step is the dissociation of the ionized drug form from an internal binding 56 site during local acidosis, since there is an increased protonation of the receptor-bound neutral drug and a subsequent increase in the fraction of recovery from block that occurs by the slow hydrophilic pathway. The direct action of H + ions on sodium channel blockers, and the ensuing potentiation of I N a blockade, provide a strategy for developing agents that are selective for ischemic myocardium. The following premises outline the chemical requirements in association with extra-cellular [H+] for achieving drug selectivity for ischemic myocardium: 1) drug distribution between ischemic and normal zones is a function drug pKa, 2) the lower the pKa the greater the distribution of the charged species in extra-cellular acid conditions, 3) ion channel blockade is potentiated for the charged species and, 4) to accommodate the influences of external pH, an external binding site is favored. These premises are not intended to reproduce or improve upon the existing pH-dependent sodium channel blockers. The purpose is to utilize the ionization properties of a compound, in association with the extra-cellular rise of H + ions that follows the onset of myocardial ischemia, as an inherent mechanism for drug distribution between ischemic and normal zones. When the charged species of the compound is localized in high concentrations in the ischemic zone, it is envisioned that the mixed ion channel blockade (preferably sodium and potassium) will prevent ischemia-induced APD shortening. Presumably, this preventative action would involve an increase in refractoriness or an increase in the threshold for tissue excitability, and thus remove the ischemic zone as a source of arrhythmic interaction. 57 1.7 Objectives RSD 1000 has previously been shown to produce blocking actions on I N a and I T 0 currents and this is believed to play a part in its antiarrhythmic actions against ischemia-induced arrhythmias in rats (Yong et al., 1999). In this thesis, a series of 15 analogues chemically and structurally related to RSD 1000 (see Figure 1 and Tables 2 and 3) was investigated to characterize and establish a structure-activity relationships associated with acute ischemic arrhythmias. The main objective of this thesis was to identify the chemical components and/or attributes involved in antiarrhythmic actions against ischemia-induced arrhythmias. The experiments varied from the level of the whole animal to the study of whole currents in isolated cells. The results suggested that certain chemical components within the RSD series of amino-2-cyclohexyl ester compounds are required for antiarrhythmic actions. The major hypothesis for this thesis was that the degree of ionization or protonation of the heterocyclic nitrogen atom plus the nature of the aromatic substituent produce a chemical entity that reduces arrhythmias most favorably (i.e., increased selectivity accompanied by minimal toxicity) when such arrhythmias are induced by early myocardial ischemia. The procedure used for investigating the hypothesis is outlined in Figure 3. The hypothesis was further subdivided into the following questions with respect to the series of amino-2-cyclohexyl ester compounds: 1. Is the reduction of ischemia-related arrhythmias a function of their mixed ion channel blockade? 2. What is the role of the aromatic substituent in determining antiarrhythmic activity? 3. What role does pKa play in controlling antiarrhythmic activity? 58 Table 1 Chemical Structure o Q O or Q o o a m , o am, o o a p e ® o ornor o 0) c o p e c o n e o a RSD formate acetate 1053 1010 1050 1051 1012 1072 1014 1019 1000 1009 1015 1025 1046 1049 General Pharmacophore ,0 . ,R2 "'N I Ri Y o 59 Table 1 A list of chemical structures for all 16 RSD compounds used in the study. Variations were made to the aryl (Ri) and amine heterocyclic (R 2) groups as denoted in the general pharmacophore. Figure 3 RSD1000 K JL ^ Chain length Phenyl-substitutions 1 -naphthalene + pKa 61 Figure 3 An overview of the chemical changes to R\ and R 2 of the general RSD pharmacophore. This systematically investigates the requirements of an aryl group, extension of the phenyl group, aromatic substitutions, positional arrangement of the naphthalene group, and pKa. 62 2. Materials and Methods The Animal Care Committee of the University of British Columbia approved all experiments and protocols. Experiments were performed using male Sprague Dawley rats weighing between 200-300 gm. The use of the rat as an appropriate animal model for the current studies is supported by the extensive data-base for this species and internal consistency of findings in this laboratory (see Johnston et al., 1983; Curtis et al., 1985c; Curtis 1987; Cheung et al., 1993). Clearly, all species have their perceived advantages and disadvantages. For the rat, these have been extensively addressed and justified (Curtis et al., 1987). Although the clinical significance of rat models and their validity in studying diseases found in man has been questioned (Fozzard, 1975), their continued role in studying cardiac arrhythmias is a testament to their utility and attests to the increasing perception that no single animal model other than human yet exists that is capable of predicting the clinical effectiveness of a drug. The reproducibility and precision of the rat models (see review by Curtis, 1998) may outweigh any disadvantages. 2.1. In Vivo Studies 2.1.1. Ischemia-induced arrhythmias Rats were anaesthetized with 60 mg/kg pentobarbital, i.p.. A tracheostomy tube (No. 14 Jelco IV catheter) was implanted and cannulation (PE-50; Becton-Dickinson) of the carotid artery and jugular vein was then performed. Body temperature was recorded with a rectal thermometer (Becton-Dickinson) and maintained at 36-38°C using a heating lamp. Under positive pressure respiration using a Palmer pump (stroke volume of 10 ml/kg at 60 cycles/min) thoracic incisions were made. Following separation of the 2 n d 63 and 3rd intercostal ribs, the heart was cradled in a pericardial sac so that a polypropylene thread (5-0, Ethicon 8720H), which was inserted into a polyethylene guide (PE-10; Becton-Dickinson), could be loosely placed around the left coronary artery (analogous to LAD in other species) as previously described (Yong et al., 1999). Rats were allowed 30 min to recover from surgery prior to random and blind drug treatment. An arterial blood serum potassium level was measured before the start of each experiment since it has been determined that arrhythmia incidence is inversely proportional to serum concentration levels (Curtis etal., 1986). A Harvard Syringe pump (model 55-2222) was used to administer vehicle/drug treatment via the venous line (using 1 or 5 cc syringes and 23G needles, Becton-Dickinson). Arterial blood samples (~0.1-0.2 ml) were taken before and after coronary artery occlusion for determination of serum potassium concentrations using a potassium ion-selective electrode (Ionetics Potassium Analyzer, Ionetics). Vehicle/drug was infused for 5 min preceding occlusion of the LAD and the infusion was maintained. A lead II configured ECG was recorded using silver-wire needle electrodes (Grass, E5SH) together with blood pressure recorded from the carotid cannula via a pressure transducer (Gould PE23ID). Paper records of the electrocardiogram and blood pressure were taken on a Grass Polygraph (model 79D) at a standard chart speed of 100 mm/sec accompanied by an oscilloscope monitor (M/Honeywell, PM-2A). The incidence and duration of arrhythmias were recorded in the 15 min post-occlusion period. At the completion of the experiment, hearts were excised and perfused by the Langendorff technique with PIPES (see section 2.2.1.2.) followed by PIPES containing cardiogreen dye (0.2 mg/L Fast 64 Green FCF) to reveal underperfused tissue in the occluded zone. The atrial appendages were removed and the underperfused ventricular tissue (occluded zone = OZ) was cut from the perfused, color-dyed tissue and OZ size was calculated as the percentage of total ventricular weight. 2.1.1.1. Exclusion criteria Prepared animals that did not comply with the following criteria were excluded from the study. Verification that an animal was inadequate for the study was made during two critical periods after surgical preparation: pre- and post-occlusion. The basis for exclusion during the pre-occlusion period include: 1) spontaneous arrhythmias (PVCs, VTs, and/or VFs) immediately following surgical preparation and up to 30 min recovery prior to the start of the experiment, 2) abnormally low mean blood pressure (from pre-surgical levels, e.g., < 85 mmHg) either due to excessive surgical trauma and/or operator error (e.g., accidental injury to aortic branch during the occlusion procedure), 3) serum potassium concentrations found to be above or below the accepted range of 2.5-4.5 mM for this species (Curtis et al., 1986). Following coronary occlusion, the exclusion criteria were: 1) arrhythmias occurring during drug-infusion, 2) an OZ size (in this species) less than 2 5% or greater than 4 5 % (Johnston et al., 1983), 3) a precipitous fall in blood pressure immediately following coronary occlusion (usually 1 - 2 min), and a post-mortem examination revealing a defective or improper ligature in the form of 4) a loose ligature, 5) overtightening, and as a result, hemorrhaging from the tissue involved, 6) accidental occlusion of the aortic branch or, 7) if the thorax cavity was found to contain blood as a 65 result of surgical intervention (the latter two conditions reduce cardiac output). Replacement of defective animals was performed immediately without disruption to the blinding code of the study. 2.1.1.2. Classification of ischemia-induced arrhythmias and the arrhythmia score (AS) The arrhythmia incidence for each animal was summarized from chart recordings and given a single cardinal score, the arrhythmia score (AS), as a means to grade the severity of the arrhythmic history of the animal. Such scoring systems are an attempt to quantify and facilitate analysis of cardiac arrhythmias. A score was assigned to each animal according to a Gaussian distributed scale that takes into account the occurrence, severity and duration of arrhythmias (Curtis and Walker, 1988). Table 2 is reproduced from Curtis and Walker (1988) and summarizes the AS used in this study. Each AS value corresponds to an arrhythmia type and additional information describing its incidence according to the guidelines established by the Lambeth Conventions (Walker et al., 1988). See Figure 5 for representative traces that show ischemia-induced arrhythmias following the onset of coronary artery ligation. Table 2: Arrhythmia Score (AS) 0 = 0 - 49 premature ventricular contractions (PVCs) 1 = 50 - 499 PVCs 2 = > 499 PVCs and/or 1 episode of spontaneously reverting VT or VF 3 = > 1 episode of VT or VF or both (<60 sec total combined duration) 4 = VT or VF or both (60-119 sec total combined duration) 5 = VT or VF or both (> 119 sec total combined duration) 6 = fatal VF starting at > 15 min after occlusion 66 7 = fatal V F starting at between 4 min and 14 min 59 sec after occlusion 8 = fatal V F starting at between 1 min and 3 min 59 sec after occlusion 9 = fatal V F starting at < 1 min after occlusion Legend: VPB = ventricular premature beats; V T = ventricular tachycardia; V F = ventricular fibrillation 2.1.1.3. Antiarrhythmic analysis The following is a detailed explanation for Figures 3-6, 11, and 12, showing the responses against ischemia-induced arrhythmias for the listed RSD compounds. These graphs are expressed as the AS for individual animals as a function of dose. At each dose, a single animal with its respective AS is represented as an open circle symbol, whereas the mean AS for each dose group is represented by an open square symbol ± SEM. Two baseline values were defined as the occurrence of arrhythmias in control animals. Vehicle-treated animals for each experiment are illustrated as filled square symbols ± SEM. The solid black thin bar across the graph represents the mean AS value for 606 control rats. These are the total numbers of control rats that have been subjected to the same coronary-ligation in over four years. This "population" mean for AS value is bordered by 95% confidence limits for both S E M and SD illustrated as dashed lines, (••••) and ( ), respectively. This allows comparison of the mean AS of the experimental group as well as values for individual animals at each dose. Dose-response curves were fit by using nonlinear square fitting programs (SlideWrite Plus® ver. 4.0, SigmaPlot® ver. 4.0) according to the logistic equation: y = a/(l+(x/x0)b) (1) where y = AS, a = ym a x (i.e., > 7), b = Hi l l coefficient, x = dose, x 0 = EC50 (in umol/kg/min). The upper two rows of numbers at the top of Figures 3-6, 11, and 12 67 indicate the number of animals reported to have episodes of V T or V F (numerator) over the total number of animals (denominator) used at that dose. These were included as an indicator of the reliability of the AS data. Conventionally, a dose-response curve, in which the dose (or concentration) is expressed on a logarithmic scale, is defined by its potency (EC50), efficacy (as defined by the maximum response) and slope (the variance of drug sensitivity). When comparing antiarrhythmic dose-response curves in this study, the same criteria apply except that efficacy is defined as AS = 0 or 100% suppression of ischemia-induced arrhythmias, slope (Hill coefficient, |b|) > 1 and potency for antiarrhythmic protection is approximately one order (on a log scale) less than EC50 doses for producing cardiac and cardiovascular responses. Antiarrhythmic dose-responses fulfilling these criteria and the corresponding RSD compound involved will be described as exhibiting antiarrhythmic "goodness ". 2.1.2. Ventricular arrhythmias induced by electrical stimulation Induction of ventricular arrhythmias by electrical stimulation was assessed in anaesthetized rats according to the method of Walker and Beatch (1988). This model was used as an indication of the ion channel blocking properties of the RSD compounds in non-ischemic myocardium. The method used to induce fibrillation electrically depends upon the concept that following normal excitation, not all cardiac fibers recover their excitability simulataneously (Moe et al., 1964). Mapping the activation of a single premature stimulus along with measurements of refractory periods at multiple sites has shown that the existence of nonuniform recovery of excitability is a major contributor to the occurrence of reentrant tachycardias (Allessie et al., 1973, 1976). It has been well 68 established that drugs that block IN 3 increase thresholds for capture of single beats and fibrillation (Winslow, 1984). In contrast, block of IK may not affect thresholds for capture but increase the cycle length for reentry by prolonging effective refractory period of the cardiac action potential (see review by Winslow, 1984). Following tracheotomy and cannulation of pentobarbitone-anaesthetized animals as described above, stimulating electrodes were implanted by a transthoracic route in the apical region of the left ventricle. The stimulating electrodes were made from Teflon-coated silver wire (Leico Medwire, AG3T) with the 2 mm desheathed segment of each wire inserted into the left ventricular free wall using a 27G (Becton-Dickinson) needle as a guide. Square wave stimulation (1 msec pulse duration at 7.5 Hz; Grass model SD9 stimulator) was used to determine threshold current (iT-uA) and pulse width (tT-ms) for induction of extrasystoles, the threshold current for induction of ventricular fibrillo-flutter (VFt-uA) at 50 Hz and effective refractory period (ERP-ms) at 7.5 Hz. The threshold current for capture (iT) is the minimum current required by the heart to follow pacing pulses generated by stimulation. This endpoint was easily observed on the E C G recording as an increase in E C G signal size (a regular rhythm at a faster rate) and a slight but sudden drop in BP. The endpoint for VFt was similar except that the mean BP drop was greater and approached 25 mmHg. The ventricle was continuously stimulated (1 msec pulse width, 50 Hz) and the intensity of the current gradually increased (at 10 (xA/sec) until flutter occurs. The high frequency of stimulation was used to determine VFt so as to increase the probability of an R-on-T phenomenon during the vulnerable repolarization period for reexcitation to initiate tachyarrhythmias 69 (N.B. , rat Q-T interval -60 msec at R - R intervals of 350-400 bpm; see also Beinfield and Lehr, 1956). The effective refractory period (ERP) of the ventricle was estimated using a paired pulse method. The ventricle was stimulated at a baseline frequency o f 7.5 H z accompanied by a single extra stimulus (of same current strength and pulse width) at a variable delay from the stimulus. The minimum delay that resulted in failure of the heart to capture the extra stimulus (observed as an extra signal on the E C G with an immediate but transient drop in BP) constituted the effective refractory period of the ventricle. Prior to drug infusion, each variable was measured three times every 5 min to consistent values. Animals were excluded i f iT values did maintain a minimum threshold of 80 - 150 u A (Winslow, 1984) and/or post-mortem examination revealed the presence of blood in the thoracic cavity, hemorrhage due to surgical trauma and/or placement of stimulating electrodes into the myocardium. Acceptable animals were randomly allocated to vehicle or drug treatment. Drug infusion (in umol/kg/min) was continuous for the duration of the experiment with successive incremental doubling of the previous infusion, each infusion level lasting 5 min. A t the end of the third minute, electrical stimulation end-point measurements (in duplicate) were taken over 2 min. Since the threshold for induction of V F t can be raised by reductions in N a + and/or K + conductances, V F t data from a select group of R S D compounds were compared with their responses against ischemia-induced arrhythmias. The effects on electrically-induced arrhythmias (i.e., VFt ) are shown in Figures 13A - D (cross hairs symbols ± S E M ) expressed as the percentage of the maximum current of 1000 p. A . Data points were fitted using nonlinear equations. Doses producing a 25% change from control (D 25%) were 70 estimated from five to seven determinations (RSD 1049 and RSD 1025). Table 5 lists the VFt values of all RSD compounds. 2.2. In V i t r o Studies 2.2.1. Isolated rat v e n t r i c u l a r myocytes 2.2.1.1. Isolat ion Hearts were excised from pentobarbital-anesthetized rats (70 mg/kg i.p. plus 1000 units heparin). Perfusion of the heart was via the aorta on a Langendorff perfusion apparatus (Cole-Palmer, 7553-20) and the initial washout period was for 8 minutes. Solutions were bubbled with 100% O2 at a constant flow rate and temperature of 37°. The washout solution (solution I) was composed of (mM): NaCl (150), KC1 (10), M g C l 2 (1.2), N a H 2 P 0 4 (1.2), HEPES (10) and glucose (11), pH at 7.35. The heart was then perfused with enzyme solution (solution II) containing 1 mg/ml collagenase (317 units/mg, Type II, Worthington Biochemicals, Freehold, NJ), 25 uM CaCl 2 , and lmg/ml B S A (bovine serum albumin, Sigma Chemical Co.). Enzyme treatment was for 15-20 minutes and the ventricles were then cut away from the atria, placed in 20 ml collagenase-free solution I with 25 uM CaCl 2 and lmg/ml B S A (solution III). The tissue was shaken and triturated gently at 37° to release the dissociated myocytes. Cells were harvested by filtering the resulting suspension through a 200 um nylon mesh and the filtrate was spun down (15 - 20 sees) using a table-top centrifuge (25 x g). The supernatant was drawn off and the pellet of cells resuspended in solution III. Myocytes were stored at room temperature and the calcium concentration was doubled successively (~10 min intervals) from 25 uM to a final concentration of 1.8 mM. Cells were allowed 71 to settle for 2 hours and those used for the experiments were rod-shaped, quiescent and exhibited clear cross striations. The cells that were used for this study met the criteria of quiescence, an initial resting membrane potential of 60 - 75 m V and the morphology of a rod-shape with clear striations. 2.2.1.2. Recording solutions For ITO'- The solution used to perfuse the cells during recording of transient outward K + current (ITO) contained (in m M ) : N a C l (137), KC1 (5.4), M g C l 2 (0.5), C a C l 2 (1.8), C d C l 2 (0.2), glucose (5.0), and H E P E S (10), adjusted to p H 7.3 with 5 M N a O H . Cadmium and tetrodotoxin (5 uM) were included in the bathing solution to block inward calcium and sodium currents, respectively. The pipette solution contained (in m M ) : KC1 (120), C a C l 2 (0.15), M g C l 2 (6), E G T A (5), N a 2 - A T P (5), and H E P E S (10), adjusted to p H 7.3 with I M K O H . For lNa: The bath solution used to study inward sodium current (IN3) contained 5.4 m M C s C l (to replace 5.4 m M KC1) and 50 m M N a C l with 87 m M Tris (to replace 137 m M NaCl) . The p H of the bath solution was titrated to 7.3 or 6.4 with 5 N HC1. 4 -AP at 5 m M was also included in the bath solution to block I T o The pipette solution contained (mM): N a C l (10), C s C l (120), E G T A (12), TES (10), M g C l 2 (1), N a 2 -A T P (5); p H was adjusted to 7.3 with I M K O H . Forlca- Experiments on inward calcium current (Ic a) used a bath concentration of the following external ionic composition (mM): Tris (137), C a C l 2 (5.5), M g C l 2 (1), C s C l (20) and glucose (5.5); p H was adjusted to 7.3 with C s O H . I N a and I T 0 currents were suppressed with external addition of 5 p M T T X and 5 m M 4-AP. The 72 pipette solution used contained (mM): CsCl (125), Mg-ATP (5), E G T A (15), TEAC1 (20) and HEPES (10); pH was adjusted to 7.3 with I M CsOH. 2.2.1.3. Data recording The procedures used were essentially those of McLarnon and Xu, (1995, 1997). Using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA) with the low-pass filter set at 1 Khz, whole-cell ITO and iNa were recorded at room temperature (21-24°C). Current measurements were elicited with command pulses generated by a 12-bit digital-to-analogue converter and controlled by pClamp 6 software (Axon Instruments). Recordings were stored on hard disk of an IBM-compatible computer. Micropipettes were from Corning 7052 glass (1.65 mm OD, 1.2 mm ID; A - M Systems, Everett, WA) and prepared with a Narishige puller (PP-830) with resistance values between 2-6 M Q when filled with the internal solution described above. When the whole-cell configuration was established, amplifier compensation for pipette resistance was made with series-resistance correction while cell capacitance was usually not fully compensated. 2.2.1.4. Current measurements ITO measurement. Under our experimental conditions, only the Ca -independent component of ITO was studied since external cadmium was added to the perfusing solution and E G T A (5 mM) was included in the pipette solution, thereby blocking IC3,L- Cells were held at -70 mV and I T O was activated with a 400 ms depolarizing step to +60 mV. 73 In rat ventricular myocytes, depolarization of the membrane to potentials more positive than -30 m V (from a holding potential > -70mV) results in a transiently activated outward current that decays exponentially to a steady-state level (Dukes and Morad, 1989). Two kinetic components have been described (Apkon and Nerbonne, 1991), the first of which activates and inactivates rapidly followed by a much slower inactivating (or non-inactivating) component of the current which remains constant for at least 10 sec during continued depolarization. In this study, only the early inactivating component of the outward current is termed ITO and the integral of this current, minus the slower sustained non-inactivating component, were analyzed (see Figure 9B, inset). The concentration-response curves were normalized to control (i.e., integral Idmg / integral Icontroi) and plotted according to equation 1. INC measurement. Cells were held at -70 m V and primed for depolarization to -20 m V (30 ms step) with a hyperpolarizing prepulse to -140 m V (60 ms step) to remove resting inactivation. The current amplitudes were normalized to control and the concentration-response curves for inhibition of IN 3 were fitted using equation 1. Use-dependent block of iNa was investigated with the application of depolarizing steps to -20 m V from a holding potential of-100 m V . The protocol consisted of a series o f 20 pulses (30 ms duration) applied at a frequency of 20 Hz . Current amplitudes were normalized to the 1 s t pulse in each episode and these were plotted as a function o f pulse number (1-20). Ica measurement. Following suppression of iNa and ITO currents, calcium currents were recorded with a single 25 ms depolarizing step to -20 m V from a holding potential o f - 7 0 m V . 74 2.2.2. Cloned sodium channels 2.2.2.1. Expression of cardiac sodium channels in Xenopus oocytes The following group of experiments was performed in cooperation with Dr. Michael Pugsley with the experiments performed by him in the laboratory of Dr. A . Goldin. These were performed according to guidelines established by the Institutional Animal Care and Use Committee of the University of California, Irvine. 2.2.2.1.1. Transcription of RNA and expression in Xenopus oocytes The plasmid pSkM2 contains the coding region for the rat cardiac (rHl) sodium channel a-subunit (Kallen et al., 1990). Plasmid D N A was linearized by digestion with Ase I, and the R N A transcript was synthesized using the message machine SP6 R N A polymerase transcription kit (Ambion, Austin, TX). Stage V oocytes were obtained from adult female Xenopus laevis frogs, defolliculated with collagenase and injected with 50 nl of in vitro transcribed R N A at a concentration sufficient to obtain current amplitudes between 1-5 uA, as previously described (Goldin and Sumikawa, 1992). The oocytes were incubated for 48 hours at 20°C in ND96 (see "Solutions and drugs" below) with 0.1 mg/ml gentamicin, 0.5 m M theophylline and 0.55 mg/ml pyruvate. 2.2.2.2. Solutions and drugs A l l oocyte experiments were performed at room temperature (20-22°C) in ND96 bath solution containing 96 m M NaCl; 2 m M KC1; 1.8 m M CaCl 2 and 5 m M HEPES, at a pH of either 7.5 or 6.5. RSD1000, RSD1025, RSD1046, and RSD1049 were solubilized in distilled water as 10 m M stock solutions prior to dilution to the final concentrations in 75 the ND96 bath solution. A low volume (0.70 ml) plexiglass recording bath allowed for efficient exchange (20-30 sec) between control and drug solutions from gravity-flow reservoirs. A suction device ensured continuous perfusion at a flow rate of 1 - 2 ml/min and maintained a constant fluid level. 2.2.2.3. Data recording Recording electrodes were prepared from borosilicate glass using a two-stage P-87 puller (Sutter Instrument Co. , Novato, C A ) . Microelectrodes were filled with filtered 3 M KC1 in 0.5% low melting point agarose. Microelectrode resistance was between 0.5-1.0 M Q . Currents were recorded using a virtual ground circuit, and the data were filtered at 3 k H z on-line and digitized at a sampling frequency o f 12.5 kHz . Currents were recorded and analyzed using p C L A M P 6.0.3 software (Axon Instruments, Foster City, C A ) . Capacitance transients and leak currents were corrected by P/4 subtraction with the depolarizations for subtraction applied after each protocol. Non-linear curve fitting was performed using SigmaPlot® (version 4.0, Jandel Scientific, San Rafael, C A ) . Data are shown as the mean ± S E M for n experiments. Statistical analyses were performed using SigmaStat® statistical software (Jandel Scientific), with p less than 0.05 being considered statistically significant. Concentration-response curves for each of the compounds were determined by measuring peak inward current for cells depolarized from -120 m V to -10 m V in the absence and presence of either RSD1000 ( l -300uM), RSD1025 (1-300 uM) , RSD1046 (1-1000 uM) , or RSD1049 (1-1000 uM). The more negative holding potential (-120 mV) was necessary for recording from the r H l channel to allow for complete recovery 76 from slow inactivation (Pugsley and Goldin, 1998). Currents were allowed to recover from slow inactivation for 10 min before beginning any electrophysiological protocol. Drugs were then added for 5 min into the bath before recording current. The resulting fractional block o f sodium current by each drug at each concentration examined was plotted against the log concentration of drug, to fit the H i l l equation: I N a = [ l + ( [ A ] / E C 5 0 ) n ] - 1 (5) where iNa is the fractional block of sodium current, [A] is the concentration of drug, and ' n ' is the H i l l coefficient. The values of these parameters are shown in Table 7. 2.3. Data analysis and statistics Arrhythmia values in the tables and figures are expressed as mean ± standard error. Statistical analysis on haemodynamic and E C G parameters (Tables 4, and 5) was performed using a paired t-test and was used to compare pre-infusion and 5 min post-infusion at each dose level prior to coronary-artery ligation (p < 0.05). Comparisons between each dose level and control groups for each compound were performed using a single-factor A N O V A followed by Dunnett's test with significance levels of p < 0.05 and p < 0.01 (Table 5). In Figures 6-9, differences in the proportionality of V T or V F incidence between vehicle- and RSD-treated animals were tested using the Fisher exact test with p < 0.05. Using Fisher's z values to transform the correlation index, r, a one-tailed analysis was performed on r 2 in the correlation matrices of Table 9 to determine significance (p < 0.05) for r 2 = 0.5 (i.e., r = 0.7). In Table 9, when two or more values had exactly the same value, the assigned rank to each of the tied ranks was the average of the ranks that would have been assigned to these ranks had they not been tied. 77 2.4. RSD compounds Nortran Pharmaceuticals Inc, Vancouver, Canada, synthesized all 16 RSD compounds used in this study. Their chemical names, structures, and some physiochemical properties are summarized in Tables 1 and Figure 3. The salt form for all of the compounds was the monohydrochloride and their percent purity was no less than 90%. Except for the formate and acetate RSD analogues, which existed in a red liquid/oil form, other compounds existed as a white crystalline powder. Molecular weights for all RSD compounds were within the range of 250 (formate) to 436 gm/mol (RSD1025). The solvent used was distilled water or a 22%o:78% mixture of ethanol:distilled water and each compound was dissolved prior to the start of each experiment. 2.4.1. pKa determination The following chemical experiments for determining pKa and log Q measurements for the tested RSD compounds were performed by Dr. R.A. Wall and Sam Tsui at the facilities of Nortran Pharmaceuticals Ltd. By definition, pKa is the negative logarithmic (-log) transformation of K a . The strength of a base (or acid) is an expression of the tendency to form the conjugate acid (or base), or to accept (or donate) protons, or to form cations (or anions) and this strength is expressed as K a . A l l RSD compounds were synthesized as monohydrochloride salts, RaM+H + A" R 3 N + A" + H* monohydrochloride salt free base 78 Molecular structures of many drugs contain groups that can be ionized in aqueous solutions to anionic and cationic forms. In the case of R S D compounds, the ionizable group in question is the nitrogen atom in position R 2 (Figure 1; Table 1). A s the monohydrochloride salt forms, the p K a values for all R S D compounds in this series were determined by the equivalence point method following base titration. Each R S D compound (~l-2 mg) was solubilized with 10 m l distilled water in a plastic titration vessel. This solution was degassed by passing a slow stream of nitrogen bubbles just below the surface of the solution. A p H electrode (Orion Ross model 8120 with Orion model 720A p H meter) was inserted into the test solution and readings were recorded at room temperature (25°C) on a chart recorder (Servogor 124). Titration was performed using 0.01 M N a O H and this was delivered from a 1 cc glass syringe at a continuous injection volume that was controlled by an infusion pump (Harvard Apparatus, Model 2681) at a rate of 0.07 ml/min. During the course of the titration, the solution was illuminated with a tightly focussed beam of light to detect precipitate of the free base form of the dissolved compound. A black board placed behind the vessel provided an additional detection method of precipitation by the Tyndall effect (the physical property by which suspended particles scatter ambient light). Both aqueous and 50% ethanol titrations were performed for some soluble compounds to ensure precision between the titration types. The equivalence point of the test solution was determined by plotting the p H of the solution versus the volume of N a O H added. From the continuous titration curve of each compound, two parallel lines tangent to the 1 s t and 2 n d curvatures of the curve were drawn, near the beginning and end of the titration curve, respectively. Using the points of 79 the two tangents as radii, two circles of the same radius are drawn intersecting the straight and steep portion (i.e., midpoint) of the titration curve. The arcs intersect each other at two points, above and below the midpoint of the titration curve. A straight line is drawn joining the two points and where the line crosses the titration curve (equivalence point) is the estimated p K a value for the test compound. One distinction must be made with regard to the ionizable nitrogen group of RSD1025 (N-methylpiperazinyl). The relative position of 1'-nitrogen to 4'-nitrogen in relation to the ester and carbonyl oxygen atoms increases the basicity strength of 4 '-nitrogen and, as such, the overall influences on the degree of ionization of the molecule. Thus, the residing positive charge of the molecule is on the 4'-nitrogen of R S D 1025 compared to the 1 '-nitrogen of other R S D analogues in the series. 2.4.2. HPLC measurement of hydrophobicity for RSD compounds A 1 mg/ml sample of R S D compound was made by transferring 4 mg of the solid sample and 4 m l of 25% methanol solution to a 4 m l vial . The solvent used was p H 4, 10 m M sodium methylsulfonate, 5 m M triefhylamine in 30% aqueous acetonitrile. A 5 u L glass syringe was rinsed with solvent 2-3 times and then filled with 2 u L solvent, 1 u L sample, and 2 u L solvent. The liquid was injected into the U V detector system (with column, Inertsil ODS2, 04 .6 x 120 mm) set at 254 nm and 0.08 A U F S . Detection peaks were recorded on a chart recorder (Servogor 124) at a chart speed of 20 cm/hr with peak retention times (f) measured in minutes. The corrected retention t' o f a compound by H P L C is related to elution time by t' = (tR-t 0)/t 0 80 where tR and to are the retention times of a retained and unretained compound, respectively (McCall, 1975). If tc = (tR - to) is the corrected elution time log t' = log t c - log to and log tc is related to hydrophobicity (or log Q) by log Q = log tc + constant. The determination of hydrophobicity by the HPLC method provides results that approximate those obtained by the traditional shake-flask method for determining octanol-water partitioning coefficients (Mirrlees et al., 1976). 81 Table 3 Apparent Calculated RSD Chemical name pKa logQ ( P H 6.4) logQ (PH 7.4) formate (±) 1,2-trans-1 -(formyloxy)-2-(4-morpholino) cyclohexane monohydrochloride 7.3 -0.1 0.5 acetate (±) 1,2-trans-1 -(acetoxy)-2-(4-morpholino) cyclohexane monohydrochloride 7.3 -0.1 0.6 1053 ' (±) 1,2-trans-1 -(benzoyloxy)-2-(4-morpholino) cyclohexane monohydrochloride 6.5 1.9 2.5 1010 (±) 1,2-trans-1 -(phenylacetoxy-2-(4-morpholino) cyclohexane monohydrochloride 6.0 1.7 2.4 1050 (±) 1,2-trans-1 -(benzoyloxy)-2-(4-morpholino) cyclohexane monohydrochloride 6.5 2.2 2.8 1051 (±) 1,2-trans-1 -(phenylbutyryloxy)-2-(4-morpholino)cyclohexane monohydrochloride 6.5 2.7 3.3 1012 (±) 1,2-trans-1 -(3,4-dichlorophenylacetoxy)-2-(4-morpholino)cyclohexane monohydrochloride 6.0 3.1 3.8 1072 (±) 1,2-trans 1 -(3,4-dimefhoxyphenylacetoxy)-2-(4-morpholino)cyclohexane monohydrochloride 6.4 1.6 2.3 1019 (±) 1,2-trans-1 -(4-bromophenylacetoxy)-2-(4-morpholino)cyclohexane monohydrochloride 6.1 2.6 3.2 1014 (±) 1,2-trans-1 -(4-nitrophenylacetoxy)-2-(4-morpholino)cyclohexane monohydrochloride 6.0 1.6 2.3 1000 (±) 1,2-trans-1 -(1 -naphfhylacetoxy)-2-(4-morpholino)cyclohexane monohydrochloride 6.1 2.9 3.6 1009 (±) 1,2-trans-1 -(2-naphfhylacetoxy)-2-(4-morpholino)cyclohexane monohydrochloride 6.2 2.9 3.6 1015 (±) 1,2-trans-1 -(2-naphthylacetoxy)-2-(4-methylpiperazinyl)cyclohexane monohydrochloride 8.9 2.4 3.4 1025 (±) 1,2-trans-1 -(1 -naphthylacetoxy)-2-(4-methylpiperazinyl)cyclohexane monohydrochloride 8.9 2.4 3.4 1046 (±) 1,2-trans-1 -(1 -naphthylacetoxy)-2-bis-N,N-(2-methoxyethyl)aminocyclohexane monohydrochloride 7.0 3.5 4.0 1049 (±) 1,2-trans-1 -(1 -naphthylacetoxy)-2-piperidinocyclohexane monohydrochloride 8.2 3.2 4.0 82 Table 3 The chemical names and values of pKa and hydrophobicity (log Q) for all 16 RSD compounds. The values for pKa and hydrophobicity (at pH 7.4 and 6.4) were measured according to protocols described in the Methods section. 83 3. Results The R S D compounds studied were chosen based on their structural similarity to R S D 1000 in order to explore chemical changes to aryl and amine heterocyclic groups (see Figure 3 and Table 1). In some cases, the lack of sufficient quantities o f compound limited the testing of some to a maximum of three infusion doses. A few of these compounds constitute intermediary steps (e.g., formate and acetate compounds; see Table 6) towards a final compound, but the majority of them are final products. Nevertheless, the core pharmacophore structure consists of a cyclohexane backbone with an ester moiety in position 1' trans to the nitrogen atom at position 2' . The present results compare this basic backbone with chemical variations to the aryl and amine heterocyclic moieties. 3.1. Effects on hemodynamic and E C G parameters in rats subjected to myocardial ischemia The effects on blood pressure (BP), heart rate (HR) and E C G variables in the presence of R S D compounds were compared in rats prior to them being subjected to coronary occlusion (Table 4 and 5). Prior to occlusion, the above variables were measured 5 min before and after 5 min of infusion. The results are summarized in Tables 4, and 5. For ease of presentation, a detailed summary of these effects is confined to the naphthalene-containing R S D compounds in Table 4 with a brief overview for all compounds in Table 5. Figures 2 A and 2B illustrate graphs of representative R S D compounds (RSD 1000, R S D 1025, and R S D 1049) that exemplify their actions on lowering blood pressure (Figure 2A) and increasing P-R interval (Figure 2B). The format of these figures was similar for the other R S D compounds for estimating D25% values for 84 all o f the hemodynamic and cardiac parameters measured. A t the doses tested, individual animals (open circles) were plotted as a percent change o f the variable under consideration from pre-infusion values (filled circles). A t each infusion dose, a single animal was chosen at random and a non-linear curve fit was applied and this was repeated for a minimum of five determinations. For each curve, interpolation of the infusion dose producing 25% change was estimated and a mean ± S E M was given (see Figures 2 A and 2B). A t the infusion doses tested, many of the compounds produced no significant changes; these were denoted " N E " for potency Not Estimated due to changes being less than 25%. In cases where there were less than 50% of the total animals at the highest infusion dose that did not exhibit a drug-induced change of 25% or more to allow for a proper estimate o f D 2 5%, the parameter in question was also deemed " N E " . Where values are indicated in Tables 4 and 5, these represent estimates of the infusion dose producing a 25% change from pre-infusion values. Antiarrhythmic infusion doses producing 50% protection ( A A ED50) against ischemia-induced arrhythmias are also listed for comparison. Compounds producing a significant (infusion dose-dependent) change from pre-infused values and versus vehicle-treated animals included R S D 1049, R S D 1025, and RSD1015. The infusion doses of these compounds produced up to 50% changes from pre-infusion and these include blood pressure and heart rate lowering, as wel l as widening E C G intervals such as PR, QRS and QT. Min imal hypotension was observed for R S D 1046, with slight increases in mean systolic and diastolic interval (not statistically significant, p > 0.05). Due to their cardiodepressant actions, high infusion 85 Table 4A: Effects on blood pressure Dose pre-drug systolic 5 min systolic pre-drug diastolic 5 min diastolic (umol/kg/min) (mmHg) (mmHg) (mmHg) (mmHg) vehicle 139 ± 1 137 ±3 119 ± 1 121 ± 1 RSD 1000 1 123 ±2 110 ± 2* 100 ±2 84 ± 2 * 2 109 ±2 99 ± 5 * 86 ±2 78 ±4 * 4 117 ± 2 89 ± 4* f 86 ±2 6 5 ± 2 *f 8 117 ± 2 81 ± 4 * t + 87 ±2 6 2 ± 3 *n vehicle 120 ±2 137 ±3 119 ± 2 119 ± 2 RSD 1046 0.5 122 ±2 123 ±3 96 ±2 97 ±2 1 139 ±3 142 ±5 107 ±2 103 ±4 2 119 ± 1 124 ±5 94 ± 1 87 ± 4 * 4 125 ± 1 116 ± 4 107 ± 1 95 ±3 8 132 ± 1 123 ±6 106 ±2 87 ± 5 * vehicle 136 ± 1 134 ± 1 117 ± 1 116 ± 3 RSD 1049 0.1 132 ± 2 134 ± 1 103 ±3 102 ±2 0.5 113 ± 3 105 ±3 91 ±2 84 ±3 1 128 ±2 117 ± 4* 97 ±2 83 ± 3 * 2 130 ±4 111 ± 4* + 107 ±3 73±3* f 4 138 ±4 91 ± 5 * n 114 ± 4 49 ± 2*t + vehicle 139 ± 1 137 ± 2 119± 1 121 ± 1 RSD 1025 0.5 131 ±2 128 ± 1 104 ±2 103 ±2 1 126 ±2 119 ± 2 * 91 ±2 80 ±2* 2 125 ±2 9 9 ± 5 * t 100 ±3 78±3* f 4 133 ±2 95 ± 2* t + 106 ± 1 5 8 ± 5 * f t vehicle 125 ± 1 127 ±2 103 ±2 103 ±2 RSD 1009 1 123 ± 1 120 ± 2 101 ±2 98 ±3 2.5 125 ±3 128 ±3 103 ±3 104 ±2 8 117 ± 3 117 ± 2 96 ±3 96 ± 1 16 120 ±2 118 ± 5 96 ±3 93 ±2 vehicle 121 ±2 119 ± 2 95 ±2 93 ±2 RSD1015 0.5 123 ± 2 124 ±2 101 ±3 104 ±2 1 119 ± 2 106 ± 4 f 95 ±3 84 ± 1 2 119 ± 3 94 ± 4* f 97 ±3 71 ± 2* + 4 123 ±2 89 ± 6* + 99 ±3 5 7 ± 4 * t t 8 120 ± 2 58 ± 3 * n 102 ±2 3 6 ± 3 * n 86 Cr I? a - , 00 2 00 C oo 00 n O 00 Ov Ov CN CN ov CO O CO o OV o (N CO O CO O CO CN Tf CN vo — OV Ov .—i CO Tf cn CO cn Tf Tf Tf CO Tf Tf Tf Tf Tf CO Tf Tf Tf Tf Tf Tf V) VI Ti in in in uo in Tf Tf in _ CN CN ~* CN _ _ CN _ _ _ _ _ _- Tf # CN +1 CN — — -H -H -H -H -H -fl -H -H +1 41 -fl -fl +1 -H +1 -H +1 -H +1 oo CN Ov O CO CN O "n O CO OV Tf CO Tf CO Tf CN CN CO CO CN CO CO CO CO CO ro CO ro ro CN CO OO ro CO ro CO CO cu 00 o le K , oo' S a T3 J3 ^ m c^ CO rvj CN CN -H -H 2 2 ^ i t f * -H J J T. JJ " " vo CN "H -H 4 1 Z Z M oo Z. t~~ m CO S CN Tf O ' fN CN CN * * OV CO 41 Z — 4 1 "fl ro C N r^ » co C N o t ^ CN co co ^ 0 0 —I iu -H O * CV * # * 1- * CN VO a\ ro — -H —* — •—1 -H CN -fl -H -H -H -H CN Tf OV 00 CO CO CN in OV CO 00 Tf in Ov CN CN CN CN 00 CN CO r» o O ro OV OV Tf CO CN OV VO OV in 00 Ov ov CN OV O Ov 00 O O OV CN -H -fl +1 -fl -fl +! -H -fl -fl -fl -H -fl -fl -fl +! -fl -H -H -H -fl -H -H -fl +i -H -fi -fi -H -H +1 -fi -H ro oo CO Tf r~ VO ro VO 00 OV Tf O OO VO O VO VO CO 00 vo Tf Tf in ro oo in Ov CO CN OO ro 00 Tf Ov ro r- Tf 00 ro o CN IO OV 00 CN CO 00 00 00 00 oo r^ r~ 00 OO oo r~ OV in VO in ov CN CO 00 00 ro Tf vo r~ CO CO CO ro ro ro ro ro ro CO ro ro ro ro ro CO CO CO ro ro ro ro ro co O « I N vf M O O o O 00 rt in — C N Tf oo J5 V VO > Tf o 5 00 rt JJ ^ in —• C N Tf ~Z> o d u ov > Tf a 00 rt JJ in C N Tf "o d u m > C N o 5 00 rt JD — in oo vo 7j C N — U CV > O o 5 oo rt JJ in — CN Tf 0O D 00 rt 87 Table 4 A and 4 B Effects on blood pressure, heart rate and E C G parameters by RSD1000, R S D 1046, RSD1049, RSD1025, RSD1009 and RSD1015 in rats prior to coronary occlusion. For each compound, the values listed include the control group and the values for each infusion dose used. Measurements were obtained after 5 min of infusion (in umol/kg/min), prior to coronary artery occlusion and are compared to pre-infusion values. B lood pressure is expressed as the averaged peak systolic and diastolic values (± S E M ) , whereas heart rate (HR) and the E C G intervals (PR, Q R S , Q-T) are listed accordingly. The (*) indicates p<0.05 from pre-infusion values by two-sample t-test, while ( |) and ( f t ) indicate p<0.05 and <0.01 from vehicle-treated animals using single-factor A N O V A followed by Dunnett's test. The sample size was 5-9 animals for each infused dose (see also antiarrhythmic dose-response figures for corresponding R S D compounds) 88 Jg Tf in ro "1 ro in vq ro C N Tf ro s % d d d d d d d d d d d d d d u +1 -H -H -H -H -H +1 41 -H -H -H -H 41 + 1 M k o oo 0 0 r— vq Ov Ov rs in oo vq ro ro Tf co Tf' ro ro' ro Tf' ro ro ro Tf' r ^ ^ -—s oq C N vq o V) C N Q * # •£ * * * * * * s ^ * w o VO Tf VO rn oq W Tf ro O ro in << d -J Tf' Tf in od Z d — _; d d -H 41 -H +1 in ro as C N ro ro fc > 2 PJ Z UJ Z w Z UJ Z w Z CN + 1 ro vd CN 41 Os ro -H U O r- T f o d d d d -H + 1 -H -H vq CM Tf' —< C N d vo oo vo d d d +1 41 41 vq — ; vn ro in (N -H ro vq Tf ro C N ro +1 -H d C N oo + 1 Tf' C N oo OS C N ro d -H 41 -H m ro uo vd Ov t~ 0 0 rN m d d d -H -H 41 ro o -H in d W Z vo -H o ro UJ z PJ z UJ z w z PJ z Ov oo CN Ov d ro d — ' -H -H -H + 1 -H CN 0 0 in Tf Os CN Tf d -H I Oi I as in ro -H vo Z Z CN in -H CN ro UJ Z uj Z UJ Z U J Z UJ Z UJ Z UJ Z UJ Z UJ Z UJ Z w z UJ z UJ z w z w z w z oo Tf Tf CN -H CN -H o -H oo oo UJ Z UJ Z UJ Z UJ Z UJ Z -H ro od UJ Z U J z Tf CN -H m UJ Z UJ Z PJ z UJ z -H m UJ Z 41 m 41 in vo ro o d d d 41 41 41 in in >n >-« CN —I UJ Z T f d 41 CN ro ro O 41 m UJ Z UJ Z ro d 41 ro oo d 41 T f CN d 41 T f d 41 P H ca l o 41 | oo vd as ro 41 >n OS CN UJ Z UJ Z UJ Z UJ Z UJ Z w z UJ z UJ z Tf d 41 Tf CN d 41 Tf CN 41 oo CO d 41 oo a Cfl o o o ro m o o © o in o m o CN © C N o T f o Ov O OS o o m o in CN o vo Tf o Ov Tf o 89 Table 5 An overall summary of 14 RSD compounds and their effects on blood pressure (BP), heart rate (HR), E C G parameters and electrical stimulation variables (iT, ERP, VFt). The values for haemodynamic and E C G parameters were estimates of the infusion dose that produced a 25% change (D25%) from pre-infusion after 5 min of infusion. The values for iT, ERP and VFt were also estimates of the infusion dose producing a 25% change from pre-drug following the cumulative infusion protocol described in the Methods section. " N E " denotes "Not Estimated" due to changes < 25% from pre-infusion levels. Antiarrhythmic dose producing 50% protection (AA ED 5 0 ) against ischemia-induced arrhythmias is also listed for comparison. Numbers in brackets denote slope factor, |b|. For some compounds, approximation of the A A E D 5 0 values were made due to insufficient doses and/or inability to produce 100% efficacy (i.e., AS = 0) with a slope factor (i.e., |b|) < 1. These values are denoted with "*". A l l values in the table are in umol/kg/min units ± SEM. 90 Figure 4A 100 o "(/) CD Q . E o CD CO CO CD o CD Q CD 13 CO CO CD TD O O CQ 75 50 H 25 + 100 75 100 5.6, 6.8, 6.8, 6.8, 7.8 = 6.8 ± 0 . 3 RSD1000 1.9,2.1,2.4, 2.7,3.1 = 2.4 ± 0 . 2 RSD1025 1.2, 1.3, 1.8,2.0,2.5 =1.8 ± 0 . 2 RSD1049 Limol/kg/min 91 Figure 4B 100 75 NE 50 25 o CO 13 c Q . E 0 CO CTJ CD b CZ CC > DC • Q_ 100 75 50 25 100 ] 0.01 o 8 -o-RSD1000 2.2,2.8,3.0,3.7, 4.2 = 3.2 ±0 .3 Limol/kg/min RSD1025 0.4, 0.6, 0.7, 1.5, 1.7 =1.0 ±0 .2 RSD1049 92 Figure 4 Effects of RSD 1000, RSD 1025 and RSD 1049 on lowering blood pressure (A) and increasing P-R interval (B) in rats prior to coronary occlusion. These are representative graphs which illustrate the determination of D25% values of RSD compounds (listed in Tables 4 and 5) after graphical expression of hemodynamic and cardiac variables (i.e., BP, HR and ECG variables). A l l graphs illustrate single animals (O) at each infused dose compared to their pre-infusion levels (•) expressed as a percent change. At each infusion dose, a single animal was chosen at random and a non-linear least squares curve fit was applied and this was repeated for a minimum of five determinations. For each curve, interpolation of the infusion dose producing 2 5 % change was estimated and a mean ± SEM was given. 93 doses of RSD1049, RSD1025, and RSD1015 were not explored. A n infusion dose of 8 umol/kg/min RSD 1015 was explored (see Table 4B) but due to its cardiodepressant actions, antiarrhythmic values for this infusion dose were not estimated. More notable were the effects of atrio-ventricular blockade and cardiac output failure (COF) produced by 4 umol/kg/min RSD 1049 8 - 10 min post-occlusion (Table 4; Figure 12B). In control animals, COF is usually observed 1-2 min post-occlusion and is associated with an occluded zone size exceeding 45% and/or an occlusion that accidentally reduced aortic blood flow. Such animals were excluded and replaced. The highest infusion dose of RSD 1049 (4 umol/kg/min) produced significant increases in both P-R and Q-T intervals (p < 0.05 and p < 0.01; Table 4B), whereas none of the other RSD compounds exhibited these effects at infusion doses equal to or greater than 4 umol/kg/min. It should be noted that at 32 umol/kg/min RSD1053, 4 out of 5 animals that were originally studied at this infusion dose had to be replaced due to drug-induced toxicity. Seven animals were used at this infusion dose and of these animals, 4 (2 from the original group and 2 replacements) exhibited COF 8 - 1 0 min post-occlusion. Similar to RSD 1049 and, to a lesser extent, RSD1015, the COF appeared as a progressive decrease in diastolic and systolic intervals (mean blood pressure decreased to < 25 mmHg) accompanied by a reduction in heart rate (364 ± 16 to 189 ± 8 bpm; p < 0.05). Occluded zone sizes were within acceptable (25-40%) limits in the 4 animals that experienced COF. It should be noted that the values listed in Table 4 and 5 were obtained from a period between pre-infusion and pre-occlusion - a total of 5 min. For this reason, the 94 interpretation of effects on cardiovascular and E C G parameters should be made with caution and the knowledge that 5 min of infusion were not sufficient to provide truly steady-state concentrations (see "Limitations" section). The parameters measured at the end of 5 min reflect an "underestimate" of true steady-state levels. A 5 min infusion regime was chosen as an adequate elapsed time to reach a "pseudo"-steady-state concentration level. For most RSD compounds, detectable effects were observed after the first 2-3 min of infusion. For the lead compound, RSD 1000, a sham occlusion was performed (Yong et al., 1999) at an effective infusion dose of 8 umol/kg/min (n = 7). At this infusion dose, it reduced the incidence of ischemia-induced arrhythmias (see Table 5; Figure 9A) to nearly 0%. The hypotensive and bradycardic actions of RSD 1000 at this infusion dose were transient but not progressive for the duration of the experiment (20 min total infusion duration). Concentrations of RSD compounds appeared to peak after 10 min from the start of the infusion based on hemodynamic and E C G parameters reaching a new steady-state level. Studies involving the metabolism of these compounds were not investigated in this thesis, although details of possible targets of chemical degradations on the RSD pharmacophore are presented in the Discussion section. 3.2. Effects of compounds on ischemia-induced arrhythmias 3.2.1. Effects of increasing aromatic chain length The antiarrhythmic effects of RSD compounds were determined in intact rats subjected to coronary-artery occlusion. Samples of original polygraph traces that show the time-course of the coronary artery-ligated model are illustrated in Figure 5. These are 96 97 Figures 5 Original polygraph traces showing the time-course of the coronary artery-ligated model for vehicle- (A) and 4 umol/kg/min RSDIOOO-treated animals (B). Traces (from upper left —> down and upper right —> down) represent "snap-shot" periods from the start o f coronary occlusion to the critical time periods of ischemia-induced arrhythmias. In each trace frame, E C G (upper), time counter (middle) and blood pressure (bottom) recordings are shown. Traces were expanded at a chart speed of 100 mm/sec. The voltage gain for recording E C G was 0.5-1 mV/cm, whereas the pressure reading for blood pressure is indicated in the upper left trace. In the vehicle-treated animal (Figure 5A), the critical period for ischemia-induced arrhythmias occurred 6-8 min post-occlusion. Few episodes of "non-fatal" P V C , V T and V F were recorded in the animal treated with 2 umol/kg/min RSD1000. Associated with each arrhythmogenic episode was a brief, prolonged, or sustained drop in blood pressure, respectively. In both examples, there was an initial increase in R-wave amplitude and a diminution of S-wave immediately following coronary artery ligation and this was followed by S-T segment elevation ~5 min post-occlusion (see Results for additional discussion). The % O Z and initial serum [K + ] were 32% and 3.9 m M for rat in Figure 5 A and 36% and 3.4 m M for rat in Figure 5B with A S of 7 and 2, respectively. 98 representative traces of recordings from the start of coronary occlusion to the critical time periods of ischemia-induced arrhythmias. In the vehicle-treated animal (Figure 5A), the critical period for ischemia-induced arrhythmias was usually 6-8 min post-occlusion. In animals treated with RSD compounds, as represented by traces in Figure 5B, the incidence of arrhythmias was dose-dependently reduced compared with control animals. In both examples, there was an initial increase in R-wave amplitude and a diminution of S-wave followed by S-T segment elevation ~5 min post-occlusion. These early electrocardiographic changes have been associated with acute myocardial ischemia but the exact mechanisms for these QRS changes remains unclear. Conduction slowing and the degree to which this is different along the transmural myocardial region have been proposed (Chang et al., 1989). Associated with each arrhythmic episode (PVC, V T and/or VF) was a brief, prolonged, or sustained drop in blood pressure, respectively. The %OZ and initial serum [K +] were 32% and 3.9 m M for rat in Figure 5A and 36% and 3.4 m M for rat in Figure 5B with AS of 7 and 2, respectively. In the case of the RSD compounds that were tested, antiarrhythmic protection against ischemia-induced arrhythmias was found to depend on both the compound and infusion dose administered. Four compounds (RSD 1053, 1010, 1050 and 1051) were used to study the effect of increasing the chain length of the phenyl group from the cyclohexyl backbone. Dose-response relationships for RSD 105 3 (A), RSD 1010 (B), RSD 1050 (C) and RSD 1051 (D) are shown in Figure 6A - 6D. A detailed description explaining the antiarrhythmic dose-response graphs is provided in the Methods section. As the chain length was increased with extra methylene groups, there was a dose-related reduction of AS (compare RSD 1053 and RSD 1051). Incidence of V T and V F as a Figure 6 A VT 10/10 VF 8/10 7 5.8±1.0; n=606 6 5 4 3 2 1 < T • 1 0 a o 2/5* 4/5 2/5* 0/3* * * * * 1/5 1/5 1/5 0/3 i i i i i i i i I I i i l l I I I I ppm I - ^ j — I 1 10 40 RSD1053 (pmol/kg/min) B VT 14/14 VF 13/14 7/7 5/7 7/7 2/7* 3/7* 4/7 3/7* 2/7* 0/7* 1/7* C/D < 7 r 6 5 4 3 2 1 0 T • 1 i i i i i i 1 1 N —o _i i i i i i i i 10 RSD1010 (umol/kg/min) 40 100 Figure 6 C co < T • 1 VT 11/15 VF 12/15 7 6 5 4 3 2 1 4/5 2/5 0/3 3/5 2/5 0/3' I 0 u a. o _i 1 1 1 1 1 1 1 o 1 RSD1050 (ymol/kg/min) - S -40 D VT 10/10 VF 8/10 CO < 5/5 4/5 1/5* 1/5* 4/5 3/5 0/5* 0/5* 1 10 RSD1051 (Mmol/kg/min) 101 F i g u r e s 6-9 RSD analogues that investigate the extension of the benzene aromatic group (RSD1053, RSD 1010, RSD 1050 and RSD 1050; 6 A - D , respectively), 3,4-phenyl substitutions (RSD1012 and RSD 1072; 7A a n d 7B) and 4-phenyl substitutions (RSD 1019 and RSD1014; 8 A a n d 8B) and positional arrangement of the naphthalene group (RSD 1000 and RSD 1009; 9A a n d 9B) were tested against ischemia-induced arrhythmias. L e g e n d The chemical structure of each compound is illustrated at the bottom of each graph for ease of reference. Each graph for each compound expresses the AS for individual animals (O) as a function of infusion dose. The mean AS for the group at each infusion dose is represented by ( • ) ± SEM, while the control group for each compound is illustrated as (•) ± SEM. Statistical analysis of the mean and individual AS values for each compound was facilitated by utilizing the AS values from 606 control rats as a representative control population that have been subjected to the same coronary artery ligation model as part of our laboratory's quality control data set. The mean AS value (5.8 ± 1.0) for the 606 control rats is represented by a solid thin black line bordered by a lower 95% confidence limit of the S E M ( ) and SD ( ). Above each graph of each RSD compound, the number of animals over the total number of animals for the group at each infusion dose that exhibited ventricular tachycardia (VT) or ventricular fibrillation (VF) is listed. The (*) denotes statistical significance from control (p < 0.05) using the Fisher exact test for testing differences in the proportions of VT or V F incidence between vehicle- and RSD-treated animals. The logistic equation used to fit the mean AS values are described in the Methods section and a list of antiarrhythmic ED50 values for all RSD compounds is summarized in Table 5. 1 0 2 Table 6 RSD Dose n= AS Mortality V T o r V F (u^mol/kg/min) Formate 8 5 7,7,5,7,4 3/5 5/5 Acetate 8 5 7,5,7,7,5 3/5 5/5 Control — 5 5,7,5,4,7 2/5 5/5 Table 6 Antiarrhythmic activity against ischemia-induced arrhythmias for the formate and acetate analogues in the coronary occluded rats. Both compounds were administered at a infusion dose of 8 umol/kg/min. The arrhythmia scores (AS) for both compounds and the vehicle-treated group are listed for each animal (n = 5 for each group). The number of mortalities and the number of animals exhibiting VF or VT over the total number of animals are also listed. A detailed description of arrhythmia scores is given in the Methods section. 103 function of infusion dose also support this relationship. In terms of antiarrhythmic "goodness" (see Methods section), increasing potency and a slope factor close to 1 were more favorable as chain length increased. RSD 105 3 has a single benzene ring as its aryl substituent without extension of the chain length. The antiarrhythmic relationship for this compound was poor with a low slope (i.e., < 1) between 2.5 and 16 (imol/kg/min. A higher potency and a dose-dependent reduction of arrhythmias was produced with RSD1051 but the variance of AS's (i.e., AS = 1-5) remained high between 2 and 16 umol/kg/min. The most notable result with RSD 1051 was that all 5 animals at 16 umol/kg/min infusion dose (compared to 32 umol/kg/min for RSD1053) survived the experiment without drug-induced complications. In the absence of a benzene ring (formate and acetate compounds), antiarrhythmic protection was not observed (Table 6). Both formate and acetate analogues were tested at 8 umol/kg/min, a welivtolerated infusion dose for these two compounds but an infusion dose in which compounds containing a phenyl reduced arrhythmia incidence by 40-50% from control. The extension and existence of the phenyl group raises possibilities of interactions with an aromatic system. 3.2.2, Effects of substitutions on the phenyl ring 3.2.2.1. 3,4-phenyl substitutions The two analogues, RSD1012 and RSD1072 (Figure 7), are 3,4-phenyl substituted derivatives of the unsubstituted phenyl compound, RSD 1010 (Figure 6B). The 3,4-phenyl substitutions of RSD 1012 and RSD 1072 may act to mimic a hydrophobic (3,4-dichloro; RSD1012) or a hydrophilic, (3,4-dimethoxy; RSD1072) "extension" in the 104 Figure 7 A < VT 11/12 VF 10/12 7 6 5 4 3 2 0 5/6 4/6 2/6* 1/6' 2/6* 2/6* 2/6* 1/6* CCD 0 1 I I 1 I I ' I I 1_ I I I I lp I I I 1 10 a RSD1012 (umol/kg/min) 40 B VT 10/10 VF 9/10 CO < 7 r 6 5 4 3 2 1 0 T I 5/5 5/5 5/5 3/5 2/5* 4/5 /5 3/5 4/5 1/5* i t i i i i 11 _J i i i i i i .OMe 1 10 RSD1072 (umol/kg/min) 40 105 same analogy as the previous compounds, whereby the chain length between the benzene group and the core pharmacophore was increased. The antiarrhythmic dose-response relationships of RSD 1012 and RSD 1072 were not that different to the unsubstituted phenyl compound, RSD1010 (Figure 6B). They both appeared to lack potency and failed to provide 100% antiarrhythmic protection (AS = 0) such as that seen with RSD 1051 Figure 6D). Thus, the antiarrhythmic effects of RSD1012 and RSD 1072 suggest that modifying the hydrophobicity or hydrophilicity of the phenyl group by 3,4-disubstitution does not appreciably influence antiarrhythmic activity against ischemic arrhythmias. 3.2.2.2. Para-substitutions: electron-withdrawing effect of the phenyl group Adequate extension of the phenyl group from the ester linkage is critical but equally important may be the nature of the aromatic group and interactions that may be associated with this group. The next step was to assess the involvement of the phenyl group and how this group influenced antiarrhythmic activity. Using para-substitutions involving a bromide (-Br) (RSD 1019) or a nitro (-N0 2) group (RSD 1014), it was possible to investigate the electron-withdrawing effect on the phenyl group from a single substitution and determine whether changes to the electron density of the phenyl group has any influence on antiarrhythmic activity. Figure 8A and 8B show that the electron-withdrawing group in RSD1019 was significantly more effective in providing a dose-dependent antiarrhythmic response than the nitro group of RSD1014. Little or no protection was provided by RSD1014 since the arrhythmia incidence and severity were independent of the infusion dose for this compound. Figure 8 A B co < VT 21/21 VF 19/21 7 6 5 4 3 2 1 VT 10/17 VF 15/17 CO < 7 r 6 5 4 3 2 1 0 T • i 7/7 4/7* 5/7 3/7* 1/7* 0/7^ 6/7 5/7 4/7 1/7* 0/7* 0/7' 10 RSD1019 (umol/kg/min) 5/5 4/5 7/7 4/5 4/5 4/5 5/5 4/7 3/5 4/5 • 1 T T J " T T • -i_ T • i i i i i t i i CXjp0LNC, RSD1014 (umol/kg/min) ^—o 40 _i i i 1 1 1 11 i i _ 40 107 Figure 9 A B VT 7/7 VF 7/7 * * 5/7 4/7 2/7 0/7 7/7 6/7 1/7* 0/7* < 7 6 5 4 3 2 1 0 T • ± RSD1000 (umol/kg/min) VT 14/15 VF 13/15 7/7 5/5 6/9 3/9 7/7 3/5 5/9 1/9* < 7 6 5 4 3 2 1 T • 40 g L _ i 1 — i — i — i i i 11 i 1 i i i i d. i i ae i i i 1 10 40 RSD1009 (umol/kg/min) O 108 3.2.3. Naphthalene position: 1- versus 2-naphthalene Based on the above results, the role of an aromatic ring system, and how its electron density is modified, may suggest evidence for the antiarrhythmic actions of the parent compound, RSD 1000 (Figure 9A). The 1-naphthyl group of RSD 1000 is a di-cyclic aromatic system. This group provides an additional four aromatic carbon atoms which, in essence, act as an extended phenyl group. Structurally, RSD 1000 possesses the same aromatic extension length as RSD 1051 but in a more rigid conformation. Moreover, when comparing the antiarrhythmic profile of the 2-naphthyl compound, RSD 1009 (Figure 9B), the positional arrangement of the naphthyl group appears to play a critical role in providing antiarrhythmic "goodness", whereby the 1-naphthyl group is preferred. The potency, slope (i.e., < 1), and efficacy for RSD 1009 were significantly less favorable than that produced by RSD 1000. Of the compounds in the series that were shown to provide antiarrhythmic protection against ischemia-induced arrhythmias as a function of their aryl side chain, RSD 1000 was shown to produce "optimal" antiarrhythmic "goodness". In an attempt to delineate the antiarrhythmic actions of RSD 1000 at a cellular level, the following experiments on RSD 1000 were performed in single voltage-clamped isolated rat ventricular cells. 3.3. Effects of R S D 1 0 0 0 on isolated rat ventricular myocytes Actions of RSD 1000 on IN 3 currents in isolated rat ventricular myocytes at pH 7.3 (•) and 6.4 ( • ) are shown in Figure 10. Original current traces of IN 3 from two different cells (Figure 10A) illustrate the current in the absence (largest current) and presence 109 30 ms -20 mV I—— -70 mV - i -140mV . 60 ms 5 m s c 0.1 1 10 50 RSDlOOO(uM) Figure 10 Effects of RSD 1000 on I N a currents in isolated rat ventricular cells. ( A ) Original I N a current traces from two different cells, one at pH 7.3 (•) and 6.4 (•), in the absence (bottom trace) and presence (top trace) of 2 uM RSD 1000. Holding potential was at -70 mV with a 30 ms depolarizing step to -20 mV following an intial prepulse to -140 mV (see inset for details on pulse protocol). ( B ) Concentration-response plot of RSD 1000 on peak I N 3 amplitude. Current amplitudes were normalized to control (C) currents and expressed as mean ± SEM. In bath solution of pH 7.3 or 6.4, data points (n = 4 - 5 cells) were fit according to the logistic function described in the Methods. The IC50 values at pH 7.3 and 6.4 were estimated to be 2.9 ± 0.3 and 0.8 ± 0.1 uM, respectively. 110 ' -20 mV 3 0 m s * \ -100 mV- x 20 at 20Hz 2 m s 1.0 c CD o "O Q) 0.5 .£>i O £ o 0.0 g ° g H o g i 5 8 0 0 H 9 o @ s 6 s a n p H 7.3 a n d 6.4 5 10 15 Pulse number 20 p H 7.3 + 5 uJvl RSD1000 p H 6.4 + 5 u.M RSD1000 Figure 11 Use-dependent inhibition of I N a by RSD 1000 in isolated rat ventricular myocytes. Currents were elicited by a series of 20 depolarizing steps to -20 mV from a holding potential of -100 mV with stimulation frequency of 20 Hz. (A) Current traces of I N a from two different cells at pH 7.3 and 6.4 in the presence of 5 uM RSD1000. The traces show responses to the 1st, 2 n d , 5 t h, 10 t h, 20 t h steps of the series. (B) Current amplitudes were normalized to the 1s t pulse in each episode and plotted as a function of episode number (1 - 20) at pH 7.3 and 6.4 in the absence (open symbols) and presence (closed symbols) of 5 p M RSD 1000. At pH 7.3 and 6.4 the peak current amplitude was reduced to a steady state level (measured at the 20 t h pulse) of 71 ± 2.2% and 39 ± 4.5%, respectively. I l l + 6 0 m V -70 m V -4 0 0 m s pH 7.3 pH6.4 2 nA 1.0 i c CD o CD o E o 0.5 0.0 100 m s 'TO L L 'KSUS 0.01 0.1 1 RSD1000 (JJM) 10 100 Figure 12 Effects of RSD 1000 on ITO currents in rat ventricular myocytes. ITO currents were recorded with a depolarizing step to +60 mV from a holding potential of -70 mV. (A) Original current traces are shown from two different cells at pH 7.3 (•) and 6.4 (O) in the absence (top trace) and presence of 2 (middle trace) and 30 uM RSD1000 (bottom trace). ( B ) Concentration-response curves for inhibition of ITO as a function of RSD1000 concentration in acid and pH 7.3. ITO currents were expressed as the integral of the inactivating component of ITO current (area of the current trace above the dashed line, see inset) normalized to the integral of control current. " C " denotes control and data points are fitted according to the logistic function described in the Methods. IC50 values were estimated as 3.3 ± 0.4 and 2.8 ± 0.1 uM at pH 6.4 and 7.3, respectively, and this is also given in Table 8. 112 30 u.M RSD1000 -> Control 4nA 25 ms Figure 13 Effects of RSD 1000 on Ic a current in isolated rat ventricular myocytes. Calcium currents were elicited by depolarizing to +30 mV from a holding potential of-70 mV before and after exposure to 30 u M RSD1000. Original traces from one cell at pH 7.3 are illustrated. 113 (smaller current) of 5 u M RSD 1000. The accompanying graph (Figure 10B) represents the concentration-dependent reduction of IN 3 expressed as the current amplitude normalized to control. The reduction of IN 3 was greater in external acid than in normal pH with IC50S of 0.8 ± 0 . 1 and 2.9 ± 0.3 uM, respectively. A similar potentiation of iNa blockade in acid versus normal pH was also found for use-dependent blockade of IN 3 by RSD 1000 (Figure 8). Original current traces of I N a from two different cells (Figure 11 A) are shown in the presence of RSD1000 at pH 7.3 (•) and 6.4 (•). The use-dependent actions of RSD 1000 as a function of 20 pulses in total are shown in Figure 1 IB for both pHs. In the absence of RSD 1000 there was no change in the current amplitude normalized to control at both pHs. In the presence of 5 uM RSD 1000 there was a decrease in current amplitude at pH 7.3 (•) to a steady-state level of 71 ± 2.2% (measured at the 20 pulse) and a greater decrease at pH 6.4 (A) to 39 ± 4.5% with increasing pulses. The blockade of ITO was also concentration-dependent but its effects were unaffected by external changes in pH (Figure 12). Original current traces of ITO are from two separate cells (Figure 12A) at pH 7.3 (•) and 6.4 (O). The reduction of peak and steady-state ITO from control (largest current) was reduced in the presence of 2 and 30 uM RSD1000. This is depicted in the lower graph (Figure 12B) for both pH's as a concentration-dependent reduction of the integral of ITO normalized to control. The estimated IC50S were 2.8 ±0 .1 and 3.3 ± 0.4 uM at pH 7.3 and 6.4, respectively. In addition to iNa and ITO currents, the effect of RSD 1000 on C a 2 + currents was also investigated and this is shown in Figure 13. While blocking both I N a and ITo currents with TTX and 4-AP, 30 uM RSD 1000 was applied with no blocking effect on C a 2 + 114 currents. In the original current traces depicting control and treated conditions for one cell, this is shown as two superimposable C a 2 + currents measured at pH 7.3. 3.4. Influence of pKa and effects on ischemia-induced arrhythmias Having established a general profile of RSD 1000 in in vivo and in vitro rat cardiac preparations, analogues closely related to RSD 1000 were subjected to similar experiments to find additional evidence associated with the antiarrhythmic profile of this compound. To take advantage of this comparison, the compounds were restricted to a small group consisting of the naphthyl moiety, either in the 1'- or 2'-position, and containing different ionizable amine groups. Results from the ischemia-induced study indicate that the nature of the aryl substituent (i.e., 1'-naphthyl) may influence antiarrhythmic activity. A second chemical moiety associated with the general pharmacophore of RSD compounds may also affect antiarrhythmic activity. The following experiments address the role of the ionizable nitrogen group (i.e., morpholinyl moiety of RSD1000). RSD1025 and RSD1015 (Figure 14) are structural homologues of RSD 1000 and RSD 1009, respectively, with an N -methyl-piperazinyl moiety in place of the morpholinyl group. Variations to the amine group at position R 2 effectively changes the dissociation constant, or pKa, of the molecule and these values have been measured (see Methods). The estimated pKa value for RSD1025 and RSD1015 was 8.9 versus 6.1 for RSD1000 and RSD1009. The antiarrhythmic dose-response relationships for the former two compounds (Figure 14) lacked the antiarrhythmic "goodness" of their parent homologues (Figure 9), irrespective of the positional arrangement of the naphthyl group. 115 Figure 14 A B VT 11/11 5/5 4/5 4/7 1/7 V F 9/11 1/5* 2/5 3/7 0/7* C/D < 7 r 6 5 4 3 2 1 0 1 ' ' i i i i i 11 -4B 1 te-l—I I I o in 1 1 0 QTVyj RSD1025 (|jmol/kg/min) O VT 12/16 7/9 7/9 5/7 2/7 V F 15/16 6/9 4/9* 4/7 1/7* 7 r 6 5 4 3 2 . 1 0 < RSD1015 (pmol/kg/min) .j i i 40 116 Figure 15 A VT 12/12 6/7 5/5 2/5 0/5 1/5 VF 11/12 4/7 4/5 2/5* 0/5* 0/5* CO < 7 6 5 4 3 2 1 0 T • O 1 [ Q j R S D 1 0 4 6 (umol/kg/min) 40 -OMe 117 Figures 14 and 15 Effects of RSD1000 analogues with different amine heterocyclic groups. RSD1025, RSD1015 (Figure 14A and 14B), RSD1046 and RSD1049 (Figure 15A and 15B) are investigated against ischemia-induced arrhythmias. At a infusion dose of 4 umol/kg/min RSD 1049, cardiac output failure (COF) was produced. A detailed description of the graph format is given in the Legend of Figures 6-9. Antiarrhythmic ED50 values are given in Table 5. 118 The role of molecular charge associated with antiarrhythmic activity was further explored by including RSD 1046 and RSD 1049 (Figures 15A and 15B), two other analogues of RSD1000 with intermediate pKa values (7.0 and 8.2, respectively) in the range between RSD 1000 and RSD 1025. Both RSD 1046 and RSD 1049 retain the 1-naphthyl group differing only in their nitrogen groups and their estimated pKa values are 7 and 8.2, respectively. A comparison of Figures 9A, 15A, 15B, and 14A indicates that as the pKa value is increased from 6.1 (RSD1000) to 8.9 (RSD1025) antiarrhythmic "goodness" becomes less favorable in terms of reduced efficacy and increased variance of antiarrhythmic response (i.e., slope < 1). 3.5. Electrically-induced arrhythmias RSD 1000, RSD 1046, RSD 1049 and RSD 1025 were investigated against arrhythmias produced by electrical stimulation in non-ischemic myocardium to determine whether the same relationship of antiarrhythmic actions and pKa existed in the setting of ischemic conditions. Arrhythmias produced by electrical stimulation were expressed as the minimum threshold current to capture ventricular fibrillo-flutter (VFt). Therefore, the more blockade of ion channels the compound exhibits in normal myocardium (i.e., versus myocardium altered by ischemia), the higher the current that is required to capture ventricular fibrillo-flutter. The minimum threshold current at each infusion dose for each compound is illustrated in Figures 16A - 16D (RSD 1000, RSD 1046, RSD 1049, and RSD 1025, respectively) as a percentage of the maximum current of 1000 uA (cross-hair symbols ± SEM). Also illustrated in the graphs for each compound are the mean arrhythmia scores (open square symbol ± SEM) reproduced from Figures 9A, 15A, 15B, RSD1046 (umol/kg/min) p K a = 7.0 RSD1025 (umol/kg/min) pKa = 8.9 121 Figure 16 Effects of RSD 1000 (A), RSD 1046 (B), RSD 1049 (C), and RSD 1025 (D) in the electrically-induced arrhythmia rat model. Current was applied using square pulses at 50 Hz to produce ventricular fibrillo-flutter (see Methods). The minimum current threshold to capture ventricular fibrillo-flutter (VFt) was determined (crossed hairs ± SEM) in the presence of RSD compound. This was expressed (right y-axis) as a percentage of the maximum current that was provided (1000 uA) as a function of infusion dose. Data points were fit using nonlinear least-square equations and infusion doses producing a 25% change from control were estimated and these are listed in Table 5. In order to compare the actions between arrhythmias produced by myocardial ischemia and electrical stimulation, the mean AS values (left y-axis) from the ischemia-induced arrhythmia rat model and their logistic lines of best fit were reproduced in these graphs for each RSD compound. 122 and 14A, respectively, to compare the actions against arrhythmias induced by electrical current versus myocardial ischemia. A l l of the RSD compounds produced a dose-dependent increase in VFt. The relationship with pKa was found to be directly proportional to the minimum current required to increase VFt. As the pKa increased, the separation of responses from electrically- and ischemia-induced arrhythmias overlapped with each other. For RSD 1049 and RSD 1025, these were the only compounds out of the series that the applied threshold currents significantly exceeded 50% of the maximum over their tested infusion dose range; these were estimated as 0.7 ± 0.3 and 2.5 ± 0.8 umol/kg/min, respectively. For ease of comparison, infusion doses producing a 25% (D25%) change from control were estimated for all four compounds (Table 5). In conclusion, both RSD 1000 and RSD 1046 were less effective than RSD 1025 and RSD 1049 in inhibiting electrical induction of arrhythmias in normal tissue. As in the ischemia-induced arrhythmia study, RSD 1049 depressed cardiac function as reflected by a decrease in mean blood pressure and heart rate at an infusion rate of 4 umol/kg/min. Therefore, VFt measurements for this compound at this rate of infusion were not obtained. This form of toxicity was not observed with any of the other agents at equal or higher infusion dose level(s). In fact, RSD 1000 at an infusion rate of 16 umol/kg/min required one-fourth of the maximum current for the induction of VFt. RSD 1025 was the only agent that required a current threshold of 1000 uA for the capture of ventricular fibrillo-flutter at its maximum tested infusion dose. Only RSD 1000 and RSD 1046 with pKa's below physiological pH exhibited significant separation between responses. 123 Figure 17 124 Figure 17 The inhibitory actions of RSD 1000, RSD 1025, RSD 1046 and RSD 1049 in cloned sodium channels. Cells were held from -120 mV and depolarized to -10 mV using a 20 ms pulse duration and the peak inward current was measured in either pH 7.3 or 6.4 external bath. The current amplitudes were normalized to control and plotted as a function of RSD concentration. Data points (± SEM) were fit according to the logistic equation described in the Methods and the estimated IC50 values are listed in Table 7. Note, the maximum soluble concentration for RSD 1000 at pH 7.3 was 300 uM. 125 Table 7 A List of IC 5o values for inhibition of I N a current in SkM2 rat heart sodium channel isoform. RSD (PKa) IC50 (HM) IC50 (uM) pH 7.3 H pH 6.4 H 1000 (6.1) 335 ±27.6 1.7 126 ± 16.2 0.8 1046 (7.0) 138 ± 14.3 0.9 92 ±4.9 1.0 1049 (8.2) 143 ± 19.8 0.9 126 ±25 0.9 1025 (8.9) 85 ±6.9 1.0 32 ±3.3 1.0 B List of IC50 values for inhibition of I N a currents in rat ventricular myocytes. RSD IC50 (uM) IC50 (uM) pH 7.3 H pH 6.4 H 1000 2.9 ±0.3 1.1 0.8 ±0.1 0.8 1009 29±3.1 1.7 — * _ _ _ * 126 Table 7 ( A ) Inhibition of heart sodium isoform channels by RSD1000, RSD1025, RSD1046 and RSD 1049 at pH 7.3 and 6.4. The pKa values for each compound is listed in the brackets. IC50 values are mean ± SEM and these were estimated from the logistic fits in Figure 17 along with their corresponding Hi l l coefficient values. (B ) In rat ventricular myocytes, the IC50 values (± SEM) for inhibition of I N a in acid and pH 7.3 by RSD1000 and RSD 1009 are also listed. These values were estimated from the logistic fits illustrated in Figures, 10 and 19, respectively. The "*" indicates that due to insufficient concentrations test for RSD 1009, no estimate of the IC50 was made. 127 3 .6 Effects of R S D 1 0 0 0 , R S D 1 0 4 6 , R S D 1 0 4 9 , and R S D 1 0 2 5 on cloned rat heart sodium channels The role of pKa was further investigated with the above four compounds on cloned rat heart sodium channels, a cell line in which there is adequate control of inward sodium currents versus channels in native cells. The channels were expressed in Xenopus oocytes and studied using a two-electrode voltage-clamp. Figure 17 shows the concentration-response curves for each drug on the sodium current from the cloned rat heart sodium channel in a clockwise direction: RSD1000, RSD1025, RSD1049 and RSD1046. Oocytes were held at -120 mV and currents were evoked by depolarizations to -10 mV. RSD1025, RSD1046 and RSD1049 blocked cardiac I N a currents with greater potency than RSD 1000. The highest soluble concentration of RSD 1000 that was possible was 300 u M at pH 7.4; thus, the reduction of I N a current at this pH was limited to 50%. When IC50 values were compared, RSD 1025 block of the heart sodium channel was 4-fold greater than for RSD 1000 (Table 7). In addition, the IC50 for block of sodium channels was 2- to 3-fold greater for RSD 1046 and RSD 1049 than for RSD 1000 (Table 7), suggesting that these structurally-related analogs of RSD 1000 exhibit a greater affinity for blockade of the cardiac isoform of the sodium channel. Unlike local anesthetic drugs that have also been studied in the oocyte expression system, such as lidocaine with an I C 5 0 of 560 ± 22 uM (Pugsley and Goldin, 1998), the RSD compounds studied herein demonstrated a higher potency of block of the rat heart sodium channel isoform. The IC50 values for producing inhibition of IN 3 by most of the RSD compounds in pH 0 7.4 and 6.4 were shown to increase with increasing pKa with greater potency in acid versus normal pH. In acid pH, the rank order of potency remained unchanged. The 128 potency shifts were greatest for RSD 1025 and RSD 1000, minimal for RSD 1046, and unchanged for RSD 1049. These data suggest that this series of 1-naphthyl analogues are potent sodium channel blockers of the rat heart isoform with a blocking potency dependent on pH and presumably a function of molecular pKa. 3.7 Effects of RSD1000, RSD1009, RSD1015 and RSD1025 on I T O currents in isolated rat ventricular myocytes It remains to be established whether the degree of ionization of a compound alone is sufficient to explain the observed potentiation of IN 3 blockade, at least in this series of compounds. Both RSD1000 and RSD1009 have the same pKa's, yet, a slight change in the naphthyl position appears to significantly change the antiarrhythmic activity against ischemia-induced arrhythmias. The next sets of experiments were designed to investigate RSD 1009 on both I N a and Ixo and to compare these results with RSD 1000. The question that was asked is whether a difference existed in the blockade of IN 3 or ITO, or both, in the presence of RSD 1009 that could account for its reduced antiarrhythmic profile in vivo compared to RSD 1000. Initial studies began with ITO currents on the assumption that the low pKa (6.1) of RSD 1009 would produce a similar potency shift as RSD 1000 on I N a currents and that the antiarrhythmic difference between these two compounds might be explained by differences in the blockade of ITO- In addition, the structural homologues, RSD 1015 and RSD 1025, were included to determine whether the naphthyl configuration and pKa are interdependent or independent variables. The blocking actions of RSD 1009, RSD 1025 and RSD 1015 were shown to accelerate the rate of decline and the peak amplitude of ITO and these are shown in Figure 129 Figure 18 (uM) 130 Figure 18 Inhibition of rat ventricular I T 0 currents by RSD 1009, RSD 1025 and RSD 1015. For each compound, current traces (insets) are from two different cells at pH 7.3 (•) and 6.4 (O) showing the control current (upper trace) and in the presence of 2 (middle trace) and 30 uM (bottom trace) concentrations. Cells were held from -70 mV and depolarized with a 400 ms pulse to +60 mV. The inhibition of ITO by each compound, at pH 7.3 and 6.4, was plotted as the integral of the fast inactivating minus the non-inactivating component of the current (for more details, see Figure 15 and Methods). Treated responses were normalized to the integral of control currents (C) and plotted as a function of RSD concentration. Data points were fitted according to the logistic function described in the Methods and the IC50 values for all three compounds were estimated and are given in Table 8. 131 Table 8 List of I C 5 0 values for inhibition of I T 0 currents by RSD 1000, RSD 1009, RSD 1025 and RSD1015. Values are mean ± S E M of number (n) of individual cells at pH 7.3 and 6.4 and the Hi l l coefficient for each RSD compound. Both sets of values were estimated from fits of the logistic equation described in the Methods section and applied to Figures 12 and 18. RSD IC 5 0 0iM) IC 5 0 (UM) pH 7.3 n H pH 6.4 n H 1025 2.1 ±0.2 6 1.1 2.6 ±0.2 7 0.8 1015 1.4 ±0.3 5 0.8 2.0 ±0.1 5 0.6 1000 3.4 ±0.3 6 0.8 4.1 ±0.2 6 1.1 1009 1.5 ±0.3 5 0.9 3.0 ±0.6 5 1.1 132 18. The effects of these three compounds on the rate of inactivation of ITO were shown to be concentration-dependent and equally potent (Figure 18; Table 8). When extracellular pH was decreased, the blocking potencies of all three compounds were slightly decreased (not statistically significant) but mainly remained unaffected compared with pH 7.3. The blocking properties of these three compounds were not different from those produced by R S D 1 0 0 0 (compare with Figure 12; Table 8). In conclusion, the ITO blocking actions of this series of R S D compounds appears to be independent of the positional arrangement of the aryl substituent and/or the nature of the amine heterocyclic group. More specifically and, perhaps, more relevant was the pH-independent actions of ITO blockade that was determined with two structurally-related compound series differing only in their dissociation constants (pKa's). 3.8. Effects of R S D 1 0 0 9 I N 3 currents in isolated rat ventricular myocytes The assumption that the low pKa (6.1) of R S D 1009 would produce a similar potency shift as R S D 1000 on I N a currents was investigated. R S D 1009 was tested in isolated rat ventricular myocytes with external pH buffers of 7.3 (•) and 6.4 ( • ) . Figure 1 9 A illustrates original current traces from one cell at both pH's and in the presence of 30 uM R S D 1009 at pH 6.4. Cells were held at -70 mV and hyperpolarized to -140 mV before depolarization to -30 mV. The inhibitory responses to IN 3 in the presence of R S D 1009 were normalized to control current amplitudes and plotted as a function of concentration (Figure 19B). The IC50 at pH 7.3 was estimated as 29 ± 3.1 u M (see also Table 7). The lack of cells and concentrations (< 10 uM) tested at pH 6.4 did not allow a proper estimate of an IC50. An "eye-ball" comparison between the concentration-133 response data of Figure 19B suggests that inhibition of IN 2 by RSD 1009 is greater in acid pH. However, a comparison of RSD1009 with RSD1000 studied in the same cell line and under similar buffer conditions (Figure 10; Table 7) indicates that the IN 3 blocking potency for RSD 1009 at p H 0 6.4 and 7.4 is less than that observed for RSD 1000. 3.9 Correlation matrices In an attempt to extend the present structure-activity comparisons and find possible relationships between the biological and chemical parameters reported in this study, correlation matrices were constructed between pKa or log Q versus AA E D 5 0 and the results are represented in Table 9. In addition, relationships between A A E D 5 0 and pKa versus cardiac and cardiovascular parameters were also performed. The values for the latter group are the same as those listed in Table 5 and they represent the infusion doses producing a 25% change from control (following a 5 min infusion) as studied in the ischemia- as well as the electrically- (i.e., iT, ERP, VFt) induced arrhythmia model. For ease of presentation, values are listed without SEMs. Since it was not possible for most RSD compounds to estimate either D 25% or A A ED50, all values, except for pKa, were transformed to orders of rank (from highest to lowest infusion dose; "NE" was considered highest) and the resulting correlations were made. An upper r value of 0.5 (r = 0.7; p < 0.05) was set and for each array, the correlation index, r 2, was calculated. Poor relationships (r2 < 0.5) were found between biological and chemical activities. In addition, A A ED50 was found to be independent of changes to cardiac parameters (e.g., ECG, as indirect indices of sodium and potassium channel blockade). A statistically significant relationship was found between pKa and effects for reducing blood pressure. 134 135 Figure 19 The effects of RSD 1009 on IN 3 currents in rat ventricular myocytes. Cells were held at -70 mV and hyperpolarized to -140 mV before depolarizing to -30 mV. Examples from one cell ( A ) illustrate original peak I N 3 current traces (arrows) at pH 7.3, 6.4, and 6.4 with 30 uM RSD 1009. Current responses were normalized to control amplitudes and plotted as concentration-response relationships (B) at pH 7.3 (•) and 6.4 ( • ) . Data points were fit using the logistic equation described in the Methods and IC50 at pH 7.3 (for n = 4 cells) was estimaed as 29 ± 3.1 uM (see also, Table 7). For details on pH 6.4 data, see Results section. Table 9 Correlation matrices A) pKa and logQ values versus A A ED 5o RSD pKa L o g Q( PH 7.4) L o g Q( PH 6.4) A A E D 5 0 rank 1000 6.1 1.87 2.49 2.5 6 1053 6.45 1.71 2.36 1 1.5 1010 6 2.16 2.81 4.6 9 1050 6.5 2.67 3.32 14 13 1051 6.45 3.13 3.78 4.6 9 1012 6 1.61 2.25 5.3 11 1072 6.35 2.59 3.24 8.8 12 1014 6 1.63 2.28 N E 14 1019 6.1 2.93 3.58 2.9 7 1009 6.2 2.93 3.58 4.6 9 1015 8.9 2.44 3.36 2 5 1025 8.9 2.44 3.36 1 1.5 1046 7 3.45 4.01 1.5 4 1049 8.2 3.15 4.03 1 1.5 r 2 for... pKa 0.14 0.4 L o g Q( PH 7.4) <0.1 <0.1 L o g Q( PH 6.4) <0.1 0.1 B) Blood pressure and heart rate versus A A ED 5o RSD 4BP rank rank A A E D 5 0 rank pKa 1000 8 4 8 5.5 2.5 6 6.1 1053 32 5.5 32 8.5 1 1.5 6.45 1010 N E 10.5 N E 12 4.6 9 6 1050 N E 10.5 N E 12 14 13 6.5 1051 N E 10.5 N E 12 4.6 9 6.45 1012 N E 10.5 N E 12 5.3 11 6 1072 N E 10.5 N E 12 8.8 12 6.35 1014 N E 10.5 8 5.5 N E 14 6 1019 N E 10.5 16 7 2.9 7 6.1 1009 N E 10.5 N E 8.5 4.6 9 6.2 1015 3 3 2 3 2 5 8.9 1025 2 1 1 1 1 1.5 8.9 1046 8 5.5 3 4 1.5 4 7 1049 2 1 1 1 1 1.5 8.2 r 2 for... BP(rank) HR(rank) 0.38 0.49 0.72* 0.47 0.66* 0.54 137 C) ECG parameters versus AA ED 5 0 RSD tP-R rank tQRS rank TQ-T rank A A E D 5 0 rank 1000 N E 9 N E 8.5 N E 11.5 2.5 4 1053 N E 9 N E 8.5 16 6 1 1 1010 N E 9 N E 8.5 N E 11.5 4.6 6 1050 N E 9 N E 8.5 N E 11.5 14 9 1051 N E 9 NE 8.5 N E 11.5 4.6 6 1012 N E 9 NE 8.5 N E 11.5 5.3 7 1072 N E 9 N E 8.5 8 5.5 8.8 8 1014 N E 9 N E 8.5 10 5 N E 10 1019 N E 9 N E 8.5 8 5.5 2.9 5 1009 N E 9 N E 8.5 N E 11.5 4.6 6 1015 5 2 6 2 5 4 2 3 1025 8 3 NE 8.5 1.5 1 1 1 1046 N E 9 N E 8.5 2.5 3 1.5 2 1049 1 1 4 1 1.5 1 1 1 r2for... P-R(rank) QRS( r ank) Q-Tjrank) 0.16 0.09 0.28 0.34 0.18 0.32 D) Electrical stimulation parameters versus AA ED50 RSD t i T rank t E R P rank t V F t rank A A E D 5 0 rank pKa 1000 15 5 11 8 8 7.5 2.5 4 6.1 1053 N E 11.5 16 11 8 7.5 1 1 6.45 1010 32 7.5 8 6 12 10.5 4.6 6 6 1050 N E 11.5 16 11 12 10.5 14 9 6.5 1051 N E 11.5 16 11 8 7.5 4.6 6 6.45 1012 N E 11.5 16 11 16 13.5 5.3 7 6 1072 N E 11.5 8 6 8 7.5 8.8 8 6.35 1014 N E 11.5 8 6 6 5 N E 10 6 1019 32 7.5 16 11 16 13.5 2.9 5 6.1 1009 18 6 17 13 15 12 4.6 6 6.2 1015 4 3.5 3 4 4 3.5 2 3 8.9 1025 2.5 2 1 5 2 2 1 1 8.9 1046 4 3.5 1 5 4 3.5 1.5 2 7 1049 1 1 0.5 1 1 1 1 1 8.2 r 2 for... i T ( r a n k ) ERP(rank) VFt ( r a n k) 0.33 0.14 0.22 0.52 0.38 0.29 0.50 0.41 0.56 138 Table 9 Correlation matrices between biological and chemical activities. The following correlation matrices investigate the dependence of antiarrhythmic E D 5 0 (AA E D 5 0 ) on pKa and log Q for 14 RSD compounds (A). (B-D) The relationships for A A E D 5 0 or pKa versus blood pressure (BP), heart rate (HR), E C G and electrical stimulation variables were also investigated. Values are those producing 25% change from control following 5 min infusion as studied in the ischemia- and electrically- (iT, ERP, VFt) induced arrhythmia models. These values are listed in Table 5 and for ease of presentation, SEMs are not shown. Except for pKa, correlation analyses was also performed on rank order (from highest to lowest; " N E " was considered highest). When two or more values had exactly the same value, the assigned rank to each of the tied ranks was the average of the ranks that would have been assigned to these ranks had they not been tied (e.g., (3+4)/2 = 3.5). Estimates of the correlation index, r 2, involving A A E D 5 0 was performed for both raw and ranked data. The correlation index, r, was transformed using Fisher's z values and a one-tailed statistical analysis was performed to determine *p < 0.05 significance. 139 4. Discussion This discussion has three major aims. The first is to consider how the RSD compounds used in this study compared with standard antiarrhythmics in terms of their effects on ischaemia-induced arrhythmias in the rat. The second aim is to consider the overall cardiovascular activity (i.e., BP, HR, ECG, electrical stimulation parameters) of the chosen RSD compounds with a view to explaining how some of the compounds were effective antiarrhythmics against ischaemia-induced arrhythmias whereas others were not. The final aim relates to the SAR of the compounds in that by an overall comparison of all compounds (as well as group and pairwise comparisons) with respect to structure and action, it might be possible to offer clues as to the chemical determinants of their actions. 4.1 Biological and chemical activity of conventional antiarrhythmics How do the biological activities and chemical structure of the RSD compounds compare with those of conventional antiarrhythmics? In order to examine the relationship between the RSD compounds used in this study and existing antiarrhythmics, the following discussion will be concerned with similarities and differences between them from both biological and chemical viewpoints. As a preamble to this discussion, it is worth summarizing the present clinical situation regarding the efficacy of conventional antiarrhythmics against ischemia-induced arrhythmias in man. To date, the role of cardiac sodium channel blockers continues to be questioned and data from clinical trials (e.g., CAST; Echt et al., 1989) indicate that the use of this class of drugs should be limited to the control of symptoms in patients who have arrhythmias and either no, or minimal, heart disease. Therefore, attention is now shifting 140 away from specific ion channel blockade (including specific I K blockers such as d-sotalol; for reference, see SWORD Trial; Waldo et al. 1995, 1996) to antiarrhythmic therapy provided by a combination of reducing sympathetic stimulation and lengthening repolarization and effective refractory period (for a review, see Singh, 1998). Agents of this class include beta-blockers and complex agents with class III properties, such as amiodarone and d,l-sotalol. The superiority of the latter two agents as compared to their specific ion channel blocker predecessors is accentuated by the results of recent controlled clinical trials (Mason et al., 1993; C A S C A D E Investigators, 1993). However, it is not clear which components of the electrophysiologic properties of amiodarone and sotalol are linked precisely to their clinical antiarrhythmic actions. This translates to non-selective modulation of ion channels and/or receptors other than those involved in arrhythmogenesis with the potential for drug-induced toxicity and proarrhythmia. Thus, the clinical utility of sotalol and amiodarone remains to be clearly defined and there is still no current antiarrhythmic, with the exception of beta-blockers (which act by a secondary mechanism, i.e., anti-ischemia by reduced sympathetic activity; Anderson et al., 1983; Yusuf et a l , 1985; Kendall et al., 1995; Kennedy, 1997a, 1997b), that prevent the occurrence of V F in the setting of myocardial ischemia and infarction without incurring unacceptable toxicity or side effects. With respect to the actions of existing antiarrhythmics against ischemia-induced arrhythmias in an experimental setting, there is an extensive literature describing results obtained with different drugs in a wide variety of models and species as discussed in the Introduction section and summarized in Table 10. There are many opposing and complementary ideas of drug action that are inherent with such an array of animals and 141 Table 10 Conventional antiarrhythmic agents against ischemia-induced arrhythmias in the early phase as tested amongst different animal species. Species Antiarrhythmic agent Year Reference dog nicorandil (K A T P channel 1998 Shinohara et al. opener) Mayuga and Singer amiodarone 1992 propranolol 1988 Ohyanagi et al. bretylium 1983 Fujimoto et al. diltiazem 1981 Fujimoto et al. disopyramide 1979 Levites and Anderson aprinidine 1977 Elharrar et al. quinidine verapamil IPNA rat a-antagonists (yohimbine, 1996 Roegel et al. prazosin) Adaikan et al. tedisamil 1992 mexilitine 1992 Igwemezie et al. atenolol 1989 Paletta et al. propranolol prazosin 1988 Kinoshita et al. atenolol desethylamiodarone 1987 Varro et al. amiodarone 1986 Varro and Rabloczky verapamil 1983 Johnston et al. lidocaine disopyramide quinidine Lepran et al." lidocaine 1983 pindolol Pig anipamil (verapamil analogue) 1995 Pugsley et al. 142 models but the general consensus underlying these theories is the importance for voltage-and time-dependent actions of any drug on cardiac ion channels (Hondeghem and Katzung, 1977). Lidocaine is one example of a drug that embodies the qualities that are optimal for voltage- and time-dependent actions on cardiac I N a which are effectively enhanced in conditions of myocardial ischemia (Kupersmith, 1979; Evans et al., 1984; Carson et al., 1986; Wendt et al., 1993). Unfortunately, its antiarrhythmic actions are not restricted to the ischemic tissue or the heart and as a consequence of its extra-cardiac toxicity (e.g., CNS and cardiovascular), its clinical utility, as with many other agents, is limited (Velebit et al., 1982; Abbott, 1981; Pentecost et al., 1981; Rademaker et al., 1986; Denaro and Benowitz, 1989; Jaffe, 1993). In view of the unknown effects of species and models on many experimental results, any meaningful comparison of the RSD compounds and existing antiarrhythmics should be confined to findings in rats using the model of ischemia-induced arrhythmias described in this thesis. These rat data will be considered by comparing and contrasting biological findings and chemical attributes of the RSD compounds with existing compounds recognizing that the latter are a heterogenous group whereas the RSD compounds are closely related, both biologically and chemically. The specific drugs considered are shown in the following table (Table 11) and are listed in accordance to their relative potencies for blockade of IN a, I T O , or both.. The discussion will focus particularly on antiarrhythmic drugs that are mixed channel blockers of I N a and I T 0 currents, i.e., those that are apparently similar to the RSD compounds with respect to their electrophysiological actions. Understandably, there are valuable insights into drug action 143 Table 11 Conventional antiarrhythmics and their potency on rat IN a and I T 0 currents. The following class I and III drugs and their structures are listed according with literature reports of their concentrations for producing 50% inhibition on IN a and I T 0 current in rat ventricular cells performed in whole-cell studies. In cases where data are not available, this is denoted as Agent Chemical structure Rat ventricular cell 1C 5 0 IN* ^TO Lidocaine (class lb) CH, 20 uM (Lee et al., 1981) 30 uM (Xu et al., 1992) — Mexilitine (class lb) C B , / Q V - O - C H 2 - C H - N H 2 ( C H , C H , 30 uM (Yatani and Akaike, 1985) — Flecainide (class Ic) OCHJCFJ x ^ ^ ^ - ^ j - O (this lab's C F J C ^ O observations) 4 U M (Slawsky et al., 1994) Quinidine (class la) O C K , 1 * C H , kJ"V> 20 uM (Lee et al., 1981) 40 uM (Ju et al., 1992) 3 uM (Jahnel et al., 1994) 6 uM (Clark et al., 1995) Dysopyramide (class la) 9 O 30 uM (Yatani and Akaike, 1985) 300 uM (Sanchez-Chapula, 1999) Tedisamil (class III; IT 0) 20-50 uM (Dukes and Morad, 1989) 5 uM (Dukes and Morad, 1989) Clofilium Cl (O/ ( C H 2 ) " _ 50 uM(Castle, 1990) 144 from other species and experimental protocols that warrant recognition and these will be mentioned wherever applicable. Except for clofilium, all other agents have been studied in the rat coronary occlusion models as used in this study (Johnston et al., 1983; Adaikan et al., 1992; Igwemezie et al., 1992; Barrett et al., 1995). The results demonstrated a limited range of antiarrhythmic protection with generally poor antiarrhythmic potency, efficacy and low therapeutic ratios. On a cellular level, these agents have been identified as blockers of I N a and/or I T 0 currents in rat ventricular cells. A quantitative association between biological and chemical activity amongst this list of agents may not be immediately clear. In fact, similar quantitative structure-activity analyses have been attempted for cardiac sodium channel blockers and the general consensus is that there is no single chemical activity or common value that can be used to definitively characterize drug-channel interaction. This appears true for most quantitative structure-activity relationships (QSAR) since the sequences leading to a response or the drug-binding site complex itself may be inherently multifaceted and require all available measures of chemical activity to unravel relationships. Courtney has been the main proponent for QSAR studies on antiarrhythmics. In a series of experiments (1980; 1987; 1990) on sodium channels blockers, attempts were made to associate physiochemical attributes with potency for blockade of IN a. In his latest model and size/solubility hypothesis (1990), he discusses how, for a given series of sodium channel blockers, their different rates of recovery from I N a block are a function of more than one of their physiochemical attributes. Thus, the rate of recovery is influenced 145 by molecular weight, end-on molecular dimensions or size, hydrophobicity (log Q or P) and pKa. Based on the classification scheme in Table 11, chemical similarities or profiles between these agents begin to emerge and a qualitative structure-activity comparison can be considered. Briefly, the blockers that are listed for I N a (la - Ic) all share several common chemical features: a 2° or 3° amine group, a lipophilic group and a linker chain connecting both groups. For local anesthetics (e.g., lidocaine and mexilitine), these subgroups and the overall profile are distinct and generic (Adamson and Bush, 1976; Strother et al., 1977; Bokesch et al., 1986; Sheldon et al., 1990). In contrast, the blockers listed for I T O do not share this profile. Instead, "bulkiness" appears to be a shared characteristic amongst the la antiarrhythmics, tedisamil and clofilium (and perhaps, flecainide) with a 3° nitrogen group. However, there are no distinct chemical features that would suggest a blocking combination for both currents (i.e., flecainide, quinidine and disopyramide). The essence of this thesis was that structural similarities and chemical modifications within the array of RSD compounds would allow an investigation into the chemical structure of a selective antiarrhythmic pharmacophore. However, there are limitations to the conclusions from the results of this thesis and these wil l be briefly discussed. 4.2. Limitations Several limitations prevented a precise determination of the chemical requirements against ischemia-induced arrhythmias and these wil l be addressed in 146 reference to both in vivo and in vitro experiments in this study. It was not possible within a limited time-frame to investigate all RSD compounds in all of the experimental protocols and procedures presented here. There was a limited availability of compounds and resources to synthesize compounds, for the purpose of exploring alternative variations within the general RSD pharmacophore. As a result, large chemical changes (i.e., different side groups while maintaining the overall pharmacophore) were studied rather than the minute and serial changes (often small changes on substituents) normally associated with conventional SAR studies. For these reasons, there are missing or unavailable parameters between pairs, or groups, of compounds, thereby limiting the discussion and comparisons of structure and function between groups of two and four compounds rather than an overall global comparison. As with the conventional antiarrhythmics discussed above, the discussion on RSD compounds will be divided into two main sections, biological and chemical activity. Making this distinction may make it possible to compare general chemical groups and their activities rather than specific groups. In doing so, the lack of precision associated with slight and sequential chemical changes may be overlooked. The use of the rat as an appropriate species for coronary artery ligation studies has been previously addressed (Curtis et al., 1987). For any in vivo experiment, the choice of test species depends to a large extent upon the goals of the investigator and the underlying aims of the study. The general aim of this study was to assess the antiarrhythmic activity of novel compounds against ischemia-induced arrhythmias in a 147 model with a minimal collateral network (see Meesman, 1982; Maxwell et al., 1987) so as to ensure minimal variance. The clinical relevance of this species to man was not an issue in this study since this is a discussion that remains unclear for all animal models. A comparative study by Bergey et al. (1982) concluded that rat, pig and dog models of myocardial ischemia are quantitatively different from each other and, it was not possible to estimate, to any degree of certainty, a measure of "clinical relevance" amongst the three groups. The high heart rate and brevity of ventricular AP are two main shortcomings when discussing the clinical relevance of the rat. Although the rat may not be identical with man in terms of cardiac anatomy or electrophysiology, this should not make it less valid than other animal models and preclude its use as a "first-line" and reproducible antiarrhythmic screen. Diagnostic measures of cardiac arrhythmias should be consistent and reproducible and this matter is discussed in detail by Curtis and Walker (1988) with respect to quantification of arrhythmias using scoring systems. The arrhythmia scores and the format used in this study were a convenient means to facilitate the analysis of ischemia-induced arrhythmias. Two measures were unequivocal as endpoints: survival with 0% incidence of arrhythmias and arrhythmia- (i.e., VF-) induced mortality. However, the precision of defining the severity of arrhythmias between these two endpoints becomes less absolute. Therefore, unless the n size is sufficiently high to compensate for this lack of precision, this clearly constitutes a possible source of variance and, perhaps, herein lies the source of the variance in estimating antiarrhythmic E D 5 0 values for most of the RSD compounds. This inherent variance remained constant throughout the arrhythmia 148 analysis since the scoring system was used consistently for each animal. The chemical stability of each RSD compound was not addressed in this study. Differences in their metabolic breakdown (see above) may reflect differences in their activity and, more importantly, in their potential for toxicity. The infusion regime used in this study may have resulted in an over- or underestimation of the effective doses producing changes in one or all parameters. However, the minimal toxicity of RSD 1000 reported in the sham occlusion as well as the high proportion of RSD compounds that produced changes in cardiac and cardiovascular parameters (BP, HR, ECG) that were 25% or less from control may reflect a tolerability of these compounds in vivo. Additional toxicity and activity from metabolites are not ruled out. Is a 5 min pre-occlusion infusion time interval sufficient for plasma concentrations to reach "active" compound levels? The aim of the infusion regime used in this study was to achieve a pseudo steady-state plasma concentration level. In a single exposure situation (i.e., bolus intravenous administration) a rapid a-phase redistribution would be expected without a sustained "peak" plasma concentration for the duration required to assess the antiarrhythmic activity of the RSD compounds. As a result, such a mode of administration would lead to increased variances particularly in light of the limited pharmacokinetic data of these compounds. Thus, the infusion protocol was chosen based on several assumptions. If the volume of redistribution is large such that the time interval to saturate the a-phase is long, the attainment of some peak plasma concentration will equally be 149 prolonged (Model B ; Figure 2 0 ) . Conversely, the time to saturate the a-phase will be reduced i f the volume of redistribution for this initial phase is small (Model A ; Figure 2 0 ) . Figure 2 0 illustrates a fictional compound that is administered as an infusion in two different models calculated with "small" and "large" volumes of distribution for both o c -and P-phases. Models with "small" and "large" Vd (Similar infusion rate and t^ 's ) • Model A; V D =0.2 (a) , 20(P) A Model B; V d = 20(a), 0.2(0) A A A A A A A A A A A A A A A A A A A A A 1 1 1 1 1 1 0 20 40 60 80 100 Time (min) Figure 20 In both models, the compound is specified with the same half-life and infusion rate (IR) and its theoretical plasma concentration (Cp) is plotted as a function of infusion time (t). The data is generated from the following equation, C p = [ V a • (1/IR) • e-V']+ [V p«(l/IR) • eV] where V a and Vp are the volumes of distribution of the model and K a and K p are the o J 6 0 c o o CD E to CO 40 "St 20 o n J 150 equilibrium constants for a- and P-phase redistribution, respectively. The assumption is that the infusion protocol used in these studies closely approximates those of Model A in which saturation of the a-phase occurs in a shorter time span than those depicted by Model B. For most RSD compounds investigated in this study, detectable effects were observed after the first 2-3 min of infusion. Where pharmacokinetic studies have been performed with similar Nortran compounds, it was found that over the period of 1 to 15 minutes following the start of in vivo infusion, serum concentrations rose 6% to 55% between sample times (personal communication from Dr. Lillian Clohs at Nortran Pharmaceuticals Inc.). At the level of the whole cell, current recordings did not address the effect of compound and/or compound structure on different channel states. Selectivity for ischemic tissue may be due to factors other than the degree of ionization and distribution differences of charged and uncharged drug forms in normal versus ischemic myocardium. The IC 5 0 values reported for those RSD compounds studied in whole cells does not take into account the possibility of affinities to different channels states. Similar to class lb antiarrhythmics, selectivity for ischemia-induced arrhythmias may be the result of a higher binding affinity for the inactivated sodium channel state. The positive correlation observed between pH-dependent actions (at a constant holding potential) on I N a in whole cell and selectivity for ischemia-induced arrhythmias for a few selected RSD compounds emphasized the role of pKa for "ischemia-selectivity" rather than on selectivity for channel states. Affinity for different channel states, however, is not ruled out and this forms the basis for future studies. 151 When the membrane potential is stepped, there is a current transient required to charge the cell membrane capacitance. The activation time of the high speed I N a current follows closely on the decay of this initial capacity current and it is usually impossible to completely separate the two (see Beeler and Reuter, 1970; Johnson and Leiberman, 1971; New and Trautwein, 1972; Noble, 1975). The time-constant required to charge the membrane is the product of membrane resistance (R.J and capacitance (C^, i.e., x = R m C m . In the "whole-cell patch" voltage clamp configuration, one micropipette is used for both voltage recording and current passing. Once whole-cell mode is achieved, the access resistance (Ra) of the pipette is in series with the cell membrane (R,,, and C . J and since R n , » R a , a good approximation of T « R a C m . The membrane potential in series with R a is equal to V c m d - I m R a and the result is an increase in the charging time (x) by the relationship, Vm(t) = V ^ l - e ^ ) , where V i n f is the steady state (or equilibrium) potential. Through electrical means afforded by the amplifier used in voltage clamp, attempts are made to compensate for R a and C m but often, with minimal success. Much of the problem lies in the large values required for R a and C m compensation (e.g., for a typical mycocyte, C m > 100 pF). Generally, current amplifier systems are limited to certain levels of compensation without the risk of large positive feedback signals that invariably produce instability of the measured current, oscillation and/or the electronic circuits to become saturated. On this basis, it was often difficult to find some resolution of the peak I N a signal from the capacitive transient. The low-pass filtering setting of 1 kHz that was used only further attenuated the signal recording of I N a . Thus, such issues raised concerns about the quality of the I N a measurement in this study. Solutions to this problem include a reduction of R a , by adjusting pipette tip diameter or the selection of smaller sized 152 myocytes to reduce C m . 4.3 RSD compounds: Biological activity This study looked at the actions of the chosen compounds on several biological activities: A) In vivo with i) antiarrhythmic activity against ischemia-induced arrhythmias, ii) electrically-induced arrhythmias and iii) cardiac and cardiovascular actions; B) In vitro with i) repolarizing, and ii) depolarizing currents in isolated rat ventricular cells (the latter was also investigated in cloned sodium channels). 4.3.1. Antiarrhythmic activity against ischemia- and electrically-induced arrhythmias Antiarrhythmic responses against ischemia-induced arrhythmias were measured as incidences of VT, V F and values of AS. The variance, efficacy and potency for these responses were dependent on the compound. Some compounds suppressed ischemia-induced arrhythmias in a predictable and distinctive manner with antiarrhythmic responses occurring at doses having minimal or no effects on blood pressure or the normal heart (ECG and electrical stimulation). The antiarrhythmic profile for these compounds fulfilled the criteria that have been used described as antiarrhythmic "goodness" (see Methods section, 2.1.1.1.). RSD 1000 was superior to all other 15 RSD compounds tested since other compounds failed to meet all criteria for antiarrhythmic "goodness". One compound, RSD 1014, exhibited no dose-dependent antiarrhythmic relationship (Figure 4B), whereas other compounds produced increasing protection with increasing dose, but the variance for such responses was high. Suppression of arrhythmias for some compounds at their highest tested dose did not reach 100%. The lack of "goodness" with some compounds 153 appeared to relate to poor antiarrhythmic protection and/or excess or cardiac or non-cardiac actions (see below). None of the compounds, in particular RSD1014, when given either pre- or post-occlusion produced any signs (as detected from the E C G or blood pressure) of rhythm disturbances or proarrhythmic actions, i.e., an increase in AS or V T / V F incidence or reduced onset time at which arrhythmias occurred. With respect to onset time, in vehicle-treated animals, the peak arrhythmic period occurred 6-10 minutes post-occlusion. This early arrhythmogenic phase is consistent with extensive data from other control animals in this laboratory (Botting et al., 1983; Johnston et al., 1983; Curtis et al., 1985; Curtis and Walker, 1987; Adaikan et al., 1992; Igwemezie et al., 1992; Barrett et al., 1995; Yong et al., 1999). Moreover, the period for late phase arrhythmias begins 1 hour post-occlusion (Johnston et al., 1983; Curtis et al., 1985; Opitz et al., 1995) and may involve mechanisms related more to infarction than to acute ischemia (Opitz et al., 1995; Patterson et al., 1998). Most compounds in this study suppressed the duration and frequency of P V C s , V T and VF, rather than the onset time of arrhythmias. In general, some control animal scores were between AS = 5 and 7, where an AS of 7 means arrhythmia-induced mortality before 15 min post-occlusion (see Table 1, Methods section). The superiority of RSD 1000 in its suppression of ischemia-induced arrhythmias appeared to be associated with its 1-naphthyl (R2) and morpholinyl (R,) substituents, based on the following observations. A positional change of the naphthyl group from the 1- to 2- position (RSD 1009) produced a compound that was slightly less potent and, perhaps, more variable in its antiarrhythmic response than RSD 1000. Based on this 154 observation and observations from other RSD analogues with modifications restricted to aromatic side groups, it was concluded that the 1-naphthyl group is the most favored aryl substituent for providing optimal "goodness" for antiarrhythmic activity. In compounds in which the 1-naphthyl group was retained (RSD 1046, RSD 1049 and RSD 1025), differences in their antiarrhythmic activity against ischemia-induced arrhythmias were associated with variations on the amine heterocyclic group. From pKa data, it was concluded that antiarrhythmic activity against ischemia-induced arrhythmias between the four compounds (RSD 1000, RSD 1046, RSD 1049 and RSD 1025) was inversely related to pKa (Figures 5A, 10A, 11A and 1 IB), i.e., antiarrhythmic "goodness" decreased with increasing pKa. To further substantiate this finding, effects against arrhythmias, independent of ischemia (e.g., raised extracellular H + ions), were tested for RSD 1000, RSD 1046, RSD 1049 and RSD 1025 in the electrically-induced arrhythmia rat model. Induction of ventricular arrhythmias by electrical stimulation was used to assess the ion channel blocking properties in non-ischemic hearts, expressed as increases in current thresholds for ventricular fibrillo-flutter (VFt). A reduction in I N a and/or repolarizing currents (e.g., I T O) increases the threshold for excitability and prolongs the effective refractory period, respectively (Winslow, 1984). RSD 1000 was the least potent of the four compounds in suppressing ventricular fibrillo-flutter. The potency rank order for increasing VFt was RSD 1025 > RSD 1049 > RSD 1046 > RSD 1000 and this correlated well with their decreasing rank order for pKa values (8.9, 8.2, 7.0 and 6.1, respectively, Figures 12A-D). Suppression of ventricular fibrillo-flutter by RSD 1025 and RSD 1049 occurred over the same dose range as that required for suppression of ischemia-induced 155 arrhythmias. This was not the case for RSD1000 and RSD1046. Thus, a reduced pKa was found to be associated with reduced potency in normal myocardium but ion channel blocking activity was maintained in myocardium exposed to ischemic conditions as reflected in the effects of ischemia-induced arrhythmias. This difference in potency, presumably as a function of pKa, for ion channel inhibition in normal tissue for the four compounds mentioned may explain the differences in their antiarrhythmic profiles against ischemia-induced arrhythmias. Differences in binding interactions due to differences in their molecular conformation or spatial requirement(s), in addition to pKa, cannot be overlooked as an explanation of their activities. It is interesting to note, however, that the influence of pKa is not limited to antiarrhythmic activity but also to the effects on cardiac function and the cardiovascular system. 4.3.2. Cardiac and cardiovascular responses Drug-induced toxicity in this study was defined as any untoward action that affects the homeostatic stasis of the animal, e.g., blood pressure lowering, asystole, proarrhythmia and/or cardiac depressant actions. As mentioned above, none of the compounds were shown to have proarrhythmic actions. A l l RSD compounds tested lowered blood pressure, decreased heart rate and increased E C G parameters to varying degrees, in a manner distinct for each compound. The therapeutic ratio in this study was defined as the ratio of doses producing a drug-induced adverse response versus the E D 5 0 for antiarrhythmic protection against ischemia-induced arrhythmias (values not shown; refer to Table 5). Of those compounds exhibiting "good" antiarrhythmic profiles, RSD 1000 and RSD 1019 both had the highest therapeutic ratios in terms of blood pressure 156 lowering, decrease in heart rate and changes to E C G parameters (Table 4). In sham coronary occlusions studies for RSD1000 (Yong et al., 1999), a high dose of 8 umol/kg/min produced an initial decrease in blood pressure and heart rate. However, following 8-10 min of infusion, there was recovery and a slightly lower steady-state level was reached and was sustained for the rest of the 20 min infusion period. RSD 1025 and RSD 1049 had the lowest therapeutic ratios (refer to Tables 4 and 5) and were equipotent in lowering blood pressure, heart rate, changing E C G variables, and antiarrhythmic doses for ischemia-induced arrhythmias. Suppression of ischemia-induced arrhythmias by RSD 1025 and RSD 1049 did not reach 100% and, for this reason, estimates of their antiarrhythmic E D 5 0 had to be made on available data. Their high potency for depressant actions on blood pressure and heart rate limited the dose which could be given and, hence, possible expression of maximum antiarrhythmic activity. On the other hand, RSD 1000, RSD 1046, RSD 1009 and RSD 1019 produced minimal effects on blood pressure and heart rate at doses that markedly reduced arrhythmia-induced V T and VF and reduced AS close to zero. The difference in the cardiac and cardiovascular effects of the above compounds, compared with RSD 1025 and RSD 1049, appeared less dependent on the aromatic side group and, analogous to their antiarrhythmic profiles, depended on the nature of the ionizable amine group or, more specifically, on pKa. The severity of toxic effects may be related to pKa but the effects on the E C G produced by RSD 1049 may suggest specific involvement of the amine heterocyclic group. By simply comparing the two most toxic compounds (RSD 1049 and RSD 1025), an increase in the P-R interval (Table 4) was produced to a greater extent by RSD 1049 than RSD 1025, a difference which might be independent of pKa. The effects of P-R 157 prolongation, accompanied by atrio-ventricular conduction failure, observed with 4 umol/kg/min RSD1049 and not with RSD1025, might suggest blockade of the slow inward C a 2 + currents. This assumption was based on evidence that the piperidinyl moiety of other compounds has been implicated in blockade of voltage-gated C a 2 + channels (Zamponi et al., 1996) where the existence of a specific binding site for piperidine-based compounds has been proposed (King et al., 1989). Electrophysiological studies on C a 2 + currents for R S D 1049 were not performed. Attempts to associate hemodynamic and cardiac physiological responses to one specific chemical and/or structural requirement were beyond the scope of this thesis. The main conclusion based on cardiac and cardiovascular effects is that the potency for depressant activity in this series of R S D compounds was enhanced when p K a was high. 4.3.3. Effects in cloned sodium channels and isolated rat ventricular cells. INa: In the above sections, the role of p K a was implicated in antiarrhythmic actions against ischemia- and electrically-induced arrhythmias, as well as, cardiovascular toxicity. In the in vitro electrophysiological studies, the availability of three 1-naphthyl derivatives (RSD1046, RSD1049 and RSD1025), with varying pKa 's , with their different amine heterocyclic groups, afforded a direct structural comparison of effects on I N a as a function of external p H . Electrophysiological studies in isolated cardiac cells have inherent problems that include inadequate voltage control of I N a and its partial separation from the decaying portion of the initial capacity transient (see Beeler and Reuter, 1970; Johnson and Leiberman, 1971; New and Trautwein, 1972; Noble, 1975). Partly for this reason, studies on the effects on I N a with R S D 1000, R S D 1046, R S D 1049 and R S D 1025 158 were performed using cloned sodium channels expressed in Xenopus oocytes. In this model, I N a is independent of overlapping currents (e.g., I C a and IK) while adequate voltage control of I N a is possible (these experiments were performed by Dr. Michael Pugsley). Using a two-electrode voltage clamp technique, potency for inhibition of I N a was found to be proportional to pKa and this was positively influenced by external pH changes, i.e., at both pH 0 7.3 and 6.4, the higher the pKa of the compounds, the lower the IC 5 0 . This could be expected from conventional local anesthetic theory that postulates unbinding of a charged and intracellular-acting species as the rate-limiting step in acid pH (Hille, 1977; Hondeghem and Katzung, 1977). At external pH 6.4, the rank order of potency remained unchanged but the IC 5 0 values for most of the RSD compounds for inhibiting I N a were increased relative to pH 0 7.3. The potency shifts were greatest for RSD1025 and RSD1000, minimal for RSD1046 and unchanged for RSD1049 (Table 7). Despite the pKa values for RSD 1046 and RSD 1049 residing between the values of RSD1000 and RSD1025, i.e., between 6 and 9, the former two did not exhibit similar potency shifts as the latter two. Additional mechanisms may be involved and a possible explanation is given in a later section. To exclude the possible role of positional arrangement of the naphthyl group as a contributing component to the pH-dependent blocking actions of I N a , RSD 1009 was also tested in isolated rat ventricular myocytes. RSD 1009 was shown to be less potent in its inhibitory actions on I N a , relative to RSD 1000. When extracellular pH was decreased, a potency shift towards a lower IC 5 0 , similar to that found for the 1-naphthyl counterparts in cloned sodium channels, was also evident with RSD 1009. The reduced potency for effects on I N a at both pH 0 6.4 and 7.3 might explain its less favorable antiarrhythmic 159 profile relative to RSD 1000 (see Figure 9). This result may indicate possible modulation of the pH-dependent blockade of INa by the naphthyl group. Electrophysiological studies involving RSD1015 (2-naphthyl), however, were not performed to confirm this observation. IT0: Up to this point, the I x o blocking component of RSD 1000 and its relative contribution to its antiarrhythmic activity have not been considered. One concern with regard to the lack of effective antiarrhythmic responses with RSD compounds structurally similar to RSD 1000 is the possibility that I x o blockade may be absent or less in such compounds. One unexpected result from the INa studies was that with a change in the positional arrangement of the naphthyl moiety (RSD 1000 versus RSD 1009), the potency for INa blockade decreased. To determine if the positional arrangement of the naphthyl group had a similar role with respect to IT 0, RSD 1009 was examined for effects on this current. Arguing from the observed potentiation of INa blockade in pH 0 6.4, it may be likely that the protonated species of a compound may be the active form and thus a positional change to the naphthyl group was expected to be expressed via this ionized form. Therefore, a predominantly charged species may be used to distinguish the relative contribution of molecular charge and the positional arrangement of the naphthyl group for IT 0 blockade. For this reason, RSD1025 and RSD1015, two predominantly charged 1-and 2-naphthyl homologues of RSD 1000 and RSD 1009, respectively, were included in the study. The results showed that RSD 1000, RSD 1009, RSD 1025 and RSD 1015 inhibited IT 0 with equal potencies and that their actions were unaffected by changes in external pH. This evidence clearly demonstrates that inhibition of ITO, for this series of 160 compounds, is independent of the molecular charge of the molecule and positional arrangement of the naphthyl group. Thus, the antiarrhythmic differences, not antiarrhythmic activity per se, observed for those RSD compounds appears to be more a function of the I N A blocking component and less on the blocking actions on I T O . 4.3.4. Summary of biological activity of RSD compounds The biological evidence discussed thus far has shown that RSD compounds with a 1-naphthyl aryl substituent (R2) and a morpholinyl amine group (R,) provide dose-dependent protection against ischemia-induced arrhythmias at doses having minimal effects on normal myocardium (ECG and electrical stimulation) and minimal cardiovascular depressant actions. The antiarrhythmic activity of the compounds tested was more sensitive to changes involving the amine heterocyclic group than the aromatic side group. The former changes influence ionization strength (i.e., pKa) of the nitrogen group and, in effect, influenced blocking activity on IN A, but not I X O , currents, as shown by direct measures of current modulation in isolated ventricular myocytes and cloned sodium channels. This observation was, in part, substantiated by effects against electrically-induced arrhythmias. 4.4. RSD compounds: Chemical activity The simplest of chemical characteristics that have been investigated for this series of compounds included pKa and lipid solubility (i.e., log Q). By definition, pKa is the negative logarithmic (-log) transformation of K a , the acid dissociation constant. The ionizable amines with a high affinity for protons form weak acids and dissociate only 161 slightly, i.e., high pKa and vice versa. B y definition, log Q is the distribution coefficient (in lipid versus water phase) of the unionized form at a particular p H (i.e., 7.4 and 6.4). For a given compound, pKa determines the ratio of ionized:unionized form in aqueous solution, whereas log Q is a measure of solubility in lipid relative to the aqueous phase. Both these measures are independent physiochemical properties that generally govern the equilibrium state of a compound between lipid and aqueous phases. The question arises as to whether there was a quantitative relationship between antiarrhythmic activity and chemical attributes. As chemical attributes of R S D compounds, was there a quantitative relationship associated with antiarrhythmic activity? Attempts to correlate pKa or log Q with antiarrhythmic E D 5 0 yielded, in each case, poor correlation indices (r2 < 0.5). This may not be surprising for a number of reasons. For some compounds, approximation of antiarrhythmic E D 5 0 values had to be made (see Table 5) since a 100% antiarrhythmic response for these agents was not achieved. Close inspection of the raw antiarrhythmic data for some of the compounds revealed a variance in the estimate of antiarrhythmic E D 5 0 of 0.4-0.6 on a log scale. This lack of precision would be unsuitable even for the most basic but precision-demanding analyses such as the Hansch analysis (1971) in the form, log 1/[C] = fh(xh) + fe(xe) + fs(xs) + constant, where fh(xh), fe(xe) and fs(xs) are the functions that characterize the influence of hydrophobic, electronic and steric properties of molecules through the corresponding parameter, x. The alternative approach was to rank antiarrhythmic E D 5 0 values (i.e., ascribe non-parametric values) but little improvement was seen for rank correlations (Table 9). 162 Such considerations raise several concerns. Are the present biological measures applicable for a quantitative structure-activity comparison? Is this data-base of RSD compounds sufficient in number to reliably detect chemical changes within the series? Is a qualitative rather than a quantitative structure-activity comparison more suited for this study? It is interesting to note that a global view of the correlation matrix (Table 9) reveals that pKa correlated highly with blood pressure and heart rate measurements as well as indices for IN a blockade (i.e., iT, blockade of cloned IN a at pH 0 7.4) . A more detailed discussion of pKa will follow in a later section but a succinct analysis that is associated with traditional QSAR studies will not be performed in light of the issues raised above. The answers to the above questions were best dealt in their entirety in the following section. 4.4.1. RSD compounds: Comparison of biological and chemical activity 4.4.1.1. Quantitative versus qualitative structure-activity analysis? Traditionally, QSAR studies require a large number of compounds with various chemical modifications and a biological test system which can reliably detect changes in activity. The required level of biological precision may be less sensitive with ion channels for the following reason. When two structurally similar compounds or, more specifically, two enantiomeric pairs act on the same voltage-gated ion channel, are the responses of great enough magnitude to distinguish between the two? With receptors, the situation is clear. A receptor is defined as a macromolecular protein that has affinity for a ligand such that the formation of a receptor-ligand complex is equated with a production of a response. This view is consistent with the earliest concept and definition of receptors 163 proposed by Elrich (1913). Analogous to classical substrate-enzyme interaction, interactions between a ligand and its receptor are so complementary that any divergence from the optimal structure, as with optical isomers or members of homologous series, results in potencies that often (but not always) vary 100- to 1000-fold. The resolution of voltage-gated ion channels, in this respect, may be less precise since ion channels are essentially pores and the agents that block them do not appear to conform to a "lock-and-key" binding scheme similar to that generally associated with receptors but, perhaps more appropriately, to a "plug-and-fill" arrangement. Contrary to models that describe the interactions between substrates and enzymes, or drugs and their receptors, there is, to date, no specific structural model for any cardiac ion channel that definitively describes binding interactions for antiarrhythmic drugs. Biochemical and molecular approaches (e.g., site-directed mutagenesis) have been used to probe the peptide segments of the principal subunits of voltage-gated ion channels that interact with extracellular and/or intracellular pore blockers (see review by Catterall, 1996). However, the structural diversity that exists amongst channel pore blockers for a particular channel and the fact that optical isomeric pairs (for an examples, see Yeh, 1980; Hi l l et al., 1988; Wang and Wang, 1992) usually produce equipotent ion channel-blockade, suggest that ion channel blocking agents may block not via a "binding site-specific" interaction but rather in a "domain-specific" fashion. Based on this rationale, any structural or chemical modification that is present within the RSD compound array may not be detectable with any precision by the measures (i.e., antiarrhythmic ED 5 0 ) used in this study or with a limited number of compounds. Moreover, a more direct measure of ion channel response, e.g., IC 5 0 for I N a , was not available for all RSD compounds. For 164 this reason, a qualitative rather than a quantitative approach to comparing the RSD compounds may be better suited to the present discussion. In view of all of the above, the following comparisons attempt to associate antiarrhythmic activity with the various chemical structures in the series. The former, however, must be defined and it may be worth analyzing the problems of earlier antiarrhythmic development that used potency as the predictor for antiarrhythmic efficacy and safety. Poor therapeutic indices have been the Achilles heel of earlier antiarrhythmic drugs. Historically, antiarrhythmic agents that have shown potency for normal cardiac tissue have generally failed to reduce mortality (e.g., CAST Trial, Echt et al., 1989). Such trials were based on a common misconception that for a given drug or compound that had a high potency for producing a response, the natural tendency was to reason that threshold for toxicity would be low. Selectivity, rather than potency, may be a more accurate determinant for the effectiveness and safety of an antiarrhythmic agent. Potency is defined as the locator (ED5 0, EC 5 0 , IC 5 0, etc.) at 50% of the maximum response on a log-dose axis. For a given drug or compound, its locator for one response relative to the locator for another response (potency ratio) defines the selectivity for that agent. Therefore, in the following discussion biological activity will be defined as the selectivity or potency (therapeutic) ratio between a toxic or untoward side-effect and the antiarrhythmic response against ischemia-induced arrhythmias namely, the antiarrhythmic ED 5 0 value. 165 4.4.2. RSD compounds: Qualitative comparison of antiarrhythmic selectivity and chemical substituents The following tables list the therapeutic ratios as they appeared according to the influence of the amine heterocyclic or aromatic groups. The ratios express effects on blood pressure (Table 12B) and effects on normal myocardium (Table 12C) versus antiarrhythmic activity against ischemia-induced arrhythmias. For ease of presentation, ratios are ranked as high or low (see Table 12 for details) according to the values listed on Table 5. A table comparing antiarrhythmic "goodness" (Table 12A) in relation to the chemical substituents is also presented. The main observation from these comparisons indicates that the therapeutic ratios are more sensitive to chemical modifications to the amine heterocyclic group than to the aryl substituent. More importantly, the highest therapeutic ratios were found when the morpholinyl group was combined with 1-, 2-naphthyl, or para-bromophenyl aryl substituents. A more detailed analysis of the aryl substituent and amine heterocyclic group in relation to antiarrhythmic activity will be presented in the following sections. 4.4.2.1. Analysis of the aromatic side group In the absence of an aryl substituent at R 2 (formate and acetate derivatives), little or no antiarrhythmic protection was observed at a dose of 8 umol/kg/min, a dose in which compounds with an aromatic side group produced modest to significant suppression of ischemia-induced arrhythmias. Once a phenyl group was added to position R 2 and extended from the cyclohexyl backbone by increasing methylene groups (RSD 1053, RSD1010, RSD1050, RSD1051), a dose-dependent suppression of ischemia-induced 166 Table 12 Therapeutic index comparison table (see text for details). A. Antiarrhythmic "goodness "for ischemia-induced arrhythmias Quali ty of "goodness" ionizable amine group aromatic group high morpholinyl or bis-methoxyethyl 1-, 2-naphthyl, or para-bromo-phenyl low N-methyl-piperazinyl or piperidinyl 1- or 2-naphthyl B. ED50(J-BP)/ED50(IA) Therapeutic ratio amine heterocyclic group aromatic group high morpholinyl or bis-methoxyethyl 1-, 2-naphthyl, or para-bromo-phenyl low N-methyl-piperazinyl or piperidinyl 1- or 2-naphthyl C. ED50(NM)/ED50(IA) Therapeutic ratio amine heterocyclic group aromatic group high morpholinyl 1-, 2-naphthyl, or para-bromo-phenyl low bis-methoxyethyl, N-methyl-piperazinyl or piperidinyl 1- or 2-naphthyl Legend: high > 3, low > 1; IA = ischemia-induced arrhythmias; N M = effects on normal myocardium (ECG, electrical stimulation parameters); BP = blood pressure. 167 arrhythmias was produced with each compound. These results implied a probable involvement or interaction with the aryl substituent. Phenylacetate analogues of RSD1010 retaining the cyclohexyl backbone and morpholinyl substituent were synthesized with 3,4- or />ara-4-phenyl substitutions in the hopes that aromatic side group extensions would mimic the elongation of the benzene group or add some element of distance with chemical "bulk". When compared with the unsubstituted-phenyl derivative, RSD 1010, there was a significant improvement in antiarrhythmic "goodness" with j?ara-4-bromo-phenyl substitution (RSD 1019), no difference with 3,4-phenyl substituted derivatives (RSD1012 and RSD1072) and an "sub-standard" response with para-4-nitro-phenyl substitution. Each of these aromatic substitutions resulted in chemical modifications to the phenyl group itself without significant alterations to the compound's lipid solubility or molecular weight (see Table 3). The result was a series of phenyl-substituted compounds that were distinctly different in their overall (antiarrhythmic) actions but other chemical factors (likely associated with these phenyl substitutions) that cannot be easily explained may be involved. This series of phenyl-substituted derivatives were studied to address the following: 1) steric hindrance, 2), hydrogen bonding interactions and 3) changes to the electron density of the benzene group. Their distinct antiarrhythmic actions may be explained according to any one or a combination of these physiochemical factors. However, when attempting to explain the antiarrhythmic actions for RSD 1000 based on these factors, there are some inconsistencies. For example, size-restriction of the aryl substituent does not appear to be a factor since the 1-naphthyl group of RSD 1000 is larger or equal in size to the largest of the phenyl-substituents, 3,4-dimethoxyl 168 (RSD 1072). It is also unlikely that hydrogen bonding interactions plays a significant role in antiarrhythmic activity since the strength of this interaction is greatest with electronegative elements such as nitrogen and oxygen (e.g., - N 0 2 and -OCH 3 substituents), intermediate with halogens (e.g., -Br and -C l ) and least with hydrocarbons (e.g., - C H 3 and -CH=CH 2 ) . Distortion of the electron density of the phenyl group is greatest in the presence of electronegative groups (e.g., -Br and -C l ) , but minimal or absent with the 1-naphthyl group suggesting that this modification is unrelated to antiarrhythmic activity. This is further supported by the stronger electron-withdrawing effect of nitro-phenyl (RSD 1014) > dichloro-phenyl (RSD1012) substitutions versus the single bromo-phenyl substitution (RSD1019) and in this series. A more accurate interpretation may not involve the magnitude of change of the electron density but rather the direction in which the electronic cloud is modified. Consider the following chemical modifications that led to improved antiarrhythmic activity against ischemia-induced arrhythmias: 1) extension of the link between the aryl substituent and the basic nitrogen (RSD 105 3 to RSD 1051), 2) an extended aryl system in a form that is rigid (RSD 1000) and 3) an electron-withdrawing />ara-substituted (RSD1019) versus an unsubstituted (RSD1010) phenyl group. A l l of these modifications imply an interaction with an aromatic system that requires some form of specificity between the electron cloud or density and a binding site. This interaction is presumed to be specific since changing the point of attachment of a naphthyl group from the 1- to the 2-position (e.g., RSD 1009), additional electron-withdrawing effect on the phenyl group (e.g., RSD1012) and the flexibility of the extended alkylene "link" (e.g., RSD1050 and 1 6 9 RSD 1051) have all shown improved but not "optimal" antiarrhythmic response as compared with RSD 1000 or RSD1019. Recent molecular biological evidence has shown that conventional local anesthetic block is largely dependent on the hydrophobicity or aromaticity of residue 1710 (usually a phenylalanine), in the transmembrane segment, IVS6, of the channel oc-subunit (Li et al., 1999). Following a series of mutations that varied the size, hydrophobicity and aromaticity of F l 710, it was proposed that local anesthetic inhibition of open or inactivated sodium channels may involve aromatic-aromatic interactions between the aromatic side chain of the amino acid (F1710) and the aromatic moiety of the drug molecule. (Li et al., 1999). Similar aromatic-aromatic interactions almost certainly exist with the RSD compounds tested. Although the site of interaction of the present RSD compounds with ion channels is not known, a related compound, RSD921 (Pugsley and Goldin, 1999) may act at an external site on the sodium channel (and presumably, at the common binding site for most local anesthetics). 4.4.2.2. lonizable amine group: pKa and its role in antiarrhythmic activity RSD 1000 and three structurally-related compounds (RSD 1000, RSD 1046, RSD 1049 and RSD 1025), which have different ionizable amine groups, such that each varied in pKa, were chosen to examine the influence of pKa on antiarrhythmic activity. A comparison of their actions against ischemia- or electrically-induced arrhythmias provided some perspectives on how pKa was associated with antiarrhythmic activity. This interaction was further studied in cells exposed to external buffer solutions of pH 0 6.4 and pH 0 7.3. The results of the study have shown that molecular charge, at least in 170 this present series of test compounds, resulted in the following observations: 1) antiarrhythmic selectivity against ischemia- versus electrically-induced arrhythmias progressively decreased as pKa increased (Figure 16 A-D), 2) blocking potency on I N a currents in ventricular myocytes and cloned sodium channels was proportional to pKa at pH 0 6.4 and pH 0 7.3 (Figure 17), 3) blockade of I N a , but not I T 0 currents, was increased in acid versus pH 0 7.3. Accordingly, the relative existence of charged versus uncharged compound species as a function of pKa and extracellular pH may offer an explanation for the observed differences in antiarrhythmic profiles against ischemia-induced arrhythmias. The relationship between degree of ionization and interactions with the inward sodium current has been documented for nerve (Hille, 1977a, 1977b) and cardiac tissue (Grant et al., 1980, 1982; Nattel et al., 1981). According to the models based on studies using local anesthetics and their quaternary analogues, the uncharged drug moiety has rapid access to an internal binding site by an aqueous open channel or a membrane-delineated pathway, whereas the charged form must transverse the former route (Hille, 1977a; Hondeghem and Katzung, 1977). During recovery from association with an internal binding site, the rate-limiting step of the protonated species, in relation to one that is uncharged, is the egress of the drug when the channel is open (Hille, 1977a). The pH-dependence of the kinetics of drug dissociation has been used to explain the greater depression of excitability in ischemic versus normal myocardium for local anesthetic-class antiarrhythmics (e.g., lidocaine; Wendt et al., 1993). 171 4.4.2.3. pH„ effects on the IN a blocking component of RSD compounds A delay in recovery kinetics may account for the observed increase in I N a blocking potency from pH 0 7.3 to 6.4 observed for RSD 1000 (Figure 10 and 17), RSD 1009 (Figure 19) and RSD1025 (Figure 17). Before discussing the role of pKa and antiarrhythmic activity for the RSD compounds, there are several experimental observations with respect to drug pKa that should be noted. These observations are based on the actions of local anesthetics that presumably act at an intracellular binding site (Strichartz, 1971; Scharwz, 1973; Hille, 1977). This study, however, has not fully established the binding profiles of any RSD compound with respect to I N a channel states and, more specifically, the binding site(s), i.e., extracellular versus intracellular. The pKa of bound drug may be shifted by the binding interaction, such that the extent and direction of the shift is dependent on the particular agent (Moorman et al., 1986). In other words, the pKa of bound drug may not be the same as pKa in free solution and this fact will have a direct influence on the rate of drug deprotonation from its binding site. Opposing evidence by Chernoff and Strichartz (1990) has shown that the pKa of bound drug is comparable to the pKa in free solution and that the lifetime of the cationic form is dependent, in addition to its deprotonation rate as a function of drug pKa, on diffusion of the cationic species directly into the lipid bilayer (see also Boulanger et al., 1981; Schreier et al., 1984; Watts and Poile, 1986) and the frequency by which sodium channel flickers open (during diastole) to release the trapped cationic form (see also Yeh and Tanguy, 1985). This observation is consistent with Courtney's proposal (1987) that both drug solubility in the cell membrane and the molecular dimensions of the drug play as important a role as drug charge in determining the overall rate of recovery. 172 The prolonged lifetime of the cationic form due to its slow rate of deprotonation under external acid pH may be adequate to explain the potentiation of I N a inhibition by a charged sodium channel blocker. The pH-dependent in vitro observations for RSD 1000, RSD1009 and RSD1025, in particular, have shown a higher potency for I N a blockade in pH 0 6.4 versus 7.3 that is consistent with observations for local anesthetics. However, several in vitro inconsistencies have been presented in this study that do not account for the different antiarrhythmic profiles (ischemia- and electrically-induced arrhythmias) exhibited by the tested RSD compounds based solely on an acid-potentiated blockade of W First, the comparison between RSD 1000 and RSD 1009 as discussed in the previous section, focussed on 1- versus 2-substitution on the naphthyl ring as a possible contributor for their different blocking properties on I N a and, as a result, for their different antiarrhythmic profiles independent of their pKa values. Secondly, the lack of pH-dependency for RSD 1046 and RSD 1049 observed in cloned sodium channels (Figure 16 and Table 7) is inconsistent with the slow "untrapping cation" model (i.e., deprotonation followed by escape of the uncharged species) in acid pH 0 . A similar finding for pH 0 -independent rate of recovery for local anesthetic drugs was also reported by Chernoff and Strichartz (1990). The third inconsistency is that the potency differences for inhibition of I N a at pH 0 6.4 and 7.3 are similar for RSD 1000 and RSD 1025 (Figure 17), yet they exhibit dissimilar antiarrhythmic profiles, particularly in their efficacy and dose-dependent response. The SkM2 cloned cardiac sodium channels used in this study (Figure 17) were the same TTX-insensitive sodium channel isoform that was first isolated by White et al. 173 (1990) and they serve as approximate surrogates to the native cardiac sodium channels in rats (White et al., 1990; Kallen et al., 1990; Satin et al., 1992). Attempts to delineate these inconsistencies may be accomplished by considering a pKa-dependent "three-stage" binding interaction: a relative distribution of charged and uncharged species, a "trapped cation" process and a "close-proximity binding" interaction. Before further explanation is offered, several assumptions must be clarified and reviewed. Using the infusion regime of 5 min (pre-occlusion) at a rate of x umol/kg/min (and assuming a nearly saturable a-phase of redistribution; see Section 4.6), a certain plasma drug concentration is reached. At concentrations directly affecting the heart tissue, it is presumed that a pseudo-equilibrium of uncharged and charged drug species is established between the intracellular and extracellular tissue compartments. Following the cessation of L A D coronary blood flow, the level of drug in the myocardial zone ahead of the occlusion should remain unchanged while drug delivery and washout in the uninvolved zone is maintained. In the rat heart, collateral coronary flow is approximately 6% (versus 16% for dog and >16% for guinea pig; Maxwell et al., 1987). Thus, it is a safe assumption that any "cross-talk" between right and left coronary beds is kept to a minimum such that rates for drug delivery and washout in the ischemic zone is presumed to be negligible. At the cellular level of the phospholipid bilayer, the membrane diffusion rate of the drug as a function of its concentration gradient will be fast for the uncharged species and relatively slow for the protonated form. For example, the rate of onset for quaternary QX-572 applied extracellularly was 5-fold slower than when applied inside the cell (Frazier et al., 1969), whereas the rate of onset for neutral benzocaine was 174 indistinguishable from either compartment (Hille, 1977). In the case of predominantly ionizable (i.e., tertiary nitrogen-containing) RSD compounds (e.g., RSD1025), this rate of diffusion may be faster compared with a quaternized analogue (not available) but, nonetheless, slower than the least ionized RSD analogue, RSD 1000. In response to myocardial ischemia, there is a decrease in intracellular pH followed by a decrease in extracellular pH (via proton efflux) and the scenario of diffusion "trapping" of the charged species into the intracellular compartment may occur. This effect may be greater with compounds of low (e.g. RSD 1000) rather than high pKa (e.g., RSD1025). It should be noted that pHj as determined by both ion-selective microelectrodes and N M R techniques initially change less than pH 0 (Vanheel et al., 1989; Koretsune and Marban, 1990; Murphy et al., 1991; Yan and Kleber, 1992). Thereafter, the changes reported for pH; vary whereas pH 0 continues to decrease and reaches a plateau range of 5.5 - 6.0 (Vanheel et al., 1989; Yan and Kleber, 1992). This is attributed to the lack of extracellular washout and the lower buffer capacity of the extracellular space (Yan and Kleber, 1992). The delay in recovery kinetics from an intracellular binding site is the underlying assumption of the "trapped cation" model as a means of explaining enhanced I N a blockade in acid pH 0 by local anesthetic-type sodium channel blockers. Without evidence as to the exact binding site(s) (intracellular or extracellular) of these RSD compounds, it is difficult to formulate an argument that explains their potency differences for I N a blockade (e.g., RSD1000, RSD1046, RSD1049 and RSD1025, see Figure 14 and Table 7) based solely on the "trapped cation" model (involving pH-dependent membrane diffusion rates by the compound). Alternatively, these differences may involve binding differences 175 between the RSD compound and ion channel at a molecular level such that recovery from channel block is determined by the dissociation rate of the compound as a function of its pKa (close-proximity binding). The higher the pKa of a compound, the stronger its ionization properties (as a function of the ionizable nitrogen group) and vice versa. If the binding of the RSD compounds to the ion channel involves some proton-donating amino acid group (e.g., R-COOH, where R represents acidic amino acids), the ionic interaction between this group and the ionizable nitrogen group of the RSD compound would be strongest (and thus, less likely to reverse) i f pKa was high (e.g., RSD1025) rather than low (e.g., RSD1000). It should be noted that the basicity between the proximal (Np) and distal (Nd) nitrogen atoms (see also Methods section, page 76) of the N-methylpiperizinyl group of RSD 1025 is strongest with the latter. The result is a strong ionizable nitrogen group that is "unhindered" by cyclo-alkyl "attachments" and "open" for binding interactions compared with the nitrogen atom proximal to the cyclohexyl ring (attached to the carbon in position 2'). RSD1000, RSD1046 and RSD1049 all have ionizable nitrogens groups proximal to the cyclohexyl ring. The result is, perhaps, a weaker ionic interaction between the ionizable nitrogen group and a proton-donating binding site since the distance for an "optimal" intramolecular interaction is "farther" and perhaps, more "hindered" for these nitrogen groups than the N d found in RSD 1025. However, this binding interaction may be more complex than a simple ionic attraction, since (with the exception of RSD 1000) it does not provide an explanation for the pH-independent I N a blocking profiles of RSD 1046 and RSD 1049 based on the proximity of the ionizable nitrogen group (but, see below) 176 The third level in this proposed "three-stage" binding interaction (that may serve as the precursor and consummates the expression of the other two binding stages) is the relative distribution differences of the charged and uncharged compound forms in ischemic (acid pH 0) versus normal myocardium (pH 0 7.4). The availability of this blocking component (presumably, the charged form) as a function of pKa, coupled with the degree to which the "trapped cation" and "close-proximity binding" processes are expressed may offer an added level of interpretation for explaining the different antiarrhythmic profiles. For example, according to the Henderson-Hasselbalch equation in the form, % ionization = 100/(1+10 p H o p K a) the percent ionization for protonated species can be calculated for a given pKa and pH 0 . So, the percentage of ionized RSD 1025 at pH 0 7.3 and 6.4 is calculated to be 97% and 99%, respectively, whereas for RSD 1000 the corresponding figures are 5% and 34%, respectively. Therefore, the amount of charged species that is present in an area will be dependent on the local pH 0 . This proposed scheme can be applied to the heart following the onset of myocardial ischemia in which there is a sudden drop in pH ; followed by a decrease in pH 0 (see Section 1.5.1.). As a result, local concentrations of a protonated compound that has a higher pKa than ambient pH 0 would be uniform in ischemic and non-ischemic myocardial tissue. Conversely, an agent with a weak ionization strength (i.e., low pKa) in acid pH 0 has a higher proportion of its charged species in ischemic than in normal tissue. The significance of a distribution difference involving the ionized form of RSD compounds becomes clear when discussing differences in action potential modulation by 177 the charged species. The charged form is assumed to be the active form based on the pH-dependent studies of the tested RSD compounds which show these results consistent with actions predominantly involving the charged form of local anesthetic sodium channel blockers. Conditions of myocardial ischemia depress conduction and excitability, which results in abnormal cardiac signal generation and configuration found in the ischemic myocardium (reviewed in Gettes et al., 1992a; 1992b; Wit and Janse, 1993; Cascio et al., 1995). These conditions produce a phase variance of cardiac signals between normal and ischemic zones and the potential for reentrant cardiac arrhythmias (El-Sherif et al., 1977a, 1977b; Janse et al., 1980; Janse and Capelle, 1982; reviewed by Lazzara and Scherlag, 1984; Kleber, 1987). Preventive measures that reduce the degeneration of cardiac signals in the ischemic zone may, in fact, reduce signal inhomogeneity between normal and ischemic regions (El-Sherif et al., 1977a, 1977b; Janse et al., 1980; Janse and Capelle, 1982). However, those same preventive mechanisms in the normal zone, in addition to the ischemic zone, may counteract effects produced in the ischemic zone and as a result increase, rather than decrease, signal dispersion between zones. Therefore, signal heterogeneity between normal and ischemic zones will be greater with high (e.g., RSD1025) than low (e.g., RSD1000) pKa compounds assuming that activity is predominantly a function of the protonated form. Most antiarrhythmics have high pKas and are primarily charged at < pH 0 7.4. It is, perhaps, not surprising to find indiscriminant cardiac ion channel blockade and drug-induced cardiac disturbances with such antiarrhythmics (Starmer, et al., 1991, 1992, 1993, 1995; Nesterenko et al., 1992; Chay, 1996). Like most ionizable local anesthetics, lidocaine exhibits pH-dependent sodium channel blockade but its blocking action on I N a is 178 further accentuated in depolarized tissue (Wendt et al., 1993). Its selective actions on ischemic myocardium to produce slowing of conduction in ischemic myocardium is not accompanied by a prolongation of refractoriness (normally associated with delayed reactivation of blocked sodium channels; Hondeghem and Katzung, 1977) but rather APD shortening due to blockade of the TTX-sensitive inward sodium current during the plateau phase (Colatsky, 1982; Carmeliet and Saikawa, 1982). This opposing effect on APD by lidocaine may increase the likelihood for signal dispersion and partly account for its limited antiarrhythmic effectiveness against models of experimental coronary occlusion (Kupersmith, 1979; Carson et al., 1986; Barrett et al., 1995) and in the clinical setting (Pentecost et a l , 1981; Abbott, 1981; Rademaker et al., 1986; Denaro and Benowitz, 1989; Jaffe, 1993). The added pH-insensitive I T O blocking component of RSD 1000, for example, may help preserve APD duration and thus, contribute to its antiarrhythmic effectiveness against ischemia-induced arrhythmias. One structural feature of the tested RSD series that is also common to most local anesthetics is the proton-attracting 3° (or 2°) nitrogen. Previous structure-activity studies have suggested that both the basic 3° nitrogen and the aromatic moieties are involved in binding interactions with the sodium channel (Courtney, 1980; Bokesch et al., 1986; Sheldon et al., 1991). Recent molecular studies have determined that phenylalanine at position 1710 in the transmembrane segment, IVS6, of the sodium channel a-subunit may be the key amino acid that offers cationic-7i or aromatic-aromatic interactions between the aromatic side chain of phenylalanine (F1710) and charged or aromatic moieties on the drug molecule (Li et al., 1999). Aromatic-aromatic interactions have been discussed for RSD compounds in a previous section. "Cationic-7t" interactions have 179 been observed in gas-phase studies (Sunner et al., 1981; Mautner and Deakyne, 1985; Taft et al., 1990) and reported as a stabilizing interaction between a cation and the electron-rich 7t-face of an aromatic ring (Petti et al., 1988; Dougherty and Stauffer, 1990). This interaction offers an explanation for nontraditional interactions involved with aromatic residues observed in protein crystal structures (Burley and Petsko, 1986; Perutz et al., 1993; Mitchell et al., 1993, 1994; Nandi et al., 1993) receptor systems (Dougherty and Stauffer, 1990; Taylor et al., 1991; Karlin, 1993) and as a possible mechanism for ion selectivity in the highly hydrophobic and aromatic amino acid sequence in the pore region of Shaker K + channels (Kumpf and Dougherty, 1993). In the case of the tested RSD compounds, the ionized amine heterocyclic group may serve as the cation that interacts with the 7t-face of an aromatic moiety in the sodium channel (but, see below) in a manner similar to that proposed for local anesthetics by L i and his group (1999). There were some unexpected findings in this study. The lack of pH-dependent I N a blocking actions of RSD 1046 and RSD 1049 demonstrated in cloned sodium channels does not appear to be consistent with the idea that differences in I N a blockade is a direct function of pKa. Other considerations must be examined, perhaps at a level that involves binding interactions with the ionizable nitrogen group. Both RSD 1000 and RSD 1025 have basic nitrogen groups with a second electronegative group (oxygen and nitrogen, respectively) that is distal to the nitrogen atom proximal to the cyclohexyl ring (attached to the cyclohexyl carbon in position 2'). The existence of this secondary electronegative group in the amine heterocyclic ring may help establish an additional close-range interaction with the binding domain that is not available with the piperidinyl or bis-methoxyethyl-amino groups of RSD 1049 or RSD 1046, respectively. Although RSD 1046 180 has two electronegative oxygen groups, the apparent flexibility and higher rotational energy barrier of the two methoxyl-ethyl groups may limit binding interaction(s) of its two oxygen groups. As previously mentioned for RSD1025, the positive charge of the molecule resides on the distal or 4'-nitrogen atom of the 4-N-methylpiperazinyl group (see also Methods; page 76). The availability of a protonated moiety at this position may offer an additional cation-anion interaction that is not available with an electronegative oxygen (morpholinyl) or alkyl group (piperidinyl). This extension of a charged moiety to a binding pocket in the sodium channel may explain the increased potency for I N a blockade that is produced by RSD 1025 compared to RSD 1000. A similar report of lengthening the arylamide-amine link (R,', optimal length was 2 carbons; see below), rather than changes in drug hydrophobicity (either with differing number of aminoalkyl substituents (R 2 ') or increasing chain length), showed increased affinity of local anesthetic block by lidocaine analogues to the sodium channel (Sheldon et al., 1991). 4.4.2.4. p H 0 effects on the I x o blocking component of RSD compounds The pR-independent blocking actions of I T O observed for RSD 1000 and RSD 1025 suggest that dose-dependent antiarrhythmic effects and selectivity for ischemia-induced arrhythmias as a function of pKa of the compounds does not appear to depend on inhibition of I T n . This result can be rationalized with reference to the amino acid 181 sequence of the pore region Shaker K + channels. The sequence is comprised of highly conserved hydrophobic amino acids namely, F (phenylalanine), Y (tyrosine) and W (tryptophan) and, as such, may involve cation-71 interaction for ion selectivity in K + channels (Durell and Guy, 1992; Bogusz et al., 1992; Kumpf and Dougherty, 1993). Incorporating these observations and with the evidence that the Shaker K + channel exists as a tetramer (Schwarz et al., 1988; MacKinnon, 1991), the selectivity filter may involve a ring of aromatic residues located at a constricted region of the K + channel pore (Kumpf and Dougherty, 1993). This model is distinctly different from the model of charged amino acid residues required for interaction of permeant ions (e.g., Na + ions), binding of externally acting pore blockers (e.g., TTX) and ion conductance and selectivity in the channel pore of sodium and calcium channels (Terlau et al., 1991; Heinemann et al., 1992). A pore region where acidic and/or basic amino acid side chains appear to modulate the affinity and/or actions of pore channel blockers is consistent with protonation and/or deprotonation binding reactions between proton-donating or attracting blockers and their binding site. Evidence in support of such reactions has been shown by augmented binding in acid conditions (Moorman et al., 1986; Chernoff and Strichartz, 1990; Wendt et al., 1993). The absence of pH effects on I T 0 inhibition by the tested RSD compounds (which have a range of acid dissociation constants) may be explained by cationic-7t interaction(s) between the charged portion of the molecule and the aromatic amino acid residue(s) in the K + channel pore. Based on this binding scheme, proton transfer would be unlikely between the electron-rich 7i-faces of the aromatic groups within the pore region and the unprotonated amino group of the test compounds. These binding 182 scenarios may offer an explanation for the pH-dependent and -independent blocking activity of RSD compounds on I N a and I x 0 , respectively. Aromatic-aromatic binding interactions between the aromatic groups of the tested RSD compounds and the 7r-faces of amino acid side chains found in channel pores should not be dismissed as possible stabilizing forces (Burley and Petsko, 1985; L i et al., 1999) for channel binding. 4.4.2.5. The role of the I x o blocking component in antiarrhythmic activity Tedisamil is a blocker of I T 0 and suppresses ischemia-induced arrhythmias in rats (Adaikan et al., 1992). However, this suppression only occurs at doses that produce significant widening of the Q-T interval (Beatch et al., 1991; Adaikan et al., 1992) and possible reduction in I x o selectivity, since Dukes et al. (1989) reported sodium channel blockade at high concentrations. There are no reports to date that have investigated the pH-dependent actions of this drug or the effects of external pH on the blocking actions of other blockers on this channel. A recent report by Wegener et al. (1998) demonstrated that the inhibition of I x 0 by external application of 4-aminopyridine (4-AP) was strongly reduced at high pH 0 in combination with low pHj. A similar reduction was shown with intrapipette application of 4-AP when pH ; was low, independent of external pH. These results suggested that 4-AP acts intracellularly in its protonated form favored by low pHj. The idea that 4-AP acts intracellularly on K + channels was first suggested by Gillespie and Hutter (1975) on the basis of their experiments on frog skeletal muscle and is further supported by other studies that demonstrate pH-dependent effects of 4-AP in other cell types (Howe and Ritchie, 1991; Choquet and Korn, 1992; Stephens et al., 1994) and reduction of 4-AP 183 potency by the presence of a cytoplasmic inactivation peptide (Stephens et al., 1994; Yao and Tseng, 1994). Unfortunately, there are no reports of 4-AP and its actions against ischemia-induced arrhythmias. Whether the I x 0 blocking component plays a primary or secondary role in the antiarrhythmic mechanism of the RSD compounds in the present study is not clear. The IC 5 0 values for I N a and I x o blockade for most of the RSD compounds tested at pH 0 7.3 were not different by a factor of 10 from each other. Thus, an approximate 1:1 blocking potency ratio for I N a and I x o may be assumed to be produced by some of the tested compounds and, furthermore, the electrophysiological studies at pH 0 6.4 show that this ratio increased for I N a versus I x o . Quinidine is similar in actions to the tested compounds in that it exhibits blocking properties on both I N a and I x 0 currents. Experimental evidence of indices of measurement of sodium channel blockade (such as dV/dt) has shown that quinidine produces greater depression of excitability in acid conditions (Nattel et al., 1981; Grant et al., 1982; Evans et al., 1984; L i and Ferrier, 1991). The value of quinidine as a selective antiarrhythmic in ischemic or depolarized myocardium is limited (see also Barrett et al., 1995) by its preference to bind to the open state form of the sodium channel and its slower binding kinetics relative to lidocaine (Hondeghem, 1987; Zilberter et al., 1994). More importantly, ischemia-induced shortening of APD remained unchanged at quinidine concentrations that produced significant changes in non-ischemic conditions (Li and Ferrier, 1991). The most recent study that investigated the I x o blocking component of quinidine in isolated rat myocytes was that of Clark et al. (1995) but these studies were confined to physiological pH 0 (7.3). No other investigation of this agent and its blockade 184 of I x 0 current in different pH media have been cited. As an agent with a protonatable nitrogen group, quinidine is predominantly ionized at the test pHs (overall molecular pKa 8.6; Avery, 1976; N.B. the ionization strength of the nitrogen in the quinuclidine cycle (pKa, = 10) is offset by the weaker ionization of the second amine group (pKs^ = 5.4)) and on the basis of its pKa and mixed I N a and I T 0 blocking properties, it would appear to have more SAR similarities to RSD 1025 (predominantly charged blocker) than to RSD 1000 (pH-dependent charged blocker). 4.5. Summary This study was designed to investigate: 1) the antiarrhythmic activity of a novel antiarrhythmic pharmacophore, 2) the chemical components and/or attributes associated with this pharmacophore that are responsible for antiarrhythmic activity against ischemia-induced arrhythmias and 3) to compare the antiarrhythmic actions of the RSD compounds of this study with standard antiarrhythmics in terms of their effects on ischaemia-induced arrhythmias in the rat. Using a series of 16 structurally-related amino-2-cyclohexyl ester compounds, it was shown that the RSD pharmacophore with morpholinyl (R^ and 1-naphthyl (R2) groups (RSD 1000) provided antiarrhythmic protection with 100% efficacy, a Hi l l coefficient = 1 and minimal adverse actions on blood pressure and normal myocardium. Electrophysiological studies in rat ventricular myocytes characterized the blocking actions of RSD 1000 and other similar RSD analogues as being a mixture of I N a and I x o with equal potencies at pH 0 7.3, but exhibiting a greater potency on I N a at pH 0 6.4. This pH-dependent blocking property for I N a was associated with the ionization strength of the tertiary nitrogen(s) in the ionizable amine group. Structurally, there are some 185 similarities between the tested RSD compounds and local anesthetics as well as several standard antiarrhythmics that offer possible connections between specific groups and their blocking interaction with specific ion channels (e.g., the tertiary nitrogen and its relationship to INa channel blockade). With reference to actions of conventional antiarrhythmic agents, the mixture of INa and I T 0 current blockade by RSD 1000 is similar in action to quinidine. One distinguishing difference between the two agents is the low pKa value of RSD1000 (6.1) versus that of quinidine (8.6). As a consequence, the INa blocking component of RSD 1000 is localized to myocardial tissue with external pH 6.4 conditions, rather than a global blockade of I N a with added pH-dependent block in ischemic tissue by quinidine. In both scenarios, signal degeneration may be prevented, but in relation to drug activity in the normal zone, where activity is greater with quinidine than RSD 1000, signal heterogeneity is increased more with the former than the latter. This study has also demonstrated that agents with mixed INa and ITO channel blocking activity may contribute to superior antiarrhythmic profiles compared with agents possessing blocking preference to only INa (e.g., lidocaine) or IT 0 (e.g., tedisamil) in the same ischemia-induced arrhythmia rat model. 4.6o Metabolism of R S D compounds: possible chemical sites Obviously, metabolism is not a factor in the in vitro experiments but could be a factor in the in vivo experiments. A brief mention of the possible sites of metabolic breakdown is imperative from the standpoint that the activity exhibited by these compounds may be, in part, a function of their lifetime in their original form. It is a possibility that some of the compounds were significantly more metabolized than others. 186 However, it should be remembered that the infusion regimen used to test the RSD compounds in vivo was chosen so as to nullify, wherever possible, metabolic effects interfering with the results. Thus it was reasoned that using the infusion protocol described in this study, it might have been possible to nearly saturate any a-phase of redistribution within the first four minutes and then maintain a quasi steady-state for the duration of the experiment. Indeed, where pharmacokinetic studies have been performed with Nortran compounds (personal communication from Dr. Lillian Clohs at Nortran Pharmaceuticals Inc.), it was found that, over the period 1 to 15 minutes following the start of in vivo infusion, serum concentrations rose at most 6% to 55% times. Similarly, in the sham occlusion and infusion of RSD 1000 reported previously (Yong et al., 1999), steady-state responses were seen five minutes into the infusion time. The metabolic degradation of these compounds is open for future investigation but, for the sake of completeness, the present section briefly considers the possible metabolism of the RSD compounds used in this study. Based on the core pharmacophore, one possible site for biochemical "attack" is the ester linkage that exists between the cyclohexyl ring and the aryl substituents (see below). ESTERASE H p ^ "'N I + ° \ C / R 2 II o This hydrolyzable function may be a likely target for hydrolytic enzymes such as carboxyl esterases which are present in plasma and in most, i f not all, tissues. Despite ubiquitous distribution in vivo, the liver microsomal fraction is considered the richest 187 source of esterases (Junge and Krisch, 1975), with as many as nine subtypes in microsomes and plasma membranes of the rat liver (Blomberg and Raftell, 1974). In fact, it is well known that in the rodent species drug metabolic activity is greatest in hepatic tissues (Litterst et al., 1975). Analogues in which the ester linkage was replaced with an amide bond were found to be more stable to oral (relative to intravenous) administration (unpublished results) than their ester counterparts. The stability of the amide bond owes its strength to the sp3-sp2 bond between the nitrogen the carbonyl group as well as to the electronegativity of the nitrogen atom. In addition, there is a lower level of non-specific amidase activity as compared to esterase (Junge and Krisch, 1975). High-performance liquid chromatography analysis on rat blood and tissues was performed on two amide derivatives to determine their tissue concentration immediately following an intravenous injection (Walker et al., 1996). Measurement of drug levels was carried out over the first 15 min of administration and redistribution of drugs from the blood to other tissue compartments was in the order of heart > brain > liver > and skeletal muscle with 70-90% extraction of the original dose. Other possible routes of metabolic degradation (predominantly by P450 systems) are illustrated in the figure below and these are denoted as 1) hydroxylation, 2) aromatic oxidation and 3) N-dealkylation with the latter found as the major pathway according to plasma extraction and analysis in a 1-naphthyl analogue of RSD 1000. 1 188 5„ References Abbott, J.A. Intravenous antiarrhythmic drugs: Newer aspects of therapy. Angiology 33, 251-258, 1982. Adaikan, G., Beatch, G.N., Lee, T.L., Ratnam, S.S., and Walker, M J . Antiarrhythmic actions of tedisamil: Studies in rats and primates. Cardiovascular Drugs and Therapy 6, 345-352, 1992. Adelman, W.J., Jr. and Palti, Y . The effects of external potassium and long duration voltage conditioning on the amplitude of sodium currents in the giant axon of the squid, loligo pealei. Journal of General Physiology. 54, 589-606, 1969. Adelman, J.P. Proteins that interact with the pore-forming subunits of voltage- gated ion channels. Current Opinion in Neurobiology 5, 286-295, 1995. Aksnes, G., Kirkeboen, K . A . , Lande, K. , and Ilebekk, A . myocardial potassium balance associated with regional ischaemia in the pig: Effects of beta-adrenoceptor blockade, duration of ischaemia and preceding ischaemic periods. Acta Physiologica Scandinavica 145, 39-48, 1992. Aldrich, R.W. Potassium channels. New channel subunits are a turn-off. Current Biology 4, 839-840, 1994. Allessie, M.A . , Bonke, F.I., and Schopman, F.J. Circus movement in rabbit atrial muscle as a mechanism of tachycardia. Circulation Research 33, 54-62: 1973. Allessie, M.A . , Bonke, F.I., and Schopman, F.J. Circus movement in rabbit atrial muscle as a mechanism of tachycardia. II. The role of nonuniform recovery of excitability in the occurrence of unidirectional block, as studied with multiple microelectrodes. Circulation Research 39, 168-177, 1976. Aimers, W., Stanfield, P. R. and Stuhmer, W. Slow changes in currents through sodium channels in frog muscle membrane. Journal of Physiology 339: 253-271, 1983. Anderson, J.L., Rodier, H.E., and Green, L.S. Comparative effects of beta-adrenergic blocking drugs on experimental ventricular fibrillation threshold. American Journal of Cardiology 51, 1196-202: 1983. Antzelevitch, C , Sicouri, S., Litovsky, S.H., Lukas, A. , Krishnan, S.C., D i Diego, J.M., Gintant, G.A., and Liu, D.W. Heterogeneity within the ventricular wall. Electrophysiology and pharmacology of epicardial, endocardial, and M cells. Circulation Research 69, 1427-49: 1991. 189 Apkon, M . and Nerbonne, J .M. Characterization of two distinct depolarization-activated K + currents in isolated adult rat ventricular myocytes. Journal of General Physiology. 97, 973-1011, 1991. Arita, M . , Kiyosue, T., Aomine, M . , and Imanishi, S. Nature of "residual fast channel" dependent action potentials and slow conduction in guinea pig ventricular muscle and its modification by isoproterenol. American Journal of Cardiology 51, 1433-1440, 1983. Armstrong, C M . Sodium channels and gating currents. Physiological Reviews 61, 644-683:1981. Arnsdorf, M.F. and Bigger, J.T., Jr. Effect of lidocaine hydrochloride on membrane conductance in mammalian cardiac Purkinje fibers. Journal of Clinical Investigation 51, 2252-63: 1972. Backx, P.H., Yue, D.T., Lawrence, J.H., Marban, E., and Tomaselli, G.F. Molecular localization of an ion-binding site within the pore of mammalian sodium channels. Science 257, 248-251, 1992. Backx, P.H. and Yue, D.T. Proton permeation through cardiac sodium channels. Biophysical Journal. 59, 26, 1991. Baldwin, T.J., Tsaur, M.L . , Lopez, G.A., Jan, Y . N . , and Jan, L . Y . Characterization of a mammalian cDNA for an inactivating voltage-sensitive K + channel. Neuron 7, 471-483, 1991. Barchi, R. L. Molecular pathology of the skeletal muscle sodium channel. Annual Review of Physiology 57: 355-85: 1995. Barchi, R.L. Protein components of the purified sodium channel from rat skeletal muscle sarcolemma. Journal of Neurochemistry 40, 1377-1385, 1983. Barhanin, J., Schmid, A. , Lombet, A. , Wheeler, K.P., Lazdunski, M . , and Ellory, J.C. Molecular size of different neurotoxin receptors on the voltage-sensitive Na + Channel. Journal of Biological Chemistry 258, 700-702, 1983. Barrett, T.D., Hayes, E.S., and Walker, M.J. Lack of Selectivity for ventricular and ischaemic tissue limits the antiarrhythmic actions of lidocaine, quinidine and flecainide against ischaemia-induced arrhythmias. European Journal of Pharmacology. 285, 229-238, 1995. Barry, D .M. , Trimmer, J.S., Merlie, J.P., and Nerbonne, J .M. Differential expression of voltage-gated K + channel subunits in adult rat heart. Relation to functional K + channels? Circulation Research 77, 361-369, 1995. 190 Bean, B.P., Cohen, C.J., and Tsien, R.W. Lidocaine block of cardiac sodium channels. Journal of General Physiology. 81,613-42: 1983. Beatch, G.N., Abraham, S., MacLeod, B.A. , Yoshida, N.R., and Walker, M J . Antiarrhythmic properties of tedisamil (KC8857), A putative transient outward K + current blocker. British Journal of Pharmacology. 102, 13-18, 1991. Beeler, G.W., Jr. and Reuter, H . Voltage clamp experiments on ventricular myocarial fibres. Journal of Physiology 207, 165-90: 1970. Beneski, D.A. and Catterall, W.A. Covalent labeling of protein components of the sodium channel with a photoactivable derivative of scorpion toxin. Proceedings of the National Academy of Sciences of the United States of America 77, 639-643, 1980. Betancourt, O J . and Dresel, P.E. Interaction of lidocaine and calcium on the electrical characteristics of rabbit atria. Journal of Pharmacology & Experimental Therapeutics 210, 64-69, 1979. Bezanilla, F. and Armstrong, C M . Inactivation of the sodium channel. I. Sodium current experiments. Journal of General Physiology. 70, 549-66: 1977. Bielen, F.V., Bosteels, S., and Verdonck, F. Consequences of C 0 2 acidosis for transmembrane Na + transport and membrane current in rabbit cardiac Purkinje fibres. Journal of Physiology 427, 325-345, 1990. Bigger, J.T., Jr. and Mandel, W.J. Effect of lidocaine on conduction in canine Purkinje fibers and at the ventricular muscle-Purkinje fiber junction. Journal of Pharmacology & Experimental Therapeutics 172, 239-54: 1970. Blair, T.A., Roberds, S.L., Tamkun, M . M . , and Hartshorne, R.P. Functional characterization of rK5, a voltage-gated K + channel cloned from the rat cardiovascular system. FEBS Letters 295, 211-213, 1991. Blomberg, F. and Raftell, M . Enzyme polymorphism in rat-liver microsomes and plasma membranes. 1. an immunochemical study of multienzyme complexes and other enzyme-active antigens. European Journal of Biochemistry 49: 21-9: 1974. Borgeat, A. , Goy, J.J., Maendly, R., Kaufmann, U . , Grbic, M . , and Sigwart, U . Flecainide versus quinidine for conversion of atrial fibrillation to sinus rhythm. American Journal of Cardiology 58,496-498, 1986. Boyett, M.R., Kirby, M.S., and Orchard, C H . The effects of acidosis on single ventricular cells isolated from ferret heart. Journal of Physiology. 407, 114P1988. 191 Braasch, W., Gudbjarnason, S., Puri, P.S., Ravens, K .G . , and Bing, R J . Early changes in energy metabolism in the myocardium following acute coronary artery occlusion in anesthetized dogs. Circulation Research 23, 429-38: 1968. Bril , A . and Man, R.Y. Effects of the potassium channel activator, BRI 34915, on the action potential characteristics of canine cardiac Purkinje fibers. Journal of Pharmacology & Experimental Therapeutics 253, 1090-1096, 1990. Brooks, C M . , Hoffman, B.F., Suckling, E.E., and Orias, O. Excitability of the heart. Grune & Stratton Inc., New York, 1955. Brown, A . M . , Lee, K.S., and Powell, T. Sodium current in single rat heart muscle cells. Journal of Physiology 318, 479-500, 1981. Brown, R.H., Jr., Cohen, I., and Noble, D. The interactions of protons, calcium and potassium ions on cardiac Purkinje fibres. Journal of Physiology 282, 345-352, 1978. Brugada, J., Boersma, L. , Kirchhof, C , and Allessie, M . Proarrhythmic effects of flecainide. experimental evidence for increased susceptibility to reentrant arrhythmias. Circulation 84, 1808-1818, 1991. Burgen, A.S., Roberts, G . C , and Feeney, J. Binding of flexible ligands to macromolecules. Nature 253, 753-755: 1975. Burley, S.K. and Petsko, G.A. Aromatic-aromatic interaction: a mechanism of protein structure stabilization. Science 229, 23-28, 1985. Butler, A . , Wei, A .G . , Baker, K. , and Salkoff, L . A family of putative potassium channel genes in Drosophila. Science 243, 943-947, 1989. Cahalan, M.D. and Aimers, W. Interactions between quaternary lidocaine, the sodium channel gates, and tetrodotoxin. Biophysical Journal 27, 39-55, 1979. Campbell, T.J. and Vaughan Williams, E . M . Voltage- and time-dependent depression of maximum rate of depolarisation of guinea-pig ventricular action potentials by two new antiarrhythmic drugs, flecainide and lorcainide. Cardiovascular Research 17, 251-258, 1983. Cannon, S. C. Ion-channel defects and aberrant excitability in myotonia and periodic paralysis. Trends in Neurosciences 19: 3-10, 1996. Carmeliet, E. Chloride ions and the membrane potential of Purkinje fibres. Journal of Physiology 156, 375-388, 1961. 192 Carmeliet, E. and Saikawa, T. Shortening of the action potential and reduction of pacemaker activity by lidocaine, quinidine, and procainamide in sheep cardiac Purkinje fibers. A n effect on Na + or K + currents? Circulation Research 50, 257-72: 1982. Carmeliet, E. Mechanisms and control of repolarization. European Heart Journal 14, Suppl H:3-13, 1993. Carson, D.L., Cardinal, R., Savard, P., Vasseur, C , Nattel, S., Lambert, C , and Nadeau, R. Relationship between an arrhythmogenic action of lidocaine and its effects on excitation patterns in acutely ischemic porcine myocardium. Journal of Cardiovascular Pharmacology 8, 126-136, 1986. C A S C A D E Investigators The C A S C A D E Study: Randomized and anti-arrhythmic drug therapy in survivors of cardiac arrest in Seattle. American Journal of Cardiology. 72, 280-287, 1993. Cascio, W.E., Johnson, T.A., and Gettes, L.S. Electrophysiologic changes in ischemic ventricular myocardium: I. influence of ionic, metabolic, and energetic changes. Journal of Cardiovascular Electrophysiology 6, 1039-1062, 1995. Catterall, W.A. Molecular properties of voltage-sensitive sodium channels. Annual Review of Biochemistry 55, 953-985: 1986. Catterall, W.A. Ion channels in plasma membrane signal transduction. Journal of Bioenergetics & Biomembranes 28, 217-8: 1996. Ceremuzynski, L. , Kleczar, E., Krzeminska-Pakula, M . , Kuch, J., Nartowicz, E., Smielak-Korombel, J., Dyduszynski, A. , Maciejewicz, J., Zaleska, T., Lazarczyk-Kedzia, E., and et al Effect of amiodarone on mortality after myocardial infarction: a double-blind, placebo-controlled, pilot study Journal of the American College of Cardiology 20, 1056-1062, 1992. Chandy, K . G . Simplified gene nomenclature. Nature 352, 26:1991. Chang, W. S., Akiyama, T., Richeson, J. F., Faillace, R. T. and Serrino, P. Origin of the giant r wave in acute transmural myocardial infarction in the pig. Japanese Heart Journal 30: 863-883, 1989. Chesnais, J .M., Coraboeuf, E., Sauviat, M.P., and Vassas, J .M. Sensitivity to H + , L f and M g 2 + ions of the slow inward sodium current in frog atrial fibres. Journal of Molecular & Cellular Cardiology 7, 627-42: 1975. Cheung, P.H., Pugsley, M.K. , and Walker, M.J. Arrhythmia models in the rat. Journal of Pharmacological & Toxicological Methods 29, 179-184, 1993. 193 Christie, M.J. , North, R.A., Osborne, P.B., Douglass, J., and Adelman, J.P. Heteropolymeric potassium channels expressed in Xenopus oocytes from cloned subunits. Neuron 4, 405-411, 1990. Clark, R.B., Giles, W.R., and Imaizumi, Y . Properties of the transient outward current in rabbit atrial cells. Journal of Physiology 405, 147-168, 1988. Clarkson, C.W. and Hondeghem, L . M . Mechanism for bupivacaine depression of cardiac conduction: fast block of sodium channels during the action potential with slow recovery from block during diastole. Anesthesiology 62, 396-405, 1985a. Clarkson, C.W. and Hondeghem, L . M . Evidence for a specific receptor site for lidocaine, quinidine, and bupivacaine associated with cardiac sodium channels in guinea pig ventricular myocardium. Circulation Research. 56, 496-506, 1985b. Cleemann, L. and Morad, M . Potassium currents in frog ventricular muscle: evidence from voltage clamp currents and extracellular K + accumulation. Journal of Physiology 286, 113-143, 1979. Cleemann, L . Heart muscle, intracellular potassium and inward-going rectification. Biophysical Journal 36, 303-310, 1981. Cohen, S.A. and Barchi, R.L. Glycoprotein characteristics of the sodium channel saxitoxin-binding component from mammalian sarcolemma. Biochimica et Biophysica Acta 645,253-261, 1981. Cohen, S.A. and Levitt, L .K . Partial characterization of the rhl sodium channel protein from rat heart using subtype-specific antibodies. Circulation Research 73, 735-742, 1993. Colatsky, T.J. Mechanisms of action of lidocaine and quinidine on action potential duration in rabbit cardiac Purkinje fibers. An effect on steady state sodium currents? Circulation Research. 50, 17-27, 1982. Comer, M.B. , Campbell, D.L., Rasmusson, R.L., Lamson, D.R., Morales, M.J. , Zhang, Y. , and Strauss, H.C. Cloning and characterization of an I T 0-like potassium channel from ferret ventricle. American Journal of Physiology 267, H1383-H1395, 1994. Cook, N.S. Potassium channels: Structure, classification, function, and therapeutic potential., Nigel, N.S., ed., Chichester, West Sussex, England: Ellis Horwood Limited, 1990. Coraboeuf, E. Ionic basis of electrical activity in cardiac tissues. American Journal of Physiology. 234, H101-H116, 1978. 194 Coraboeuf, E. and Carmeliet, E. Existence of two transient outward currents in sheep cardiac Purkinje fibers. Pflugers Archiv - European Journal of Physiology 392, 352-359, 1982. Coronel, R., Fiolet, J.W., Wilms-Schopman, F.J., Schaapherder, A.F. , Johnson, T.A., Gettes, L.S., and Janse, M.J. Distribution of extracellular potassium and its relation to electrophysiologic changes during acute myocardial ischemia in the isolated perfused porcine heart. Circulation 77, 1125-1138, 1988. Corr, P.B. and Crafford, W.A., Jr. Enhanced alpha-adrenergic responsiveness in ischemic myocardium: role of alpha-adrenergic blockade. American Heart Journal 102, (Pt 2):605-612, 1981. Costa, M.R. and Catterall, W.A. Phosphorylation of the alpha subunit of the sodium channel by protein kinase c. Cellular & Molecular Neurobiology 4, 291-297, 1984. Courtney, K.R. Mechanism of frequency-dependent inhibition of sodium currents in frog myelinated nerve by the lidocaine derivative GEA968. Journal of Pharmacology & Experimental Therapeutics 195, 225-236, 1975. Courtney, K.R. Why do some drugs preferentially block open sodium channels? Journal of Molecular & Cellular Cardiology. 20,461-464, 1988. Courtney, K.R. Sodium channel blockers: the size/solubility hypothesis revisited. Molecular Pharmacology 37, 855-859, 1990. Cranefield, P.F. and Aronson, R.S. Torsade de pointes and other pause-induced ventricular tachycardias: the short-long-short sequence and early afterdepolarizations. Pacing & Clinical Electrophysiology 11, (Pt l):670-678, 1988. Crijns, H.J., van Wijk, L . M . , van Gilst, W.H., Kingma, J.H., van Gelder, L C , and Lie, K.I. Acute conversion of atrial fibrillation to sinus rhythm: clinical efficacy of flecainide acetate, comparison of two regimens. European Heart Journal 9, 634-638, 1988. Cross, H.R., Clarke, K. , Opie, L .H. , and Radda, G.K. Is lactate-induced myocardial ischaemic injury mediated by decreased ph or increased intracellular lactate? Journal of Molecular & Cellular Cardiology 27, 1369-1381, 1995. Curtis, M.J. , Johnston, K . M . , MacLeod, B.A. , and Walker, M.J. The actions of felodipine on arrhythmias and other responses to myocardial ischaemia in conscious rats. European Journal of Pharmacology 117, 169-178, 1985a. Curtis, M.J. , Johnston, K . M . , and Walker, M.J. Arrhythmias and serum potassium during myocardial ischaemia. IRCSMed. Sci. 13, 688-689,1985b. 195 Curtis, M.J. , MacLeod, B.A. , and Walker, M.J. The effects of ablations in the central nervous system on arrhythmias induced by coronary occlusion in the rat. British Journal of Pharmacology 86, 663-670, 1985c. Curtis, M.J. , MacLeod, B.A. , and Walker, M.J. Models for the study of arrhythmias in myocardial ischaemia and infarction: the use of the rat. Journal of Molecular & Cellular Cardiology 19, 399-419: 1987. Curtis, M.J. and Walker, M.J. Quantification of arrhythmias using scoring systems: an examination of seven scores in an in vivo model of regional myocardial ischaemia. Cardiovascular Research. 22, 656-665, 1988. Curtis, M.J. The rabbit dual coronary perfusion model: a new method for assessing the pathological relevance of individual products of the ischaemic milieu: role of potassium in arrhythmogenesis. Cardiovascular Research 25, 1010-1022, 1991. Curtis, M.J. Characterisation, utilisation and clinical relevance of isolated perfused heart models of ischaemia-induced ventricular fibrillation. Cardiovascular Research 39, 194-215, 1998. Darbon, H. , Jover, E., Couraud, F., and Rochat, H . Photoaffinity labeling of alpha- and beta- scorpion toxin receptors associated with rat brain sodium channel. Biochemical & Biophysical Research Communications 115, 415-422, 1983. Deal, K . K . , England, S.K., and Tamkun, M . M . Molecular physiology of cardiac potassium channels. Physiological Reviews 76, 49-67, 1996. Deck, K . A . and Trautwein, W. Ionic currents in cardiac excitation. Pflug. Archiv. ges. Physiol. 280, 63-80, 1964. Denaro, C P . and Benowitz, N .L . Poisoning due to class lb antiarrhythmic drugs, lignocaine, mexiletine and tocainide. Medical Toxicology & Adverse Drug Experience 4, 412-428, 1989. DiFrancesco, D. and Noble, D. A model of cardiac electrical activity incorporating ionic pumps and concentration changes. Philosophical Transactions of the Royal Society of London - Series B: Biological Sciences 307, 353-98: 1985. Dixon, J.E. and McKinnon, D. Quantitative analysis of potassium channel mRNA expression in atrial and ventricular muscle of rats. Circulation Research 75, 252-260, 1994. Dixon, J.E., Shi, W., Wang, H.S., McDonald, C , Yu , H. , Wymore, R.S., Cohen, I.S., and McKinnon, D. Role of the Kv4.3 K + channel in ventricular muscle. A molecular correlate for the transient outward current. Circulation Research. 19, 659-668, 1996. 196 Dougherty, D.A. and Stauffer, D.A. Acetylcholine binding by a synthetic receptor: implications for biological recognition. Science 250, 1558-1560, 1990. Downar, E., Janse, M.J. , and Durrer, D. The effect of "ischemic" blood on transmembrane potentials of normal porcine ventricular myocardium. Circulation 55, 455-462, 1977. Duff, H.J., Rahmberg, M . , and Sheldon, R.S. Role of quinidine in the mexiletine-quinidine interaction: electrophysiologic correlates of enhanced antiarrhythmic efficacy. Journal of Cardiovascular Pharmacology 16, 685-692, 1990. Duff, H.J., Mitchell, L .B. , Wyse, D.G., Gillis, A . M . , and Sheldon, R.S. Mexiletine/quinidine combination therapy: electrophysiologic correlates of anti-arrhythmic efficacy. Clinical. &. Investigative. Medicine - Medecine. Clinique. et. Experimentale. 14(5), 476-483, 1991. Durell, S.R. and Guy, H.R. Atomic scale structure and functional models of voltage-gated potassium channels. Biophysical Journal 62, 238-247, 1992. Echt, D.S., Liebson, P.R., Mitchell, L .B . , Peters, R.W., Obias-Manno, D., Barker, A .H . , Arensberg, D., Baker, A . , Friedman, L. , Greene, H.L., Huther, M.L-., Richardson, D.W., and and the CAST Investigators Mortality and morbidity in patients receiving encainide, flecainide, or placebo. The Cardiac Arrhythmia Suppression Trial New England Journal of Medicine 324, 781-788, 1991. Ehrlich, P. Chemo therapeutics: scientific principles, methods and results. Lancet, ii, 445-451. El-Sherif, N . , Hope, R.R., Scherlag, B.J., and Lazzara, R. Re-entrant ventricular arrhythmias in the late myocardial infarction period. 2. Patterns of initiation and termination of re-entry. Circulation 55, 702-719, 1977a. El-Sherif, N . , Scherlag, B.J., Lazzara, R., and Hope, R.R. Re-entrant ventricular arrhythmias in the late myocardial infarction period. 1. Conduction characteristics in the infarction zone. Circulation 55, 686-702, 1977b. Escande, D., Coraboeuf, E., and Planche, C. Abnormal pacemaking is modulated by sarcoplasmic reticulum in partially-depolarized myocardium from dilated right atria in humans. Journal of Molecular & Cellular Cardiology 19, 231-241, 1987a. Escande, D., Coulombe, A. , Faivre, J.F., Deroubaix, E., and Coraboeuf, E. Two types of transient outward currents in adult human atrial cells. American Journal of Physiology 252, (Pt 2):H142-H148, 1987b. 197 Fermini, B. , Wang, Z., Duan, D., and Nattel, S. Differences in rate dependence of transient outward current in rabbit and human atrium. American Journal of Physiology 263, Pt 2):H1747-54, 1992. Fiset, C , Clark, R.B., Shimoni, Y . , and Giles, W.R. Shal-type channels contribute to the Ca2+-independent transient outward K + current in rat ventricle. Journal of Physiology 500, 51-64, 1997. Fozzard, H.A. and Arnsdorf, M.F. Cardiac electrophysiology. In: The heart and cardiovascular system. H.A. Fozzard, ed., New York: Raven Press, pp. 1-30. 1986. Frazier, D.T., Narahashi, T., and Yamada, M . The site of action and active form of local anesthetics. II. Experiments with quaternary compounds. Journal of Pharmacology & Experimental Therapeutics. 171, 45-51, 1970. Freedberg, A.S., Papp, J.G., and Williams, E . M . The effect of altered thyroid state on atrial intracellular potentials. Journal of Physiology 207, 357-69: 1970. Frey, W. Weitere Erfhrungen mit Chinidin bei absoluter Herzunregelmassigkeit. Wien Klin. Wochenschr. 55, 849-853, 1918. Garlick, P.B., Radda, G.K., and Seeley, P.J. Studies of acidosis in the ischaemic heart by phosphorus nuclear magnetic resonance. Biochemical Journal 184, 547-554, 1979. Gettes, L . and Surawicz, B. Effects of low and high concentrations of potassium on the simultaneously recorded Purkinje and ventricular action potentials of the perfused pig moderator band. Circulation Research 23, 717-29: 1968. Gettes, L.S. Electrolyte abnormalities underlying lethal and ventricular arrhythmias. Circulation 85,170-176, 1992. Gettes, L.S. and Cascio, W.E. Effect of acute ischemia on cardiac electrophysiology. In: The heart and cardiovascular system. H.A. Fozzard, R.B. Jennings, E. Haber, A . M . Katz, and H.E. Morgan, eds., New York: Raven Press, pp. 2021-2054, 1992. Giardina, E.G. and Bigger, J.T., Jr. Procaine amide against re-entrant ventricular arrhythmias, lengthening R-V intervals of coupled ventricular premature depolarization as an insight into the mechanism of action of procaine amide. Circulation 48, 959-70: 1973. Gintant, G.A., Cohen, I.S., Datyner, N.B. , and Kline, R.P. Time-dependent outward currents in the heart. In The heart and cardiovascular system. H.A. Fozzard, ed., New York: Raven Press, pp. 1121-1169, 1992. 198 Glowka, M . L . , Dargie, R.L., and Codding, P.W. Spatial requirements of the Na + channel binding site for class I antiarrhythmics as derived from the crystal structures of 4-substituted 2,6-bis(l-pyrrolidinylmethyl)phenols. Journal of Medicinal Chemistry 34, 2678-2684, 1991. Goldin, A . L . and Sumikawa, K . Preparation of R N A for injection into Xenopus oocytes. Methods in Enzymology 207, 279-97: 1992. Goldman, D.E. Potential, impedance, and rectification in membranes. Journal of General Physiology. 27, 37-60, 1943. Grant, A.O., Strauss, L.J. , Wallace, A.G. , and Strauss, H.C. The influence of pH on the electrophysiological effects of lidocaine in guinea pig ventricular myocardium. Circulation Research. 47, 542-550, 1980. Guo, P., Pugsley, M . K . , Yong, S.L., and Walker, M J . Cardiac transplantation does not effect ischaemia-induced arrhythmias in rats. Cardiovascular Research. 43, 930-938, 1999. Hall, A.E. , Hutter, O.F., and Noble, D. Current-voltage relations of Purkinje fibres in sodium-deficien solutions. Journal of Physiology 166, 225-240, 1963. Harris, A.S. Delayed development of ventricular ectopic rhythm following experimental coronary occlusion. Circulation 1, 1318-1328, 1950. Harris, A.S., Bisteni, A. , Russell, R.A., Brigham, J .C , and Firestone, J.E. Excitatory factors in ventricular tachycardia resulting from myocardial ischemia: Potassium a major excitant. Science 199, 200-203, 1954. Hartmann, H.A., Kirsch, G.E., Drewe, J.A., Taglialatela, M . , Joho, R.H., and Brown, A . M . Exchange of conduction pathways between two related K + channels. Science 251, 942-944, 1991. Heger, J.J., Prystowsky, E.N., Jackman, W.M. , Naccarelli, G.V., Warfel, K .A . , Rinkenberger, R.L., and Zipes, D.P. Clinical efficacy and electrophysiology during long-term therapy for recurrent ventricular tachycardia or ventricular fibrillation. New England Journal of 'Medicine 305, 539-545, 1981. Heidbuchel, H. , Vereecke, J., and Carmeliet, E. Three different potassium channels in human atrium. Contribution to the basal potassium conductance. Circulation Research 66, 1277-1286, 1990. Heinemann, S.H., Terlau, H. , Stuhmer, W., Imoto, K. , and Numa, S. Calcium channel characteristics conferred on the sodium channel by single mutations. Nature 356, 441-443, 1992. 199 Hellestrand, K . J. Efficacy and safety of long-term oral flecainide acetate in patients with responsive supraventricular tachycardia. American Journal of Cardiology 7 7 , 83A-88A, 1996. Herre, J .M., Sauve, M.J. , Malone, P., Griffin, J.C., Helmy, I., Langberg, J.J., Goldberg, FL, and Scheinman, M . M . Long-term results of amiodarone therapy in patients with recurrent sustained ventricular tachycardia or ventricular fibrillation. Journal of the American College of Cardiology 1 3 , 442-449, 1989. Hi l l , J.L. and Gettes, L.S. Effect of acute coronary artery occlusion on local myocardial extracellular K + activity in swine. Circulation 6 1 , 768-778, 1980. Hille, B . The pH-dependent rate of action of local anesthetics on the node of Ranvier. Journal of General Physiology. 6 9 , 475-496, 1977a. Hille, B. Local anesthetics: hydrophilic and hydrophobic pathways for the drug-receptor reaction. Journal of General Physiology. 6 9 , 497-515, 1977b. Hille, B. Ionic channels of excitable membranes. In: Hille, B. , ed., Sunderland, Mass. USA: Sinauer Associates, Inc., 1984. Hille, B. Ionic channels of excitable membranes. In: Hille, B. , ed., Sunderland, Mass. USA: Sinauer Associates, Inc., 1992. Hilvert, D., Hatanaka, Y . , and Kaiser, E.T. A highly active thermophilic semisynthetic flavoenzyme. Journal of American Chemical Society. 1 1 0 , 682-689, 1988. Hiraoka, M . and Kawano, S. Calcium-sensitive and insensitive transient outward current in rabbit ventricular myocytes. Journal of Physiology 410,187-212, 1989. Ho, K. , Nichols, C.G., Lederer, W.J., Lytton, J., Vassilev, P .M. , Kanazirska, M.V. , and Hebert, S.C. Cloning and expression of an inwardly rectifying ATP-regulated potassium channel. Nature 3 6 2 , 31-38, 1993. Hodgkin, A . L . and Huxley, A.F. A quantitative description of membrane current and its application to conduction and excitation in nerve. Journal of General Physiology. 6 9 , 497-515, 1952a. Hodgkin, A . L . and Huxley, A.F. A quantitative description of membrane current and its application to conduction and excitation in nerve. Journal of Physiology. 1 1 7 , 500-544, 1952b. Hodgkin, A . L . and Katz, B. The effect of sodium ions on the electrical activity of the giant axon of the squid. Journal of Physiology. (Lond) 1 0 8 , 37-77, 1949. 200 Hoffman, B.F. and Cranefield, P.F. Electrophysiology of the heart. McGraw Hi l l , New York, 1960. Hondeghem, L . M . Antiarrhythmic agents: modulated receptor applications. Circulation 75,514-520, 1987. Hondeghem, L . M . and Katzung, B.G. Time- and voltage-dependent interactions of antiarrhythmic drugs with cardiac sodium channels. Biochimica. et. Biophysica. Acta 472, 373-398, 1977. Hondeghem, L . M . and Snyders, D J . Class III antiarrhythmic agents have a lot of potential but a long way to go. Reduced effectiveness and dangers of reverse use dependence. Circulation 81, 686-690, 1990. Hoshi, T., Zagotta, W.N., and Aldrich, R.W. Biophysical and molecular mechanisms of Shaker potassium channel inactivation. Science 250, 533-538, 1990. Hoshi, T., Zagotta, W.N., and Aldrich, R.W. Two types of inactivation in shaker K + channels: effects of alterations in the carboxy-terminal region. Neuron 7, 547-556, 1991. Hoshi, T. and Zagotta, W.N. Recent advances in the understanding of potassium channel function. Current Opinion in Neurobiology 3, 283-290: 1993. Hutter, O.F. and Noble, D. Rectifying properties of cardiac muscle. Nature (Lond.) 188, 495, 1960. Igwemezie, L .N . , Beatch, G.N., McErlane, K . M . , and Walker, M J . Mexiletine's antifibrillatory actions are limited by the occurrence of convulsions in conscious animals. European Journal of Pharmacology 210, 271-277, 1992. Imaizumi, Y . and Giles, W.R. Quinidine-induced inhibition of transient outward current in cardiac muscle. American Journal of Physiology 253, (Pt 2):H704-H708, 1987. Imanishi, S., Arita, M . , Kiyosue, T., and Aomine, M . Effects of SG-75 (nicorandil) on electrical activity of canine cardiac Purkinje fibers: possible increase in potassium conductance. Journal of Pharmacology & Experimental Therapeutics 225, 198-205, 1983. Imanishi, S., Arita, M . , Aomine, M . , and Kiyosue, T. Antiarrhythmic effects of nicorandil on canine cardiac purkinje fibers. Journal of Cardiovascular Pharmacology 6, 772-779, 1984. Irisawa, H. and Sato, R. Intra- and extracellular actions of proton on the calcium current of isolated guinea pig ventricular cells. Circulation Research 59, 348-355, 1986. Isom, L .L . , De Jongh, K.S., and Catterall, W.A. Auxiliary subunits of voltage-gated ion channels. Neuron 12, 1183-94: 1994. 201 Ito, H. , Tung, R.T., Sugimoto, T., Kobayashi, I., Takahashi, K. , Katada, T., U i , M . , and Kurachi, Y . On the mechanism of G protein beta gamma subunit activation of the muscarinic K + channel in guinea pig atrial cell membrane, comparison with the ATP-sensitive K + channel. Journal of General Physiology. 99, 961-983, 1992. Jaffe, A.S. The use of antiarrhythmics in advanced cardiac life support. Annals of Emergency Medicine 22, (Pt 2):307-316, 1993. James, M.A . , Papouchado, P., and Jones, J.V. Combined anti-arrhythmic therapy with class I and class III drugs. Postgraduate Medical Journal 65, 519-524, 1989. Jan, L . Y . and Jan, Y . N . A superfamily of ion channels. Nature 345, 672:1990a. Jan, L . Y . and Jan, Y . N . How might the diversity of potassium channels be generated?. Trends in Neurosciences 13, 415-419, 1990b. Jan, L . Y . and Jan, Y . N . Tracing the roots of ion channels. Cell 69, 715-718, 1992. Janse, M.J. , Cinca, J., Morena, H. , Fiolet, J.W., Kleber, A .G . , de Vries, G.P., Becker, A.E. , and Durrer, D. The "border zone" in myocardial ischemia, an electrophysiological, metabolic, and histochemical correlation in the pig heart. Circulation Research. 44, 576-588: 1979. Janse, M.J. , van Capelle, F.J., Morsink, H. , Kleber, A .G . , Wilms-Schopman, F., Cardinal, R., d'Alnoncourt, C.N., and Durrer, D. Flow of "injury" current and patterns of excitation during early ventricular arrhythmias in acute regional myocardial ischemia in isolated porcine and canine hearts. Evidence for two different arrhythmogenic mechanisms. Circulation Research. 47, 151-165, 1980a. Janse, M.J. , van Capelle, F.J., Morsink, H. , Kleber, A .G . , Wilms-Schopman, F., Cardinal, R., d'Alnoncourt, C.N., and Durrer, D. Flow of "injury" current and patterns of excitation during early ventricular arrhythmias in acute regional myocardial ischemia in isolated porcine and canine hearts, evidence for two different arrhythmogenic mechanisms. Circulation Research. 47, 151-65: 1980b. Janse, M.J. and Kleber, A . G . Electrophysiological changes and ventricular arrhythmias in the early phase of regional myocardial ischemia. Circulation Research. 49, 1069-1081, 1981a. Janse, M.J. and Kleber, A . G . Electrophysiological changes and ventricular arrhythmias in the early phase of regional myocardial ischemia. Circulation Research 49, 1069-1081: 1981b. 202 Janse, M.J . and van Capelle, F.J. Electrotonic interactions across an inexcitable region as a cause of ectopic activity in acute regional myocardial ischemia, a study in intact porcine and canine hearts and computer models. Circulation Research 50 , 527-537, 1982. Johnson, E.A. and McKinnon, M . G . The differential effect of quinidine and pyrilamine on the myocardial action potential at various rates of stimulation. Journal of Pharmacology & Experimental Therapeutics. 120, 460-468, 1957. Johnson, E.A. and Lieberman, M . Heart: excitation and contraction. Annual Review of Physiology 33, 479-532: 1971. Johnson, P.N., Freedberg, A.S., and Marshall, J .M. Action of thyroid hormone on the transmembrane potentials from sinoatrial node cells and atrial muscle cells in isolated atria of rabbits. Cardiology 58 , 273-89: 1973. Josephson, I.R., Sanchez-Chapula, J., and Brown, A . M . Early outward current in rat single ventricular cells. Circulation Research 54 , 157-162, 1984a. Josephson, I.R., Sanchez-Chapula, J., and Brown, A . M . A comparison of calcium currents in rat and guinea pig single ventricular cells. Circulation Research 54 , 144-156, 1984b. Junge, W. and Krisch, K . The carboxylesterases/amidases of mammalian liver and their possible significance. CRC Critical Reviews in Toxicology 3: 371-435: 1975. Kallen, R.G., Sheng, Z.H. , Yang, J., Chen, L.Q., Rogart, R.B., and Barchi, R.L. Primary structure and expression of a sodium channel characteristic of denervated and immature rat skeletal muscle. Neuron 4, 233-242, 1990. Kamb, A. , Tseng-Crank, J., and Tanouye, M . A . Multiple products of the Drosophila Shaker gene may contribute to potassium channel diversity. Neuron 1, 421-430, 1988. Kantor, P.F., Coetzee, W.A., Carmeliet, E.E., Dennis, S.C., and Opie, L . H . Reduction of ischemic K + loss and arrhythmias in rat hearts, effect of glibenclamide, a sulfonylurea. Circulation Research 66, 478-485, 1990. Karlin, A . Structure of nicotinic acetylcholine receptors. Current Opinion in Neurobiology 3, 299-309, 1993. Kennedy, H.L. Beta blockade, ventricular arrhythmias, and sudden cardiac death. American Journal of Cardiology 80 , 29J-34J, 1997a. Kennedy, H.L. Beta-blocker prevention of proarrhythmia and proischemia: clues from CAST, C A M I A T , and EMIAT. Cardiac Arrhythmia Suppression Trial. Canadian Amiodarone Myocardial Infarction Arrhythmia Trial. European Myocardial Infarct Amiodarone Trial. American Journal of Cardiology 80 , 1208-11: 1997b. 203 Kenyon, J.L. and Gibbons, W.R. Influence of chloride, potassium, and tetraethylammonium on the early outward current of sheep cardiac Purkinje fibers. Journal of General Physiology. 73, 117-138, 1979. Kingsley, P.B., Sako, E.Y. , Yang, M.Q., Zimmer, S.D., Ugurbil, K. , Foker, J.E., and From, A . H . Ischemic contracture begins when anaerobic glycolysis stops: a 3 1 P - N M R study of isolated rat hearts. American Journal of Physiology 261, (Pt 2):H469-H478, 1991. Kleber, A . G . Resting membrane potential, extracellular potassium activity, and intracellular sodium activity during acute global ischemia in isolated perfused guinea pig hearts. Circulation Research 52, 442-450, 1983. Kleber, A . G . Extracellular potassium accumulation in acute myocardial ischemia. Journal of Molecular & Cellular Cardiology 16, 389-394, 1984. Kleber, A .G . , Janse, M.J. , Wilms-Schopmann, F.J., Wilde, A . A . , and Coronel, R. Changes in conduction velocity during acute ischemia in ventricular myocardium of the isolated porcine heart. Circulation 73, 189-198, 1986. Kleber, A . G . Conduction of the impulse in the ischemic myocardium—implications for malignant ventricular arrhythmias. Experientia 43, 1056-1061, 1987. Kleber, A .G . , Riegger, C.B., and Janse, M.J. Electrical uncoupling and increase of extracellular resistance after induction of ischemia in isolated, arterially perfused rabbit papillary muscle. Circulation Research 61, 271-279, 1987a. Kleber, A .G . , Riegger, C.B., and Janse, M.J. Extracellular K + and FT shifts in early ischemia: mechanisms and relation to changes in impulse propagation. Journal of Molecular & Cellular Cardiology 19, Suppl 5:35-44, 1987b. Kohlhardt, M . , Haap, K. , and Figulla, H.R. Influence of low extracellular ph upon the ca inward current and isometric contractile force in mammalian ventricular myocardium. Pflugers Archiv - European Journal of Physiology 366, 31-38, 1976. Koren, G., Liman, E.R., Logothetis, D.E., Nadal-Ginard, B. , and Hess, P. Gating mechanism of a cloned potassium channel expressed in frog oocytes and mammalian cells. Neuron 4, 39-51, 1990. Koretsune, Y . and Marban, E. Mechanism of ischemic contracture in ferret hearts: relative roles of [Ca2+]j elevation and atp depletion. American Journal of Physiology 258: Pt2):H9-16, 1990. Kubo, Y . , Baldwin, T.J., Jan, Y . N . , and Jan, L . Y . Primary structure and functional expression of a mouse inward rectifier potassium channel Nature 362, 127-133, 1993a. 204 Kubo, Y . , Reuveny, E., Slesinger, P.A., Jan, Y . N . , and Jan, L . Y . Primary structure and functional expression of a rat g-protein-coupled muscarinic potassium channel Nature 364, 802-806, 1993b. Kukushkin, N.L , Gainullin, R.Z., and Sosunov, E.A. Transient outward current and rate dependence of action potential duration in rabbit cardiac ventricular muscle. Pflugers Archiv - European Journal of Physiology 399, 87-92, 1983. Kumpf, R.A. and Dougherty, D.A. A mechanism for ion selectivity in potassium channels: computational studies of cation-7r interactions. Science 261, 1708-1710, 1993. Kupersmith, J., Antman, E .M. , and Hoffman, B.F. In vivo electrophysiological effects of lidocaine in canine acute myocardial infarction. Circ. Res. 36, 84-91, 1975. Kurachi, Y . The effects of intracellular protons on the electrical activity of single ventricular cells. Pflugers Archiv - European Journal of Physiology 394, 264-270, 1982. Langenfeld, H. , Weirich, J., Kohler, C , and Kochsiek, K . Comparative analysis of the action of class i antiarrhythmic drugs (lidocaine, quinidine, and prajmaline) in rabbit atrial and ventricular myocardium. Journal of Cardiovascular Pharmacology 15, 338-345, 1990. Lathrop, D.A., Nanasi, P.P., and Varro, A . In vitro cardiac models of dog purkinje fibre triggered and spontaneous electrical activity: effects of nicorandil. British Journal of Pharmacology 99, 119-123, 1990. Lee, J.H. and Rosen, M.R. Use-dependent actions and effects on transmembrane action potentials of flecainide, encainide, and ethmozine in canine Purkinje fibers. Journal of Cardiovascular Pharmacology 18, 285-292, 1991. L i , H.L., Galue, A. , Meadows, L. , and Ragsdale, D.S. A molecular basis for the different local anesthetic affinities of resting versus open and inactivated states of the sodium channel. Molecular Pharmacology 55, 134-141, 1999. Litterst, C. L. , Mimnaugh, E. G., Reagan, R. L. and Gram, T. E. Comparison of in vitro drug metabolism by lung, liver, and kidney of several common laboratory species. Drug Metabolism & Disposition 3: 259-265, 1975. Logothetis, D.E., Kammen, B.F., Lindpaintner, K. , Bisbas, D., and Nadal-Ginard, B. Gating charge differences between two voltage-gated K + channels are due to the specific charge content of their respective S4 regions. Neuron 10, 1121-1129, 1993. Lombet, A. , Norman, R.L, and Lazdunski, M . Affinity labelling of the tetrodotoxin-binding component of the Na + channel. Biochemical & Biophysical Research Communications 114, 126-130, 1983. 205 Luo, C H . and Rudy, Y . A dynamic model of the cardiac ventricular action potential. II. afterdepolarizations, triggered activity, and potentiation. Circulation Research 74, 1097-1113,1994. MacKinnon, R., Latorre, R., and Miller, C. Role of surface electrostatics in the operation of a high-conductance Ca2+-activated K + channel. Biochemistry 28, 8092-8099, 1989. MacKinnon, R. and Miller, C. Mutant potassium channels with altered binding of charybdotoxin, a pore-blocking peptide inhibitor. Science 245, 1382-1385, 1989. MacKinnon, R. and Yellen, G. Mutations affecting tea blockade and ion permeation in voltage- activated K + channels. Science 250, 276-279, 1990. MacKinnon, R. Determination of the subunit stoichiometry of a voltage-activated potassium channel. Nature 350, 232-235, 1991. Marrer, S. Molecular modeling of anti-arrhythmia agents in class la and lb - model conceptions of different binding receptors. Pharmaceutica Acta Helvetiae 64, 338-344, 1989. Mason, J.W. Amiodarone. New England Journal of Medicine 316, 455-66: 1987. Mautner, M . and Deakyne, C A . Unconventional ionic hydrogen bonds. I. C H + " X . Complexes of quaternary ions with n- and rc-donors. Journal of American Chemical Society. 107, 469-474, 1985. Maxwell, M.P., Hearse, D.J., and Yellon, D . M . Species variation in the coronary collateral circulation during regional myocardial ischaemia: a critical determinant of the rate of evolution and extent of myocardial infarction. Cardiovascular Research 21, 737-746, 1987. McAllister, R.E. and Noble, D. The time and voltage dependence of the slow outward current in cardiac Purkinje fibres. Journal of Physiology 186, 632-662, 1966. McCall, J .M. Liquid-liquid partition coefficients by high-pressure liquid chromatography. Journal of Medicinal Chemistry 18, 549-552, 1975. McLarnon, J.G. and Xu, R. Actions of the benzopyran compound terikalant on macroscopic currents in rat ventricular myocytes. Journal of Pharmacology & Experimental Therapeutics 215, 389-396, 1995. McPhee, J .C , Ragsdale, D.S., Scheuer, T., and Catterall, W.A. A mutation in segment ivs6 disrupts fast inactivation of sodium channels. Proceedings of the National Academy of Sciences of the United States of America 91, 12346-12350, 1994. 206 Meesman, W. (1982). Early arrhythmias and primary ventricular fibrillation after acute myocardial ischaemia in relation to pre-existing coronary collaterals. In Early arrhythmias resulting from myocardial ischaemia. J.R. Parratt, ed. (London: MacMillan Press Ltd.), pp. 93-112. Messner, D.J. and Catterall, W.A. The sodium channel from rat brain. Separation and characterization of subunits. Journal of Biological Chemistry 260, 10597-10604, 1985. Miller, J.A., Agnew, W.S., and Levinson, S.R. Principal glycopeptide of the tetrodotoxin/saxitoxin binding protein from electrophorus electricus: isolation and partial chemical and physical characterization. Biochemistry 22, 462-470, 1983. Mirrlees, M.S., Moulton, S.J., Murphy, C.T., and Taylor, P.J. Direct measurement of octanol-water partition coefficients by high-pressure liquid chromatography. Journal of Medicinal Chemistry 19, 615-619, 1976. Mitchell, J.B., Nandi, C.L., McDonald, I.K., Thornton, J .M., and Price, S.L. Amino/aromatic interactions in proteins: Is the evidence stacked against hydrogen bonding? Journal of Molecular Biology 239, 315-331, 1994. Moe, G., Rheinboldt, W.C., and Abildskov, J.A. A computer model of atrial fibrillation. American Heart Journal. 67,200-220, 1964. Moorman, J.R., Yee, R., Bjornsson, T., Starmer, C.F., Grant, A.O., and Strauss, H.C. Pka does not predict pH potentiation of sodium channel blockade by lidocaine and w6211 in guinea pig ventricular myocardium. Journal of Pharmacology & Experimental Therapeutics 238, 159-166, 1986. Morena, FL, Janse, M.J. , Fiolet, J.W., Krieger, W.J., Crijns, FL, and Durrer, D. Comparison of the effects of regional ischemia, hypoxia, hyperkalemia, and acidosis on intracellular and extracellular potentials and metabolism in the isolated porcine heart. Circulation Research 46, 634-46: 1980. Morganroth, J. Comparative efficacy and safety of oral mexiletine and quinidine in benign or potentially lethal ventricular arrhythmias. American Journal of Cardiology 60, 1276-1281, 1987. Murphy, E., Perlman, M . , London, R. E. and Steenbergen, C. Amiloride delays the ischemia-induced rise in cytosolic free calcium. Circulation Research 68: 1250-1258, 1991. Nademanee, K. , Singh, B.N. , Cannom, D.S., Weiss, J., Feld, G., and Stevenson, W.G. Control of sudden recurrent arrhythmic deaths: role of amiodarone. American Heart Journal 106, (Pt 2):895-901, 1983. 207 Nademanee, K. , Piwonka, R.W., Singh, B .N. , and Hershman, J .M. Amiodarone and thyroid function. Progress in Cardiovascular Diseases 31,427-437, 1989. Nakamura, T.Y., Coetzee, W.A., Vega-Saenz De Miera, E., Airman, M . , and Rudy, B. Modulation of Kv4 channels, key components of rat ventricular transient outward K + current, by P K C . American Journal of Physiology 273, H1775-H1786, 1997. Narahashi, T. Chemicals as tools in the study of excitable membranes. Physiological Reviews 54: 813-89: 1974. Narahashi, T. and Frazier, D.T. Site of action and active form of local anesthetics. Neurosciences Research 4, 65-99: 1971. Nattel, S., Elharrar, V . , Zipes, D.P., and Bailey, J.C. pH-dependent electrophysiological effects of quinidine and lidocaine on canine cardiac Purkinje fibers. Circulation. Research. 48, 55-61, 1981. Nattel, S., Talajic, M . , Quantz, M . , and DeRoode, M . Frequency-dependent effects of amiodarone on atrioventricular nodal function and slow-channel action potentials: evidence for calcium channel-blocking activity. Circulation 76, 442-449, 1987. Nattel, S., Talajic, M . , Beaudoin, D., Matthews, C , and Roy, D. Absence of pharmacokinetic interaction between amiodarone and lidocaine. American Journal of Cardiology 73, 92-94: 1994. New, W. and Trautwein, W. Inward membrane currents in mammalian myocardium. Pflugers Archiv - European Journal of Physiology 334, 1-23: 1972. Noble, D. Electrical properties of cardiac muscle attributable to inward-going (anomalous) rectification. Journal Cellular Comparative Physiology. Suppl 2, 127-136, 1965. Noble, D. The initiation of the heartbeat. In: Noble, D., ed., Oxford: Clarendon Press, 1975. Noble, D. The surprising heart: a review of recent progress in cardiac electrophysiology. Journal of Physiology. (Lond.) 353, 1-50, 1984. Noda, M . , Shimizu, S., Tanabe, T., Takai, T., Kayano, T., Ikeda, T., Takahashi, H. , Nakayama, FL, Kanaoka, Y . , Minamino, N . , and et al Primary structure of electrophorus electricus sodium channel deduced from cDNA sequence. Nature 312, 121-127, 1984. Noda, M . , Suzuki, FL, Numa, S., and Stuhmer, W. A single point mutation confers tetrodotoxin and saxitoxin insensitivity on the sodium channel i i . FEBS Letters 259, 213-216, 1989. 208 Nokin, P., Clinet, M . , and Schoenfeld, P. Cardiac beta-adrenoceptor modulation by amiodarone. Biochemical Pharmacology 32, 2473-2477, 1983. Nordrehaug, J.E. and von der Lippe, G. Hypokalemia and ventricular fibrillation in acute myocardial infarction. British Heart Journal 50, 525-529, 1983. Nordrehaug, J.E. and von der Lippe, G. Serum potassium concentrations are inversely related to ventricular, but not to atrial, arrhythmias in acute myocardial infarction. European Heart Journal 7, 204-209, 1986. Ohyanagi, M . , Matsumori, Y . , and Iwasaki, T. Beta-adrenergic receptors in ischemic and nonischemic canine myocardium: relation to ventricular fibrillation and effects of pretreatment with propranolol and hexamefhonium. Journal of Cardiovascular Pharmacology 11, 107-114, 1988. Opie, L . H . Effects of regional ischemia on metabolism of glucose and fatty acids, relative rates of aerobic and anaerobic energy production during myocardial infarction and comparison with effects of anoxia. Circulation Research 38, (Suppl l):I52-74, 1976. Opitz, C.F., Mitchell, G.F., Pfeffer, M.A . , and Pfeffer, J .M. Arrhythmias and death after coronary artery occlusion in the rat. continuous telemetric E C G monitoring in conscious, untethered rats. Circulation 92, 253-261, 1995. Oshita, S., Sada, H. , Kojima, M . , and Ban, T. Effects of tocainide and lidocaine on the transmembrane action potentials as related to external potassium and calcium concentrations in guinea-pig papillary muscles. Naunyn-Schmiedebergs Archives of Pharmacology 314, 67-82, 1980. Paletta, M.J. , Abraham, S., Beatch, G.N., and Walker, M.J. Mechanisms underlying the antiarrhythmic properties of beta-adrenoceptor blockade against ischaemia-induced arrhythmias in acutely prepared rats. British Journal of Pharmacology 98, 87-94, 1989. Papazian, D .M. , Schwarz, T.L., Tempel, B.L. , Jan, Y . N . , and Jan, L . Y . Cloning of genomic and complementary D N A from Shaker, a putative potassium channel gene from Drosophila. Science 237, 749-753, 1987. Papazian, D .M. , Timpe, L.C. , Jan, Y . N . , and Jan, L . Y . Alteration of voltage-dependence of Shaker potassium channel by mutations in the S4 sequence. Nature 349, 305-310, 1991. Pardo, L .A . and Stuhmer, W. Extracellular potassium modulates a transient potassium current in rat atrial cells. In Ion channels in the cardiovascular system: Function and dysfunction. P .M. Spooner, A . M . Brown, W.A. Catterall, G.J. Kaczorowski, and H.C. Strauss, eds. Futura Publising Company, Armonk, New York, pp. 341-355, 1994. 209 Patterson, E., Kalcich, M . , and Scherlag, B.J. Phase Ib ventricular arrhythmia in the dog: localized reentry within the mid-myocardium. Journal of Interventional Cardiac Electrophysiology!, 145-161, 1998. Pentecost, B.L. , De Giovanni, J.V., Lamb, P., Cadigan, P.J., Evemy, K . L . , and Flint, E.J. Reappraisal of lignocaine therapy in management of myocardial infarction. British Heart Journal 45, 42-47, 1981. Perier, F., Radeke, C M . , and Vandenberg, C A . Primary structure and characterization of a small-conductance inwardly rectifying potassium channel from human hippocampus. Proceedings of the National Academy of Sciences of the United States of America 91, 6240-6244, 1994. Perutz, M.F. , Fermi, G., Abraham, D.J., Poyart, C , and Bursaux, E. Hemoglobin as a receptor of drugs and peptides: X-ray studies of the stereochemistry of binding. Journal of American Chemical Society. 108, 1064-1078, 1986. Po, S., Snyders, D.J., Baker, R., Tamkun, M . M . , and Bennett, P.B. Functional expression of an inactivating potassium channel cloned from human heart. Circulation Research 71, 732-736, 1992. Podrid, P.J. Potassium and ventricular arrhythmias. American Journal of Cardiology 65, 33E-44E, 1990. Polster, P. and Broekhuysen, J. The adrenergic antagonism of amiodarone. Biochemical Pharmacology IS, 131-134: 1976. Pongs, O., Kecskemethy, N . , Muller, R., Krah-Jentgens, I., Baumann, A. , Kiltz, H.H., Canal, I., Llamazares, S., and Ferrus, A . Shaker encodes a family of putative potassium channel proteins in the nervous system of Drosophila. EMBO Journal 7, 1087-1096, 1988. Pongs, O. Molecular biology of voltage-dependent potassium channels. Physiological Reviews 72, (Suppl):S69-88, 1992a. Pongs, O. Structural basis of voltage-gated K + channel pharmacology. Trends in Pharmacological Sciences 13, 359-365, 1992b. Pugsley, M.K. , Ries, C.R., Guppy, L.J. , Harvie, C.J., and Walker, M.J . Effects of anipamil, a long acting analog of verapamil, in pigs subjected to myocardial ischemia. Life Sciences 57, 1219-1231, 1995. Pugsley, M . K . and Goldin, A . L . Effects of bisaramil, a novel class I antiarrhythmic agent, on heart, skeletal muscle and brain Na + channels. European Journal of Pharmacology 342, 93-104, 1998. 2 1 0 Pugsley, M . K . and Goldin, A . L . Molecular analysis of the Na + channel blocking actions of the novel class I anti-arrhythmic agent RSD 921. British Journal of Pharmacology 127, 9-18, 1999. Rademaker, A.W., Kellen, J., Tarn, Y . K . , and Wyse, D.G. Character of adverse effects of prophylactic lidocaine in the coronary care unit. Clinical Pharmacology & Therapeutics 40, 71-80, 1986. Ragsdale, D.S., McPhee, J .C , Scheuer, T., and Catterall, W.A. Molecular determinants of state-dependent block of Na + channels by local anesthetics. Science 265, 1724-1728, 1994. Rehm, H. and Lazdunski, M . Purification and subunit structure of a putative K+-channel protein identified by its binding properties for dendrotoxin I. Proceedings of the National Academy of Sciences of the United States of America 85, 4919-4923, 1988. Remko, M . , Scheiner, S., and Rode, B . M . Molecular modelling of the antiarrhythmic-receptor interaction. Journal of Molecular Structure. (Theochem) 307, 35-46, 1994. Roberds, S.L. and Tamkun, M . M . Cloning and tissue-specific expression of five voltage-gated potassium channel cDNAs expressed in rat heart. Proceedings of the National Academy of Sciences of the United States of America 88, 1798-1802, 1991. Roden, D .M. , Woosley, R.L., and Primm, R.K. Incidence and clinical features of the quinidine-associated long qt syndrome: implications for patient care. American Heart Journal 111, 1088-1093, 1986. Roden, D .M. , Lazzara, R., Rosen, M . , Schwartz, P.J., Towbin, J., and Vincent, G .M. Multiple mechanisms in the long-QT syndrome, current knowledge, gaps, and future directions, the SADS foundation task force on LQTS. Circulation 94, 1996-2012, 1996. Roden, D. M . and George, A . L.,Jr. Structure and function of cardiac sodium and potassium channels. American Journal of Physiology 273: Pt 2):H511-H525, 1997. Rozanski, G.J. and Witt, R . C Early afterdepolarizations and triggered activity in rabbit cardiac Purkinje fibers recovering from ischemic-like conditions. Role of acidosis. Circulation 83, 1352-1360, 1991. Ruben, P. C , Starkus, J. G. and Rayner, M . D. Steady-state availability of sodium channels: interactions between activation and slow inactivation. Biophysical Journal 61: 941-955, 1992. Ruff, R. L. , Simoncini, L . and Stuhmer, W. Slow sodium channel inactivation in mammalian muscle: a possible role in regulating excitability. Muscle & Nerve 11: 502-510, 1988. 211 Saint, K . M . , Abraham, S., MacLeod, B.A. , McGough, J., Yoshida, N . , and Walker, M J . Ischemic but not reperfusion arrhythmias depend upon serum potassium concentration. Journal of Molecular & Cellular Cardiology 24, 701-709, 1992. Salkoff, L . , Baker, K. , Butler, A. , Covarrubias, M . , Pak, M.D. , and Wei, A . A n essential 'set' of K + channels conserved in flies, mice and humans. Trends in Neurosciences 15, 161-166, 1992. Sanchez-Chapula, J.A. Mechanism of transient outward K + channel block by disopyramide. Journal of Pharmacology & Experimental Therapeutics 290, 515-523, 1999. Satin, J., Kyle, J. W., Chen, M . , Rogart, R. B . and Fozzard, H . A . The cloned cardiac Na + channel alpha-subunit expressed in Xenopus oocytes show gating and blocking properties of native channels. Journal of Membrane Biology 130: 11-22, 1992. Schmidt, R.R. and Betz, H. The beta-bungarotoxin-binding protein from chick brain: binding sites for different neuronal K + channel ligands co-fractionate upon partial purification. FEBS Letters 240, 65-70, 1988. Schmitt, F.O. and Erlanger, J. Directional differences in the conduction of the impulse through heart muscle and their possible relation to extrasystolic and fibrillary contractions. American Journal of Physiology. 87, 326-347, 1929. Schomig, A. , Richardt, G., and Kurz, T. Sympatho-adrenergic activation of the ischemic myocardium and its arrhythmogenie impact. FLerz 20, 169-186, 1995. Schwarz, W., Palade, P.T., and Hille, B. Local anesthetics. Effect of pH on use-dependent block of sodium channels in frog muscle. Biophysical Journal 20, 343-368, 1977. Shaikh, N .A. and Downar, E. Effects of chronic amiodarone treatment on cat myocardial phospholipid content and on in vitro phospholipid catabolism. Molecular & Cellular Biochemistry IS, 17-25, 1987. Sharkey, R.G., Beneski, D.A., and Catterall, W.A. Differential labeling of the alpha and beta 1 subunits of the sodium channel by photoreactive derivatives of scorpion toxin. Biochemistry 23, 6078-6086, 1984. Sheldon, R.S., Cannon, N.J., and Duff, H.J. A receptor for type I antiarrhythmic drugs associated with rat cardiac sodium channels. Circulation Research 61, 492-497, 1987. Sheldon, R.S., Hi l l , R.J., and Duff, H J . Antiarrhythmic drugs and the cardiac sodium channel: current models. Clinical Chemistry 35, 748-754, 1989. 212 Sheldon, R.S., Duff, H.J., and Hi l l , R.J. Class I anti-arrhythmic drugs: structure and function at the cardiac sodium channel. Clinical & Investigative Medicine - Medecine Clinique et Experimentale 14, 458-465, 1991a. Sheldon, R.S., Hi l l , R.J., Taouis, M . , and Wilson, L . M . Aminoalkyl structural requirements for interaction of lidocaine with the class I antiarrhythmic drug receptor on rat cardiac myocytes. Molecular Pharmacology 39, 609-614, 1991b. Sheldon, R.S., Duff, H.J., Thakore, E., and Hi l l , R.J. Class I antiarrhythmic drugs: allosteric inhibitors of [3H] batrachotoxinin binding to rat cardiac sodium channels. Journal of Pharmacology & Experimental Therapeutics 268, 187-194, 1994. Shibata, E.F., Drury, T., Refsum, H. , Aldrete, V. , and Giles, W. Contributions of a transient outward current to repolarization in human atrium. American Journal of Physiology 257, (Pt 2):H1773-1781, 1989. Simoncini, L. and Stuhmer, W. Slow sodium channel inactivation in rat fast-twitch muscle. Journal of Physiology 383: 327-337, 1987. Singh, B.N. , Nademanee, K. , Josephson, M.A . , Ikeda, N . , Venkatesh, N . , and Kannan, R. The electrophysiology and pharmacology of verapamil, flecainide, and amiodarone: correlations with clinical effects and antiarrhythmic actions. Annals of the New York Academy of Sciences 432, 210-235, 1984. Singh, B.N. , Venkatesh, N . , Nademanee, K. , Josephson, M.A. , and Kannan, R. The historical development, cellular electrophysiology and pharmacology of amiodarone. Progress in Cardiovascular Diseases 31, 249-280: 1989. Singh, B .N . Is class III antiarrhythmic activity important?. Cardiovascular Drugs & Therapy 4, Suppl 3:597-602, 1990. Singh, B .N. Antiarrhythmic drugs: A reorientation in light of recent developments in the control of disorders of rhythm. American Journal of Cardiology 81, 3D-13D, 1998. Slesinger, P.A., Jan, Y . N . , and Jan, L . Y . The S4-S5 loop contributes to the ion-selective pore of potassium channels. Neuron 11, 739-749, 1993. Snyders, D.J. and Yeola, S.W. Determinants of antiarrhythmic drug action, electrostatic and hydrophobic components of block of the human cardiac hKvl.5 channel. Circulation Research 11, 575-583, 1995. Sobel, B.E., Corr, P.B., Robison, A . K . , Goldstein, R.A., Witkowski, F .X. , and Klein, M.S. Accumulation of lysophosphoglycerides with arrhythmogenic properties in ischemic myocardium. Journal of Clinical Investigation 62, 546-553, 1978. 213 Steenbergen, C , Hi l l , M.L . , and Jennings, R.B. Volume regulation and plasma membrane injury in aerobic, anaerobic, and ischemic myocardium in vitro. Effects of osmotic cell swelling on plasma membrane integrity. Circulation Research 57, 864-875, 1985. Steinberg, M . L , Sullivan, M.E. , Wiest, S.A., Rockhold, F.W., and Molloy, B .B. Cellular electrophysiology of clofilium, a new antifibrillatory agent, in normal and ischemic canine Purkinje fibers. Journal of Cardiovascular Pharmacology 3, 881-895, 1981. Steinberg, M . L , Ertel, P., Smallwood, J.K., Wyss, V. , and Zimmerman, K . The relation between vascular relaxant and cardiac electrophysiological effects of pinacidil. Journal of Cardiovascular Pharmacology 12, Suppl 2:S30-40, 1988. Stengl, M . , Carmeliet, E., Mubagwa, K. , and Flameng, W. Modulation of transient outward current by extracellular protons and C d 2 + in rat and human ventricular myocytes. Journal of Physiology 511, 827-836, 1998. Stephenson, S.E., Cole, R.K., Parrish, T.F., Bauer, F .M. , Johnson, I.T., Kochtitzky, M . , Anderson, J.S., Hibbitt, L .L . , McCarty, J.E., Young, E.R., Wilson, J.R., Meiers, H.N., Meador, C.K., Ball, C O . , and Meneely, G.R. Ventricular fibrillation during and after coronary artery occlusion. Incidence and protection afforded by various drugs. American Journal of Cardiology. 5, 77-87, 1960. Stocker, M . , Pongs, O., Hoth, M . , Heinemann, S.H., Stuhmer, W., Schroter, K . H . , and Ruppersberg, J.P. Swapping of functional domains in voltage-gated K + channels. Proceedings of the Royal Society of London - Series B: Biological Sciences 245, 101-107, 1991. Strichartz, G.R. The inhibition of sodium currents in myelinated nerve by quaternary derivatives of lidocaine. Journal of General Physiology. 62, 37-57, 1973. Stuhmer, W., Conti, F., Suzuki, H . , Wang, X.D. , Noda, M . , Yahagi, N . , Kubo, H. , and Numa, S. Structural parts involved in activation and inactivation of the sodium channel. Nature 339, 597-603, 1989. Sunner, J., Nishizawa, K. , and Kebarle, P. Ion-solvent molecule interaction in the gas phase. The potassium ion and benzene. Journal of Physical Chemistry. 85, 1814-1820, 1981. Surawicz, B. Electrophysiologic substrate of torsade de pointes: Dispersion of repolarization or early afterdepolarizations?. Journal of the American College of Cardiology 14, 172-184, 1989. Surawicz, B . Role of potassium channels in cycle length dependent regulation of action potential duration in mammalian cardiac Purkinje and ventricular muscle fibres. Cardiovascular Research 26, 1021-1029, 1992. 214 Suttorp, M.J. , Kingma, J.H., Jessurun, E.R., Lie-A-Huen, L. , van Hemel, N . M . , and Lie, K.L The value of class Ic antiarrhythmic drugs for acute conversion of paroxysmal atrial fibrillation or flutter to sinus rhythm. Journal of the American College of Cardiology 16, 1722-1727, 1990. Taft, R.W., Anvia, F., Gal, T.-F., Walsh, S., Capon, M . , Holmes, M . C , Hosn, K. , Oloumi, G., Vasanwala, R., and Yazdani, S. Free energies of cation-molecule complex formation and of cation-solvent transfers. Pure & Applied Chemistry. 62, 17-23, 1990. Takumi, T., Ohkubo, H. , and Nakanishi, S. Cloning of a membrane protein that induces a slow voltage-gated potassium current. Science 242, 1042-1045, 1988. Tamkun, M . M . , Knoth, K . M . , Walbridge, J.A., Kroemer, H. , Roden, D .M. , and Glover, D . M . Molecular cloning and characterization of two voltage-gated K + channel cDNAs from human ventricle. FASEB Journal 5, 331-337, 1991. Tanabe, T., Takeshima, H. , Mikami, A. , Flockerzi, V . , Takahashi, H. , Kangawa, K. , Kojima, M . , Matsuo, H. , Hirose, T., and Numa, S. Primary structure of the receptor for calcium channel blockers from skeletal muscle. Nature 328, 313-318, 1987. Tani, M . and Neely, J.R. Role of intracellular Na + in Ca 2 + overload and depressed recovery of ventricular function of reperfused ischemic rat hearts. Possible involvement of H + -Na + and Na + -Ca 2 + exchange. Circulation Research 65, 1045-1056, 1989. Task Force of the Working Group on Arrhythmias of the European Society of Cardiology. The Sicilian Gambit: a new approach to the classification of antiarrhythmic drugs based on their actions on arrhythmogenic mechanisms. Circulation 84, 1831-1851, 1991. Taylor, P., Abramson, S.N., Johnson, D.A., Valenzuela, C.F., and Herz, J. Distinctions in ligand binding sites on the nicotinic acetylcholine receptor. Annals of the New York Academy of Sciences 625, 568-587, 1991. Terlau, H. , Heinemann, S.H., Stuhmer, W., Pusch, M . , Conti, F., Imoto, K. , and Numa, S. Mapping the site of block by tetrodotoxin and saxitoxin of sodium channel II. FEBS Letters 293, 93-96, 1991. Terlau, H . and Stuhmer, W. Structure and function of voltage-gated ion channels. Naturwissenschaften 85, 437-444, 1998. Timpe, L . C , Jan, Y . N . , and Jan, L . Y . Four cDNA clones from the Shaker locus of Drosophila induce kinetically distinct a-type potassium currents in Xenopus oocytes. Neuron 1,659-667, 1988. Trautwein, W. and Kassebaum, D.G. On the mechanism of spontaneous generation in the pacemaker of the heart. Journal of General Physiology 45, 317-330, 1961. 215 Tseng-Crank, J.C., Tseng, G.N., Schwartz, A., and Tanouye, M.A. Molecular cloning and functional expression of a potassium channel cDNA isolated from a rat cardiac library. FEBSLetters 268, 63-68, 1990. Tsushima, R.G., Li , R.A., and Backx, P.H. Altered ionic selectivity of the sodium channel revealed by cysteine mutations within the pore. Journal of General Physiology. 109,463-475,1997. Valenzuela, C, Snyders, D.J., Bennett, P.B., Tamargo, J., and Hondeghem, L.M. Stereoselective block of cardiac sodium channels by bupivacaine in guinea pig ventricular myocytes. Circulation 92, 3014-3024, 1995. Valenzuela, C. and Bennett, P. B.,Jr. Gating of cardiac Na + channels in excised membrane patches after modification by alpha-chymotrypsin. Biophysical Journal 67: 161-171, 1994. Vanheel, B., Leybaert, L., de Hemptinne, A. and Leusen, I. Simulated ischemia and intracellular pH in isolated ventricular muscle. American Journal of Physiology 257: Pt 1):C365-C376, 1989. Vanoli, E., Priori, S.G., Nakagawa, H., Hirao, K., Napolitano, C, Diehl, L., Lazzara, R., and Schwartz, PJ. Sympathetic activation, ventricular repolarization and i K r blockade: implications for the antifibrillatory efficacy of potassium channel blocking agents. Journal of the American College of Cardiology 25, 1609-1614, 1995. Varro, A., Elharrar, V., and Surawicz, B. Frequency-dependent effects of several class i antiarrhythmic drugs on vmax of action potential upstroke in canine cardiac Purkinje fibers. Journal of Cardiovascular Pharmacology 7, 482-492, 1985. Varro, A., Bodi, I., Jednakovits, A., and Rabloczky, G. Frequency-dependent effects of class I, III and IV antiarrhythmic drugs on the conduction and excitability in rabbit ventricular muscle. Archives Internationales de Pharmacodynamic et de Therapie 292, 157-165, 1988. Varro, A., Knilans, T.K., Nanasi, P.P., Rabloczky, G., and Lathrop, D.A. Concentration-and rate-dependent electrophysiological effects of restacorin on isolated canine Purkinje fibres. Naunyn-Schmiedebergs Archives of Pharmacology 342, 691-697, 1990. Vaughan Williams, E.M. Classification of antidysrhythmic drugs. Pharmacol. Ther. 12, 115-138, 1975. Velebit, V., Podrid, P., Lown, B., Cohen, B.H., and Graboys, T.B. Aggravation and provocation of ventricular arrhythmias by antiarrhythmic drugs. Circulation 65, 886-894, 1982. 216 Vogel, S. and Sperelakis, N . Blockade of myocardial slow inward current at low pH. American Journal of Physiology 233, C99-103, 1977. Waldo, A . L . , Camm, A.J . , deRuyter, H. , Friedman, P.L., MacNeil, D.J., Pitt, B. , Pratt, C M . , Rodda, B.E. , and Schwartz, P.J. for the SWORD Investigators. Preliminary mortality results from the Survival With Oral d-Soltalol (SWORD) trial. Journal of the American College of Cardiology. 25, 15 A , 1995. Waldo, A . L . , Camm, A.J . , deRuyter, H. , Friedman, P.L., MacNeil, D.J., Pauls, J.F., Pitt, B. , Pratt, C M . , Schwartz, P.J., and Veltri, E.P. Effect of d-sotalol on mortality in patients with left ventricular dysfunction after recent and remote myocardial infarction. The SWORD Investigators. Survival With Oral d-Sotalol. Lancet 348, 7-12, 1996. Walker, M.J. , Curtis, M.J. , Hearse, D.J., Campbell, R.W., Janse, M.J. , Yellon, D .M. , Cobbe, S.M., Coker, S.J., Harness, J.B., Harron, D.W., Higgins, A.J . , Julian, D.G., Lab, M.J., Manning, A.S., Northover, B.J., Parratt, J.R., Riemersma, R.A., Riva, E., Russel, D . C , Sheridan, D.J., Winslow, E., and Woodward, B. The Lambeth Conventions: guidelines for the study of arrhythmias in ischaemia infarction, and reperfusion. Cardiovasc. Res. 22, 447-455, 1988. Walker, M.J . and Beatch, G.N. Electrically induced arrhythmias in the rat. Proceedings of the Western Pharmacological Society. 31, 167-170, 1988. Wallert, M . A . and Frohlich, O. Na + -H + exchange in isolated myocytes from adult rat heart. American Journal of Physiology 257, (Pt 1):C207-213, 1989. Wang, D. W., Yazawa, K. , George, A . L.,Jr. and Bennett, P. B. Characterization of human cardiac Na + channel mutations in the congenital long qt syndrome. Proceedings of the National Academy of Sciences of the United States of America 93: 13200-13205, 1996. Wang, G. K . and Wang, S. Y . Altered stereoselectivity of cocaine and bupivacaine isomers in normal and batrachotoxin-modified Na + channels. Journal of General Physiology 100: 1003-1020, 1992. Wang, Q., Shen, J., L i , Z., Timothy, K. , Vincent, G. M . , Priori, S. G., Schwartz, P. J. and Keating, M . T. Cardiac sodium channel mutations in patients with long qt syndrome, an inherited cardiac arrhythmia. Human Molecular Genetics 4: 1603-1607, 1995a. Wang, Q., Shen, J., Splawski, I., Atkinson, D., L i , Z., Robinson, J. L . , Moss, A . J., Towbin, J. A . and Keating, M . T. SCN5A mutations associated with an inherited cardiac arrhythmia, long qt syndrome. Cell 80: 805-811, 1995b. Warmke, J.W. and Ganetzky, B. A family of potassium channel genes related to eag in Drosophila and mammals. Proceedings of the National Academy of Sciences of the United States of America 91, 3438-3442, 1994. 217 Watson, C L . and Gold, M.R. Modulation of cardiac IN a by acidosis. Biophysical Journal. 66, A204, 1994. Watson, C L . and Gold, M.R. Effect of intracellular and extracellular acidosis on sodium current in ventricular myocytes. American Journal of Physiology 268, (Pt 2):H1749-1756, 1995. Wei, A., Covarrubias, M . , Butler, A., Baker, K., Pak, M . , and Salkoff, L. K + current diversity is produced by an extended gene family conserved in Drosophila and mouse. Science 248, 599-603, 1990. Wei, A., Rubin, H. , Cooperman, B.S., and Christianson, D.W. Crystal structure of an uncleaved serpin reveals the conformation of an inhibitory reactive loop. Nature Structural Biology 1, 251-258, 1994. Weidmann, S. Effects of calcium ions and local anaesthetics on electrical properties of Purkinje fibres. Journal of Physiology. 129, 568-582, 1955. Wendt, D.J., Starmer, C.F., and Grant, A .O. pH dependence of kinetics and steady-state block of cardiac sodium channels by lidocaine. American Journal of Physiology. 264, (Pt 2):H1588-1598, 1993. West, J.W., Patton, D.E., Scheuer, T., Wang, Y. , Goldin, A.L. , and Catterall, W.A. A cluster of hydrophobic amino acid residues required for fast Na+-channel inactivation. Proceedings of the National Academy of Sciences of the United States of America 89, 10910-10914, 1992. White, M . M . , Chen, L. Q., Kleinfield, R., Kallen, R. G. and Barchi, R. L. Skm2, a Na + channel cDNA clone from denervated skeletal muscle, encodes a tetrodotoxin-insensitive Na + channel. Molecular Pharmacology 39: 604-608, 1991. Winslow, E. Methods for the detection and assessment of antiarrhythmic activity. Pharmacol. Ther. 24,401-433, 1984. Winslow, E. , Martorana, M . , and Bell, P. Electrophysiologic effects of org 7797, a new steroidal antiarrhythmic agent: Comparison with class la, Ib, and Ic drugs. Journal of Cardiovascular Pharmacology 14, 205-212, 1989. Wit, A .L . , Cranefield, P.F., and Hoffman, B.F. Slow conduction and reentry in the ventricular conducting system. II. single and sustained circus movement in networks of canine and bovine Purkinje fibers. Circulation Research 30, 11-22, 1972a. Wit, A .L . , Hoffman, B.F., and Cranefield, P.F. Slow conduction and reentry in the ventricular conducting system. I. return extrasystole in canine Purkinje fibers. Circulation Research 30, 1-10: 1972b. 218 Wit, A . L . and Janse, M . J . Ventricular arrhythmias in the acute phase of myocardial ischemia and infarction. In: The ventricular arrhythmias of ischemia and infarction: Electrophysiological mechanisms. Wit , A . L . Janse, M . J . , eds., Mount Kisko , New York: Futura Publishing Company, pp. 174-260, 1993. Woodhull, A . M . Ionic blockage of sodium channels in nerve. Journal of General Physiology. 61, 687-708: 1973. X u , Z . and Rozanski, G.J . Proton inhibition of transient outward potassium current in rat ventricular myocytes. Journal of Molecular & Cellular Cardiology 29, 481-490, 1997. Yan , G . X . and Kleber, A . G . Changes in extracellular and intracellular p H in ischemic rabbit papillary muscle. Circulation Research 71: 460-470, 1992. Yatani, A . , Brown, A . M . , and Akaike, N . Effect of extracellular p H on sodium current in isolated, single rat ventricular cells. Journal of Membrane Biology 78, 163-168, 1984. Yeola, S.W. and Snyders, D.J . Electrophysiological and pharmacological correspondence between Kv4.2 current and rat cardiac transient outward current. Cardiovascular Research 33, 540-547, 1997. Yeh , J. Z . Blockade of sodium channels by stereoisomers of local anesthetics. In Molecular Mechanisms of Anesthesia, B . R. Fink ,ed., pp. 35-44, Raven Press, New York, 1980. Yeh, J.Z. and Tanguy, J. Internal and external protons modulate voltage dependence of receovery from use-dependent block of lidocaine. Biophysical Journal. 59, 26, 1991. Yellen, G . , Jurman, M . E . , Abramson, T., and MacKinnon, R. Mutations affecting internal tea blockade identify the probable pore-forming region of a K + channel. Science 251, 939-942, 1991. Yong, S.L., X u , R., McLarnon, J .G., Zolotoy, A . B . , Beatch, G . N . , and Walker, M . J . R S D 1000: a novel antiarrhythmic agent with increased potency under acidic and high-potassium conditions. Journal of Pharmacology & Experimental Therapeutics 289, 236-244, 1999. Y o o l , A . J . and Schwarz, T . L . Alteration of ionic selectivity of a K + channel by mutation of the h5 region Nature 349, 700-704, 1991. Zagotta, W . N . and Aldrich, R . W . Voltage-dependent gating of Shaker A-type potassium channels in Drosophila muscle. Journal of General Physiology 95, 29-60, 1990. Zhang, J.F. and Siegelbaum, S.A. Effects of external protons on single cardiac sodium channels from guinea pig ventricular myocytes. Journal of General Physiology. 98, 1065-1083, 1991. 219 Zygmunt, A.C. Intracellular calcium activates a chloride current in canine ventricular myocytes. American Journal of Physiology 267, (Pt 2):H1984-H1995, 1994.