Anions and Arrhythmias in Experimental Heart Disease

ANIONS AND ARRHYTHMIAS IN EXPERIMENTAL HEART DISEASE

PAUL DAMIEN RIDLEY

M.D. ProQuest Number: U542518

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a com plete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest

ProQuest U542518

Published by ProQuest LLC(2017). Copyright of the Dissertation is held by the Author.

ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 2

ABSIBAGI

Ischaemic heart disease can cause sudden cardiac death from ventricular arrhythmias occuring as a consequence of ischaemia itself, or as a consequence of subsequent reperfusion of ischaemic tissue. Current agents designed to prevent these arrhythmias do not provide adequate protection. Conventionally, antiarrhythmic agents are subdivided into four classes on the basis of actions on (3-adrenoceptors, or on membrane currents carried by cations. We have explored the modulation of anions as a possible new approach to the prevention of sudden cardiac death. Using an isolated rat heart model, we showed that substitution of the chloride anion by nitrate protects against ischaemia- and reperfusion-induced arrhythmias in a concentration-dependent manner without deleterious haemodynamic consequences. Studies of the site of action demonstrated that (i) protection against ischaemia-induced VF resulted largely from an action in the ischaemic zone, and (ii) protection against reperfusion-induced VF resulted principally from an action occurring during reperfusion and within the reperfused tissue. Significant reductions in ventricular arrhythmias were produced by anions with membrane permeabilities greater than chloride and the anion, methylsulphate, which is less permeable than chloride was proarrhythmic. These findings strongly support the hypothesis that altered membrane permeability may contribute to the antiarrhythmic activity of some anion surrogates. In conclusion, anions appear to play a hitherto unrecognized role in arrhythmogenesis in ischaemia and reperfusion. Their manipulation represents a novel target for antiarrhythmic agents. 3

CONTENTS

PAGE

TABLES...... 8

FIGURES...... 10

INTRODUCTION...... 16 1. Ischaemia-induced arrhythmias ...... 16 A. Re-entry B. Abnormal automaticity C. Flow of injury current

2 .Reperfusion-induced arrhythmias ...... 19

3. Anions and the heart...... 21 A. Alteration of membrane conductance B. Alteration of intracellular cyclic nucleotides C. Alteration of intracellular pH D. Inotropic effects 4. Clinical considerations ...... 28 A. Prevention of sudden cardiac death B. Epidemiology of sudden cardiac death 5. Objectives ...... 30

STUDY 1 NITRATE SUBSTITUTION OF CHLORIDE (NCS1...... 31 1. Introduction ...... 32

2 . Materials and methods...... 32 A. Animals and experimental methods B. Verification of coronary occlusion and reperfusion C. Composition of perfusion solutions D. Experimental protocol E. Diagnosis and quantification of arrhythmias F. Electrogram (ECG) analysis G. Exclusion criteria H. Statistics 4

3. Results...... 38 A. Ischaemia-induced arrhythmias B. Reperfusion-induced arrhythmias C. Heart rate, coronary flow and occluded zone size D. Electrogram (ECG) changes 4. Discussion ...... 51 A. Action of NCS on ischaemia-induced arrhythmias B. Action of NCS on reperfusion-induced arrhythmias C. Insight into mechanisms of action from changes in ECG D. Cellular mechanism of action: hypotheses

STUDY 2. INVESTIGATION OF THE SITE OF ANTIARRHYTHMIC ACTION OF NCS...... 59

2 u l CROSS-OVER STUDY IN REGIONAL ISCHAEMIA: SITE OF ANTIARRHYTHMIC ACTION OF NCS...... 60 1. Introduction ...... 60

2 . Materials and Methods...... 60 A. Composition of perfusion solutions B. Experimental protocol C. Exclusion criteria 3. Results...... 62 4. Discussion ...... 62

2,2. A NEW MODEL OF GLOBAL ISCHAEMIA: APPLICATION TO THE STUDY OF SYNCYTIAL MECHANISMS OF ARRHYTHMOGENESIS AND SITE OF ANTIARRHYTHMIC ACTION OF NCS...... 65 1. Introduction ...... 65 2. Animals and methods...... 67 A. Global ischaemia: a new model B. Composition of perfusion solutions C. Experimental protocol D. Exclusion criteria 5

3. Results...... 69 A. Occluded zone sizes B. Ischaemia-induced arrhythmias C. Reperfusion-induced arrhythmias D. Antiarrhythmic action of NCS E. Haemodynamics 4. Discussion ...... 76 A. A new model of global ischaemia B. Occluded zone size and ischaemia- induced arrhythmias; implications for 'injury current' theory of arrhythmogenesis C. Reperfusion arrhythmias D. Application of the model to NCS

Z2. CROSS-OVER STUDIES IN THE GLOBALLY ISCHAEMIC HEART; ANTIARRHYTHMIC ACTION OF NCS DURING REPERFUSION...... 81 1. Introduction ...... 81 2. Materials and methods...... 81 A. Animals and experimental methods B. Composition of perfusion solutions C. Experimental protocol D. Exclusion criteria 3. Results...... 83 4. Discussion ...... 83

STUDY 3. POSSIBLE CELLULAR MECHANISMS OF ACTION

OF ANIQN SUBSIIIU T IQ J iQ N ARRHYTHMIAS...... 8 6

2l± modification of membrane resistance ...... 87 1. Introduction...... 87

2 . Materials and methods...... 87 A. Animals and experimental methods B. Composition of perfusion solutions C. Experimental protocol 6

D. Exclusion criteria

3. Results...... 8 8 A. Ischaemia-induced arrhythmias B. Reperfusion-induced arrhythmias C. Heart rate, coronary flow and occluded zone size D. ECG changes 4. Discussion ...... 100 A. Actions of anion substitution on ischaemia- induced arrhythmias B. Actions of anion substitution on reperfusion- induced arrhythmias C. Insight into mechanism of action from changes in ECG D. Insight into mechanism of action from haemodynamic variables E. Importance of the membrane permeability of anion surrogates

2*2 MODIFICATION OF INTRACELLULAR CYCLIC - NUCLEOTIDES...... 102 1. Introduction ...... 102 2. Materials and methods...... 102 A. Animals and experimental methods B. Composition of perfusion solutions C. Experimental protocol D. Assay techniques E. Exclusion criteria and statistics 3. Results...... 104 A. Sham ligation B. Effect of duration of regional ischaemia C. Effect of anion surrogates 4. Discussion ...... 109

2*2 MODIFICATION OF INTRACELLULAR pH...... 111 1. Introduction ...... 111

2 . Materials and methods...... 111 7

A. Animals and experimental methods B. Composition of perfusion solutions C. Experimental protocol D. Statistical analysis 3. Results...... 113 A. Metabolic effects B. Effects on pH 4. Discussion ...... 118

2 A MODIFICATION OF ISOCHORIC VENTRICULAR CONTRACTION...... 119 1. Introduction ...... 119

2 . Materials and methods...... 119 A. Animals and experimental methods B. Composition of perfusion solutions C. Experimental protocol D. Exclusion criteria E. Statistics and data analysis 3. Results...... 121 A. Systolic function B. Diastolic function C. Heart rate and coronary flow 4. Discussion ...... 126

DISCUSSION...... 128 1. NCS...... 128

2 . Site of antiarrhythmic action of NCS ...... 128 3. Possible cellular mechanisms of action of NCS ...... 130 4. Clinical implications ...... 132 5. Species considerations and future progress ...... 133

6 . Conclusion ...... 135

ACKNOWLEDGEMENTS...... 136

REFERENCES...... 137 8

TABLES

STUDY 2.1

Table 1 : Incidences (%) of ischaemia-induced arrhythmias in the 30 min regional ischaemia cross-over study. See text for experimental protocol for

groups 1-4 (n=1 2/group). Abbreviations: Cl- = chloride-containing solution, NO 3- = nitrate-containing solution, NZ = non-occluded zone, OZ = occluded zone, VPB = ventricular premature beat, BG = bigeminy, S = salvo, VT = ventricular tachycardia, VF = ventricular fibrillation. * indicates p< 0.05 compared to group 1 (chloride-containing solution throughout).

STUDY 2.2

Table 2 : Heart rate (beats/ min). Groups are identified by their occluded zone size expressed as a % of total ventricular weight. (* indicates nitrate perfusion). There were no significant differences between groups before or during ischaemia or during reperfusion.

STUDY 2.3

Table 3: Incidence (%) of reperfusion-induced ventricular tachycardia (VT) and ventricular fibrillation (VF) in the 10 min global ischaemia cross-over study. See text for protocol for groups A-B. Abbreviations: Cl- = chloride-containing solution, N03- = nitrate-containing solution, IZ = ischaemic zone, RZ = reperfused zone. * indicates p< 0.05 compared to group A (chloride-containing solution throughout).

STUDY 3.3

Table 4: Mean contents (pmoI/heart) ± SEM of intracellular phosphate (Pi), phosphocreatine (PCr) and adenosine triphosphate (ATP) in 2 groups of hearts (n=4/group) perfused with either standard chloride-containing solution or solution modified by complete isotonic substitution of nitrate for chloride. * indicates p< 0.05 vs. chloride control. 9

STUDY 3.4

Table 5: Measures of sytolic function before and after switching from chloride solution to one of 7 interventions. Ees = slope (mmHg/ pi) relationship between peak systolic pressure and intraventricular volume. Developed pressure (Devel P) = systolic pressure - diastolic pressure. All values expressed as mean ± SEM. * indicates p < 0.05 for paired t test comparing value before and after cross-over of solution.

Table 6 : Measures of diastolic function. Diastolic pressure before and after switching chloride solution to one of seven interventions. All values expressed as mean ± SEM. * indicates p < 0.05 for paired t test comparing value before and after cross-over of solution.

Table 7: Heart rate (beats/min) and coronary flow (ml/min/g wet ventricular weight) in the seven groups. All values are mean ± SEM. Values were recorded 25 minutes before intervention (-2 5 min), 5 min after intervention ( 5 min) and 25 min after intervention ( 25 min). * indicates p< 0.05 vs. chloride group. 10

FIGURES

STUDY 1

Fig. 1 : Measurement of QRST width at 90% repolarization (QRSTg0). No separate T wave is seen in the rat ECG so that conventional measurement of QT interval was not practicable. A Ag-AgCI wire electrode was inserted into the left ventricular muscle mass to record an electrogram (ECG) which was recorded against a lead attached to the aortic cannula. This arrangement gave a clear P wave and ventricular complex.

Fig. 2: Ischaemia-induced arrhythmias during 30 min regional ischaemia. Arrhythmia incidences were reduced (p<0.05) in the case of VF (with chloride: nitrate ratios of ^ 50: 50), VT (with ratios ^ 25: 75), S and BG (0:100 only). VPB incidence was not reduced by nitrate substitution of chloride (NCS).

Fig. 3: Reperfusion-induced ventricular fibrillation (VF). Group incidences in hearts reperfused after 5, 10, 15 or 30 min regional ischaemia are shown.

Fig. 4: Reperfusion-induced ventricular tachycardia (VT). Group incidences in hearts reperfused after 5, 10, 15 or 30 min regional ischaemia are shown.

Fig. 5: Reperfusion-induced sustained ventricular fibrillation (SVF). 'Binned' group incidences (data for hearts subjected to 5, 10, 15 and 30 min ischaemia having been combined) are shown.

Fig. 6 : Reperfusion-induced ventricular premature beat (VPB). Group incidences in hearts reperfused after 5, 10, 15 or 30 min regional ischaemia are shown.

Fig. 7: Onset times of first reperfusion-induced arrhythmia in hearts reperfused after 5, 10, 15 or 30 min regional ischaemia. Values are mean ± SEM (logarithmic scale).

Fig. 8 : Heart rate 1 min before the start of reperfusion. Values are mean ± SEM beats/min. 11

Fig. 9: Calculated coronary flow in the non-ischaemic region recorded 1 min before the start of reperfusion. Values are mean ± SEM ml/min/g wet wt.

Fig. 10: Calculated recovery of coronary flow in the reperfused region recorded

1 min after the start of reperfusion. Values are mean ± SEM ml/min/g wet wt.

Fig.1 1 : Occluded zone size (% total ventricular weight) expressed as mean ± SEM of total ventricular wt. There were no statistically significant differences between groups.

Fig. 1 2 : ECG changes in hearts subjected to 30 min of regional ischaemia.

Width (msec) was expressed as mean ±SEM; (a) QRSTg0 (width of the QRST complex measured at 90% repolarization), (b) PR interval. Statistical significance 'stars' have been omitted for clarity.

Fig. 13: The 'bell shaped' time-susceptibility curves of reperfusion-induced arrhythmias. Changes in the shape or location of the bell shaped curve give an indication as to the mechanism of action of an antiarrhythmic intervention. For example, bradycardia shifts the curve to the right, indicative of a delay in development of susceptibility to arrhythmias. Drugs which shift the curve downwards can be regarded as having a specific action on the arrhythmogenic stimulus of reperfusion (most likely an electrophysiological action). In the present case NCS appeared to shift the curve substantially downward and to the left.

STUDY 2.2

Fig. 14: Relationship of % incidence of ischaemia-induced ventricular tachycardia (VT) and ventricular fibrillation (VF) to occluded zone size (% of total ventricular weight).

Fig. 15: Incidence (%) of VPB in consecutive 5 min periods during 30 min ischaemia for different groups with different occluded zone size (% of total ventricular weight). Sham =0%, low = 21 ± 0.8%, high = 47 ± 1.0% and global = 1 2

100%. The incidence of VPBs during sham ligation were low, but not zero, as the ligature itself had a minor proarrhythmic effect.

Fig. 16: Relationship of the % incidence of reperfusion-induced ventricular tachycardia (VT) and ventricular fibrillation (VF) to occluded zone size (% of total ventricular weight).

Fig. 17: Incidence (%) of VPB in consecutive 5 min periods during 30 min global ischaemia in hearts initially perfused with either standard chloride-containing solution or solution modified by substitution of chloride with nitrate.

STUDY 3.1

Fig. 18: Incidence (%) of ischaemia-induced ventricular fibrillation (VF).

Fig. 19: Incidence (%) of organized ischaemia-induced ventricular ectopy. Abbreviations; VPB = ventricular premature beat, BG = bigeminy, VT = ventricular tachycardia.

Fig. 20: Onset time of first ischaemia-induced arrhythmia. Note that the timescale is logarithmic.

Fig. 2 1 : Incidence (%) of reperfusion-induced VF. The high incidence of ischaemia-induced arrhythmias at 30 min in the methylsulphate group made study of reperfusion-induced arrhythmias impracticable in this group.

Fig. 22: Incidence (%) of reperfusion-induced organized ventricular ectopy. Abbreviations; VPB =ventricular premature beats, BG = bigeminy, VT = ventricular tachycardia. The high incidence of ischaemia-induced arrhythmias at 30 min in the methylsulphate group made study of reperfusion-induced arrhythmias impracticable in this group.

Fig. 23: Onset time of first reperfusion-induced arrhythmias. Note that the time scale is logarithmic. The high incidence of ischaemia-induced arrhythmias at 30 min in the methylsulphate group made study of reperfusion-induced arrhythmias 13 impracticable in this group.

Fig. 24: Heart rate (beats/ min) in all groups throughout time course of experiment.

Fig 25: Composite figure showing coronary flow in all groups during the course of the experiment. Flow 1 min before occlusion is expressed as ml/min/g total ventricular weight. Flow at 1 min after occlusion and 1 min before reperfusion is expressed as ml/min/g non-ischaemic ventricular weight. Flow 1 min after reperfusion is expressed as ml/min/g ischaemic ventricular weight and represents recovery of flow in the reperfusion zone.

Fig. 26: Occluded zone size as a % of total ventricular weight.

Fig. 27: Width in msec of QRST complex at 90% repolarization (QRSTgo) throughout time course of experiment.

STUDY .3,2

Fig. 28: Mean concentration ± SEM of cAMP (nmole/g protein), cGMP (pmole/g protein) and molar ratio of cAMP/cGMP at 5 min (n=4), 15 min (n= 6 ) and 30 min (n=4) of regional ischaemia. Myocardium was biopsied from both ischaemic (isch) and nonischaemic (nonisch) zones.

Fig. 29: Concentrations ± SEM of cAMP and cGMP and molar ratio of cAMP/cGMP in ischaemic (isch) and non-ischaemic (nonisch) zones of hearts biopsied after 15 min regional ischaemia. Groups of hearts (n=6 /group) were perfused with either standard chloride-containing solution (chlor) or with one of 4 isotonic anion surrogates; methylsulphate (MethS), bromide (Brom), iodide or nitrate. A separate group of hearts (n=4) were sham ligated and perfused for 15 min without regional ischaemia.

Fig. 30: Concentration (nmole/g) of cAMP in the non-ischaemic zone of hearts plotted against % incidence of VF in similar groups (Study 2.1). Hearts were rendered regionally ischaemic for 15 min before biopsy. Groups (n= 1 2 /group) 14 were perfused with standard chloride-containing solution or one of 4 other solutions with isotonic substitution of chloride with methylsulphate, bromide, iodide or nitrate. Considerable overlap of values occurred between groups revealing little evidence of correlation between incidence of VF and concentration of cAMP in the non-ischaemic zone of hearts rendered regionally ischaemic.

Fig. 31: Mean concentrations ± SEM of cAMP (nmole/g protein), cGMP (pmole/g protein) and molar ratio of cAMP/cGMP in reperfused zone of 5 groups of hearts

(n=6 /group) perfused with methylsulphate (MethS), chloride (Chlor), bromide (Brom), iodide or nitrate. Hearts were regionally ischaemic for 15 min and then reperfused. Reperfused tissue was biopsied 1 min after reperfusion. These values are compared to a control group (sham) in which the heart was not rendered ischaemic.

STUDY 3.3

Fig. 32: Illustrative NMR spectra of a heart perfused with standard chloride-containing solution and rendered globally ischaemic for 10 min. First spectrum recorded 1 min before onset of ischaemia (1-1) and 9 spectra taken at

1 min intervals during the course of global ischaemia. Spectral peaks for the external standard methylenediphosphonate (MDP) and myocardial phosphate (Pi), phosphocreatine (PCr) and adenosine triphosphate- (ATP) are identified.

Fig. 33: Illustrative NMR spectra of a heart perfused with solution modified by nitrate substitution of chloride and rendered globally ischaemic for 1 0 min. First spectrum recorded 1 min before onset of ischaemia (1-1) and 9 spectra taken at

Fig. 34: Alteration in mean intracellular pH ± SEM in hearts (n=4/group) perfused with either standard chloride-containing solution (chi) or solution modified by complete nitrate substitution of chloride (nitr). pH was measured 1 min before the onset of global ischaemia and at 1 min intervals during the global 1 5 ischaemia. * indicates p< 0.05 compared to time-matched chloride-containing hearts. 16

INTRODUCTION

Ischaemia-induced ventricular fibrillation (VF) is a major cause of sudden death (Campbell, Murray, and Julian 1981, Davies 1981, Lo et al. 1988, Rissanen, Romo, and Siltanen 1978, Roberts and Jones 1979). Ischaemia-induced arrhythmias have been shown to respond to numerous drugs in animal models. For example, class I, II, III and IV antiarrhythmics as well as alpha adrenoceptor antagonists, 5-Hydroxytryptamine receptor antagonists and numerous other classes of drugs have all been reported to have antiarrhythmic activity (Parratt 1982). However, owing to various side effects, none have proven useful as routine prophylactic agents for patients with coronary artery disease at risk of myocardial ischaemia; ironically, the most effective antifibrillatory agents, the class III drugs (Bacaner et al. 1986, Vaughan Williams 1985), appear to have the greatest potential for causing side effects (Singh 1983). For this reason it is apparent that there is an acute need for a new approach to the prevention of ischaemia-induced VF. Although the incidence of reperfusion-induced VF in sudden death is not established, it is recognised that spontaneous reperfusion (occurring, for example, in vasospasm) is capable of eliciting VF in man (Tzivoni et al. 1983). Therefore, it would be advantageous if novel means for prevention of reperfusion-induced VF could also be developed. Modification of cation homeostasis has proven a logical target for intervention in the setting of ischaemia and reperfusion. Ischaemia-induced arrhythmias are susceptible to inhibition by drugs which inhibit sodium, calcium and potassium currents (Bacaner et al. 1986, Curtis 1986) and reperfusion-induced arrhythmias can be inhibited by lowering extracellular calcium concentration (Tosaki and Hearse 1987) or raising extracellular potassium concentration (Curtis and Hearse 1989a). By contrast, homeostasis of the major extracellular anion, chloride, has been neglected as a possible target for intervention and little information is available concerning the antiarrhythmic potential of manipulation of anion homeostasis during ischaemia or reperfusion.

1. Ischaemia-induced arrhythmias 17

Susceptibility to ischaemia-induced VF is greatest during the first few hours of ischaemia and then declines exponentially with time (Adgey et al. 1971, Campbell, Murray, and Julian 1981). The exact mechanisms of ischaemia-induced arrhythmias are not fully established (Janse and Wit 1989, Parratt 1982). Possible pathophysiological mechanisms which may play a role in the initiation of ischaemia-induced arrhythmias include accumulation of catecholamines (Sheridan et al. 1980), lysophosphatides (Corr, Gross, and Sobel 1984), thromboxane (Coker et al. 1980), leukocytes (Koltai et al. 1982) and elevated intracellular cyclic adenosine 3’: 5' monophosphate (cAMP) (Corr, Witkowski, and Sobel 1978, Dobson Jr. and Mayer 1973, Podzuweit and Lubbe 1977, Podzuweit, Lubbe, and Opie 1976, Podzuweit et al. 1978). It is possible that these mechanisms might influence anions. For example, a cAMP-mediated activation of a Cl- current has recently been demonstrated to be responsible for a catecholamine-induced increase in membrane conductance in ventricular cells (Bahinski et al. 1989, Harvey and Hume 1989, Matsuoka, Ehara, and Noma 1990). In addition to pathophysiological mechanisms, syncytial mechanisms of arrhythmogenesis during ischaemia should also be considered. Animal experiments have provided useful information in this regard. In the dog, it has been shown that susceptibility to arrhythmias is critically dependent upon heart rate (Redwood, Smith, and Epstein 1972), occluded zone size (Endo et al. 1983) and coronary collateral flow (Bolli, Fisher, and Entmann 1986). Coronary collateral flow can be considerable in the dog and its abundance is associated with a reduced incidence of arrhythmias (Bolli, Fisher, and Entmann 1986, Parratt 1982). Using the rat it has been shown that heart rate is also critically important (Bernier, Curtis, and Hearse 1989) but, as collateral flow is low in the rat (Maxwell, Hearse, and Yellon 1978, Wexler et al. 1988), susceptibility to arrhythmias is uniformly high within a wide range of heart rates (Curtis, MacLeod, and Walker 1987) making it a useful species for studying arrhythmias. In the rat the maximal susceptibility to ischaemia-induced arrhythmias occurs 10-15 min after coronary occlusion suggesting that the electrophysiological dysfunction is maximal at this time. The conditions present in the ischaemic tissue 10-15 min after coronary occlusion include shortening of the action potential duration (Carmeleit 1978), 1 8 depolarisation, variable changes in refractoriness, slow conduction (Williams et al. 1974), reduced intracellular and extracellular pH (Garlick, Radda, and Seely 1979) and elevated extracellular potassium concentration (Hirche et al. 1980, Janse and Wit 1989, Parratt 1982). These conditions are believed to predispose to the initiation of ventricular arrhythmias. Susceptibility to ischaemia-induced arrhythmias declines as the cell damage develops and the cells become electrically quiescent. Three syncytial mechanisms of arrhythmogenesis have been proposed; re-entry, automaticity and flow of injury current.

A. Re-entrv

For many years re-entrant excitation has been considered to be the most important mechanism of ischaemia-induced arrhythmias (El-Sherif, Scherlag, and Lazzara 1975, El-Sherif et al. 1977, Han, Goel, and Hanson 1970). Re-entrant excitation requires a region of unidirectional block. Conduction must occur in tissue beyond the block (via adjacent uninvolved tissue) with a sufficiently long delay so that the tissue proximal to the block has recovered its excitability and can be re-excited by a wave front passing retrogradely. Thus, the circulating wave travels around an area of refractory tissue, the length of the circuit being determined by the product of the conduction velocity and the refractory period. Factors promoting re-entry are slow conduction, short refractory periods and inhomogeneities in refractory periods of adjacent areas (Mines 1913, Schmidt and Erlanger 1928). The presence of an anatomical obstacle, around which an excitatory wave may perpetually circulate, is an additional factor enhancing the likelihood of re-entry. Such obstacles may be present in the late stages of ischaemia and infarction when strands of connective tissue intermingle with surviving myocardial fibres. In acutely ischaemic myocardium, no anatomical obstacles are present but this does not exclude a re-entry mechanism because re-entry has been described without the involvement of an anatomical obstacle (Allessie, Bonke, and Schopman 1977).

B. Abnormal automaticity

Normally, pacemaker activity in Purkinje fibres occurs at a slow rate, and is masked by the normal sinus impulse, which is four to five times as rapid. 19

Abnormal automaticity occurring in Purkinje fibres and ventricular muscle that is partially depolarized during ischaemia might increase the pacemaker rate in Purkinje fibres making them a focus for ventricular premature beats (VPB). However, this mechanism is unlikely to be important in ischaemic tissue (Sherlag et al. 1970) as the high extracellular K+ tends to suppress abnormal automaticity (Hoffman and Rosen 1981).

C. Flow of injury current

This more recent theory is believed to be an important mechanism for the initiation of arrhythmias in early regional ischaemia. Ischaemic tissue is depolarized with respect to adjacent uninvolved tissue and slow conduction is present, putting the action potential in the adjacent ischaemic and non- ischaemic zones out of phase; as a consequence, injury current may flow from the ischaemic zone prematurely re-exciting the non-ischaemic tissue (Janse et al. 1979, Janse et al. 1980). A single exclusive mechanism need not necessarily be responsible for the initiation of all ventricular arrhythmias; for example, it has been suggested that re-entry may be initiated by the flow of injury current (Pogwizd and Corr 1990).

2. Reperfusion-induced arrhythmias

It has long been established that reperfusion of ischaemic myocardium may result in arrhythmias (Tennant and Wiggers 1935). Interest in reperfusion-induced arrhythmias has been stimulated by evidence suggesting that they might be responsible for some sudden cardiac deaths in the community (Maseri, Severi, and Marzullo 1982, Tzivoni et al. 1983) and in recent years the increasing use of fibrinolysis in the acute management of myocardial infarction (Yusuf et al. 1985) has made iatrogenic reperfusion arrhythmias more common; indeed, ventricular arrhythmias have been described as a marker of successful reperfusion during thrombolytic therapy in the management of acute myocardial infarction (Goldberg et al. 1983). The electrophysiological mechanisms underlying reperfusion-induced arrhythmias are unresolved (Manning and Hearse 1984, Witkowski and Corr 20

1984) and it remains unclear whether they differ substantially from the mechanisms causing ischaemia-induced arrhythmias. Any attempt to directly assess these mechanisms is fraught with difficulties owing to the inherently short lag-time between the beginning of reperfusion and the onset of arrhythmias. In intact hearts in vivo and in vitro reperfusion-induced arrhythmias develop within seconds of reperfusion in both rats and dogs (Blumgart, Gilligan, and Schlesinger 1941, Curtis and Hearse 1989b, Fujimoto et al. 1983, Harris 1948, Murdock et al. 1980). Consequently the anecdotal nature of some data has been associated with contradictory reports and hypotheses (Fujimoto et al. 1983, Murdock et al. 1980, Penkoske, Sobel, and Corr 1978). In order to permit electrophysiological studies it is necessary to use less pathophysiologically relevant models than the intact heart, such as superfused Purkinje fibres (Ferrier, Moffat, and Lukas 1985). The disadvantages of the superfused Purkinje fibre model are that (i) ischaemia is simulated rather than real, (ii) electrophysiological dysfunction develops slowly (over minutes rather than seconds) and (iii) ischaemia is simulated globally rather than regionally, precluding assessment of syncytial electrophysiological dysfunction (arrhythmias). These considerations indicate that the potent but short lived, arrhythmogenic stimulus of reperfusion may be more profitably investigated using a semi-empirical approach whereby mechanisms are probed by examining the effectiveness of interventions (Curtis and Hearse 1989a). Such an empirical approach has revealed that reperfusion-induced arrhythmias appear to be resistant to many drugs which are considered to be of benefit in ischaemia (Manning and Hearse 1984, Naito et al. 1981). Furthermore, it has been shown that the effects of K+ on.ischaemia-induced VF were different from its effects on reperfusion-induced VF. These findings suggest that ischaemia- and reperfusion-induced VF are unlikely to be initiated by a common mechanism. However, the mechanism of maintenance of ischaemia-induced and reperfusion-induced VF might be common, since the tendency for spontaneous defibrillation to occur as reflected in the incidence of sustained VF (SVF) is equally sensitive to extracellular K+ in both settings (Curtis and Hearse 1989a). When considering reperfusion-induced arrhythmias, the relationship between reperfusion and the preceding period of ischaemia must be taken into account. The incidence of reperfusion-induced VF is related to the period of 21 preceding ischaemia, the time-susceptibility relationship typically shows a bell shaped distribution which in the isolated rat heart shows maximal susceptibility at 10-15 mins of ischaemia (Manning and Hearse 1984). The peak period of susceptibility varies between species; for example, in the dog it is after 20-30 mins of ischaemia (Balke et al. 1981). It appears that the occurrence of reperfusion-induced arrhythmias is not determined by the extent of arrhythmias occurring during the preceding period of ischaemia (Curtis and Hearse 1989a, Shiki and Hearse 1987). Suggested pathophysiological mechanisms responsible for the production of reperfusion-induced arrhythmias include disturbance of cyclic nucleotide metabolism (Podzuweit et al. 1989), activation of adrenoceptors (Sheridan et al. 1980), accumulation of amphiphiles (Sobel et al. 1978) and production of free radicals (Woodward and Zakaria 1985). Although all of these potential mechanisms have been linked with electrophysiological disturbance (Witkowski and Corr 1984), it remains unclear which plays the most important role in arrhythmogenesis. Indeed, the fact that it is difficult to abolish reperfusion-induced arrhythmias with any single intervention suggests that multiple pathophysiological mechanisms might be involved (Manning and Hearse 1984) and that other, as yet unidentified mechanisms may also contribute.

3. Anions and the heart

Anion manipulation can affect many biochemical and electrophysiological processes and alterations may, in theory, affect susceptibility to arrhythmias, thereby representing a new focus for therapeutic interventions.

A. Alteration of Membrane conductance

Intracellular microelectrode recordings of transmembrane potentials (Em) reveal that the intracellular compartment of cardiac muscle is negatively charged relative to the extracellular compartment. This negative resting potential varies in different areas of the heart but at rest across the membranes of Purkinje and ventricular cells it is in the region of -90 mV (Winslow 1984). The intracellular and extracellular ionic compositions are well known with Na+ and Cl- the major 22

extracellular cation and anion respectively, and K+the major intracellular cation with the negative intracellular charge being carried by CI-, organic acids, proteins, phosphates and sulphates (Honig 1981).The prime determinant of Em is the selective permeability of the various ions on each side of the cell membrane. In cardiac muscle during diastole, K+ has the greatest permeability and hence primarily determines the diastolic membrane potential; that is, the

equilibrium potential for K+ (E k+) most closely approximates to that of Em. The

cardiac cell membrane has relatively slight permeability to Na+ and even less to C l-. The equilibrium potential for Cl- is approximately -37 mV (De Mello 1963). The tendency of ions to move across the cell membrane down the concentration gradient is opposed by the electrical gradient. When these charges are balanced no net movement of ions occurs. Given the intracellular and extracellular ionic concentrations, the Nernst equation predicts the equilibrium potential (Em) if only one ion is involved;

Em=RT/ZF In [Co]/[Ci]

where R is the gas constant, T the absolute temperature, F the Faraday, Z the valence of the ion, [Co] the concentration of the ion outside the cell, and [Ci] the concentration of the ion inside the cell.

The membrane is of course permeable to more than one ion and in order to calculate the approximate combined effects on resting transmembrane potential, the Goldman-Hodgkin-Katz Constant Field Equation (Goldman 1943, Hodgkin and Katz 1949) is employed;

(Pk+ [M + PNa+[Nao+] + Pcr[Clr]) Em = RT/F In ------(P k + IM + PNa+[Nai+] + Pci-[Clo-])

where Pk+, PNa+. and Per represent membrane permeabilities to K+, Na+ and Cl- respectively and the values within brackets [ ] represent ion concentrations of K+, Na+ and CI-. The symbols i and o represent inside and outside respectively.

The electrochemical gradients are maintained by energy dependent ionic 23

pumps (e.g. the K+-Na+ ATP-dependent pump). The intracellular Cl- concentration is approximately 3 times higher than would be expected from a passive distribution at the resting membrane potential (Desilets and Baumgarten 1986, Vaughan-Jones 1979a) implying active transport of chloride into the cells. The mechanism of active transport remains controversial. In cultured chick heart cells a K+-CI- cotransport has been described (Piwinica-

Worms et al. 1985) and in mammalian cardiac muscle a reversible C I-HO 3 - counter-exchange pump has been described (Vaughan-Jones 1979b).

In mammalian skeletal muscle Cl- conductance contributes 6 8 % of total membrane conductance (Gm) (Hodgkin and Horowicz 1959). The magnitude of this contribution to membrane conductance is in striking contrast to the lower Cl- conductance seen in resting cardiac muscle where Cl- permeability is low. However, in a major review of chloride channels in muscle, Bretag (Bretag 1987) emphasized the considerable evidence from membrane potential changes and conductance measurements for a small but significant chloride conductance (gCI-) in cardiac muscle. Relative conductance ratios between K+, Na+ and Cl- in the rabbit sinoatrial (SA) node have been calculated to be 1.0: 0.58: 0.15 respectively (Seyama 1977) with Cl- contributing about 9% of Gm. Hutter and Noble (Hutter and Noble 1961) pointed out that in the Purkinje fibre, Cl- ions carry, at the most, 30% of Gm in the resting condition and Carmeleit (Carmeleit 1961) also concluded that Cl- contribution to Gm of sheep Purkinje cells is small. Fozzard and Lee (Fozzard and Lee 1976) estimated that the ratio between K+ and Cl- conductance is 1.0: 0.15 in rabbit ventricular myocardium. A small but significant gCI- has also been demonstrated in rabbit atrial fibres (De Mello 1963). Furthermore, chloride may become a more important charge carrier during depolarization (Carmeleit 1961, De Mello 1963, Hutter and Noble 1961). Monovalent anions may be regarded as Cl- surrogates. A number of Cl- surrogates are available. Seyama (Seyama 1979) substituted nitrate, iodide and bromide for chloride and studied their conductance by performing voltage clamp experiments on quiescent rabbit SA node cells. In quiescent cells all caused transient hyperpolarization, the order of membrane permeability being nitrate > iodide> bromide > chloride (the same authors found that the less permeable anion acetate causes transient depolarization in Na+ free solution).This sequence is inversely proportional to the naked radius of these 24

anions rather than the hydrated radius. Seyama (Seyama 1979) speculates that the SA node cell may not discriminate permeability of anions by the size of the hydrated ion radius but may do so by the difference in free energy between the binding site and the hydration of ions (Diamond and Wright 1969, Eisenman 1961). Seyama (Seyama 1979) calculated the relative membrane conductance of nitrate: iodide: bromide: chloride to be 1.5:1.15:1.07:1.0. In an earlier study of the effects of anion surrogates on strips of Purkinje tissue from several mammals (sheep, dog, cat and rabbit), Hutter and Noble (Hutter and Noble 1961) found a similar order of anion permeability, although they described a reversed order for, nitrate and iodide, the two most permeable anions they studied (iodide> nitrate> bromide> chloride> methylsulphate). Theoretical considerations (Goldman constant field equation) would predict that the effect of extracellular substitution of a permeable anion for Cl- would be time-dependent. Hyperpolarization of cardiac sarcolemma might be expected as the permeable anion entered the cell down its concentration gradient. Subsequently, as Cl- left the cell down its concentration gradient Em would fall to a new resting potential. This equilibrium Em would be more

influenced by the more permeable anion; that is closer to Eanj 0 n, resulting in possible depolarization. However, the influence of anion surrogates on active transport mechanisms is unknown as is the actual effect of permeable anion surrogates on the whole heart. It has been suggested that the effect of Cl- substitution might be mediated

via cations such as Ca 2+ or K+. Kenyon and Gibbons (Kenyon and Gibbons 1977) suggested that substitution of Cl-with impermeable anions might reduce Ca+activity in the external medium. However, Seyama (Seyama 1979) demonstrated that doubling the Ca2+ concentration failed to modify the response of rabbit SA node to acetate solution. Therefore changes in the electrical activity on substitution of Cl- with large anions are not due to the reduction of Ca+ activity but are due to the elimination of the Cl- contribution to the electrical activity of the SA node. A reduction of K+ permeability has been described following the replacement of extracellular Cl- by impermeable, but not permeable, anions (Carmeleit and Verdonck 1977). This mechanism might partly explain the reduction of membrane conductance after replacement of Cl-with impermeable anions but Hutter and Noble (Hutter and Noble 1961) found no reason to suspect that the effects of nitrate and iodide are brought about indirectly through 25 an increase in potassium permeability, for although such an action might explain the initial slowing or arrest, it cannot explain the eventual acceleration observed in dog Purkinje fibres. However, anions may influence K+ release from myocytes during ischaemia. Following coronary artery ligation, cells in the ischaemic zone depolarize (Czarnecka, Lewartowski, and Propopezyk 1973, Downar, Janse, and Durrer 1977, Kardesh, Hogancamp, and Bing 1958, Russell, Oliver, and Wojtczak 1977, Samson and Sher 1960) as myocardial cells release K+ into the extracellular space (Harris et al. 1954). Potassium loss cannot be due simply to the cell membrane becoming more permeable to potassium. If that were so, the membrane potential would initially become more negative and approach the

equilibrium potential for K+ (E k +). In hypoxia and ischaemia the membrane is initially depolarized (Poole-Wilson 1975) and the membrane potential only equals Ek+ later. Evidence from studies in the arterially perfused interventricular septum of the rabbit has shown that hypoxia resulted in an increased efflux with no evidence of altered influx of K+ suggesting that extrusion of accumulated anions from the myocardium could be the major determinant of the early potassium loss during hypoxia and ischaemia (Gaspardone et al. 1986).

B. Alteration of intracellular cyclic nucleotides

An association between ventricular fibrillation and an increased level of cAMP has been reported in animal models of arrhythmias (Opie, Nathan, and Lubbe 1979), and a general case may be made that arrhythmias are more frequent in conditions in which cAMP is elevated such as hyperthyroidism, or high autonomic sympathetic background. Elevated myocardial cAMP is associated with ischaemia in the baboon (Podzuweit et al. 1978), the dog (Wollenberger, Krausse, and Heier 1969), the cat (Corr, Witkowski, and'Sobel 1978), the pig (Podzuweit and Lubbe 1977) and the rat (Dobson Jr. and Mayer 1973). Dibutyryl cAMP (which crosses the cell membrane) lowers the ventricular fibrillation threshold in isolated rat hearts (Lubbe et al. 1976, Lubbe et al. 1978). It has therfore been proposed that cAMP may have a major role in the initiation of VF (Podzuweit, Lubbe, and Opie 1976). Theoretical mechanisms by which cAMP might be arrhythmogenic include stimulation of the slow influx of Ca+. metabolic effects (acceleration of glycogenolysis, lactate accumulation or 26

accelerated lypolysis) and accelerated pacemaker activity (Podzuweit, Lubbe, and Opie 1976). Furthermore, it has been suggested that reperfusion-induced arrhythmias in the pig may be initiated by unmasking the effects of cAMP which accumulates in the ischaemic tissue during the preceding period of ischaemia (Podzuweit et al. 1989). In contrast to cAMP, little attention has been given to the potential role of cyclic guanosine 3’:5’ monophosphate (cGMP) in arrhythmogenesis. In recent years, few discoveries have provoked more interest than the identification of endothelial-derived relaxing factor (EDRF) as NO (Palmer, Ferrige, and Moncada 1987), its synthesis from L-arginine (Palmer, Ashton, and Moncada 1988) and its ability to stimulate the enzyme guanylate cyclase to raise cellular levels of cGMP causing profound vasodilatation (Editorial 1987, Editorial 1988, Griffith and Randall 1989). There is evidence that this system may be ubiquitous throughout the body (Garthwaite, Charles, and Chess-Williams 1988), and there has been continuing interest in cGMP with the demonstration that its elevation is fundamental to the mechanism of action of nitrovasodilators (Bennett et al. 1988, Forstermann et al. 1986, Rapaport, Drazin, and Murad 1983). Thus theoretically, anion substitution with nitrate might be expected to increase cGMP. cGMP may be the second messenger of vagal stimulation released from guanyl cyclase by acetyl choline (ACh) stimulation (Forstermann et al. 1986, Ignarro et al. 1984) which decreases the inward calcium current (Ikemoto and Goto 1977) opposing one of the actions of cAMP which is associated with ischaemia-induced arrhythmias (Clusin, Buchbinder, and Harrison 1983). Thus, there is a theoretical basis for nitrate to oppose the action of cAMP, via stimulation of cGMP.

C. Alteration of Intracellular pH

Myocardial ischaemia lowers intracellular pH (Hirche et al. 1980). This effect may be related to the production of lactic acid during anaerobic metabolism and impairment of wash out of potassium (Gaspardone et al. 1986). The time-course of development of acidosis is rapid (Hirche et al. 1980), occurring during the period in which ischaemia-induced arrhythmias and susceptibility to reperfusion-induced arrhythmias develop. 27

A transmembrane reversible chloride-bicarbonate (Cl- -HCO3-) exchange process across the sarcolemma is thought to operate under normal physiological conditions. This exchange mechanism can be selectively inhibited by the amine-reactive drug SITS (4-acetamido-4'-isothiocyanato-stilbene-2,2- disulphonic acid) which increases intracellular pH from 7.08 to 7.25 in quiescent sheep Purkinje fibres (Vaughan-Jones 1979b). SITS has also recently been shown to inhibit ischaemia-induced and reperfusion-induced arrhythmias in the isolated rat heart (Curtis 1989).

D. Inotropic effects

Many antiarrhythmic agents have negative inotropic actions which limit their clinical usefulness in the prevention of ischaemia-induced and reperfusion-induced arrhythmias (Singh et al. 1987). This is particularly the case for some calcium antagonists which are effective in animal models but ineffective in man owing to the upper limit imposed on dosage by their negative inotropic action (Curtis, Walker, and Hearse 1989). Reduction of [CI-] has been shown to be positively inotropic in non-mammalian (Anderson and Foulks 1973, Horackova and Vassort 1982) and mammalian (Nosek 1979) isolated heart muscle preparations. Some of the mechanisms discussed above might also affect inotropy either by indirect or direct cellular actions. For example, in guinea pig and rabbit heart muscle preparations, a fall in intracellular pH impairs and an elevation increases contractility (Fry and Poole-Wilson 1981). Therefore, a reduction of extracellular

Cl- resulting in intracellular HCO 3 - accumulation and an elevation in pH

(Vaughan-Jones 1979b) would be predicted to increase myofibrillar Ca 2 + sensitivity and entry during the action potential and its subsequent accumulation and release by the sarcoplasmic reticulum (Allen and Orchard 1987), thus increasing the force of myocardial contraction. Anion manipulation might also act indirectly. Firstly, withdrawal of chloride might inhibit the recently described chloride-sensitive current induced by beta agonists such as isoprenaline and adrenaline (Bahinski et al. 1989, Harvey and Hume 1989, Matsuoka, Ehara, and Noma 1990). This current results in a shortened action potential duration as seen during myocardial ischaemia. A reduction in extracellular chloride with associated prolongation of 28 action potential duration (Basingthwaite, Fry, and McGuigan 1976) might therefore be expected to exert a positive inotropic effect by promoting Ca+ influx (Anderson and Foulks 1973, Horackova and Vassort 1982). Secondly, in vitro studies suggest that adenosine 5'-triphosphate-creatine phosphotransferase might be inhibited by substrate-anion complexes so that anions might influence intracellular adenosine triphosphate (ATP) metabolism which would be expected to affect myocardial contractility (Milner-White and Watts 1971).

4. Clinical considerations

A. Prevention of sudden cardiac death

Because of its acute nature, the clinical classification and aetiology of sudden death remains controversial (Hinkle Jr and Thaler 1982, Lovegrove and Thompson 1978). However, there is strong evidence that ischaemia-induced VF is a major cause(Campbell, Murray, and Julian 1981, Davies 1981, Lo et al. 1988, Rissanen, Romo, and Siltanen 1978, Roberts and Jones 1979). This is supported by the recently described efficacy of the internal cardiac defibrillator (Tchou et al. 1988). Frequent and multifocal ventricular arrhythmias are predictive of future sudden death in patients suffering from ischaemic heart disease (Bigger Jr et al. 1984, Ruberman et al. 1977, Temesy-Armos et al.

1988). Anti-ischaemic interventions such as 6 -blockers (Norwegian Multicentre Study Group 1981), coronary artery surgery and antiplatelet drugs (Lewis Jr et al. 1983) appear to reduce the incidence of sudden arrhythmic death in specific subgroups of patients with coronary artery disease (Nattel and Waters 1990) and in patients with ventricular failure, treatment to improve cardiac function may also reduce arrhythmias and mortality (CONSENSUS Trial Study Group 1987). However, the use of antiarrhythmic agents which reduce the incidence of ventricular ectopic activity has not, to date, been proven to reduce the incidence of sudden cardiac death; in some cases they have actually increased mortality (CAST Investigators 1989) in patients with ischaemic heart disease (Goldstein 1990) and sudden cardiac death remains the major mode of death in the United Kingdom.

B. Epidemiology of sudden cardiac death 29

In Britain there is a strong geographical gradient in the risk of major ischaemic heart disease events and residence appears to be a more important determinant of risk than place of birth (Elford et al. 1989). It has been suggested that water quality might influence these regional variations (Pocock et al. 1980). The United Kingdom is endeavouring to comply with the European Community Directive 80/778/EEC on the quality of water intended for human consumption which came into force in 1985 and specifies a maximium nitrate concentration of 50mg/L (Report on a WHO meeting 1984, The working party on nitrate and related compounds 1987). The possible deliterious effects of nitrates in drinking water have received much attention. There is a proven relationship between elevated nitrate in drinking water and infantile methaemoglobinaemia in bottle fed babies. However, in Britain, the Department of Health has received reports on only 14 cases of infantile methaemoglobinaemia attributable to nitrates in drinking water in the past 35 years. The last reported death was in 1950 and the last confirmed case was in 1972. Nitrates can be converted in vivo to the nitrite ion which can combine with secondary or tertiary amines to form N-nitroso derivatives. Certain N-nitroso compounds have been shown to produce cancers in a variety of laboratory animals. However, epidemiological reviews, directed mainly at the incidence of gastric cancer, have not demonstrated a relationship between elevated nitrate levels in drinking water and malignancy (Beresford 1985, Forman, Al-Dabbagh, and Doll 1985). But is there any evidence that nitrates in drinking water might be beneficial? The British Regional Heart Study (Pocock et al. 1980) was a major epidemiological project which related cardiovascular mortality to water quality in 253 large urban centres throughout Great Britain. It demonstrated a strong negative correlation between cardiovascular mortality and the nitrate content of drinking water; the relationship was as strong as that between water hardness and cardiovascular mortality. As in all such epidemiological studies, such correlations cannot be interpreted as showing a causal link. However the evidence (presented in this thesis) that nitrates may confer specific antiarrhythmic activity heightens interest in the relationship between cardiovascular disease and nitrate consumption. As we have pointed out (Ridley and Curtis 1990c) further epidemiological data on the incidence of ischaemic 30 heart disease, sudden cardiac death and their relationship to nitrate content of drinking water would be of great interest.

5. Objectives

In this thesis the potential role of anions in arrhythmogenesis is examined in isolated rat heart models to determine whether modification of anion homeostasis may represent a potential new target for therapeutic intervention. Initially we aim to establish that nitrate substitution of the chloride anion in the extracellular milieu is capable of profoundly affecting the incidence of ischaemia- and reperfusion-induced arrhythmias. Having established this phenomenon, sites of antiarrhythmic action are compared in ischaemia- and reperfusion-induced arrhythmias; differences in sites of action strongly suggest different mechanisms of arrhythmogenesis. Finally, possible cellular mechanisms to explain the antiarrhythmic action of anion manipulation are explored together with consideration of the effect of anion manipulation on inotropy in the whole heart. 31

STUDY 1

NITRATE SUBSTITUTION OF CHLORIDE (NCS1 32

1. INTRODUCTION

As an initial approach to investigating the role of anions in ischaemia- and reperfusion-induced arrhythmias, we replaced the principal extracellular anion, chloride, with the nitrate anion. This had been shown in pilot studies to have antiarrhythmic potential. In this study 5 solutions with different ratios of chloride to nitrate and 4 different periods of regional ischaemia were employed to investigate the concentration-dependence and (in the case of reperfusion-induced arrhythmias) time-dependence of antiarrhythmic activity. Possible indirect mechanisms responsible for any antiarrhythmic effects were evaluated by contiguous assessment of relevant antecedent hemodynamic variables. Brief reports of this work have already been published (Curtis, Ridley, and Hearse 1990, Ridley and Curtis 1990a, Ridley, Hearse, and Curtis 1989).

2. MATERIALS AND METHODS

A. Animals and experimental methods

Male Wistar rats (220-260 g, Banting and Kingman) were anaesthetized with diethyl ether, and 200 IU sodium heparin was administered intravenously. After 30 sec each heart was quickly excised and immersed in ice-cold perfusion solution (constituents described below) to induce rapid arrest. The aorta was cannulated for Langendorff perfusion (Langendorff 1895) employing a constant

aortic root pressure of 100 cm H 2 O. An Ag-AgCI wire electrode was inserted into

the left ventricular muscle mass to record an electrogram (ECG) which was recorded against a lead attached to the aortic cannula. This electrode

arrangement gave a clear P wave and ventricular complex (Fig. 1 ). The ECG was continuously displayed on a Gould pen recorder at 5 mm/sec chart speed

and a Gould digital storage type 1421 oscilloscope at 1 0 0 mm/sec sweep speed. Permanent chart recordings at fast speed (100 mm/sec) were taken (i) periodically (for estimating heart rate and ECG analysis), (ii) during the onset of

arrhythmias, and (iii) continuously for the period 1 0 sec prior to the start of reperfusion until 30 sec after the start of reperfusion. A traction-type coronary occluder (Rushmer et al. 1963) consisting of a silk suture threaded through a polyethylene guide cannula was used for coronary occlusion. The suture was 33

positioned loosely around the left main coronary artery according to Heimberger (Heimberger 1946) as modified for in vitro experimentation (Kannengeisser, Lubbe, and Opie 1975). Each heart was allowed to stabilize for 10 min before coronary occlusion. After the designated period of ischaemia, the occluder was released and the heart was reperfused for 3 min.

B. Verification of coronary occlusion and reperfusion

To validate the experiment, proof of occlusion and reperfusion was required for each heart. Two independent methods were used to verify occlusion during regional ischaemia, and to delineate the ischaemic (occluded zone) tissue from the non-ischaemic tissue as described previously (Curtis and Hearse 1989a). First, coronary flow was recorded by timed collection of the coronary effluent and occlusion was verified by comparing flow 1 min before occlusion with flow 1 min after occlusion. Second, at the end of each experiment (3 min after the start of reperfusion) the perfusion solution was replaced by a solution of sulphan blue dye B.P.C. (Disulphine blue, ICI). Each heart received 5 ml of diluted dye (500 mg in 1 litre 0.9% saline) and the coronary occluder was retightened. Perfusion solution was then re-introduced for 3 min, leaving dye trapped in the occluded zone. After excision of the atria and excess mediastinal tissue, the occluded zone was dissected from the non-dyed non-ischaemic zone. The occluded zone was weighed and quantified as a percentage of the total ventricular weight. Reperfusion was also verified by two independent methods as described previously (Curtis and Hearse 1989a). First, uniform staining of the occluded zone by dye was taken to indicate effective reperfusion. Second, coronary flow during reperfusion was measured and compared with pre-occlusion flow. Rat atria receive their blood supply almost exclusively from extra-coronary vessels, whereas the ventricles are supplied by the left and right coronary arteries (Halpern 1957); therefore, coronary flow in the rat Langendorff preparation represents, almost exclusively, ventricular flow. For this reason occluded zone size determined by the dye method, correlates with occluded zone size determined by the flow reduction method (Curtis and Hearse 1989b). For the sake of simplicity we have expressed occluded zone size on the basis of the dye method only. 34

DIGITIZED ECG

MEASUREMENT OF QRSTg0

0 % Repolarization

" " r " 90 % Repolarization 100 % Repolarization

Q R S ^ o

Fig. 1 : Measurement of QRST width at 90% repolarization (QRST90). No separate T wave is seen in the rat ECG so that conventional measurement of QT interval was not practicable. A Ag-AgCI wire electrode was inserted into the left ventricular muscle mass to record an Electrogram (ECG) which was recorded against a lead attached to the aortic cannula. This arrangement gave a clear P wave and ventricular complex. 35

C. Composition of perfusion solutions

Standard chloride-containing solution (constituents in mM: NaCI 118.5, NaHC 0 3

25.0, KCI 4.0, MgS0 4 1 .2 , CaCl2 1.4 and glucose 11.1) was modified to contain chloride:nitrate ratios of 100:0, 75:25, 50:50, 25:75, 0:100 by isotonic

substitution of chloride salts with nitrate salts (KNO 3 , Ca(N0 3 ) 2 and NaNOs). The composition of the other constituents was fixed. Calculated osmolarity was identical in each group. Temperature of the solutions was maintained at 37°C.

The pH was maintained at 7.4 by carbogenation (bubbling with 5% CO 2 , 95%

O2 ).

D. Experimental protocol

Hearts were exposed to 30 min of regional ischaemia for examining the effect of anion manipulation on susceptibility to ischaemia-induced arrhythmias. In the isolated rat heart, early arrhythmias wax and begin to wane during this period (Curtis and Hearse 1989a). Thus 30 min ischaemia allows the investigation of the effects of interventions on early arrhythmias which may play a major role in sudden cardiac death (Campbell, Murray, and Julian 1981). The severity of reperfusion-induced arrhythmias is highly dependent upon the duration of preceding ischemia (Balke et al. 1981, Manning and Hearse 1984). We have therefore used four durations of ischaemia (5, 10, 15 and 30 min) for examining reperfusion arrhythmias (the 30 min group was also used for the study of

ischaemia-induced arrhythmias; see above) thus, a total of 2 0 groups of hearts were used. Each heart was used only once, and 12 hearts were used per group as in previous studies (Curtis and Hearse 1989a). Thus, a total of 240 hearts were used. We used a hierarchical, nested, randomized block design (Armitage and Berry 1987), as described previously (Curtis and Hearse 1989a). Five randomization tables were constructed. Each table contained four 'experimental

units' corresponding to the four ischaemic durations arranged into 1 2 'blocks'. The choice of solution was determined by reference to a sixth randomization table from which two solutions were chosen for each experimental run (typically

8 rats). The experimental operator (P.D.R.) was blinded to the nature of the solution. Records were identified by a number code only, and therefore analysis

| ! i 36 was carried out without knowledge of the solution employed. Codes were broken on completion of analysis.

E. Diagnosis and quantification of arrhythmias

Diagnosis and quantification of arrhythmias conformed with the guidelines of The Lambeth Conventions (Walker et al. 1988). The incidence of VPB, bigeminy (BG), salvos (S), ventricular tachycardia (VT) and VF were recorded. VT was defined as a run of 4 or more VPB. Individual deflections in a run of VT were not included as VPB. VF was defined as ventricular rhythm with no recognisable QRS complex, in which signal morphology changed from cycle to cycle, and for which it was impossible to estimate heart rate. Signal morphology was evaluated from fast chart speed pen recordings in conjunction with an oscilloscope. In order to evaluate the effects of interventions on VF maintenance, the incidence of sustained VF (SVF) was determined. SVF was defined as an episode of VF lasting for longer than 2 min, as described previously (Curtis and Hearse 1989b): 2 min of continous VF is uniformly fatal in rats in vivo, even if defibrillation is subsequently successful (Curtis 1986). It should be noted that VF duration is an inappropriate variable to quantify, since mean ± SEM values are meaningless given that VF can be absent, transient or irreversible.

F. Electroaram (ECG1 analysis

Two measurements were made to analyse the electrogram, (i) the duration between onset of the P wave and onset of the ventricular complex (PR interval) and (ii) the width of the ventricular complex. The T wave is superimposed on the terminal portion of the QRS segment in the rat ECG. Thus conventional measurement of QT interval is impracticable. Therefore the width of the ventricular complex at 90% re polarization (QRSTgo) was measured (Fig. 1 ).

G. Exclusion criteria

In order to ensure that the model functioned appropriately as a bioassay, rigorous exclusion criteria were devised and applied, as recommended (Walker 37

et al. 1988). A total of 268 hearts were used for this study of which 240 were retained. Twenty-eight preparations were excluded for the reasons listed below.

(i) Stability criteria

Unstable preparations were excluded. A stable preparation was defined as having (5 min before occlusion) a sinus rate of at least 290 beats/min and a coronary flow of at least 9 ml/min and an absence of arrhythmias. Stability criteria were not fulfilled by 5 hearts.

(ii) Censoring

t Twenty-three hearts not in sinus rhythm at the moment of reperfusion were excluded from the reperfusion study and replaced, as described previously (Curtis and Hearse 1989b), since it would erstwhile have been impossible to | determine whether ischaemia or reperfusion was responsible for arrhythmias | occurring during reperfusion. The resultant censoring and selection generated | | two subsets of hearts: (i) a subset suitable for the study of ischaemia-induced I arrhythmias (the hearts used in the first attempt to fill each 'cell' of the | randomization table, some of which were not in sinus rhythm at the time of reperfusion); (ii) a subset suitable for the study of reperfusion-induced arrhythmias (the hearts which were in sinus rhythm at the moment of reperfusion).

H. Statistics

Statistical analyses were based on previously published guidelines (Walker et al. 1988, Wallenstein, Zuker, and Fleiss 1980). Gaussian-distributed variables were expressed as mean ± SEM and were subjected to analysis of variance. If treatment constituted a significant source of variance, each group of hearts with differing chlorideinitrate ratios was compared with the control standard chloride-only group using Dunnett's test. Data transformation was necessary to permit analysis of means for variables not Gaussian distributed. The time of onset of reperfusion-induced arrhythmias are not Gaussian distributed and were therefore log-io-transformed as described previously (Curtis and Hearse 1989a, i i 38

Curtis, MacLeod, and Walker 1987, Johnston, MacLeod, and Walker 1983). Only those hearts that exhibited these rhythm disturbances were used in the calculation of means ± SEM values. Group % incidences of arrhythmias were compared using Mainland's contingency tables (Mainland, Harrera, and Sutcliffe 1956). We took p<0.05 to indicate a statistically significant difference.

3. RESULTS

For the sake of brevity ‘nitrate substitution by chloride' has been abbreviated to NCS from here on.

A. ischaemia-induced arrhythmias

Statistically significant concentration-dependent reductions in the incidences of ischaemia-induced arrhythmias were produced by NCS. The incidence of VF was reduced by 50% by perfusion with solution containing a chloridernitrate ratio of 75:25; thus, only a modest NCS produced a large antiarrhythmic effect. Higher proportions of nitrate abolished VF (Fig. 2). The incidences of other ventricular arrhythmias were also reduced in a concentration-dependent manner. If arrhythmias are ranked in a subjective order of severity (VF>VT>S>BG>VPB) it can be seen that nitrate substitution was proportionately more effective against the more severe arrhythmias, with increasingly higher NCS ratios required for inhibition of less severe arrhythmias (Fig. 2).

B. Reperfusion-induced arrhythmias

NCS led to statistically significant concentration-dependent reductions in the incidences of reperfusion-induced VF (Fig. 3) and VT (Fig. 4) in hearts subjected to 10, 15, and 30 min of ischaemia. The incidence of SVF was also reduced by NCS in a concentration-dependent manner. To reveal the statistical significance of the latter effect it was necessary to ‘bin’ data from hearts subjected to different durations of ischaemia, owing to the low incidence of VF with high concentrations of nitrate (Fig. 5). The less severe reperfusion-induced arrhythmias appeared to be increased by NCS (VPB are shown in Fig. 6 ). 39

VF E3 VT S3 Salvo Bigeminy Eg VPB

00 71 CD O 7 c 80 0) TO O 7 c . 60 aj 7 E 40 -£Z 7 x: 20 0 7

100:0 75:25 50:50 25:75 0:100

Ratio of chloride to nitrate

Fig. 2: Ischaemia-induced arrhythmias during 30 min regional ischaemia. Arrhythmia incidences were reduced (p<0.05) in the case of VF (with chloride: nitrate ratios of < 50: 50), VT (with ratios < 25:75), S and BG (0:100 only). VPB incidence was not reduced by nitrate substitution of chloride (NCS). 40

However, this may have resulted in part from ‘unmasking’, by the inhibition of severe arrhythmias (VT, VF), rather than from a genuine pro-arrhythmic effect. Reperfusion-induced arrhythmias in hearts subjected to 5 min of ischaemia were affected differently. First, it should be noted that in all-chloride controls the incidence of arrhythmias was lower than in hearts reperfused after longer durations of ischaemia as expected (Curtis and Hearse 1989a). Second, NCS appeared to have a concentration-dependent proarrhythmic effect. This trend was seen for all classes of arrhythmias including VF (Fig. 3) and VT (Fig. 4) but was most pronounced in the case of the least severe arrhythmias. The proarrhythmic effect was statistically significant only in the case of VPB incidence, which is shown in Fig. 6 . This proarrhythmic effect on VPB incidence cannot be attributed to 'unmasking' via inhibition of severe arrhythmias (VT and VF), since severe arrhythmias were not prevented by NCS (they could not be as they were almost absent in the all-chloride control group). NCS delayed arrhythmia onset; the latency of onset of the first reperfusion-induced arrhythmia was increased in a concentration-dependent manner by NCS in hearts reperfused after 10 or 15 min of ischaemia (Fig. 7). However, in hearts reperfused after 5 min of ischaemia the opposite effect was seen, and NCS accelerated arrhythmia onset (Fig. 7).

C. Heart rate coronary flow and occluded zone size

In order to determine whether the antiarrhythmic effects of NCS were some consequence of altered hemodynamics we measured heart rate and coronary flow throughout the experiment and occluded zone size at the end. Heart rate fell slightly during the course of the experiment in all groups (Fig. 8 ). NCS had little effect on heart rate, and chloride:nitrate ratio did not correlate with heart rate at any stage in the protocol (Fig. 8 ). Not surprisingly, therefore, heart rate did not correlate with arrhythmia incidence or antiarrhythmic activity of NCS. NCS had a time-dependent effect on coronary flow, causing a progressive fall in the non-ischaemic region. This effect was concentration-dependent (Fig. 9). Although the fall in flow occasionally reached statistical significance there was no correlation between flow and arrhythmias (p:NS). 41

o o o o o co o o o in

® ® E:>:v 2 '///✓ 2 ,-tr 2 5 :7 5

25 :7 5 c in . .. E to o o o k_ o ® CD o ® o ■U ID */> cfl O o o ID ID x : ■o JC o CD o CO m 3 55 o ID o H E CM o o CD tr o CCJ 'E o 0)

o 100:0 c/> i— r o o o o o O o o o o o o o CO CD -c j- CM 8 O OO CD ^ CM c CD -JA peonpuj-uojsnpedej ■g j / \ peonpui-uojsnjjedej jo (%) eoueppui o jo (%) eoueppuj c CL3 o <5 LL > 8 c o o '■*—< CD

7 5 2 5 <0 QC -£Z QC c o I

100:0 to CD CL 1— I— I CD o o o o o o o o o o o □c O OO CO r t CM o CO CO CM CO JA peonpuj-uojsnjjedej JA peonpuj-uopnpedej d) jo (%) eoueppuj jo (%) eoueppui shown. are ischaemia regional min 42

o o o o '1— *— ra o

® © 2? 2 2 5 :7 5 2 5 :7 5 c c o o ® o © o ;o in m LO o o o O O m in o s z a V o h— TO Cl o in o in E o CM o ro© CM in to JZ in 10 h- o C'- DC IT c o o E _ o o o o o CO I— I— I— I— I— I o o o o o o O CO CO t* CM

1A peonpui-uojsnpedej 1A paonpuj-uopnjjedej jo (%) eoueppui jo (%) eoueppui

o o o o T— 1— o o

1—I 1 1 1 1 1...... 1 o o o o o o o o o o o o O CO CD CM o CO co ■c* CM 1A peonpuj-uojsnpedej 1A poonpuj-uojsnpedej Fig. 4: Reperfusion-inducecJ ventricular tachycardia (VT). Group incidences in hearts reperfused 5, 10,15 after reperfused 30 hearts or in incidences Group tachycardia (VT). ventricular 4: Reperfusion-inducecJ Fig. jo (%) eoueppui jo (%) eoueppui shown. are ischaemia regional min 43

p<0.05

o U)

100:0 75:25 50:50 25:75

Ratio of chloride to nitrate

Fig. 5: Reperfusion-induced sustained ventricular fibrillation (SVF). 'Binned' group incidences (data for hearts subjected to 5, 10, 15 and 30 min ischaemia having been combined) are shown. 44

o 92 o O o o CO O LO

in in r^ e - in ® in lO CM CM 2 cd 4=r c ca in o o TO o © o © CD d in ;g m TO CO v d d 3 CL in o in t SZ 03 o CL o o CD o o in in CO eg © CM © tr in CC © DC ca in CD E C'- m © SZ © c SZ o CO c o o 8 o o c E 03 o o ■g CO o r ~ r ~ i— i— i— i O O o o m m^ CM o o o o CL o CO CO CM =3 O QdA peonpuj-uojsnpedej QdA peonpuj-uopnjjedej O jo (%) eoueppui jo (%) eoueppui CO CL o > o ca d JOCD CD . □ k _ 3 M "ca © © E ra 2 5 :7 5 ra CD .t; ,*s c c Q. _ □ U k — i5 © © TO !S

5 0 :5 0 o -C XI o o is O O n H— o ______o in > - 5 o o o d © ra ■ § 2 5 DC o ca E v 7 5 :2 5 DC CL t o ca ra© SZ o . □ o d 2 ? c o co to E m t - T I I I I i— i— i— r i i o o o o o o o o o o CC o CO CO CM o 00 CO CM O) CD 9? QdA poonpuj-uopnpedej QdA peonpuj-uopnjjedej jo (%) eoueppui d ) .E jo (%) eoueppui iZ E 45

o o ■*— o as c o LO CD CD cd CD LO . t r L. CM c c o ® 'E O “O o LO o CO s z o OLO S o O o m v o CL LO CO 03 o c - o o CD E o CO o T~ 13 T= i i i i Q) CM O CM CL CM 0) CO CO ^ co k _ (oes) BjtuqiAqjJB uojsnjjedej (oes) BjUJiijXqjjB uojsnjjedej tr jsjij jo eiu;i jesuo cd isjjj |o eoiji lasuo CD SZ

cd E ~ o SZ 0) o y o f s cd o TO E LO © § f LO TO CO CM 2 C CD c T O o d Cl^ ® - ZE x> §5 *c 3 w o t s z CD _i_| o CL 4 1 CD C LO o •“ 03 o PI o *-> CD / LO 0 CM CO LO OC I E 1 / h- in cd o cd CD CD CD V\ o bc _130 y o / o ~ Cd / T“ 0) > c cd T CM O E CM CM1------O1------O1 J------t1- • • CD CO CO co » co cd (oes) BiLuqiXqjJB uojsnjjedej (oes) B!Lui|iAqjJB uojsnjjedej O)■ s O z j s d j jo eoiji lesuo jsjjj jo etuij lesuo Ll .S2 46

Recovery of flow in the reperfused zone was calculated by measuring total coronary flow and expressing the increase occurring during reperfusion as a function of occluded zone size, as described previously (Curtis and Hearse 1989b). In controls, a hyperaemic response was observed in hearts reperfused after 5 or 10 min ischaemia (compare values in Fig. 10 with equivalent values in the non-ischaemic region shown in Fig. 9), with a progressively smaller hyperaemic response in hearts reperfused after 15 or 30 min (Fig. 10). NCS caused a concentration-dependent inhibition of the hyperaemic response, such that recovery of coronary flow was significantly reduced by 1 0 0 % NCS in hearts reperfused after 5 min of ischaemia (Fig. 10). Owing to the decline in hyperaemic response in controls with progressive increments in the duration of ischaemia, the anti-hyperaemic effect of NCS was no longer apparent in hearts reperfused after 15 or 30 min ischaemia (Fig. 10). Despite being statistically significant, the variations in recovery of flow can be regarded as pathophysiologically unimportant with respect to the mechanism by which NCS inhibits reperfusion-induced VF, since there was no correlation between mean flow recovery and VF incidence (p:NS). Furthermore, there was no correlation between recovery of flow and latency to onset of the first reperfusion-induced arrhythmia; although recovery of flow (Fig. 10) and arrhythmia onset (Fig. 7) were both delayed by NCS in hearts reperfused after 1 0 min of ischaemia, the opposite was true in hearts reperfused after 5 min of ischaemia with arrhythmia onset occurring sooner despite a reduction in recovery of flow.

NCS had no effect on occluded zone size (Fig. 1 1 ).

D. Electrogram (ECG1 changes

NCS had a triphasic effect on QRSTgo (Fig. 1 2 a) which was dependent upon when the values were recorded. Before the onset of ischaemia no changes were recorded. During the first minute of ischaemia, width remained constant in the 1 0 0 : 0 chloride:nitrate group, but was shortened in a concentration-dependent manner by NCS. The effect was statistically significant when the 1 0 0 : 0 chloride: nitrate group was compared with the 0 : 1 0 0 chloride:nitrate group (p<0.05).Thereafter there was a marked increase in the width of the interval which was, again,concentration-dependent. Widening was statistically significant throughout the remainder of the experiment when the 47

o < < < < < o o : < < < < i o 1— <<<<<< .<<<<< i o <<<<<< o

CD CD in in ca in in CM c CM c o o CD CD TO T3 o ‘ i_ o in O in O d o sz. in ■6 m o o o o <<<<<<< o £ in <<<<<< In ca <<<<<<< OJ 03 <<<<<< CM ca CD in CC in r r JO n - < < < < < n - LU CO o [ o o o o -H o o c -r— ca T~ CD E CD i— I— r i i r k _ o o o o o o o ca o o o o o o co CO C\J CO CM *- CD uoisnpedej ejojeq ujiu i uojsnpadej ejojeq u;ui [ _=> ca (ujiu/siBeq) eiBi pbbh (ujiu/siBeq) eiBJ jjb0H > d o CO o 3 o o t: T— o CD ■*— Q. d d CD

_ 0) in ca in tr r^- k_ p CD eg ca in c in k_ co CM CM CD o ‘c sz -- o CD CD o P k- O CD o in o in a> d sz d 'k_ in o in o n **—. c o kl— E in V-*o in o CM 03 CM o in or 3 in 03 ca h- k - or tr ca p p CD m d d x o in o 00 T- T“ 6 ) 1— I— I I— I— i— i o o o o o o o o o o o o CO CO CM T- uojsnpedej ejojeq uiiu j. uojsnjjedej ajojeq ujiu i. (ujiu/sjBeq) 0 jbj PB8H (ujiu/s}Baq) eiBJ viBen 48

+1 c cd CD E CD cd 0) CO cd CD -*—* =3 U f a c cd o > CD CD c in p o o p V. o CO v o o Z3 CL x : sz Tr m o % o CD H— CL o o CD ro o o E ’■*—« 4—• O 0 03 03 ro DC t r s z QC cd o 0 CO c CD x r E o CD ■»— o CD n c o in in O in in E CM CM T— 0uoz ojiuegosi-uou jo 0UOZ O!lU0ljOS!-UOU JO "O CD ;m J 9 M 6/ujiu/iiu 1M }0M 6/lNLU/ILU ■o o o CD o *— o c= T— o o CJ) CD o m CD E r - cd CD cd in -*—« sz CM c o o CO CD cz o T3 CD o U) ■O c= o o CD U) x : in XI — o o - t o o cr o V Cl ■ o S in o o ro CM *4-* o 03 4—* *4— Fro in DC 03 ro QC cd sz o c 0 o c o o S _ o E o -O CD in o CD cd CD Z3 C I C) 1 1 b o m in o cd CM i— o in in O CM b 03 euoz o!iu0gos!-uou jo 0UOZ O!LU0ljOS!-UOU JO 2 CJ) LU JM J0M 6/U!lU/|lU JM 10M 6/U!lU/|lU Li. CO 4 9

O O ->— o

__ Cl) ■4—» U) 0 - CO -*—> m OJ c _ o -*—1 Cl) o T3 55 in O o SZ u ; o o LO o O U) O t\J 03 03 v in cr C l E N- ro0) JZ o o CO c o o E i — o T_ T o loin o m o o in in c \ j OJcvj -r- T- auoz pasnjjadaj jo jm jom B/ujuj / iuj auoz pasnjjadaj jo jm ja/vi 6 /ujiu/iiu m o |j Ajbuojoo jo AjaAooay MO|j Ajbuojoo jo AjaAooay

LO o o' v a.

o in in o in in OJ OJ auoz pasnjjadaj jo jm ;a/v\ 6 /uiuu/|uj auoz pasnjjadaj jo jm jaM 6 /ujUJ/|iu Fig. 10: Calculated recovery of coronary flow in the reperfused region recorded 1 min after the Values of reperfusion. start the after min 1 recorded region the reperfused in flow coronary of recovery Calculated 10: Fig. are mean ± SEM wt. wet g ± ml/min/ SEM mean are mo|j Xjbuojoo jo /OaAOoay mo|j Ajbuojoo jo AjaAooay 50

+1 CO c CD CO O E E E CD C c o in o LO t — i — CO E E CO 0) □ S3 0 S3 CO tt= ■o TD o (0 CD C CD CT CO t= > O) "cO CD o £ to A—' CO CO o to o o c c CD 0) > q3 "eg $ o CD ** * CD vPoN -C H CD . N CO % CCD CO o o N o 'k-A—> -o c CO CD CD ■o CL o > 3 o to O o o D) O -*—* C CD o CD jl]6!0m jB|nouju0A ibjoi % ^ £ O) LLI CD Li- CO _Q 51

1 0 0 : 0 chloride: nitrate group was compared with the 0 : 1 0 0 chloride: nitrate group (p<0.05). The groups perfused with intermediate proportions of nitrate exhibited intermediate intervals (which were prolonged vs the 1 0 0 : 0 chloride: nitrate group from 20 min after the start of ischaemia; p<0.05), but it was not possible to demonstrate a graded concentration dependence. Consequently, the incidences of ischaemia-induced arrhythmias did not correlate with interval width measured during peak arrhythmia susceptibility (10 or 15 min after occlusion). It is important to stress that NCS primarily affected interval width only after the tissue had been made ischaemic and that this effect was small before the onset of ischaemia. A small widening of PR interval produced by NCS (significant in the 0:100 chloride: nitrate group at 30 min after occlusion) failed to correlate with ischaemia-induced arrhythmia incidence (Fig. 12b) Assessment of the relationship between ECG shape and reperfusion arrhythmias was complicated by the fact that arrhythmias appeared within a few seconds of the start of reperfusion and were sustained for variable periods; such censoring gave rise to data selection and unequal group sizes for ECG data. By

1 min after the start of reperfusion there was considerable variation in the numbers of hearts in which sinus rhythm had resumed. However, in the case of QRSTgo width, clear differences between groups were evident and values during reperfusion have been included in Fig. 12a. Although reperfusion caused a rapid reversal of QRSTgo widening, values remained significantly increased by NCS compared with pre-ischaemic values (particularly in the case of the 0:100 chloride: nitrate group). Furthermore, QRSTgo width correlated inversely with the incidence of reperfusion-induced VF (p<0.05). In the case of PR interval, values during reperfusion have been omitted from Fig. 12b owing to the large standard errors associated with diminished group size (especially in the all-chloride controls), as explained above. ECG intervals can be expected to be influenced by heart rate. We made no correction for heart rate in the present study; none was required since NCS had no effect on heart rate (Fig. 8 ).

4. DISCUSSION

This is the first report of a potentially new class of antiarrhythmic activity, anion 52

o>o 60 - C k N Q j h“ CO A 100:0 oc o 50 - O 75:25 • 50:50 40 - □ 25:75 ■ 0:100

30 0 5 10 15 20 25 30 Time after occlusion (min)

0 5 10 15 20 25 30

Time after occlusion (min)

Fig. 12: ECG changes in hearts subjected to 30 min of regional ischaemia. Width (msec) was expressed as mean + SEM; (a) QRST 90 (width of the QRST complex measured at 90% repolarization), (b) PR interval. Statistical significance 'stars' have been omitted for clarity. 53 manipulation by NCS. VF, elicited either by ischaemia or by reperfusion could be almost fully abolished by NCS. The effects were not secondary to changes in heart rate, coronary flow or occluded zone size, indicating an action mediated at the cellular level. The precise mechanism of action remains to be determined.

A. Action of NCS on ischaemia-induced arrhythmias

In the isolated rat heart with regional ischaemia it was shown that NCS causes a reduction in the incidence of ischaemia-induced arrhythmias. This effect was concentration-dependent (i.e., proportional to the relative amount of chloride and nitrate) and was most pronounced with the most severe arrhythmias (VT and VF). The ability of VF to sustain once it occurred was also inhibited by NCS, an action previously demonstrated in this model with raised extracellular potassium and with Class III agents (Curtis and Hearse 1989a, Tsushihashi and Curtis 1990). The dependence of the antiarrhythmic effects on changes in antecedent variables was investigated by contiguous assessment of coronary flow and heart rate and subsequent measurement of occluded zone size as these variables play an important role in determining arrhythmogenesis in this model (Bernier, Curtis, and Hearse 1989, Curtis and Hearse 1989a, Curtis and Hearse 1989b, Curtis, MacLeod, and Walker 1987). No relationship between these variables and arrhythmias was found. Therefore, antiarrhythmic activity cannot be attributed to bradycardia, coronary vasodilatation or a reduction in the size of the ischaemic zone. These data indicate that NCS inhibits ischaemia-induced arrhythmias by directly affecting ischaemia-induced electrophysiological dysfunction, not by influencing the underlying causes of this dysfunction, or the timecourse of its development.

B. Action of NCS on reperfusion-induced arrhythmia

DOWNWARD SHIFT Fundamental antiarrhythmic effect on reperfusion- induced arrhythmias

Cl Duration of ischaemia

SHIFT TO THE RIGHT Delay in 100 development of susceptibility to arrhythmias

Q. Duration of ischaemia

SHIFT TO THE LEFT

1 0 0 n Enhancement of development of susceptibility to arrhythmias CL Duration of ischaemia Fig. 13: The 'bell shaped' time-susceptibility curves of reperfusion-induced arrhythmias. Changes in the shape or location of the curve give an indication as to the mechanism of action of an antiarrhythmic intervention. For example, bradycardia shifts the curve to the right, indicative of a delay in development of susceptibility to arrhythmias. Drugs which shift the curve downwards can be regarded as having a specific action on the arrhythmogenic stimulus of reperfusion (most likely an electrophysiological action). In the present case nitrate substitution of chloride (NCS) appeared to shift the curve substantially downward and to the left. 56 electrophysiological action); the new antiarrhythmic R56865 appears to fall into this class (Garner, Bernier, and Hearse 1989). In the present case NCS appeared to shift the curve substantially downward and to the left. There are two ways of interpreting this; either NCS speeds up the time course of development of susceptibility to reperfusion-induced arrhythmias or it has a selective effect on different arrhythmogenic triggers, inhibiting those operative 10 or more minutes after the onset of ischaemia but facilitating those operating with shorter (5 min) durations of ischaemia. Further work is required to elucidate which hypothesis is correct. It should be noted that incidences of the less severe arrhythmias (VPB, BG and S) tended to be increased by NCS in hearts reperfused after 10, 15 and 30 min ischaemia, as well as after 5 min. However, it may not be appropriate to attribute this to a true proarrhythmic effect since it may simply represent unmasking, as a result of amelioration of the more severe arrhythmias (VT and VF).

C. Insight into mechanism of action from changes in the ECG

During ischaemia, NCS widened QRSTgo interval. This js consistent with an observation, made almost 30 years ago, that NCS can prolong action potential duration in papillary muscle (Hutter and Noble 1961). By implication, this would indicate that antiarrhythmic activity may have resulted from an associated prolongation of refractory period. However, this is unlikely to be a causal relationship because (i) drugs which produce significant widening of the QRST interval, such as amiodarone, are not necessarily effective in preventing VF during ischaemia or reperfusion in the rat heart (Riva and Hearse 1989) and (ii) it was recently shown, using a model identical to that of the present study, that tedisamil, a transient outward current blocker (lto ) (Dukes and Morad 1989) has no effect on the incidence of VF in ischaemia or reperfusion (Tsushihashi and Curtis 1990), despite its ability to widen QRST interval in the rat heart to an even greater extent than did NCS in the present study.Therefore we conclude that widening of the QRST and prevention of VF by NCS in the rat heart are casually rather than causally related and that the antifibrillatory effects of NCS results from some mechanism unrelated to its ‘Class III effect’ on QRST. On the other hand, QRST widening may have played a role in the 57

defibrillatory effect of NCS (its ability to reduce the incidence of SVF) since amiodarone (Riva and Hearse 1989) and tedisamil (Tsushihashi and Curtis 1990) both possess defibrillatory activity in the rat heart. The mechanism by which NCS mimics the ECG effects of amiodarone and tedisamil remains to be determined. It is of interest to note, however, that these ECG effects occurred primarily during ischaemia, and no significant change in configuration was present prior to ischaemia onset. Changes in the PR interval were slight and unrelated to the antiarrhythmic effects. Since PR interval is prolonged by drugs which inhibit the fast and slow inward current, this would appear to rule out inhibition of these currents as the mechanism of antiarrhythmic action of NCS.

D. Cellular mechanism of action: hypotheses

The present study does not permit a direct assessment of the cellular mechanism of action of NCS. However, some possibilities warrant discussion. First, NCS can be expected to increase Gm. Although chloride conductance is an order of magnitude lower than potassium conductance in ventricular tissue (Fozzard and Lee 1976), NCS will increase Gm as a consequence of the permeability difference between the anions (Hutter and Noble 1961, Seyama 1979). The resultant fall in excitability would disfavour propagation of premature ectopic activation. To test this hypothesis we have examined the antiarrhythmic effects of chloride surrogates with differing membrane permeabilities (Study 3.1) Second, nitrate may elevate intracellular cGMP (Katsuki, Arnold, and Murad 1977). The effects of such a change on arrhythmogenesis are unknown. However since cAMP appears to play a role as a trigger for arrhythmias, especially during reperfusion (Podzuweit et al. 1989), elevations of intracellular cGMP may conceivably have an opposite effect. This hypothesis was studied by measuring cyclic nucleotides in ischaemic and nonischaemic myocardium (Study 3.2). Third, NCS may inhibit chloride-bicarbonate exchange. Recent studies have shown that SITS can inhibit both ischaemia- and reperfusion-induced arrhythmias in the isolated rat heart (Curtis 1989). This action occurs at a concentration effective in inhibiting chloride-bicarbonate exchange, but not at lower concentrations. Inhibition of chloride-bicarbonate exchange will elevate

[ 58 intracellular pH (Vaughan-Jones 1979b), opposing the fall in intracellular pH subsequently occurring during ischaemia. This fall in pH is closely associated with development of susceptibility to arrhythmias (Hirche et al. 1980). A fall in intracellular pH shortens action potential duration (Sato et al. 1985). Therefore evidence that NCS widens QRST (this study) and prolongs action potential duration (Hutter and Noble 1961) is consistent with the hypothesis that blockade of chloride-bicarbonate exchange may play a role in the antiarrhythmic effects of NCS. To test this hypothesis we measured changes in intracellular pH in beating hearts by phosphate NMR spectroscopy (Study 3.3). 59

STUDY 2

INVESTIGATION OF THE SITE OF ANTIARRHYTHMIC ACTION OF NCS 60

STUDY 2.1

CROSS-OVER STUDY IN REGIONAL ISCHAEMIA: SITE OF ANTIARRHYTHMIC ACTION OF NCS

1. INTRODUCTION

Although NCS has been shown to have impressive antiarrhythmic action against ischaemia-induced arrhythmias, no insight was obtained from the initial study (Study 1) about the site of action of NCS. Identification of the site of action may assist to elucidate the mechanism of action.The effect on ischaemia-induced arrhythmias might be mediated by an action on the ischaemic or the non-ischaemic tissue. This question was investigated by crossover experiments which permitted selective manipulation of the extracellular milieu before and during ischaemia. This was achieved by varying the composition of the solution delivered to the zone to become occluded and the non-occluded zone in the same heart. A brief report of this work has already been published (Ridley and Curtis 1990b).

2. MATERIALS AND METHODS

Experimental methods, verification of coronary occlusion, diagnosis and quantification of arrhythmias, ECG analysis and statistical analysis were as described in Study 1.

A. Composition of perfusion solutions

Two solutions were employed in this study, i) standard chloride-containing solution as described in Study 1 with a chloride:nitrate ratio of 100:0, and ii) solution modified by complete NCS to produce a chloride:nitrate ratio of 0:100.

B. Experimental protocol

C. Exclusion Criteria

One heart was excluded as it failed to meet the stability criteria described in Study 1.

3. RESULTS

In hearts with chloride trapped in the occluded zone and nitrate delivered to the non-ischemic zone, the incidences of VT and VF were reduced to 17% (p< 0.05) and 0% (p<0.05) from control (chloride throughout) values of 100% and 67% respectively. This implied that the antiarrhythmic effect was mediated via an action in the non-ischaemic zone. However, in hearts with nitrate trapped in the occluded zone and chloride delivered to the non-ischaemic zone VF incidence was also reduced to 17% (p<0.05) although the incidence of VT was unchanged (100%). Thus, in the case of VF, NCS appeared to produce its antiarrhythmic effect partly as a consequence of an action in the ischaemic tissue and partly by an action in the non-ischaemic tissue. On the contrary, inhibition of VT appeared to result exclusively from an action on the non-ischaemic zone. Of the less 'severe' classes of arrhythmias, the incidence of VPB was high in all groups but the incidence of BG and S followed a similar pattern to VT being significantly reduced only in groups (3 and 4) in which nitrate was delivered to the non-occluded zone (Table 1).

4. DISCUSSION

Partial inhibition of ischaemia-induced VF occurred both in hearts with entrapment of nitrate in the occluded zone and in hearts with delivery of nitrate to the non-ischaemic zone. From this study, it would appear that the site of action of NCS against ischaemia-induced VF is in part the ischaemic tissue and in part the surrounding non-ischaemic tissue. By contrast, this initial study suggested that other arrhythmias (BG, S and VT) were inhibited only if NCS was delivered to the non-ischaemic zone. The approach employed in this study relied upon the known low coronary collateral flow (Maxwell, Hearse, and Yellon 1978, Winkler et al. 1984) which allows reproducible high incidences of ischaemia- and reperfusion-induced arrhythmias in the rat heart. However, 63

Group NZ OZ RRRHVTHMIR

UPB BG S UT UF

1 Cl- Cl- 100 100 92 100 67

2 Cl- N03- 100 100 100 100 17*

3 N03- Cl- 83 42* 25* 17* 0*

4 N03- N03- 100 25* 0* 25* 0*

Table 1: Incidences (%) of ischaemia-induced arrhythmias in the 30 min regional ischaemia cross-over study. See text for experimental protocol for groups 1-4 (n=12/group). Abbreviations: Cl- = chloride-containing solution, N03- = nitrate-containing solution, NZ = non-occluded zone, OZ = occluded zone, VPB = ventricular premature beat, BG = bigeminy, S = salvo, VT = ventricular tachycardia, VF = ventricular fibrillation. * indicates p< 0.05 compared to group 1 (chloride-containing solution throughout). 64 although low, there is some collateral flow in the rat heart; it has been estimated as 6.1 ± 0.7 % of flow in the normal zone by Maxwell (Maxwell, Hearse, and Yellon 1978) and less than 0.01 ml/min/g by Winkler (Winkler et al. 1984). Significant collateral flow would be expected to wash nitrate out of the ischaemic region and might lead to underestimation of its effect in the ischaemic region. This problem was overcome with a model of global ischaemia (Study 2.2). STUDY 2.2

A NEW MODEL OF GLOBAL ISCHAEMIA: APPLICATION TO THE STUDY OF SYNCYTIAL MECHANISMS OF ARRHYTHMOGENES1S AND SITE OF ANTIARRHYTHMIC ACTION OF NCS

1. INTRODUCTION

Consideration of the relationship between susceptibility to arrhythmias and the amount of tissue subject to injury (e.g. by ischaemia or reperfusion) has been employed previously to gain insight into the syncytial mechanisms responsible for arrhythmogenesis in particular settings; recent studies have shown that the induction of VF in hearts reperfused after a short (10 min) period of ischaemia peaks when approximately 40% of the ventricle is involved, and that susceptibility is maintained if occluded zone size is increased to 100% of total ventricular weight (Curtis and Hearse 1989b). This finding was interpreted to indicate that an interaction between reperfused and adjacent uninvolved tissue is unnecessary for arrhythmogenesis in reperfusion after short periods of ischaemia, since with 100% (global) ischaemia there is no uninvolved region. This indicates that the injury current model of arrhythmogenesis favoured in acute regional ischaemia (Janse et al. 1979, Janse et al. 1980), need not be invoked as a mechanism of arrhythmogenesis in reperfusion. In the present study, we have focused our attention on arrhythmogenesis in early ischaemia and during reperfusion after more sustained periods of ischaemia. If Janse’s injury current hypothesis applies, global ischaemia (in which scope for injury current between ischaemic and non-ischaemic zones is minimized) should be associated with a low susceptibility to arrhythmias. The objective, therefore, was to examine the relationship between incidence of ischaemia-induced arrhythmias and occluded zone size over a range including 100% of total ventricular weight (global ischaemia). Global ischaemia rapidly causes sinus bradycardia, atrioventricular (AV) block and asystole (Curtis and Hearse 1989b). To achieve the objective we required a model of global ischaemia in which asystole could be circumvented. This is because asystole both slows the rate of development of susceptibility to arrhythmias (Bernier, Curtis, and Hearse 1989) and renders ventricles quiescent 66 precluding both re-entry and flow of injury current from occurring and operating as arrhythmogenic mechanisms. Conceivably, one might obviate asystole by pacing ventricles. However the technique is inadvisable since it creates an ectopic focus in the ventricle which, in ischaemia and reperfusion, is intrinsically arrhythmogenic (Nakata, Hearse, and Curtis 1990) rendering interpretation of data impossible. An alternative approach, pacing of atria, is inadvisable as a means of maintaining normal ventricular beating rate during global ischaemia, since ischaemia prolongs AV nodal refractory period; pacing of atria during global ischaemia at normal sinus rate converts AV block to complete AV dissociation (Curtis, unpublished observation). These technical considerations have hitherto confounded all attempts to develop a model of complete (no flow) global ischaemia for investigation of arrhythmogenesis, rendering it impossible to test the injury current model by the technique of relating arrhythmogenesis to occluded zone size. In the present study an attempt was made to overcome the problem of asystole during ischaemia by right atrial intracavity superfusion. This was found to maintain sinus rhythm and ventricular beating rate within the normal range and permitted an examination of the relationship between arrhythmogenesis and occluded zone size over the range 0-100% of the total ventricular weight. In previous studies a variant of the new model was used to examine reperfusion-induced arrhythmias after a brief period of ischaemia (10 min) (Curtis and Hearse 1989b). However, reperfusion-induced arrhythmias are known to show marked dependence on the duration of preceding ischaemia (Manning and Hearse 1984) and it has been inferred that distinct mechanisms of arrhythmogenesis operate after brief versus sustained ischaemia (Hearse and Tosaki 1988, Tosaki and Hearse 1987). Thus, we have also used the new model to examine the relationship between occluded zone and incidence of reperfusion-induced arrhythmias in hearts reperfused after a sustained (30 min) period of ischaemia. Finally, the global ischaemia model was employed to examine the antiarrhythmic site of action of NCS and compared to the site of action indicated by cross-over studies of regional ischaemia. The value of the new model in determining aspects of the mechanism of antiarrhythmic action of anion manipulation are discussed. 67

2. ANIMALS AND METHODS

Experimental methods, verification of coronary occlusion and reperfusion, diagnosis and quantification of arrhythmias, ECG analysis and statistical analysis were as described in Study 1 except the weight of the rats was 250-300g and they were anaesthetized with halothane. Halothane does not affect the incidence of ventricular arrhythmias in the Langendorff perfused rat heart (MacLeod et al. 1989). The method of producing regional ischaemia was as described in Study 1.

A. Global ischemia: a new model

A modification of the conventional no-flow global ischaemia model was used. Each heart was allowed to stabilize during 10 min of control perfusion. During this period the heart was prepared for right atrial intracavity superfusion. A small hole was carefully made in the right atrium away from the region of the sinoatrial node; this was achieved by slightly enlarging the inferior vena caval orifice with a single fine scissor cut. A fine polyethylene cannula fashioned from a size 12FG nasogastric feeding tube, cut to a convenient length (approximately 5cm) and attached to a side arm of a standard constant-pressure Langendorff perfusion apparatus, was gently positioned within the right atrium and attached proximally to a bar on the perfusion rig to avoid displacement. This procedure could be performed quickly and manipulation of the heart was kept to a minimum. Fine toothed forceps were minimally traumatic when manipulating the heart and were preferred to non-toothed forceps. Following the 10 min period of control perfusion, the right atrial cannula was opened to allow right atrial intracavity superfusion to begin. Simultaneously the aortic cannula was cross-clamped to terminated aortic flow and induce no-flow global ischaemia.The procedure of right atrial intracavity superfusion was designed to maintain normal sinus rate and prevent AV block. At no time was pressure allowed to develop within the right atrium and the solution drained readily around the cannula from the dependent hole at the site of the inferior vena cava. The superfusion flow could be controlled with a 3-way tap. In preliminary studies, the end point in determining this was heart rate during global ischaemia. Typically an adequate flow was approximately 15 ml/min, but this 68 was not critical; within limits, the rate in each heart could be titrated against the flow. After 30 min global ischaemia, the right atrial cannula was gently removed, great care being taken to avoid disturbing the preparation, and the aortic root cannula declamped to permit reperfusion.

B. Composition of perfusion solutions

Standard chloride-containing solution was used in four groups of hearts. In one group the perfusing solution was modified by complete isotonic substitution of chloride salts with nitrate salts.

C. Experimental protocol

Hearts (n=12/ group) were randomized to groups. Regional ischaemia to produce large and small occluded zones was achieved by deliberate high or low ligation of the left coronary artery respectively as described previously (Curtis and Hearse 1989b). In a third group a suture was passed around the left coronary artery but was not tightened (sham ligation). In a fourth group hearts were subjected to global ischaemia according to the technique described above. Standard chloride-containing solution was used in these four groups. The antiarrhythmic effect of NCS was also studied by the inclusion of a fifth group with global ischaemia in which perfusion and reperfusion employed a solution with complete NCS. Coronary flows in ml/min/g ventricular weight of perfused tissue (total ventricular tissue prior to ischaemia, the uninvolved tissue, if any, during ischaemia) and of the reperfused tissue during reperfusion together with heart rate were recorded in all experiments.

D. Exclusion criteria i) Stability criteria

Stability criteria, as described in Study 1, were not fullfilled by 5 hearts which were replaced immediately. 69

ii) Ventricular rate

During global ischemia, any heart with a ventricular rate of less than 150 beats/min was excluded. Previous observations show that this rate is a sine qua non for the occurrence of reperfusion-induced VF (Curtis and Hearse 1989b). This led to exclusion and replacement of 1 heart.

iii) Censoring

Three hearts not in sinus rhythm at the moment of reperfusion were excluded I from the reperfusion study and replaced, as described in Study 1. |

3. RESULTS

A. Occluded zone sizes

Proximal left and distal left coronary occlusion produced occluded zones of 47 ± 1.0% and 21 ± 0.8 % respectively, with values in sham ligation and global | ischaemia being 0% and 100%. Perfusion with sulphan blue dye at the end of t each experiment revealed occluded zones to be transmural and uniformly stained.

B. Ischaemia-induced arrhythmias

Incidences of VT and VF were critically dependent on occluded zone size. Susceptibility to VT increased to a maximum with proximal left coronary occlusion and the maximum persisted in globally ischaemic hearts (Fig.14). Also, susceptibility to VF peaked with proximal left coronary occlusion. However, peak susceptibility did not persist with global ischaemia but declined to a level equivalent to that in the distal occlusion group (Fig. 14). As the incidence of ischaemia-induced VPBs during the 30 min period of ischaemia was high (100%) in all groups except the sham group, we analysed the incidence of VPBs in consecutive 5 min time periods during the entire 30 min period of ischaemia in order to examine the differences between groups. This showed susceptibility 70 to VPBs increased to a maximum with proximal left coronary occlusion and that this persisted during global ischaemia (Fig. 15). Thus, occluded zone size influenced VPB and VT in a similar manner. Both VPB and VT differed from VF in showing no reduction in incidence with global ischaemia.

C. Reperfusion-induced arrhythmias

Susceptibility to reperfusion-induced VT was related to occluded zone size in a qualitatively similar manner to susceptibility to ischaemia-induced VT (compare Fig. 14 to Fig. 16), with no decline in maximal susceptibility, once reached, with increasing occluded zone size. During reperfusion the relation between occluded zone size and susceptibility to VF was qualitatively different from the situation during ischaemia since susceptibility to VF did not decline in hearts reperfused after global ischaemia (Fig. 16); 92% of hearts developed VF when reperfused after proximal left occlusion and 92% did the same when reperfused after global ischaemia.

D. Antiarrhythmic action of NCS

We compared hearts perfused with standard chloride-containing solution with hearts perfused with nitrate in place of chloride using the new global ischaemia model. The incidence of ischaemia-induced VT was reduced from 100% to 58% (p<0.05) and VF from 17% to 0% (p:NS) by NCS. The incidence of VPB was analysed in consecutive 5 min periods throughout the 30 min period of global ischaemia. This revealed a reduction in susceptibility to VPB in the nitrate group compared with the chloride group (Fig. 17). Reperfusion-induced VT was reduced from 100% to 58% (p <0.05) and VF from 92% to 0% (p< 0.05) by NCS.

E. Haemodynamics

Heart rates are shown in Table 2. There were no significant differences between the distal and proximal occlusion, global ischaemia, global ischaemia nitrate and time-matched sham-ligated controls at any time during the experiment. Also, AV block did not occur, even in the global ischaemia groups. Changes in 71

100

03 'E 60 -E CO mo— 40

OCD c CD - o 20 ‘O c

0 20 40 60 80 100 Occluded zone size

Fig 14: Relationship of % incidence of ischaemia-induced ventricular tachycardia (VT) and ventricular fibrillation (VF) to occluded zone size (% of total ventricular weight). 72

■Cro OS o»— — o (/) _ j X o

S**i*KiK*K*K*KtiK*KK*K«K*«^ o " ' """* ' '""' '

R5S^ CD in “ Q> o o o o o o 00 VO

100

Occluded zone size

Fig. 16: Relationship of the % incidence of reperfusion-induced ventricular tachycardia (VT) and ventricular fibrillation (VF) to occluded zone size (% of total ventricular weight). 74

od * + - • cd o _ Q CO jQ O) O) c g 'g <\j ro E iS XJo oJ_> c£> O 8 § f g 10 ■°CO Io P CD O P ~ *v . —j tw CD -g O - E C .E iS -d E co-| 0, CO £ **— ° d T3 O Q) c 8 W O •E-C? EPQ_ 8.S co > >».£> O £ w —xo .EF d ^ ^<0-0 m P $ O dj § ® "8 po|Jdd duuu u|uu g ipea -II ui QdA (50 aouappui N ’■ CO o .g > o 2 L L P O 75

Occluded 1 min before 1 min after 1 min before zone size ischaemia ischaemia reperfusion

0% 301.3 ± 2.1 299.3 + 2.5 278.7 ± 7.1

21 ± 0.8% 298.7 ± 4.7 280.8 ± 7.8 254.0 ± 4.5

47 ± 1.0% 302.7 ± 4.5 277.3 ± 5.4 269.3 ± 12.4

100% 299.2 ± 7.1 261.3 ±13.7 247.1 ± 14.3

*100% 315.2 ± 5.8 282.7 ± 11.1 231.7 ±11.3

Table 2: Heart rate (beats/ min). Groups are identified by their occluded zone size expressed as % of total ventricular weight. (* indicates nitrate perfusion). There were no significant differences between groups before or during ischaemia or during reperfusion. 76 coronary flow with regional ischaemia in the isolated rat heart have been examined in detail previously (Study 1); changes observed in the present study were equivalent to those previously reported and have been omitted here for the sake of brevity. In the global ischaemia groups coronary flow during ischaemia was set at zero. Values 1 min after reperfusion were 7.3 ± 0.8 ml/min/g ventricular weight in the chloride group and 7.1 ± 0.6 ml/min/g in the nitrate group (p:NS). The onset of the first reperfusion arrhythmia was delayed in the global nitrate vs. the global chloride group (1.42 ± 0.1 logio sec and 1.24 ± 0.17 logio sec respectively, p< 0.05). As noted in previous studies (Curtis and Hearse 1989b), incidence of reperfusion-induced arrhythmias and onset time of first reperfusion arrhythmia did not correlate with recovery of coronary flow on reperfusion. Electrical activity was usually well maintained during 30 min global ischaemia but observation of the heart, as expected, revealed a progressive loss of contractile activity during this time.

4. DLSCUS51QM

Global ischaemia rapidly produces asystole in isolated rat heart preparations making study of ischaemia-induced arrhythmias impracticable. A recent model of global ischaemia in which ventricular rate is maintained by atrial superfusion with warm perfusion solution is a considerable improvement (Curtis and Hearse 1989b). This mimics the pulmonary effluent which superfuses the sinoatrial node in the conventional rat Langendorff preparation. It provides a reliable model for the study of reperfusion-induced arrhythmias so long as the heart rate remains above 150 beats/min during global ischaemia. Unfortunately, it is difficult to maintain ventricular rate close to normal for long periods with this model; the heart tends to develop complete heart block followed by absence of ventricular contractions. As a result it is seldom possible to use the model for periods longer than 10 min.

A. A new model of global ischaemia

We have described a new model of global ischaemia, employing right atrial intracavity superfusion to maintain normal sinus rate and 1:1 AV conduction 77 during ischaemia. It was designed to be simple to use and suitable for studies of ischaemia-induced arrhythmias. The latter was achieved by obviating the requirement for ventricular pacing to maintain normal ventricular beat rate during ischaemia; ventricular pacing is intrinsically arrhythmogenic during ischaemia and renders interpretation of data impossible (Nakata, Hearse, and Curtis 1990). The new model is simple and practicable to set up with small modifications to standard perfusion apparatus. Although attention to detail is required, exceptional dexterity is not necessary in positioning the right atrial intracavity cannula.

B. Occluded zone size and ischaemia-induced arrhythmias; implications for ‘injury current’ theory of arrhythmogenesis

A graded increase in the incidence of ischaemia-induced VF occurring in the sham, low occlusion and high occlusion groups is consistent with the mechanism of initiation of arrhythmogenesis being dependent on the extent of ischaemic involvement of ventricular muscle mass (Austin et al. 1982, Bolli, Fisher, and Entmann 1986, Curtis, MacLeod, and Walker 1987, Endo et al. 1983). However, a decline in VF incidence with global ischaemia was observed. This was not secondary to bradycardia or AV block, since the model precluded such dysfunction from developing (as intended). This decline in VF susceptibility strongly supports the injury current hypothesis of arrhythmogenesis in ischaemia (Austin et al. 1982, Janse et al. 1980). In contrast, a persistence of a high incidence of VPB and VT with global ischaemia argues against involvement of injury current between the ischaemic and non-ischaemic zones in the initiation of these arrhythmias. Studies of regional ischaemia in the dog (Bolli, Fisher, and Entmann 1986) report that the incidence of organized ventricular ectopy (VPB and VT) exhibits only a weak correlation with the occluded bed size during regional ischaemia, in contrast to VF. This and our data suggests that the mechanism of VPB and VT initiation differs from that of VF. In the dog the importance of collateral flow has been emphasized as a covariant determinant of arrhythmogenesis in regional ischaemia (Bolli, Fisher, and Entmann 1986). An attractive feature of the rat heart for the study of arrhythmias is its paucity of coronary collaterals (Maxwell, Hearse, and Yellon 78

1978) which allows easier consideration of ischaemic zone size as a determinant of arrhythmogenesis. The present study indicates a graded relationship between susceptibility to VPB and VT and occluded zone size. In the dog (Bolli, Fisher, and Entmann 1986), an increase in the occluded zone size above 40% results in a very marked increase in susceptibility to VT. This was not observed in the rat owing to the high incidence of VT when occluded zone size was less than 40%. This difference in 'potency' of occluded zone to ellicit VT may reflect a lack of protection by collateral flow in the rat vs. the dog (Maxwell, Hearse, and Yellon 1978). The VF which occurred during global ischaemia (in 2 hearts) was sustained for longer than 2 minutes suggesting that although global ischaemia reduces the incidence of VF, it does not facilitate spontaneous defibrillation. The fact that a low incidence (17%) of ischaemia-induced VF was seen in the global ischaemia group suggests that a second injury current-independent arrhythmogenic mechanism (re-entry or abnormal automaticity) may function during early ischaemia, albeit with considerably less importance than the injury current-dependent mechanism. Alternatively, varying degrees of workload might result in patchy areas within the ischaemic ventricular tissue with severely ischaemic areas adjacent to less ischaemic areas (Steenbergen et al. 1977) which might give rise to occasional flow of injury current.

C. Reperfusion arrhythmias

VF and VT increased in incidence in proportion to the size of the reperfused zone. The incidence of VF and VT had a saturating (rather than bell-shaped) relationship with reperfused zone size and did not diminish in hearts reperfused following global ischaemia. Thus, reperfusion-induced VF and VT are initiated from within the reperfused tissue rather than as a consequence of an interaction between that tissue and adjacent uninvolved tissue (e.g., via flow of injury current between these regions). Previously, similar findings in hearts reperfused after a brief (10 min) period of ischaemia led to similar conclusions (Curtis and Hearse 1989b). Thus, the syncytial mechanism responsible for VF in hearts reperfused after 30 min ischaemia may be the same as the mechanism operating in hearts reperfused after 10 min ischaemia. The underlying pathophysiological mechanisms may however, vary according to the duration of 79 preceding ischaemia (Hearse and Tosaki 1988, Tosaki and Hearse 1987).

D. Application of the model to NCS

Previous studies using coronary ligation (regional ischaemia) have shown that NCS in perfusion solution ameliorates both ischaemia- and reperfusion-induced arrhythmias (Study 1). Thus far it has not been possible to test NCS in globally ischaemic rat hearts in order to better assess the ischaemic zone as a possible site of antiarrhythmic action. In addition, absolute proof of uninvolvement of improvement of collateral flow in the antiarrhythmic effect has hitherto been lacking. It remains possible, although unlikely, that incrementation of collateral flow may have played some role in the beneficial effects of nitrate in previous studies. One obvious advantage of the present model is that zero flow ischaemia precludes all possibility of any antiarrhythmic effect being mediated by incrementation of collateral flow. In addition, it can be argued that any benefit with an intervention is independent of attenuation of flow of injury current between ischaemic and adjacent uninvolved tissue, since in our model, there is no uninvolved tissue. In view of this we applied the model to the study of the mechanism of antiarrhythmic action of NCS. We found this intervention to be effective against both ischaemia- and reperfusion-induced VT and reperfusion-induced VF in the new model. The findings were similar to those from the regional ischaemia and reperfusion study (Study 1). We deduce, therefore, that an action within the ischaemic tissue contributes to the effect of NCS on ischaemia-induced VT. This action is independent of any modification of flow of injury current between ischaemic and uninvolved tissue, and is not a consequence of incrementation of collateral flow. This finding strongly suggests that the lack of apparent effect of NCS via the ischaemic zone in cross-over studies of regional ischaemia (Study 2.1) might have been due to wash out of nitrate from the ischaemic zone by a small, but significant, collateral flow. The model did not permit assessment of effects of NCS on ischaemia-induced VF since control incidences of VF were low. Our data suggests that the presence of non-ischaemic as well as ischaemic tissue is necessary in most cases for ischaemia to elicit VF (in contrast to VT) and thus 80 regional ischaemia models would be preferred for the study of antifibrillatory interventions such as NCS. Nevertheless, abolition of ischaemia-induced VF from a control incidence of 17% and abolition of reperfusion-induced VF from a control incidence of 92% (p< 0.05) by NCS is consistent with an antifibrillatory property mediated by a mechanism independent of injury current or augmentation of collateral flow 81

STUDY 2.3

CROSS-OVER STUDIES IN THE GLOBALLY ISCHAEMIC HEART: ANTIARRHYTHMIC ACTION OF NCS DURING REPERFUSION

1. INTRODUCTION

Although substantial antiarrhythmic action was demonstrated for NCS on reperfusion-induced arrhythmias, no insight was obtained from the initial study (Study 1) about the site of action of NCS. The reduction of reperfusion-induced arrhythmias might be mediated by an action during reperfusion (direct action) or from an action during ischaemia (indirect action). Therefore, this question was investigated by a cross-over experiment which selectively manipulated the extracellular milieu before ischaemia and upon reperfusion using a global ischaemia model. A brief report of this work has already been published (Ridley and Curtis 1990b).

2. MATERIALS AND METHODS

Diagnosis and quantification of arrhythmias, ECG analysis and statistical analysis were as described in Study 1. The rig was modified for cross-over perfusion as described in Study 2.1.

A. Animals and experimental methods

The arterial supply of the rat ventricle has been shown to possess few functional collaterals. However, as discussed above (Studies 2.1 and 2.2) even a small amount of collateral flow might cause wash out of solution trapped in the ischaemic zone by the solution delivered to the non-ischaemic zone. In order to overcome this potential problem, we utilized a model of global ischaemia to study the site of antiarrhythmic action of NCS on reperfusion-induced arrhythmias. As relatively short periods (10 min) of global ischaemia were required the established model of Curtis and Hearse (Curtis and Hearse 1989b) rather than the modified model (Study 2.2) was employed for this study. Global ischaemia 82 was achieved by occluding the aortic inflow line. In order to prevent sinus bradycardia during global ischaemia, heart rate was maintained by superfusing the atria (not the right atrial cavity) with the perfusion solution at 37°C (Curtis and Hearse 1989b). Sinoatrial superfusion mimics the situation during Langendorff perfusion and regional ischaemia in which the sinoatrial node is superfused by the pulmonary artery effluent.

B. Composition of perfusion solutions

Two solutions were employed in this study, i) standard chloride-containing solution as described in Study 1, and ii) solution modified by complete NCS c. Experimental protocol

Four groups of 12 rats were studied. In the first group, chloride-containing solution was delivered during a 10 min control period which was followed by 10 min of global ischaemia with continuous atrial superfusion. Each heart was then reperfused with chloride-containing solution. Thus, in group A chloride-containing solution was trapped in the ischaemic tissue and also delivered during reperfusion. In group B, chloride-containing solution was employed for the initial 10 min control perfusion and was trapped in the heart during global ischaemia, whereas nitrate-containing solution was delivered during reperfusion. In group C, nitrate-containing solution was delivered during the initial 10 min control perfusion and trapped in the heart during global ischaemia, whereas chloride-containing solution was delivered during reperfusion. In group D, nitrate-containing solution was delivered during control perfusion (and trapped in the ischaemic tissue) and also during reperfusion. Hearts were randomized to treatment and analysis of records was carried out blind.

D. Exclusion Criteria

Exclusion criteria were as described in Study 1. In addition, in the global ischaemia groups any heart with a ventricular rate less than 150 beats/min 9 min after the beginning of global ischaemia and 1 min before reperfusion was 83 excluded in accordance with previous observations in this model in which it was found that in control hearts this criterion was a sine qua non for occurrence of reperfusion-induced VF (Curtis and Hearse 1989b). This led to exclusion and replacement of 40 hearts.

3. RESULTS

There were no significant differences betwen the four groups in the incidence of reperfusion-induced VT (Table 3). In hearts perfused before ischaemia with chloride-containing solution and reperfused with this solution, the incidence was 100%, and in hearts perfused and reperfused with nitrate-containing solution the value was 83%. Similarly values in hearts perfused with chloride and reperfused with nitrate or vice versa, were 75% and 92% respectively. However, the incidence of VF was reduced by NCS (Table 3). Furthermore this required that nitrate be present during reperfusion. The VF incidence in the chloride perfusion, chloride reperfusion group was 92% and was. significantly reduced to 8% (p<0.05) in the nitrate perfusion, nitrate reperfusion group. A significant reduction also occurred in the chloride perfusion, nitrate reperfusion group, to 50% (p<0.05), but in the group perfused with nitrate and reperfused with chloride there was no significant reduction in VF incidence (75%). The effects on the incidences of minor arrhythmias (VPB, BG and salvos) are difficult to interpret because of the masking effect of the major arrhythmias (VT, VF).

4. DISCUSSION

In cross-over studies performed to assess whether reperfusion-induced arrhythmias were inhibited by an action of NCS during reperfusion or an action during the preceding period of ischaemia, we simplified experimental design by using an established model of global ischaemia and employed a relatively short (10 min) period of global ischaemia. We were able to do this because reperfusion arrhythmias arise from within the reperfused zone rather than from an interaction between it and the adjacent non-ischaemic zone (Curtis and Hearse 1989b). We found that inhibition of reperfusion-induced arrhythmias required that nitrate be present during reperfusion. This would imply a specific antiarrhythmic action on the arrhythmogenic triggers operating during 84

Group IZ RZ RRRHYTHMIR

UT OF

n Cl- Cl- 100 92

B Cl- N03- 75 50*

C N03- Cl- 92 75

D N03- N03- 83 8*

POSSIBLE CELLULAR MECHANISMS OF ACTION OF ANION

SUBSTITUTION ON ARRHYTHMIAS 87

STUDY 3.1

MODIFICATION OF MEMBRANE RESISTANCE

1. INTRODUCTION

In this study we have tested the hypothesis that anion substitutions inhibit ventricular arrhythmias by increasing membrane conductance due to their increased membrane permeability relative to chloride. This was achieved by isotonic replacement of the major extracellular anion (chloride) in perfusion media with a surrogate anion. The relative cardiac sarcolemmal membrane permeabilities of the anions studied are nitrate > iodide > bromide > chloride > I methylsulphate (Seyama 1979). A brief report of this work has already been published (Ridley, Yacoub, and Curtis 1990).

2. MATERIALS AND METHODS

In this study male Wistar rats (250-300 g, Harlan Olac, Bicester) were anaesthetized with halothane before excision of the heart for Langendorff perfusion.

B. Composition of perfusion solutions

Standard chloride-containing solution was employed in a control group and in other groups it was modified by complete isotonic substitution of chloride with one of four anion surrogates (nitrate, iodide, bromide, methylsulphate).

C. Experimental protocol 88

A different anion was used for each of 5 groups of hearts. Each heart was used only once, and 12 hearts were used per group as in previous studies. The choice of solution was determined by reference to a randomization table from which two solutions were chosen for each experimental run (typically 8 rats). The experimental operator was blinded to the nature of the solution. Records were identified by a number code only, and therefore analysis was carried out without knowledge of the solution employed.

D. Exclusion criteria

A total of 67 hearts were used for these studies of which 60 were retained. Stability criteria (as described in Studyl) were not fulfilled by 2 hearts. Five hearts not in sinus rhythm at the moment of reperfusion were excluded from the reperfusion study and replaced (as described in Study 1). In the methylsulphate group there was a very high incidence of hearts which were not in sinus rhythm at the moment of reperfusion. Assessment of reperfusion arrhythmias was therefore impracticable in this group and was abandoned.

3. RESULTS

A. Ischemia-induced arrhythmias

Compared to the standard chloride solution, the incidence of VF in bromide, nitrate and iodide groups was significantly reduced (Fig. 18). By contrast, the incidence of VF was increased in the methylsulphate group (Fig. 18). Furthermore, in those hearts having ischaemia-induced VF, methylsulphate was associated with SVF in 83% (8/12) compared to 36% (3/8) in the chloride group. Significant reductions in the incidence of VT were seen in the nitrate and iodide groups and significant reductions in BG and salvos were seen in the nitrate group (Fig. 19). The mean latency to onset of the first ischaemia-induced arrhythmia (usually VPB) was not found to be altered by anion substitution (Fig. 20).

B. Reperfusion-induced arrhythmias 89

100

4; 80 - p<0.05

' o a> ci o T 40 -

20 —

0 - I

MethylS Chloride Bromide Iodide Nitrate

Fig. 18: Incidence (%) of ischaemia-induced ventricular fibrillation (VF). 90

inq o in o CL o Q.

CL o o o o O o o o Cn | § 8 9 8 o CO

T> x> in q o CL

CD X)

CD o o o § 8 ? ° CD 8dA peonpuj-Biuueeijos! soa |b s peonpui-BjweBipsi beat, BG beat, = BG bigeminy, VT = ventricular tachycardia. jo (%) eouepjou| jo (%) eoueppu| 91

1000 -|

100 -

10 -

MethylS Chloride Bromide Iodide Nitrate

Fig. 20: Onset time of first ischaemia-induced arrhythmia. Note that the timescale is logarithmic 92

Significant reductions in the incidences of reperfusion-induced VF occurred in the nitrate and iodide groups but not in the bromide group (Fig. 21). A significant reduction in the incidence of reperfusion-induced VT was produced only in the nitrate group. The incidences of other arrhythmias (VPB, BG, S) were not affected by anion substitution (Fig. 22). No difference was seen in the incidence of SVF between the chloride and bromide groups; in each case it occurred in 64% (7/11); the 2 cases of reperfusion-induced VF in the iodide group were not sustained. There was a significant increase in the time to onset of reperfusion-induced arrhythmias in the nitrate, iodide and bromide groups compared to the chloride group (Fig. 23).

C. Heart rate, coronary flow and occluded zone size

In order to determine whether the effects of anion substitution on arrhythmias were some consequence of altered haemodynamics we measured coronary flow throughout the experiment and occluded zone size at the end. Heart rate fell slightly during the course of the experiment in all groups (Fig. 24). Anion substitution had little effect on heart rate which therefore did not correlate with arrhythmia incidence. Throughout the time course of the experiment, no significant difference was identified between the coronary flows in the chloride group and the groups with anion substitution. There was no correlation between flow and arrhythmias (Fig. 25). Recovery of flow in the reperfused zone was calculated by measuring total coronary flow and expressing the increase as a function of occluded zone size. No significant alteration was seen in recovery of flow compared to the chloride group (Fig. 25). There was no relationship between the recovery of flow and onset time of first reperfusion-induced arrhythmia. Anion substitution had no significant effect on occluded zone size (Fig. 26).

D. ECG changes

Ischaemia had a triphasic effect on QRSTgo (Fig. 27) which was dependent upon when the values were recorded. Before the onset of ischaemia, changes between groups were small. During the first minute of ischaemia, the width 93

p<0.05

Chloride Bromide Iodide Nitrate

Fig. 21: Incidence (%) of reperfusion-induced VF. The high incidence of ischaemia-induced arrhythmias at 30 min in the methylsulphate group made study of reperfusion-induced arrhythmias impracticable in this group. 94

0) JD - 8 ■ 2 t> r? T3 2

■ = 8 O. c 3 E CD T3 — > C CO II “ .S3 co ~ E CL E -c > in c o CO * 4-* CO • 5 ° 8 I 8 g Si? C <2o o > • £ co .c ^ o.2> © I------1--1-----1------1---1 I 1---- 1---- 1---- 1---- 1 CD § § 8 $ 8 ° w » jo (%) eouappuj g « * ■ jo (%) ooueppui C P 3 ^ CO to

C «S E CO k_ Q . O)_C0 3 “ O S '- 51 §13 .2 H i ,_ % 3 c h c/) .2 > >» § to = p . 2 t g: g - CD CD ® .2 > £ o -g c ^ 0 c CQ E CD . o O CO c CO 03 CO 03 CO . C/> g- I— I— I— I— I— I oc 2^ o O Q o O o . . 3 | I O 00 § § OJ ° O 00 <5 "T CM CO ^ j= O) -tr * CO s o a |b s peonpuj-uojsnjjedej CVJ g x: .c 8dA poonpin-uojsnjjedej jo (%) eoueppu| 0 3 a? T I *- jo (%) eouepjoui L L C L co . 2 95

100 p< 0.05

Chloride Bromide Iodide Nitrate

9lBy UB0H 97

CL O) CL

(/)

m m . m

V/////// / // // // / /~7~7\

WWW in o m CM CM

JM J9M 6/U!LU/|UJ U| MO|J A j b u o jo q 98

50 nj o 40 -

cain 30 - w cd a> ZJ C O o d 20 H N c ■g 0) T3 > .2 10 o o O 0-1

M ethylS Chloride brom ide Iodide Nitrate

Fig. 26: Occluded zone size as a % of total ventricular weight. 99

c ( 0 CD CD CD T > T 5 CD k _ CD E xz o E T > CO 'u CD O T 3 CD j C O 2 Q. O m X CD

• ■ o □ < o COk _ 3 8 CD E 3 nO O) O3

(0 cc O o •■S3 CO N _C0 O a>Q. >p O0s o> 4—• CO 0)X a. E 8

I------1------1------.------1------1------1------1------o o o o o co co in tj-

O0SUJ u| iflp!m 06 JLSUO 100 shortened in all groups. By 5 minutes of ischaemia there was a marked increase in width but this was only sustained throughout the remaining period of ischaemia in the 2 groups with the highest membrane permeability.

4. DISCUSSION

A. Actions of anion substitution on ischaemia-induced arrhythmias

In the isolated rat heart with regional ischaemia, it was shown that anions with membrane permeability greater than chloride, caused a reduction in the incidence of ischaemia-induced arrhythmias. Furthermore, this effect correlated well with the relative permeabilities of the anion surrogates studied and was most pronounced with the most severe arrhythmias (VT, VF). In the two groups (chloride, methylsulphate) which had a sufficient incidence of ischaemia-induced VF, it was also possible to investigate a second proarrhythmic measure; the ability of VF to sustain.This suggested that methylsulphate, in addition to increasing the incidence of VF, also promoted SVF. Raised extracellular potassium and Class III agents have been shown to inhibit the ability of VF to sustain in this model (Curtis and Hearse 1989b, Tsushihashi and Curtis 1990), suggesting that the action of methylsulphate may be related to its ability to shorten the QT interval.

B. Actions of anion substitution on reperfusion-induced arrhythmias

Reduction of reperfusion-induced arrhythmias correlated with increasing membrane permeability of anions with significant amelioration with the most permeable anions (nitrate and iodide). However the predicted small increase in membrane permeability in the bromide group (Seyama 1979), although sufficient to reduce ischaemia-induced VF, did not affect reperfusion-induced arrhythmias. This may relate to the hypothesis that the mechanisms of arrhythmogenesis differ between ischaemia- and reperfusion-induced arrhythmias (Study 2) or that reperfusion is a more 'potent' arrhythmogenic stimulus, requiring more 'potent' antiarrhythmic interventions.

C. Insight into mechanism of action from changes in ECG 101

The initial reduction in the width of QRSTgo is consistent with previous observations of shortening of action potential duration in ischaemic tissue (Carmeleit 1978). Subsequent prolongation of the QRST90 was sustained only in the groups with anions which had greatest permeability (nitrate, iodide) and fewest arrhythmias suggesting a prolongation of refractory period might be involved in the antiarrhythmic effects. However, as discussed in Study 1, whether this relationship is coincidental or causal in reducing the incidence of ventricular arrhythmias remains unclear.

D. Insight into mechanism of action from haemodvnamic variables

No significant alterations in heart rate, coronary flow or occluded zone size were seen between groups, indicating that the mechanism of action of chloride surrogates was not secondary to an effect on haemodynamic variables. This is consistent with a direct cellular effect.

E. The importance of the membrane permeability of anion surrogates

The correlation between the membrane permeability and the antiarrhythmic (nitrate, iodide, bromide) or (in the case of methylsulphate) proarrhythmic actions of anion surrogates strongly suggests that their mechanism of action involves an action on the sarcolemmal membrane to increase or decrease membrane conductance. 102

STUDY 3.2

MODIFICATION OF INTRACELLULAR CYCLIC NUCLEOTIDES

1. INTRODUCTION

Elevated tissue levels of cAMP during myocardial ischaemia have been associated with arrhythmias VF (Lubbe et al. 1976). The antiarrhythmic action of NCS might feasibly involve stimulation of production of cGMP (Bennett et al. 1988) to oppose the action of cAMP (Ikemoto and Goto 1977). In this study we test the hypothesis that the action of anions on arrhythmogenesis might be mediated by manipulation of cyclic nucleotides. Anion surrogates with a known order of antiarrhythmic action (nitrate> iodide> bromide> chloride> methylsulphate) were employed (Study 3.1).

2. MATERIALS AND METHODS

A. Animals and experimental methods

These were as described in Study 1. Male Wistar rats (250-300 g, Harlan Olac, Bicester) were anaesthetized with halothane before excision of the heart for Langendorff perfusion. After rendering the heart regionally ischaemic for a specified period of time, myocardial biopsies were taken from the ischaemic zone (readily identifiable due to its dusky colour) and the nonischaemic zone. A biopsy was also taken from the reperfused zone 1 min after reperfusion.

B. Composition of perfusion solutions

Standard chloride-containing solution was employed in a control group and in other groups it was modified (as in Study 3.1) by complete isotonic substitution of chloride with one of four anion surrogates (nitrate, iodide, bromide, methylsulphate).

C. Experimental protocol 103

A different anion was used for each of 5 groups of hearts (n=6/group). The choice of solution was determined by reference to a randomization table. Each heart was allowed to stabilize for 10 minutes and then the left coronary artery occluder was tightened to produce regional ischaemia. After 15 min regional ischaemia, myocardial biopsies of ischaemic and non-ischaemic tissue were taken. The heart was then reperfused and 1 minute after reperfusion a further biopsy was taken from the reperfused region. The 15 min time period was chosen as this is the time of maximal susceptibility to both ischaemia- and reperfusion-induced arrhythmias in the isolated rat heart (Manning and Hearse 1984). To determine the effect of time on the level of cyclic nucleotides a further 2 groups of hearts (4 hearts/group) were studied during 5 and 30 min regional ischaemia and biopsies taken from ischaemic and nonischaemic regions. It is possible that the presence of regional ischaemia might influence the levels of cyclic nucleotide in the non-ischaemic tissue. Therefore sham ligation was performed in a further group of 4 hearts which were biopsied after 15 minutes.

D. Assay techniques

Immediately after biopsy the myocardial specimen was flash frozen in liquid nitrogen and stored at -70QC for batch assay. The specimens were crushed with a stainless steel pestle and mortar and the crushed tissue was rapidly transferred to a cold solution of 2.2% perchloric acid and shaken for a period of 2 min and then centrifuged at 40C for 2min. The centrifugation concentrated the debris pellet which was available for subsequent protein assay using the Lowry method (Lowry et al. 1951). A 400 pi aliquot of supernatant was then removed and kept on ice during neutralization to pH 7.4 with a solution of 1.1 M K3PO4.

The volume of solution added (typically 75pl) was noted for calculation of dilution factor. A flocculent precipitate formed and the specimen was again centrifuged at 40C for 2 min and the supernatant employed for assay of cyclic nucleotides with kits provided by Amersham International pic. cAMP was assayed with the cAMP [125|] scintillation proximity assay (SPA) system [code RPA 538]. cGMP was assayed with cGMP [125|] SPA system [code RPA540]. 104

E. Exclusion criteria and statistics

Unstable preparations as described in Study 1 led to the* exclusion of 2 hearts. Statistical analysis was as described in Study 1.

3. RESULTS

A. Sham ligation

In the group subjected to 15 min sham ligation the mean concentrations of cAMP and cGMP were 7.28 ± 1.45 nmole/g protein and 78.94 ± 13.327 pmole/ g protein respectively. The mean molar ratio of cAMP/cGMP was 103 ± 29.

B. Effect of duration of regional ischaemia

The effect of different periods (5,15 and 30 min) of regional ischaemia on cAMP, cGMP and the molar ratio of cAMP/cGMP are demonstrated in Fig. 28. Small increases in the mean cAMP concentrations compared to control (sham ligation) were noted; the mean cAMP concentration was maximal in the non-ischaemic zone after 15 min regional ischaemia. However, these increases failed to reach statistical significance. Similarly differences in the mean cGMP concentration and the molar ratio of cAMP/cGMP compared to sham ligated controls failed to reach statistical significance.

C. Effect of anion surrogates

The mean concentrations of cAMP, cGMP and the molar ratio of cAMP/cGMP in the ischaemic and non-ischaemic zones (Fig. 29) demonstrated no statistically significant difference from the sham ligation group. cAMP was not reduced by the anions which were most antiarrhythmic and in fact increased (in the non-ischaemic zone) with increasingly antiarrhythmic anions. However, this is likely to be a casual, rather than causal, association as, i) differences between groups are small, ii) the trend is opposite to the proposed theory that cAMP promotes VF, and iii) there is poor correlation between cAMP concentrations and the known incidences of ischaemia-induced VF resulting from these anion 105

o ■ cAMP isch E a □ cAMP nonisch

C l JZ

Sham 5 15 Duration of regional Ischaemia 200

■ isch cGMP □ nonisch cGMP

Sham 5 15 30 Duration of regional Ischaemia (min) 400 i

O. <2 3oo i \ CL H < ° 200 ■ ratio isch □ ratio nonisch

00 -

Sham 5 15 30 Duration of regional Ischaemia (min) Fig. 28: Mean concentration ± SEM of cAMP (nmole/g protein), cGMP (pmole/g protein) and molar ratio of cAMP/cGMP at 5 min (n=4), 15 min (n=6) and 30 min (n=4) of regional ischaemia. Myocardium was biopsied from both ischaemic (isch) and nonischaemic (nonisch) zones. 1 06

20 -

\o> a> "o isch 10 - nonisch

< o

Shem MethS Chlor Brom Iodide Nitrete

200 -

o isch E 100 " CL nonisch -- X CL 51 CD o

Sham MethS Chlor Brom Iodide Nitrate

CL H 200 - CD \O Q_ z : < isch o nonisch ° 100 -

o 51

Sham MethS Chlor Brom Iodide Nitrate Fig. 29: Concentrations ± SEM of cAMP and cGMP and molar ratio of cAMP/cGMP in ischaemic (isch) and non-ischaemic (nonisch) zones of hearts biopsied after 15 min regional ischaemia. Groups of hearts (n=6/group) were perfused with either standard chloride-containing solution (chlor) or with one of 4 isotonic anion surrogates; methylsulphate (MethS), bromide (Brom), iodide or nitrate. A separate group of hearts (n=4) were sham ligated and perfused for 15 min without regional ischaemia 107

100 *1

80 “

60 - ee a> o t z 40 1 X3a> oa

I— h - e - a H 1 T "T I 12 16 20 24

cAMP In nonischaemic zone (nmole/g)

Fig. 30: Concentration (nmole/g) of cAMP in the non-ischaemic zone of hearts plotted against % incidence of VF in similar groups (Study 2.1). Hearts were rendered regionally ischaemic for 15 min before biopsy.Groups (n=12/group) were perfused with standard chloride-containing solution or one of 4 other solutions with isotonic substitution of chloride with methylsulphate, bromide, iodide or nitrate. Considerable overlap of values occurred between groups revealing little evidence of a correlation between incidence of VF and concentration of cAMP in the non-ischaemic zone of hearts rendered regionally ischaemic 108

20 -

CT> a> o Ea

CL n

Sham MethS Chlor Brom Iodide Nitrate

l 40 " 120 "

100 -

80 -

60 - CL r 40 - too

20 -

Sham MethS Chi Brom Iodide Nitrate

CL 200 - Y. too \ CL x:

o n

Sham MethS Chi Brom Iodide Nitrate

Fig 31: Mean concentrations ± SEM of cAMP (nmole/g protein), cGMP (pmole/g protein) and molar ratio of cAMP/cGMP in reperfused zone of 5 groups of hearts (n=6/group) perfused with methylsulphate (MethS), chloride (Chi), bromide (Brom), iodide or nitrate. Hearts were regionally ischaemic for 15 min and then reperfused. Reperfused tissue was biopsied 1 min after reperfusion. These values are compared to a control group (sham) in which the heart was not rendered ischaemic. surrogates (Fig. 30). cGMP was not significantly elevated by nitrate, nor was the molar ratio of cAMP/cGMP reduced. Similarly the other anion surrogates failed to significantly alter these variables. Anion substitution caused no significant alteration in cAMP and cGMP concentrations or in the molar ratio of cAMP/cGMP in the reperfused tissue(Fig. 31).

4. DISCUSSION

The findings do not support the hypothesis that the antiarrhythmic activity of NCS is mediated by changes in cyclic nucleotide homeostasis. NCS failed to significantly increase cGMP or reduce cAMP and there is thus no evidence to support the hypothesis that its antiarrhythmic action might involve an inhibition of cAMP by elevated levels of cGMP. This is further emphasised by the lack of significant reduction in the molar ratios of cAMP/cGMP in ischaemic, non-ischaemic or reperfused tissue with NCS. Earlier studies expressing concentrations of cyclic nucleotides in terms of wet (Dobson Jr. and Mayer 1973, Lubbe et al. 1978, Podzuweit et al. 1978, Wollenberger, Krausse, and Heier 1969) and dry (Podzuweit et al. 1989) ventricular weight, and in terms of protein (Corr, Witkowski, and Sobel 1978) have shown that changes elicited by ischaemia are not always high and not necessarily restricted to the ischaemic or reperfused zones. Podzuweit et al (Podzuweit et al. 1978) found an elevation of cAMP in the ischaemic zone of approximately 60% in baboons (n=6) with high left coronary ligation which preceded the onset of VF; low left coronary ligation did not alter cAMP. Wollenberger et. al. (Wollenberger, Krausse, and Heier 1969) describe a cAMP elevation of approximately 100% in the ischaemic zone of the dog (n=13) which reached a maximum within 5 min of the onset of ischaemia and persisted for 20 min. Corr et al (Corr, Witkowski, and Sobel 1978) describe a cAMP elevation of approximately 90% in the ischaemic zone of the cat (n=6) which reached a peak at 15 min but returned to pre-ischaemic levels by 40 min. However, Podzuweit et al (Podzuweit et al. 1989) describe a cAMP increase of only 20% in the ischaemic zone in the pig (n=10) after 30 min regional ischaemia and the level declined to approximately 30% below the pre-ischaemic level by 90 min of 110 ischaemia. Dobson and Mayer (Dobson Jr. and Mayer 1973) described a cAMP elevation of only 40% in the working globally ischaemic rat heart (n=5) after 5 min of ischaemia, and cAMP had returned to pre-ischaemic levels within 10 min of the onset of ischaemia. In the non-working globally ischaemic rat heart (n=5), the same authors (Dobson Jr. and Mayer 1973) describe a very small (20%) elevation in cAMP requiring 20 min of ischaemia to become manifest. Furthermore, the cAMP concentration fell to pre-ischaemic levels by 30 min and then rose to 50% of pre-ischaemic levels by 60 min. Importantly, this study did not demonstrate significant elevation of cAMP in the rat heart at 15 min (time of maximal susceptibility to VF) in either of the models of global ischaemia. Our findings are consistent with these observations in that although there was a trend toward cAMP elevation following regional ischaemia, this failed to reach statistical significance. This study does not suggest a major role for cyclic nucleotides in arrhythmogenesis in the isolated rat heart during ischaemia or reperfusion Furthermore, it has been shown that forskolin exerts an antiarrhythmic effect in the isolated perfused rat heart despite its ability to increase cAMP. Since ischaemia elevates cAMP and elicits arrhythmias, whereas forskolin elevates cAMP and inhibits ischaemia-induced arrhythmias there seems to be no basis for the hypothesis that cAMP is linked to arrhythmias in ischaemic heart disease. It has been suggested by others that cAMP and ischaemia-induced VF are casually rather than causally related (Manning et al. 1985). In the present study, the trend towards elevation of cAMP seen during regional ischaemia was as pronounced in the non-ischaemic as the ischaemic zone of the heart (although it did not reach statistical significance in either zone). In previous studies attempts to differentiate between ischaemic and non-ischaemic tissue in the rat heart have not been made (Dobson Jr. and Mayer 1973, Lubbe et al. 1976, Manning et al. 1985, O’Brien and Strange 1975) but in other species this has been done and findings have varied. In the baboon (Podzuweit et al. 1978), the dog (Wollenberger, Krausse, and Heier 1969) and the pig (Podzuweit et al. 1989), elevation in cAMP was identified only in the ischaemic zone. However in other studies in the dog (Ziegelhoffer et al. 1976) and in the cat (Corr, Witkowski, and Sobel 1978), regional ischaemia has been reported to cause elevation of cAMP within both the ischaemic and non-ischaemic zones of the myocardium. 111

STUDY 3.3

MODIFICATION OF INTRACELLULAR pH

1. INTRODUCTION

It was possible that the antiarrhythmic effect of NCS was mediated by modulating intracellular pH by an action on the sarcolemmal chloride-bicarbonate exchange mechanism (Curtis 1989, Vaughan-Jones 1979b). Elevation of pH via this mechanism might exert its antiarrhythmic action by inhibiting the fall in pH which occurs in ischaemic myocardium (Hirche et al. 1980). We tested this hypothesis by measuring the changes in the position of the intracellular inorganic phosphate peak (characteristically altered by changes in intracellular pH) using phosphate nuclear magnetic resonance (31P-NMR) spectroscopy. The principle of intracellular pH measurements is based on examining the 3ip spectrum and considering the position of the inorganic phosphate (Pi) peak with respect to that of phosphocreatine (PCr); the relative position of the peaks shifts in proportion to the intracellular pH (Moon and Richards 1973). The nature of the NMR system means that any ischaemia/ reperfusion experiments are most easily performed using global rather than regional ischaemia. Global ischaemia also has a positive advantage over regional ischaemia in that there is no technical necessity for differentiation between non-ischaemic and ischaemic or reperfused tissue during ischaema and reperfusion. Since NMR studies are designed to examine cellular events rather than syncytial antecedants of ischaemia and reperfusion there are no obvious inherent disadvantages to global ( as opposed to regional) manipulation in this context.

2. MATERIALS AND METHODS

A. Animals and experimental methods

Employing the technique described in Study 1, hearts were excised from male Wistar rats ( 280-310g, Banting and Kingman) and perfused by the Langendorff 112 technique in a double-walled chamber in a system which was interfaced with a Bruker AM-400 NMR spectrometer. The temperature of the fluid as it entered the heart was 37°C and the perfusion chamber was maintained at 37°C by a flow of thermostatted air through the bore of the magnet. With this technique the heart was horizontal and supported within the chamber by perfusion solution from the pulmonary effluent.

B. Composition of perfusion solutions

Two solutions were employed in this study; standard chloride-containing solution, or solution modified by complete NCS. Phosphate does not affect the incidence of arrhythmias in the isolated rat heart model (Curtis 1991) and it was not employed in the solutions used for the studies in this thesis. It is of particular importance that phosphate is omitted from solutions employed in 3ip NMR studies.

C. Experimental protocol

Two groups of hearts (n=4/ group) were randomized to perfusion with either standard chloride-containing solution or solution modified by complete NCS. Fully relaxed 31P-NMR quantification spectra (40 scans using 90° pulses with a 15s repetition time) were collected during 65 min of aerobic perfusion and used for quantification of heart metabolites. The repetition time of 15s, used for all the quantification spectra, was sufficient for complete relaxation of the external standard, methylenediphosphonate (MDP) and of all the heart metabolites, since the longest t-i, that of PCr, is 3s (Lamprecht et al. 1974).

During 10 min global ischaemia partially saturated spectra (48 scans using 70° pulses with a 1s repetition time) were collected every minute. Such partially saturated spectra are adequate for calculation of pH. The NMR-visible amounts of ATP, PCr and Pj in fully relaxed spectra were quantified using the external standard, MDP, in the annulus of the perfusion chamber. Intracellular pH was calculated from the difference in the chemical shifts of the Pj and PCr peaks (3 ) according to the formula below

(Bailey et al. 1981); 113 pH = 6.72 + log 10 {(3 - 3.27)/(5.69 - d)}

D. Statistical analysis

All results were expressed as mean ± SEM. Statistical comparison of groups was carried out by Student's t test. p< 0.05 was considered to be statistically significant.

3. RESULTS a. Metabolic,'effects

The pre-ischaemic NMR-visible amounts of Pi, PCr and ATP were measured with fully relaxed spectra in all hearts and are shown as mean contents (pmole/heart) ± SEM for the two groups of hearts (Table 4). Significant elevation of Pi and significant reduction in PCr was seen with NCS. ATP levels were not significantly changed by NCS. Illustrative examples of spectra from a standard chloride-perfused heart is shown (Fig. 32) and may be compared to an illustrative example of a heart perfused with complete NCS solution (Fig. 33). The initial spectrum is taken fully relaxed before the onset of global ischaemia (1-1) and shows a lower broader Pi peak and a lower PCr peak following NCS. During the time course of global ischaemia, NCS leads to a reduction in ATP peak (from approximately 5-6 min), soon after the PCr peak disappears (4-5 min). By contrast, ATP is preserved in the chloride-perfused heart and the PCr peak is preserved until approximately 6-7 min of global ischaemia. Thus, the initial reduction in PCr by NCS appears to lead to its more rapid consumption during global ischaemia followed by consumption of ATP.

B. Effects on pH

Pre-ischaemic mean intracellular pH was reduced by NCS and remained lower throughout the period of global ischaemia (Fig. 34). As noted above, the Pi peak (Fig. 33) was broader following NCS and in 3 of the 4 hearts contained 2 or 3 minor peaks in the pre-ischaemic spectrum and during the first 2 minutes of 114

Chloride N itra te

Pi 1.920 ± 0.256 3.896 ± 0.431 *

PCr 6.960 ± 0.805 3.621 ± 0.349 *

ATP 3.356 ± 0.608 2.897 ± 0.460

Table 4: Mean contents (pmol/heart) ± SEM of intracellular phosphate (Pi), phosphocreatine (PCr) and adenosine triphosphate (ATP) in 2 groups of hearts (n=4/group) perfused with either standard chloride-containing solution or solution modified by complete isotonic substitution of nitrate for chloride. * indicates p< 0.05 vs.chloride control. 115

8-9 min > ^ V rV W W W ^ ^ ^

Fig. 32: Illustrative NMR 7-8 min spectra of a heart perfused with standard chloride-containing solution and rendered globally ischaemic for 10 A ' /I A 6-7 mln min. First spectrum V-1' ^ recorded 1 min before onset of ischaemia (1-1) and 9 spectra taken at 1 min intervals during the / ’ II » 5-6 min course of global ischaemia. Spectral peaks for the external standard > I methylenediphosphonate „ j ' i . ../. ,!\ J 4-5 min (MDP) and the myocardial phosphate (Pi), phosphocreatine (PCr) and adenosine triphosphate (ATP) are identified. W V < LjJJw JLv.. 3 -4 m ,n

2-3 min

1-2 min .%

0-1 min v w v ^ PCr

MDP Pi i i ATP .-. • Pl ! 1 | ^ 1-1 min 'V’v'V^Ad 116

min Fig. 33: Illustrative NMR spectra of a heart perfused with solution modified by nitrate substitution of chloride and rendered globally y^-v* ischaemic for 10 min. First spectrum recorded 1 min before onset of ischaemia (1-1) and 9 6-7 min spectra taken at 1 min intervals during the course of global ischaemia. Spectral peaks for the external i 5-6 min standard .^^K'AV- vVn V-^'^ W V * ^ ' V>v^ a/.^ ^ .^ v^ methylenediphosphona te (MDP) and myocardial phosphate (Pi), phosphocreatine (PCr) and adenosine ,;.w U lv> a. 4 *5 m In triphosphate (ATP) are V^V'V/ ^v,wvw^v^^ identified.

\ ' J l a 3-4 min

2-3 min

1-2 min

/v^/*oA/^'vV'1i'M

0-1 min

PCr

MDP Pi I , ATP 1-1 min

/,^' 117

7.2

7.0

6.8

EJ Chi

6.6 NCS X Q.

6.4

6.2

6.0 1------1------1------1------r -2 0 2 4 6 8

Duration of global Ischaemia

Fig. 34: Alteration in mean intracellular pH + SEM in hearts (n=4/group) perfused with either standard chloride-containing solution (chi) or solution modified by complete nitrate substitution of chloride (nitr). pH was measured 1 min before the onset of global ischaemia and at 1 min intervals during the global ischaemia. * indicates p< 0.05 compared to time-matched chloride-containing hearts. 118 ischaemia suggesting heterogenous intracellular pH. In all cases these sub-peaks indicated a pH less than all the time-matched pH measurements in the chloride-perfused hearts. For the purposes of data analysis the pH value for each heart was taken as the mean of the value calculated from each sub-peak.

4. DISCUSSION

This study demonstrates that NCS reduces intracellular pH, rather than elevating it. Thus the theory that NCS might exert its antiarrhythmic action via inhibition of the CI-HCO3 exchange mechanism cannot be sustained as such an inhibition would predict an elevation in pH (Vaughan-Jones 1979b). Complete NCS appears to cause heterogenous areas of low intracellular pH, the significance of which is unknown. Evidence of impaired metabolic function is gained from the significant rise in NMR-visible Pi. It is possible that such changes might be partially explained by a reduction in the NMR-invisibility of Pi following NCS (Humphrey and Garlick 1991). However, the combination of an elevation of Pi together with a reduction of PCr suggests impairment of metabolic function by NCS. 119

STUDY 3.4

MODIFICATION OF ISOCHORIC VENTRICULAR CONTRACTION

1. INTRODUCTION

Many antiarrhythmic agents have negative inotropic actions which detract from their therapeutic efficacy (Singh et al. 1987). If antiarrhythmic effectiveness of chloride substitution is associated with negative inotropic effects, then this would indicate that the potential beneficial effect is intimately linked to a potentially detrimental action which, as with most antiarrhythmic agents, would jeopardize its potential therapeutic usefulness. However, evidence from studies in isolated myocardial tissue preparations suggests that substitution of chloride for other anions may actually be positively inotropic (Anderson and Foulks 1973, Horackova and Vassort 1982, Nosek 1979). In preceding experiments we have shown that anion manipulation has little effect on the haemodynamic variables of heart rate and coronary flow. However, there was evidence from the NMR study (Study 3.3) that complete nitrate substitution had a deleterious effect on intracellular metabolism. In this study we have investigated the effect of anion manipulation on left ventricular contractility employing an isochoric intraventricular balloon technique (Curtis et al. 1986).

2. MATERIALS AND METHODS

A. Animals and experimental methods

Male Wistar rats (280-350g, Harlan Olac, Bicester) were anaesthetized with halothane and the hearts set up for Langendorff perfusion employing a constant aortic root pressure of 100 mm Hg. A small compliant, but nonelastic balloon was made from wrapping film (Clingfilm) and connected via a length of polypropylene tube to a pressure transducer (Spectramed P23XL, Statham). The balloon was made slightly more capacious than the left ventricle (typically 0.4 ml) to ensure that, as the balloon was filled, any increase in the measured pressure was due to an increase in ventricular wall tension, not balloon wall 120 tension. An opening was cut in the left atrium and employed for insertion of the balloon catheter into the left ventricle. The balloon catheter was inflated with distilled water. With the balloon inflated isochoric pressure recordings of left ventricular systolic and diastolic pressures were made before and after intervention.

B. Composition of perfusion solutions

A total of 7 solutions were employed. A standard chloride-containing solution was used as a control. In this study the standard chloride-containing solution was modified by i) complete substitution of chloride with either methylsulphate, bromide, iodide or nitrate or ii) addition of 10 pmolar adrenaline or 1 pmolar verapamil. The latter interventions were standards for testing the sensitivity of the preparation as a model for detecting negative and positive inotropic activity.

C. Experimental protocol

A randomized design was employed to study the 7 interventions and separate groups of isolated rat hearts (n=9 / group) were used for each intervention. In each case the heart was perfused initially with standard chloride-containing solution. Each heart was allowed to stabilize for 15 minutes during which time the intraventricular balloon was inserted atraumatically taking care not to disturb the preparation. The balloon was then inflated in 10 pi increments and diastolic and systolic pressures were measured (Curtis et al. 1986) for each increment. The volume required to achieve a diastolic pressure of 10 mmHg was noted. At 30 minutes the solution was switched either to an identical chloride-containing solution or to one of the 6 other interventions. The heart was again allowed to stabilize for 15 minutes and the balloon was reinflated in 10 pi increments and the diastolic and systolic pressures were measured. The* heart rate and coronary flow were recorded at 5, 35 and 55 minutes; that is, during stabilization (-25 min) and 5 and 25 minutes after switching to the intervention.

D. Exclusion criteria

Unstable preparations were excluded. A stable preparation was defined as 121 having at 5 min (during the stabilization period) a sinus rate of at least 290/min and a coronary flow of at least 10 ml/min and which, when pressure measurements were taken before switching solutions, had a developed pressure (systolic pressure - diastolic pressure) of greater than 100 mmHg with a diastolic pressure of 10 mmHg. Stability criteria resulted in the exclusion of 9 hearts out of an original 72 entered into the study.

E. Statistics and data analysis

There is considerable evidence that the slope of the peak isochoric pressure-volume relationship (E es) is a reliable measure of left ventricular systolic function (Anderson et al. 1990., Sagawa 1978, Sagawa 1981, Wexler et al. 1988). Ees was readily plotted as a best fit linear regression line. In all hearts the correlation coefficient for this line was > 0.93 confirming the linear relationship of end systolic pressure and volume (Apstein Jr, Mueller, and Hood 1977). A second measure of systolic function (contractility) was calculated as the developed pressure (systolic pressure - diastolic pressure) at a set end diastolic pressure (Anderson et al. 1990., Curtis et al. 1986). Developed pressure was recorded at a diastolic pressure of 10 mmHg during initial perfusion with chloride-containing solution. This pressure was generated by balloon volumes associated with the linear portion of the pressure-volume relationship. After switching solutions the pulse pressure was again measured with the same intraventricular volume. Diastolic pressure reflects left ventricular diastolic function (compliance) (Apstein Jr, Mueller, and Hood 1977, Wexler et al. 1988). The diastolic pressure after switching solutions was recorded at constant intraventricular volume (that required to produce a diastolic pressure of 10 mmHg during the initial chloride perfusion). All values were expressed as mean ± SEM. A paired t test was employed to compare E e s , developed pressure and diastolic pressure before and after switching solutions.

3. RESULTS

A. SystQliC function (Table 5) 122

The mean developed pressure at a diastolic pressure of 10 mmHg remained constant when solution was "switched" from chloride to an identical chloride solution. Thus the act of switching solutions was not a significant source of variance. Verapamil caused a large and significant reduction (of approximately 86 mmHg) in mean developed pressure and adrenaline caused a large and significant increase (of approximately 121 mmHg) in developed pressure. In the other groups the changes in developed pressure were small and significant (p< 0.05) only in the methylsulphate group in which there was an elevation of approximately 21 mmHg. There were no significant changes in Ees when chloride solution was "switched " to chloride. Verapamil caused a significant reduction in Ees of approximately 0.24 mmHg/pl and adrenaline caused a significant increase of approximately 0.8 mmHg/ pi. Of the other interventions, only nitrate caused a significant reduction of approximately 0.11 mmHg/pl.

B. Diastolic function (Table 6)

In all hearts the diastolic pressure of 10 mmHg was chosen prior to switching solutions. After switching solutions the values in the group "switched" to chloride-containing solution did not change significantly. Verapamil elevated mean diastolic pressure by approximately 30 mmHg (p< 0.05) and adrenaline reduced it by approximately 23 mmHg (p< 0.05). In the other groups (nitrate, iodide, bromide, methylsulphate) any changes were significant only in the iodide group in which there was a significant elevation of approximately 11 mmHg (p< 0.05).

C. Heart rate and coronary flow (Table 7)

The mean heart rate was similar in all groups 20 min before intervention. A small decline in heart rate occurred with time in the chloride control group. Thus values 5 and 25 min after switching to interventions are time-matched with the chloride group for comparison. In the verapamil group there was a significant fall in heart rate (of approximately 85 beats/min) and in the adrenaline group a significant elevation (of approximately 42 beats/min) 5 min after intervention with changes persisting at the 25 min time point. None of the chloride surrogates 123

© * * * k _ CM VO m D • CM ON VO O o • to in ■*- CO VO in ro to c © o flu +i +i +i (/) «- v_ +1 +1 +1 +i .9 ql 2 2 © © in in o *-i= © • CM• ro• 5 .9 U *4-» in t t d O © 00 00 CM o 8 -2 a K) CM ro ro ro © "D © CM © *■“ t/am !9-8 ^ ^ 2 o w k_ a ‘a -9 © 2 2- o * * * O 2 . 0.0 k. vO SO in m 00 to in © O O o o o © o o o o o d o w o a : Ul +1 +i +1 +1 +i +i >“» w U l +1 *5)5 « v ON c © n - 00 VO 00 ON Lc n Q. K) ro TT TT in CM CM © • 4-»O •—QL o o o o o o CL 0> — © ■ (/> c © o c *- o © ~ ^ c« Q3 1 9 ° 3 CD o T 5 *“ 2 . © u ON vo vo ON c pr* • • CM• r - © a . in tt co zL S m 2 ■s d r - TT T f ro 2 co q ■d +i 2 o> o © +i +i +i +1 +1 +i > © 5 1 a2 J+1 o a 00 n - ro n - ro• n - n - © © • • • • • 8 £ u § « o Q o CM o VO 2 © 2 L_ CM ro to CM CM c £ a. p o CJ *“ *- .9 © o t o © w u . © > © © k* 1 1 Q n - to to in VO in ro 3 © Q2 to c © E B > © © © E © 2 © 22 5 c § < o> > © ■o I© © © > *5> CL c © © .E © u j= m L . - u E o ■4-* © © c 3 © ■o o © u k. © © Q_ G r o u p u JZ o © TO ■O © E z CO CJ T a s * c © _O © CL o 124

Interuention Diastolic pressure Diastolic pressure before crossouer after crossouer

Nitrate 10 13.0 ± 1.5

Iodide 10 20.78 ± 2.8*

Bromide 10 8.3 ± 0.8

Chloride 10 10.7 ± 1.1

MethylS 10 9.1 ± 0.9

Uerapamil 10 40.2 ± 4.0*

Adrenaline 10 -12.6 ± 3.7*

Table 6: Measures of diastolic function. Diastolic pressure before and after switching chloride solution to one of seven interventions. All values expressed as mean ± SEM. * indicates p < 0.05 for paired t test comparing value before and after crossover of solution. 125

i n 0 0 so m CM Os • • » ■ o © o o *“ o +1 + i +1 + i +1 +1 +1 Ui _ o m T f 00 OS 00 CL in r o in i n r o in tt O ® qna O ) « c CL C C 3 e tT m SI 2 Tt* o> m m ro 8 « n - so *— r - + i * - CD CM +i + i + i +1 + i + i in s ? s oo CM 0 0 Os o 0 0 in CM c ' B | r o so r - r r CM K> CM CM CM * o CM CM *? cm . 9 > ~ > t o j O in i n Os 0 0 to CO CD ° so • • o 3 > V d o O o o s Q- +1 + i +1 + i +1 +1 + i E r « a SO c - ® o 0 0 OS Ti to i n • • r * O *“ * HI CM O) CO * E CM > CD ^ OS 0 0 *— T f oo to m | S £ +1 + i +1 + i + i +1 +1 E c e r - f * o 0 0 T f ro CM P E in TT o m 0 0 0 0 m — in ™ CM CM CM CM CM ro * £ cm c 0-0 0 **- CD ^ 3 g so 0 0 0 0 n - n - 0 0 • • r>> g 8 1 • • • o • d o o o o o o } « a + i +1 + i o + i + i +1 + i O8 w«®E ^ CM _ m T f ro • • n - T J |® c • ■ ■ n - • c > ^ SO SO TT m CO CO 10 E •“ *“ ^ CD C in .EOF CM P co c CM so CM CM T f r o i i n •E > in cc +1 +1 +1 +1 + i +1 + i <2 . cm 3= SS-o m O OS V- 0 0 r o SO £ LLI £ o o Os OS o o os r o r o CM CM ro CM r o © +|c C 2? cco E o ■ c c d m ♦«* a) co p — CD C c c X 0) 0 0 ) a) 0) E 0 ) ■a TO a a> 3 ) <0 CO ^ C CL u CO TO k. xz c 0) E CO a) 0>m WCD 2 > O k_ ■o o 0 ) k . -Q ^ Q o s . JZ a) • o z — QQ u T © oc (5 2 1 126 affected heart rate. Verapamil caused a slight increase in coronary flow which was statistically significant when compared with the chloride group (in which flow declined slightly during the course of the experiment).

4. DISCUSSION

In a model which detected positive inotropic effects of adrenaline and negative inotropic effects of verapamil, together with differential effects on systolic and diastolic function, it was found that chloride surrogates had, in general, relatively minor effects on contractile function. A significant alteration in diastolic pressure by anion substitution was seen only in the iodide group. Iodide produced a significant elevation in mean diastolic pressure indicating an impairment of diastolic ventricular function. Bretag (Bretag 1987) has pointed out that the myotonic effect of iodide in frog skeletal muscle has long been known (Stockman and Charteris 1901) and that myotonia is also readily observed in rat skeletal muscle in vivo following Cl- channel blockade by 2,4- dichlorophenoxyacetate (2,4-D) (Eyzaguirre et al. 1948). Studies of twitch tension in strips of frog ventricular muscle (Anderson and Foulks 1973) have shown that anion substitution affects not only the rate of rise of tension but also the rate of relaxation; relaxation was slowed in the presence of sulphate, and increased in the presence of thiosulphate. This study suggests that iodide impairs diastolic relaxation in the whole heart, although to a lesser extent than pmolar verapamil! No other anions showed evidence of impairing diastolic function. A significant alteration in mean developed pressure by anion substitution was seen only in the methylsulphate group. This would be consistent with previous data from isolated heart muscle studies which have shown a positive inotropic effect of chloride substitution in non-mammalian (Anderson and Foulks 1973, Horackova and Vassort 1982) and mammalian (Nosek 1979) isolated heart muscle preparations. However, no significant change in E es was seen in this group, indicating that augmentation of systolic function by methylsulphate is probably small in whole heart preparations.

A significant reduction in mean E es by anion substitution was seen only in the nitrate group indicating some impairment of systolic function by NCS, although this was small compared to pmolar verapamil. As the developed 127 pressure was not affected, the effect of nitrate on inotropy appears to be small in the physiological range and would be unlikely to be a major factor in the non-working Langendorff model we have employed to study arrhythmias (Studyl). 128

DISCUSSION

1.NCS

The hypothesis that anions might play an important role in arrhythmogenesis and their manipulation might be antiarrhythmic was initially tested by substitution of the major extracellular anion, chloride, with the surrogate anion, nitrate. Nitrate was shown to have concentration-dependent antiarrhythmic action on both ischaemia- and reperfusion-induced ventricular arrhythmias which was greatest in the most severe arrhythmias. The actions were not secondary to changes in heart rate, coronary flow or occluded zone size and the inhibition of reperfusion-induced arrhythmias was present at 10, 15 and 30 min indicating a direct reduction in these arrhythmias rather than a delay in the development of sensitivity to them; that is, there was no rightward shift of the time susceptibility curve of reperfusion-induced arrhythmias (Fig. 13). There was evidence that NCS actually had a minor proarrhythmic role on reperfusion-induced arrhythmias after short (5 min) periods of ischaemia (leftward shift of time susceptibility curve). However these changes were small, only reaching significance in the case of VPB and the overwhelming effect of NCS was a reduction in the incidence of both ischaemia- and reperfusion-induced arrhythmias. Thus, a novel antiarrhythmic action (anion manipulation) was described and characterized in isolated rat heart preparations. The mechanism of action appeared to be mediated directly at the cellular level rather than an effect secondary to altered haemodynamic function.

2. Site of antiarrhythmic action of NCS

Having established the phenomenon of the antiarrhythmic action of NCS, further insight was gained by attempts to study its site of action. Initial studies of regionally ischaemic rat hearts (Study 2.1) in which nitrate was 'trapped' in the ischaemic zone suggested that the effect of NCS in reducing VF was mediated in part in the ischaemic zone but that its effect on other ventricular arrhythmias did not involve the ischaemic zone. This study was based on the premise that collateral flow in the rat heart is negligible (Johns and Olsen 1954, Maxwell, Hearse, and Yellon 1978, Winkler et al. 1984). This premise was tested in a 129 model of global ischaemia in which collateral flow was eliminated. Furthermore, this model allowed assessment of the mechanism of action of an antiarrhythmic intervention such as NCS by allowing effects mediated via amelioration of coronary flow to be discriminated from other mechanisms a priori. NCS caused a significant reduction in VT and VPB (Study 2.2) Thus, the site of antiarrhythmic action of NCS during ischaemia appeared to be mediated largely in the ischaemic zone as the presence of non-ischaemic tissue was not essential to observe its effect. Therefore, although small, it appears that the collateral flow in the rat heart is large enough to make its use in regional ischaemia cross-over studies unreliable. That a reduction was seen in VF with NCS trapping in the ischaemic zone, despite presumed wash out by chloride-containing solution, is interesting in that it suggests that NCS is most effective against the most severe ventricular arrhythmia, VF and that the arrhythmogenesis of VF may differ from other ventricular arrhythmias. The new model of global ischaemia (Study 2.2) allowed consideration of the relationship between ischaemia-induced VF and ischaemic zone sizes up to 100% (no-flow global ischaemia) and provided strong evidence to support the importance of a current of injury at the ’border zone' between ischaemic and non-ischaemic tissue in the genesis of ischaemia-induced VF. In contrast, other mechanisms appeared to be responsible for ischaemia-induced VPB and VT as these were not reduced in the new model of global ischaemia. Using the new model of global ischaemia (right intra-atrial superfusion), the incidences of reperfusion-induced VT and VF were as high after 30 min global ischaemia as after regional ischaemia, indicating that reperfusion-induced VT and VF are initiated within the reperfused tissue and excluding the possibility of an interaction with adjacent non-ischaemic tissue (e.g. current of injury theory). This was consistent with an earlier study (Curtis and Hearse 1989b) employing 10 min global ischaemia which also found that VF was initiated within the reperfused tissue. Although indicating similar syncytial mechanisms of action with reperfusion after 10 and 30 min of ischaemia, this does not necessarily mean that the underlying pathophysiological mechanisms are the same; there is evidence that duration of ischaemia can alter the response of reperfusion-induced arrhythmias to interventions. Hearse and Tosaki (Hearse and Tosaki 1988) demonstrated that 130 the spin trap agent N-tert-butyl-alpha-phenylnitrone (PBN) caused a significant reduction in reperfusion-induced VF after 10 but not 20 min ischaemia, whereas calcium reduction caused a significant reduction in VF after 20 but not 10 min ischaemia. Employing an established model of global ischaemia (atrial, not intra-atrial, superfusion) (Curtis and Hearse 1989b) cross-over studies (Study 2.3) indicated that the effect of NCS was mediated by a direct action of the reperfusate on the reperfused tissue rather than an indirect action on the ischaemic tissue prior to reperfusion. In summary, studies of the site of action of NCS indicated that ischaemia-induced VF was inhibited largely by an action in the ischaemic zone and that reperfusion-induced arrhythmias were reduced largely by a direct action of the reperfusate on the reperfused zone.

3. Possible cellular mechanisms of action of NCS

Of three possible mechanisms proposed to explain the antiarrhythmic action of NCS, we have presented data which does not support a major role for modification of pH or cyclic nucleotide homeostasis but which does support modification of membrane resistance as a possible mechanism of action. The inhibition of ischaemia- and reperfusion-induced ventricular arrhythmias by NCS can also be produced by other anion substitutions. It has been shown that the antiarrhythmic action of chloride surrogates correlates with their known order of sarcolemmal permeability demonstrating this to be a determinant of their antiarrhythmic activity (Study 3.1). Indeed, the anion, methylsulphate, with membrane permeability lower than chloride was shown to be proarrhythmic. Anion substitution appears to inhibit ventricular arrhythmias by a direct cellular action which is associated with, but not necessarily caused by, a mechanism causing sustained prolongation of the QRST width reminiscent of a class III effect (Fig. 27). Although nitrate and iodide are well documented to be the most permeable of the anions studied, their rank order is controversial. Hutter and Noble (Hutter and Noble 1961) found iodide to be more permeable. Our findings would be more consistent with Seyawa’s more recent findings that nitrate is 131 more permeable than iodide (Seyama 1979) as, although both abolished ischaemia-induced VF, nitrate was more effective against ischaemia-induced salvos, BG and VT and marginally more effective against reperfusion-induced VT and VF. Investigation of cAMP and cGMP (Study 3.2) did not suggest that changes in myocardial cyclic nucleotide homeostasis contributed to the antiarrhythmic effects of anion manipulation. In addition, this study did not suggest a major role for cyclic nucleotides in arrhythmogenesis in the isolated rat heart during ischaemia or reperfusion. Whether cAMP or cGMP elevation occurs in response to anion manipulation in other species is not established from the present study. However, we suggest that any such changes, should they occur, are not a necessary mechanism for the prevention of arrhythmias, as shown by the findings in the rat in the present study. The theory that pH elevation might be involved in the mechanism of the antiarrhythmic effect of NCS by blocking the chloride-bicarbonate exchange mechanism (Curtis 1989, Vaughan-Jones 1979b) was not supported by our NMR study (Study 3.3). NCS actually caused a significant decrease in pH together with evidence of heterogenous regions of low intracellular pH. Although primarily designed to consider the pH effects of NCS, this study also gave insight into the metabolic effect of NCS. The reduction in PCr and elevation of Pi indicated that NCS impaired metabolic function although no insight was gained into the functional significance of this effect on the whole heart. Little consideration has previously been given to the effect of anion manipulation in the whole heart although studies in isolated ventricular muscle strips suggested that anion substitution might actually be positively inotropic (Anderson and Foulks 1973, Horackova and Vassort 1982, Nosek 1979). Our initial study of NCS (Study 1) did not reveal major alterations in the haemodynamic parameters measured (coronary flow, heart rate), but a more specific measure of ventricular function was required. We therefore studied the effect of anion surrogates on the systolic and diastolic function of isolated rat hearts using isochoric intraventricular balloons (Study 3.4). This model demonstrated marked impairment and improvement (in both diastolic and systolic function) with verapamil and adrenaline, respectively. By contrast, changes in ventricular function with complete isotonic anion substitution were relatively small. 132

Overall, isochoric balloon studies (Study 3.4) did not support the hypothesis that anions might represent potential positively inotropic, antiarrhythmic agents. However, of the anions which have been shown to be antiarrhythmic, bromide did not demonstrate any negative inotropic effect, iodide demonstrated impaired diastolic but not systolic activity and nitrate showed a reduction in only one of the two measures of ventricular systolic function (E e s )- Complete NCS has little negative inotropic effect at physiologic diastolic pressures but has an effect on the slope relationship between ventricular volume and systolic pressure. The latter is unlikely to have any bearing on arrhythmogenesis in these studies as the hearts in the Langendorff mode were non-working. In essence, the antiarrhythmic chloride surrogates can ameliorate arrhythmias while producing relatively inconsequential effects on contractile function. Our primary concern that antiarrhythmic activity is intimately connected with substantial negative inotropic activity was shown to be unfounded. Thus the principle of anion manipulation as an antiarrhythmic intervention remains viable.

4. Clinical implications

An understanding of the mechanisms of ischaemia- and reperfusion-induced ventricular arrhythmias is important as, to date, there is no pharmacological agent available for routine use in the prevention of ischaemia-induced ventricular fibrillation. 3-blockers may confer some protection following myocardial infarction (Norwegian Multicentre Study Group 1981) and the automatic implantable defibrillator (Tchou et al. 1988) represents a new and effective, although invasive and expensive, method of managing life-threatening VF in some groups of patients. However, the incidence of sudden death in patients with ischaemic heart disease remains high and acute ventricular arrhythmias are probably responsible for most of these deaths (Gordan and Kannel 1971, Richards et al. 1991). Although, much attention has been devoted to the development of antiarrhythmic agents which act by modifying cation homeostasis, little attention has been given to the possible therapeutic effects of anion manipulation. To our knowledge, the only therapeutic antiarrhythmic agent which was considered to act on anion homeostasis was the clonidine derivative, alinidine which has been suggested to exert its bradycardic and antiarrhythmic effects via anion 133 antagonism (Millar and Vaughan Williams 1981a) and thus to represent a fifth class of antiarrhythmic action (Millar and Vaughan Williams 1981b). However, the mechanism of action of this agent remains controversial. As pointed out by Bretag (Bretag 1987) alinidine is structurally unlike any known chloride channel blocker and its guanidium moiety might be more expected to interact with sodium channels. Furthermore, the bradycardic effect of alinidine is not seen with anion substitution in our studies (Study 1) and an alternative hypothesis suggests that alinidine may inhibit i f , the hyperpolarization-activated Na+-K+ current implicated in pacemaker activity in the sinoatrial node. The clinical role of anions in arrhythmogenesis during myocardial ischaemia and reperfusion remains unknown. It would be premature to suggest that the studies described in this thesis will have a clinical application. However, they represent a potentially novel approach for the development of new drugs to control arrhythmias which are outside the four classes of antiarrhythmic action described by Vaughan Williams (Vaughan Williams 1970, Vaughan Williams 1984). Similarly the optimum desirable nitrate content of drinking water is unknown. Nitrates are generally considered to be undesirable in drinking water but it is possible that nitrates in drinking water might be cardioprotective (Pocock et al. 1980, Ridley and Curtis 1990c).

5. Species considerations and future progress

As in all studies based on animal experimentation, the extent to which results can be applied to man is unknown. Obvious ethical constraints limit the extent to which arrhythmias can be studied clinically. The demonstration of an effect in more than one species, although not definite proof of applicability to man, lends greater significance to the effect. The systematic investigation of the genesis and control of arrhythmias during ischaemia and reperfusion requires a reproducible preparation in which arrhythmias can be induced with relative ease. The rat heart is used for investigating arrhythmias resulting from ischaemia (Curtis, MacLeod, and Walker 1987) and reperfusion (Manning and Hearse 1984) because its paucity of collaterals, compared to other species (Johns and Olsen 1954, Maxwell, Hearse, and Yellon 1978), leads to reproducible zones of ischaemia upon ligation of a coronary artery and because large numbers of animals can be used 134 to generate quantitative information, without inordinate financial cost. The rat offers many advantages but is criticized on the basis of its atypical action potential (absence of separate T wave) and an additional small animal model would be desirable. The guinea-pig, because of its highly collateralized coronary circulation, cannot be used for induction of regional ischaemia and the rabbit is often dismissed on the grounds that it has a low incidence of arrhythmias. However, if perfusion conditions (duration of ischaemia and extracellular K+ concentration) are selected then the rabbit heart has been shown to exhibit reproducible susceptibility to reperfusion arrhythmias (MacConaill 1987). Our current facilities preclude the use of dogs, cats or pigs. Present Home Office rules makes the aquisition of new licences for such work difficult. We would also be disinclined to use dogs owing to the limited use of this species as a bioassay as a consequence of its variable collateral anatomy (Trolese- Mongheal et al. 1985) or pigs owing to inexplicable variability in arrhythmia susceptibility (Walker and MacLeod, personal communication). In addition, cost would prohibit the use of these species even if facilities were available. We plan to do some Langendorff cat studies in collaboration with a non-UK institute for which such studies are legal, inexpensive and feasible. However, with respect to the present, we do now have some rabbit data from Langendorff perfused rabbits with double coronary ligation to encourage ventricular arrhythmias. Unfortunately, we found that no hearts developed VF with ischaemia or reperfusion in the control group, thus precluding assessment of the antifibrillatory effect of NCS (VF incidence was zero also with NCS). The study was randomized to lines; after 6 lines had been completed and it had become apparent that no VF was occurring we abandoned the study and broke the codes. The main finding was that classes of arrhythmias for which control incidence was high (VPB during ischemia, VPB and VT during reperfusion) exhibited suppression when NCS was used. For example, the incidence of ischaemic VPBs was reduced from 100% to 50%, and reperfusion VT was reduced from 66% to 17%, similar effects to those observed in rats. Occluded zone sizes (42±5 and 41 ±6) were also similar to those in rat studies. Thus, NCS appears to work in rabbits. However, we feel reluctant to include this data in the main body of this thesis since we cannot explain the lack of VF in the rabbit controls. We are planning to look at this question soon, but there is no evidence 135 that it relates to anion manipulation and it is therefore not relevant to this thesis. There is considerable scope for further investigation into the effects of anion manipulation on the heart. Currently in vivo studies of arrhythmias in rats fed on a high nitrate-low chloride diet are in progress under the supervision of Dr M.J. Curtis (Cardiovascular Research Laboratories, Pharmacology Group Division of Medical Sciences, King's College, University of London) to determine if antiarrhythmic activity can be demonstrated in vivo and be affected by diet in this animal model. It is also hoped that further collaboration with Dr Pamela Garlick (Department of Radiological Sciences, Guy's Hospital, St. Thomas' Street, London SE1) will be possible to employ NMR spectroscopy to investigate the concentration-dependent pH and metabolic changes caused by NCS and also to study the effect of other anion surrogates on pH and metabolic changes. Some anions are not suitable chloride surrogates (for example, the insolubility of calcium fluoride means that substitution with fluoride is impracticable) but the literature contains numerous examples of other potential monovalent anion surrogates which might be screened in the isolated rat heart for antiarrhythmic activity; from the present studies we would expect those anions with greatest permeability to be the most antiarrhythmic.

6. Conclusion

We have demonstrated the concentration-dependent antiarrhythmic action of NCS. This effect appears to be independent of haemodynamic variables of rate and flow. Ischaemia-induced arrhythmias are inhibited largely by an action in the ischaemic zone. Reperfusion-induced arrhythmias are inhibited largely by a direct action of reperfusate on the reperfused tissue. We have presented evidence which disfavours a role for pH or cyclic nucleotide manipulation in the mechanism of action of NCS. Substitution with other anions suggests that the mechanism of antiarrhythmic action of NCS and other anions (iodide and bromide) is dependent upon their membrane permeabilities; the greater the membrane permeability, the greater the antiarrhythmic effect. Therefore, their mechanism of action probably involves reduction of membrane resistance. We have described a novel mechanism of antiarrhythmic activity (anion manipulation) which might be used to develop new therapeutic options in the management of sudden cardiac death. 136

ACKNOWLEDGEMENTS

I am grateful for the assistance of the following;

Dr. Michael J. Curtis (Cardiovascular Research Laboratories, Pharmacology Group Division of Medical Sciences, King's College, University of London, Manresa Road, London. SW3 6LX) for supervision at all stages of this thesis.

Dr. Pamela Garlick (Department of Radiological Sciences, Guy's Hospital, St. Thomas' Street, London SE1) for assistance with the NMR study.

Prof. David J. Hearse (The Rayne Institute, St Thomas' Hospital, London. SE1 7EH) for provision of facilities to perform study 1.

Prof. Magdi H. Yacoub (Department of Cardiac Surgery, Harefield Hospital, Harefield, Middlesex. UB9 6JH) for provision of facilities and finance to perform subsequent studies.

British Heart Foundation for support with a one year junior research fellowship (F219) in the department of cardiac surgery at Harefield Hospital. 137

REFERENCES

Adgey, AAJ, JD Allen, JS Geddes, RGG James, SW Webb, SA Zaidl, and JF Pantridge. 1971. Acute phase of myocardial infarction. Lancet i : 501-504.

Allen, DG, and CH Orchard. 1987. Myocardial contractile function during ischaemia and hypoxia. Circ Res 60 :153-168.

Allessie, MA, FI Bonke, and FJG Schopman. 1977. Circus movement in rabbit atrial muscle as a mechanism of tachycardia. III. The 'leading circle' concept: a new model of circus movement in cardiac tissue without the involvement of an anatomical obstacle. Circ Res 41 : 9-18.

Anderson, ER , and JG Foulks. 1973. Effects of anions on frog ventricle. Can J Physiol Pharmacol 51 : 709-726.

Anderson, PG, MF Allard, GD Thomas, SP Bishop, and SB Digerness. 1990. Increased ischaemic injury but decreased hypoxic injury in hypertrophied rat hearts. Circ Res. 67. : 948-959.

Apstein Jr, CS, M Mueller, and WB Hood. 1977. Ventricular contracture and compliance changes with global ischaemia and reperfusion, and their effect on coronary resistance in the rat. Circ Res 41 : 206-217.

Armitage, F, and G Berry. 1987. Statistical methods in medical research. 239- 243. Oxford: Blackwell Scientific Publications.

Austin, M, TL Wenger, FE Harrell Jr, FA Luzzi, and HC Strauss. 1982. Effect of myocardium at risk on outcome after coronary artery occlusion and release. Am J Phvsiol 243 : H340-H345.

Bacaner, MB, JR Clay, A Schrier, and RM Brochu. 1986. Potassium channel blockade: A mechanism for suppressing ventricular fibrillation. Proc Natl Acad Sci USA 83 : 2223-2227. 138

Bahinski, A, AC Nairn, P Greengard, and DC Gadsby. 1989. Chloride conductance regulated by cyclic AMP-dependent protein kinase in cardiac myocytes. Nature 340 : 718-721.

Bailey, IA, SR Williams, GK Radda, and DG Gadian. 1981. Activity of phosphorylase in total global ischaemia in the rat heart. A phosphorus-31 nuclear magnetic resonance study. Biochem J 196 :171-178.

Balke, CW, E Kaplinsky, EL Michelson, M Naito, and LS Dreifus. 1981. Reperfusion ventricular tachyarrhythmias: correlation with antecedent coronary artery occlusion tachyarrhythmias and duration of myocardial ischaemia. Am Heart J 101 : 449-456.

Basingthwaite, JB, CH Fry, and JAS McGuigan. 1976. Relationships between internal calcium and and outward current in mammalian ventricular muscle; a mechanism for the control of the action potential duration? J Physiol 262 :15-37.

Bennett, BM, H Schroder, LD Hayward, SA Waldman, and F. Murad. 1988. Effect of in vitro organic nitrate tolerance on relaxation, cyclic GMP accumulation, and guanylate cyclase activation by glyceryl trinitrate and the enantiomers of isoidide dinitrate. Circ Res 63 : 693-701.

Beresford, SAA. 1985. Is nitrate in the drinking water associated with the risk of cancer in the urban UK? Int J Epidemiol 14 : 57-63.

Bernier, M, MJ Curtis, and DJ Hearse. 1989. Ischaemia-induced and reperfusion- induced arrhythmias: importance of heart rate. Am J Physiol 256 : H21-H31.

Bigger Jr, JT, JL Fleiss, R Kleiger, JP Miller, LM Rolnitsky, and The Multicentre Post-Infarction Research Group. 1984. The relationships among ventricular arrhythmias, left ventricular dysfunction and mortality in the 2 years after myocardial infarction. Circulation 69 : 250-258.

Blumgart, HL, DR Gilligan, and MJ Schlesinger. 1941. Experimental studies on the effect of temporary occlusion of coronary arteries. Am Heart J 22 : 374-388. 139

Bolli, R, DJ Fisher, and ML Entmann. 1986. Factors that determine the occurrence of arrhythmias during acute myocardial ischaemia. Am Heart J 111 : 261-270.

Bretag, AH. 1987. Muscle chloride conductance. Physiol Rev 67 : 618-724.

Campbell, RWF, A Murray, and DG Julian. 1981. Ventricular arrhythmias in the first 12 hours of acute myocardial infarction: natural history study. Br Heart J 46 : 351-357.

Carmeleit, E. 1961. Chloride ions and the membrane potential of Purkinje fibres. J Phvsiol 156 :375-388.

Carmeleit, E. 1978. Cardiac transmembrane potentials and metabolism. Circ Res 42 : 577-587.

Carmeleit, E, and F Verdonck. 1977. Reduction of potassium permeability by chloride substitution in cardiac cells. J Phvsiol 265 : 193-206.

CAST Investigators. 1989. Preliminary report: Effect of encainide and flecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. N Ena J Med 321 : 406-412. *

Clusin, WT, M Buchbinder, and DC Harrison. 1983. Calcium overload, "injury current” and ischaemic cardiac arrhythmias-a direct connection. Lancet i : 272- 274.

Coker, SJ, JR Parratt, IM Ledingham, and IJ Zeitlin. 1980. Thromboxane and prostacycline release from ischaemic myocardium in relation to arrhythmias. Nauture 291 :323-324.

CONSENSUS Trial Study Group. 1987. Effects of enalapril on mortality in severe congestive heart failure: Results of the Co-operative North Scandinavian Enalapril Survival Study (CONSENSUS). N Eng J Med 316 : 1429-1435. 140

Corr, PB, RW Gross, and BE Sobel. 1984. Amphipathic metabolites and membrane dysfunction in ischaemic myocardium. Circ Res 55 :135-154.

Corr, PB, FX Witkowski, and BE Sobel. 1978. Mechanisms contributing to malignant dysrhythmias induced by ischaemia in the cat. J Clin Invest 61 :109- 119.

Curtis, MJ. 1986. The actions of calcium antagonists on arrhythmias and other responses to myocardial ischaemia in the rat. PhD Thesis. University of British Columbia. Canada

Curtis, MJ. 1989. Inhibition of chloride-bicarbonate exchange: a new approach to the attenuation of ischaemia-induced arrhythmias. J. Mol Cell Cardiol 21 (Suppl II) : 421.

Curtis, MJ. 1991. Extracellular phosphate and arrhythmogenesis. Br J Pharmacol 102 : P200.

Curtis, MJ, and DJ Hearse. 1989a. Ischaemia-induced and reperfusion-induced arrhythmias differ in their sensitivity to potassium: implications for mechanism of initiation and maintenance of ventricullar fibrillation. J Mol Cell Cardiol 21 : 21- 40.

Curtis, MJ, and DJ Hearse. 1989b. Reperfusion-induced arrhythmias are critically dependent upon occluded zone size: relevance to the mechanism of arrhythmogenesis. J Mol Cell Cardiol 21 : 625-637.

Curtis, MJ, BA MacLeod, R Tabrizchi, and MJA Walker. 1986. An improved apparatus for small animal hearts. J Pharm Methods 15 : 87-94.

Curtis, MJ, BA MacLeod, and MJA Walker. 1987. Models for the study of arrhythmias in myocardial ischaemia and infarction: the use of the rat. J Mol Cell Cardiol 19 : 399-419. 141

Curtis, MJ, PD Ridley, and DJ Hearse. 1990. Anions and arrhythmogenesis: proarrhythmic action of chloride substitution by nitrate on arrhythmias induced by reperfusion after a short period of ischaemia. J. Mol Cell Cardiol 22 (Suppl III) : S.75.

Curtis, MJ, MJA Walker, and DJ Hearse. 1989. Calcium antagonists in myocardial ischaemia and reperfusion. Eur J Clin Pharmacol 36 : A20-A21.

Czarnecka, M, B Lewartowski, and A Propopezyk. 1973. Intracellular recording from the in situ working heart in physiological conditions and during acute ischaemia and fibrillation. Acta Phvsiol Pol 24 : 331-337.

Davies, MJ. 1981. Pathological view of sudden cardiac death. Br Heart J 45 : 88- 96.

De Mello, WC. 1963. Role of chloride ions in cardiac action and pacemaker potentials. Am J Phvsiol 205 : 567-575.

Desilets, M, and CM Baumgarten. 1986. K+, Na+ and Cl- activities in ventricular myocytes isolated from rabbit heart. Am J Phvsiol 251 : C197-C208.

Diamond, JM, and EM Wright. 1969. Biological membranes: the physical basis of ion and nonelectrolyte selectivity. Ann Rev Physiol 31 : 581-646.

Dobson Jr, JG, and SE Mayer. 1973. Mechanisms of activation of cardiac glycogen phosphorylase in ischaemia and anoxia. Circ Res 33 : 412-420.

Downar, E, MJ Janse, and D Durrer. 1977. The effect of acute coronary artery occlusion on subepicardial transmembrane potentials in the intact porcine heart.

Circulation 5 6 : 2 1 7 -2 2 4 .

Dukes, ID, and M Morad. 1989. Tedisamil inactivates transient outward K+ current in rat ventricular myocytes. Am J Phvsiol 257 : H1746-H1749.

Editorial. 1987. EDRF. Lancet ii : 137-138. 142

Editorial. 1988. Yin and yang of vasomotor control. Lancet ii : 19-20.

Eisenman, G. 1961. On the elementary atomic origin of equilibrium ionic specificity. In Symposium on membrane transport and metabolism. Edited by A. Kleinzeller and A. Kotyk. 163-179. New York: Academic Press.

El-Sherif, N, BJ Scherlag, and R Lazzara. 1975. Electrode catheter recordings during malignant ventricular arrhythmias following experimental acute myocardial ischaemia. Evidence for re-entry due to conduction delay and block in ischaemic myocardium. Circulation 51 : 1003-1014.

El-Sherif, N, BJ Scherlag, R Lazzara, and RR Hope. 1977. Re-entrant ventricular arrhythmias in the late myocardial infarction period: 1.Conduction characteristics in the infarction zone. Circulation 55 : 686-702.

Elford, J, AN Phillips, AG Thomson, and AG Shaper. 1989. Migration and geographic variations in ischaemic heart disease in Great Britain. Lancet 1 : 343- 346.

Endo, T, LGT Ribeiro, WM Cheung, DB Faria, M Petranto, and PR Maroko. 1983. Relationship between the extent of the hypoperfused zone of the myocardium and the occurrence of ventricular fibrillation. Am Heart J 105 : 915-920.

Eyzaguirre, C, BP Folk, Kl Zierler, and JL Lilienthal. 1948. Experimental myotonia and repetetive phenonena: the veratrinic effects of 2,4- dichlorphenoxyacetate (2,4-D) in the rat. Am J Physiol 155 : 69-77.

Ferrier, GR, MP Moffat, and A Lukas. 1985. Possible mechanisms of ventricular arrhythmias elicited by ischaemia followed by reperfusion. Studies on isolated canine ventricular tissue. Circ Res 56 :184-194.

Forman, D, S Al-Dabbagh, and R Doll. 1985. Nitrates, nitrites and gastric cancer in Great Britain. Nature 313 : 620-625. 143

Forstermann, U, A Mulsch, E Bohme, and R Busse. 1986. Stimulation of soluble guanylate cyclase by an acetylcholine-induced endothelium-derived factor from rabbit and canine arteries. Circ Res 58 : 531-538.

Fozzard, HA, and CO Lee. 1976. Influence of changes in external potassium and chloride ions on membrane potential and intracellular potassium ion activity in rabbit ventricular muscle. J Physiol 256 : 663-689.

Fry, CH, and PA Poole-Wilson. 1981. Effects of acid-base exchange on excitation-contraction coupling in guinea-pig and rabbit cardiac ventricular muscle. J Physiol 313 : 141-160.

Fujimoto, T, T Peter, H Hamamoto, and WJ Mandel. 1983. Electrophysiological observations on ventricular tachyarrhythmias following reperfusion. Am Heart J 105 : 201-209.

Garlick, PB, GK Radda, and PJ Seely. 1979. Studies of acidosis in the ischaemic heart by phosphorus nuclear magnetic resonance. Biochem J 184 : 547-554.

Garner, JA, M Bernier, and DJ Hearse. 1989. R56865 abolishes reperfusion- induced arrhythmias in vivo without altering hemodynamics. Circulation 80 (Suppl II) : 1993.

Garthwaite, J, SL Charles, and R Chess-Williams. 1988. Endothelium-derived relaxing factor release on activation of NMDA receptors suggests role as intercellular messenger in the brain. Nature 336 : 385-388.

Gaspardone, A, Kl Shine, SR Seabrooke, and PA Poole-Wilson. 1986. Potassium loss from rabbit myocardium during hypoxia: Evidence for passive efflux linked to anion extrusion. J Mol Cell Cardiol 18 : 389-399.

Goldberg, S, AJ Greenspan, PL Urban, B Muza, B Berger, P Walinsky, and PR Maroko. 1983. Reperfusion arrhythmia: A marker of restoration of antegrade flow during intracoronary thrombolysis for acute myocardial infarction. Am Heart J 144

105:26-32.

Goldman, DE. 1943. Potential impedence and rectification in membranes. J Gen PhySiQl 27 : 37-60.

Goldstein, S. 1990. Toward a new understanding of the -mechanism and prevention of sudden death in coronary heart disease. Circulation 82 : 284-288.

Gordan, T, and WB Kannel. 1971. Premature mortality from coronary heart disease. The Framingham Study. J Am Med Ass 215 : 1617-1625.

Griffith, T, and M Randall. 1989. Nitric oxide comes of age. Lancet ii : 875-876.

Halpern, MH. 1957. The dual supply of the heart. Am J Anat 101 :1-16.

Han, J, BG Goel, and CS Hanson. 1970. Re-entrant beats induced in the ventricle during coronary occlusion. Am Heart J 80 : 778-784.

Harris, AS. 1948. Terminal electrocardiographic patterns in experimental anoxia, coronary occlusion and haemorrhagic shock. Am Heart J 35 : 895-909.

Harris, AS, A Bisteni, RA Russell, JC Brigham, and JE Firestone. 1954. Excitatory factors in ventricular tachycardia resulting from myocardial ischaemia: Potassium a major excitant. Science 119: 200-203.

Harvey, RD, and JR Hume. 1989. Autonomic regulation of chloride current in heart. Science 244 : 983-985.

Hearse, DJ, and A Tosaki. 1988. Free radicals and calcium: Simultaneous interacting triggers as determinants of vulnerability to reperfusion-induced arrhythmias in the rat heart. J Mol Cell Cardiol 20 : 213-223.

Heimberger, RF. 1946. Injection into pericardial sac and'ligation of coronary artery of rat. Arch Surg 52. : 677-689. 145

Hinkle Jr, LE, and HT Thaler. 1982. Clinical classification of cardiac deaths. Circulation 65 (3) : 457-464.

Hirche, HJ, C Franz, L Bos, R Bissig, R Lang, and M Schramm. 1980. Myocardial extracellular K+ and H+ increase and noradrenaline release as possible cause of early arrhythmias following acute coronary artery occlusion in pigs. J Mol Cell Cardiol 12 : 579-593.

Hodgkin, AL, and P Horowicz. 1959. The influence of potassium and chloride ions on the membrane potential of single muscle fibres. Physiol 148 :127-160.

Hodgkin, AL, and B Katz. 1949. The effect of sodium loss on the electrical acticity of the axon of the giant squid. J Physiol 108 : 37-77.

Hoffman, BF, and MR Rosen. 1981. Cellular mechanisms for cardiac arrhythmias. Circ Res 49 :1-15.

Honig, CR. 1981. Modern cardiovascular physiology. 1st ed. Boston,USA: Little, Brown and Company.

Horackova, M, and G Vassort. 1982. Effects of chloride replacement in cardiac muscle: Mechanisms of positive inotropy. J Mol Cell Cardiol 14 :123-130.

Humphrey, SH, and PB Garlick. 1991. NMR-visible ATP and Pi in normoxic and reperfused rat hearts: a quantitative study. Am J Physiol 260 : H6-H12.

Hutter, OF, and D Noble. 1961. Anion conductance of cardiac muscle. J Physiol 157 : 335-350.

Ignarro, LJ, TM Burke, KS Wood, MS Wolin, and PJ Kadowitz. 1984. Association between cyclic GMP accumulation and acetylcholine-elicited relaxation of bovine intrapulmonary artery. J Pharmacol Exp Ther 228 : 682-690.

Ikemoto, Y, and M Goto. 1977. Effects of ACh on slow inward current and tension components of the bullfrog atrium. J Mol Cel Cardiol 9 : 313-326. 146

Janse, MJ, J Cinca, H Morena, JWT Fiolet, AG Kleber, GP deVries, AE Becker, and D Durrer. 1979. The “border zone” in myocardial ischaemia: An electrophysiological, metabolic and histochemical correlation in the pig heart. Circ Res 44 : 576-588.

Janse, MJ, FJL Van Capelle, H Morsink, AG Kleber, F Wilms-Schopman, R Cardinal, CN Naumann D'Alnoncourt, and D Durrer. 1980. Flow of "injury" current and patterns of excitation during early ventricular arrhythmias in acute regional myocardial ischaemia in isolated porcine and canine hearts. Circ Res 47 : 151-165.

Janse, MJ, and AL Wit. 1989. Electrophysiological mechanisms of ventricular arrhythmias resulting from myocardial ischaemia and infarction. Phys Rev 69 : 1049-1169.

Johns, TNP, and BJ Olsen. 1954. Experimental myocardial infarction. A method of coronary occlusion in small animals. Ann Surg 140 : 675-682.

Johnston, KM, BA MacLeod, and MJA Walker. 1983. Responses to ligation of a coronary artery in conscious rats and the actions of antiarrhythmics. Can J Phvsiol Pharmacol 61 : 1340-1353.

Kannengeisser, GJ, WF Lubbe, and LH Opie. 1975. Experimental myocardial infarction with left ventricular failure in the isolated perfused rat heart. Effects of isoproterenol and pacing. J Mol Cell Cardiol 7 : 135-151.

Kardesh, M, CE Hogancamp, and RJ Bing. 1958. The effect of complete ischaemia on the intracellular electrical activity of the whole mammalian heart. Circ Res 6 : 715-720.

Katsuki, S, WP Arnold, and F Murad. 1977. Effects of sodium nitroprusside, nitroglycerin and sodium azide on levels of cyclic nucleotides and mechanical activity of various tissues. J Cycl Nuc Res 3 : 239-247. 147

Kenyon, JL, and WR Gibbons. 1977. Effects of low-chloride solutions on the action potentials of sheep cardiac Purkinge fibres. J Gen Phys 70 : 635-660.

Koltai, M, I Lepran, G Nemecz, and L Szekeres. 1982. Experimental myocardial infarction in adjuvant arthritic rats. .J Cardiovasc Pharmacol. 4 :363-369.

Lamprecht, W, P Stein, F Heinz, and H Weisser. 1974. Creatine phosphate; determination with creatine kinase, hexokinase and glucose-6-phosphate dehydrogenase. In Methods of Enzvmatic Analysis (2nd edition). Edited by H. Bergmeyer. 1777-1785. New York: Academic Press.

Langendorff, O. 1895. An examination of the viability of the perfused heart. Pflug Arch Ges Phvsiol 61 : 291-332.

Lewis Jr, HD, JW Davis, DG Archibald, WE Steinke, TC Smitherman, J Doherty, HW Schnaper, MM LeWinter, E Linares, JM Pouget, SC Sabharwal, E Chesler, and H DeMots. 1983. Protective effects of aspirin against acute myocardial infarction and death in men with unstable angina. N Eng J Med 309 : 396-403.

Lo, YSA, JE Cutler, A Wright, J Kron, K Blake, and CD Swerdlow. 1988. Long- segment coronary ulcerations in survivors of sudden cardiac death. Am Heart J 116:1444-1447.

Lovegrove, T, and P Thompson. 1978. The role of acute myocardial infarction in sudden death - a statistician's nightmare. Am Heart J 96 : 711-713.

Lowry, OH, NJ Rosebrough, AL Farr, and RJ. Randall. 1951. Protein measurement with the Folin phenol reagent. J Biol Chem 193 : 265-275.

Lubbe, WF, OL Bricknell, T Podzuweit, and LH Opie. 1976. Cyclic AMP as a determinant of vulnerability to ventricular fibrillation in the isolated rat heart. Cardiovasc Res 10 : 697-702.

Lubbe, WF, T Podzuweit, PS Daries, and LH. Opie. 1978. The role of cyclic adenosine monophosphate in adrenergic effects on ventricular vulnerability to 148 fibrillation in the isolated perfused rat heart. J Clin Invest 61 :1260-1269.

MacConaill, M. 1987. Ventricular fibrillation thresholds in Langendorff perfused rabbit hearts: all or none effect of low potassium concentration. Cardiovasc Res 21 : 463-468.

MacLeod, BA, R McGroarty, RH Morton, and MJ Walker. 1989. Effects of halothane on arrhythmias induced by myocardial ischaemia. Can J Anaesth 36 : 289-294.

Mainland, D, L Harrera, and Ml Sutcliffe. 1956. Statistical tables for use with binomial samples - contingency tests.confidence limits and sample size estimates. New York: University College of Medicine Publications.

Manning, AS, and DJ Hearse. 1984. Reperfusion-induced arrhythmias: Mechanism and prevention. J Mol Cell Cardiol 16 : 497-518.

Manning, AS, K Kinishita, E Buschmans, DJ Coltart, and DJ Hearse. 1985. The genesis of arrhythmias during myocardial ischaemia: Dissociation between changes in cyclic adenosine monophosphate and electrical instability in the rat. Circ Res 57 : 668-675.

Maseri, A, S Severi, and P Marzullo. 1982. Role of coronary arterial spasm in sudden coronary ischaemic death. Ann N Y Acad Sci 382 : 204-216.

Matsuoka, S, T Ehara, and A. Noma. 1990. Chloride-sensitive nature of the adrenaline-induced current in guinea-pig cardiac myocyte. J Physiol 425 : 579- 598.

Maxwell, MP, DJ Hearse, and DM Yellon. 1978. Species variation in the coronary collateral circulation during regional myocardial ischemia: a critical determinant of the rate of evolution and extent of myocardial infarction. Cardiovasc Res 21 : 737-746.

Millar, JS, and EM Vaughan Williams. 1981a. Anion antagonism-a fifth class of 149 antiarrhythmic action? Lancet i : 1291-1293.

Millar, JS, and EM Vaughan Williams. 1981b. Pacemaker selectivity: influence on rabbit atria of ionic environment and of alinidine, a possible anion antagonist. Cardiovasc Res 15 : 335-350.

Milner-White, EJ, and DC Watts. 1971. Inhibition of adenosine 5’-triphosphate- creatine phosphotransferase by substrate-anion complexes. Biochem J 122 : 727-740.

Mines, RG. 1913. On dynamic equilibrium in the heart. J Physiol 46 : 350-383.

Moon, RB, and JH Richards. 1973. Determination of intracellular pH by phosphorus nuclear magnetic resonance. J Biol Chem 248 : 726-728.

Murdock, DK, JM Loeb, DE Euler, and WC Randall. 1980. Electrophysiology of coronary reperfusion. A mechanism for reperfusion arrhythmias. Circ 61 :175- 182.

Naito, M, E Michelson, J Kmetzo, E Kaplinsky, and L Dreifus. 1981. Failure of antiarrhythmic drugs to prevent experimental reperfusion ventricular fibrillation. Circulation 63 : 70-79.

Nakata, T, DJ Hearse, and MJ Curtis. 1990. Are reperfusion-induced arrhythmias caused by disinhibition of an arrhythmogenic component of ischaemia? J Mol Cell Cardiol 22 : 843-858.

Nattel, S, and D Waters. 1990. What is an antiarrhythmic drug? From clinical trials to fundamental concepts. Am J Cardiol 66 : 96-99.

Norwegian Multicentre Study Group. 1981. Timolol-induced reduction in mortality and reinfarction in patients surviving acute myocardial infarction. N Eng J Med 304:801-807.

Nosek, TM. 1979. Effects of chloride on the electrical and mechanical properties 150 of guinea pig ventricle. Pflugers Archiv. Eur J Physiol 381 :171-177.

O’Brien, JA, and RC Strange. 1975. The release of adenosine 3’:5’-cyclic monophosphate from the isolated perfused rat heart. Biochem J 152 : 429-432.

Opie, LH, D Nathan, and WF Lubbe. 1979. Biochemical aspects of arrhythmogenesis and ventricular fibrillation. Am J Cardiol 43 :131-148.

Palmer, RMJ, DS Ashton, and S Moncada. 1988. Vascular endothelial cells synthesize nitric oxide from L-arginine. Nature 333 : 664-666.

Palmer, RMJ, AG Ferrige, and S Moncada. 1987. Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor. Nature 327 : 524-526.

Parratt, JR. 1982. Early arrhythmias resulting from myocardial ischaemia. London, UK.: McMillan.

Penkoske, PA, BE Sobel, and PB Corr. 1978. Disparate electrophysiological alterations accompanying dysrhythmias due to coronary occlusion and reperfusion in the cat. Circulation 58 : 1023-1035.

Piwinica-Worms, D, R Jacob, CR Horres, and M Lieberman. 1985. Potassium- chloride cotransport in cultured chick heart cells. Am J Phys 249 : C337-C344.

Pocock, SJ, AG Shaper, DG Cook, RF Packham, RF Lacey, P Powell, and PF Russell. 1980. British Regional Heart Study: geographic variations in cardiovascular mortality, and the role of water quality. Br Med J 280 :1243- 1249.

Podzuweit, T, and WF Lubbe. 1977. Relation between post-ligation arrhythmias and myocardial cyclic AMP levels in the pig. J Mol Cell Cardiol 9 : 40.

Podzuweit, T, WF Lubbe, and LH Opie. 1976. Cyclic adenosine monophosphate, ventricular fibrillation and antiarrhythmic drugs. Lancet 1 : 341- 151

342.

Podzuweit, T, K-H Binz, P Nennsteil, and W Flaig. 1989. The anti-arrhythmic effects of myocardial ischaemia. Relation to reperfusion arrhythmias? Cardiovasc Res 23 : 81 -90.

Podzuweit, T, AJ Dalby, GW Cherry, and LH Opie. 1978. Cyclic AMP levels in ischaemic and non-ischaemic myocardium following coronary artery ligation: Relation to ventricular fibrillation. J Mol Cell Cardiol 10 : 81-94.

Pogwizd, SM, and PB. Corr. 1990. Mechanisms underlying the development of ventricular fibrillation during early myocardial ischaemia. Circ Res 66 : 672-695.

Poole-Wilson, PA. 1975. Potassium and the heart. Clin Endocrinol Metab 2 : 249- 268.

Rapaport, RM, MB Drazin, and F Murad. 1983. Endothelium-dependent relaxation in rat aorta may be mediated through cyclic GMP-dependent protein phosphorylation. Nature 306: : 174-176.

Redwood, DR, ER Smith, and SE Epstein. 1972. Coronary artery occlusion in the conscious dog. Effects of alteration in heart rate and arterial pressure on the degree of myocardial ischaemia. Circulation 46 : 323-332.

Report on a WHO meeting. 1984. Environmental health: health hazards from nitrates in drinking-water. WHO, Copenhagen.

Richards, DAB, K Byth, DLR Ross, and JB Uther. 1991. What is the best predictor of spontaneous ventricular tachycardia and sudden death after myocardial infarction. Circulation 83 : 756-763.

Ridley, PD, and MJ Curtis. 1990a. Effects of chloride substitution on PR and QRS complex: relevance to antiarrhythmic activity. J Mol Cell Cardiol 22 (Suppl III): S86. 152

Ridley, PD, and MJ Curtis. 1990b. Antiarrhythmic effects of chloride substitution by nitrate are mediated at specific sites in ischaemic and reperfused myocardium. Circulation 82 (Suppl III) : 2541.

Ridley, PD, and MJ Curtis. 1990c. Nitrates in drinking water: relationship to cardiac death. Eur Heart J 11 :1135-1136.

Ridley, PD, DJ Hearse, and MJ Curtis. 1989. Anions and arrhythmogenesis: chloride substitution by nitrate inhibits ischaemia- and reperfusion-induced arrhythmias. Circulartion 80 (Suppl II) : 59.

Ridley, PD, M Yacoub, and MJ Curtis. 1990. Membrane Permeability is a determinant of arrhythmogenesis following anion substitution. Br J Pharmacol 101 (Proceedings suppl) : 511 P.

Rissanen, V, M Romo, and P Siltanen. 1978. Prehospital sudden death from ischaemic heart disease. A postmortem study. Br Heart J 40 :1025-1033.

Riva, E, and DJ Hearse. 1989. Antiarrhythmic effects of amiodarone and desethylamiodarone on malignant ventricular arrhythmias arising as a consequence of ischaemia and reperfusion in the anaesthetized rat. Cardiovasc Res 23 :331-339.

Roberts, WC, and AA Jones. 1979. Quantitation of coronary arterial narrowing at necropsy in sudden cardiac death. Analysis of 31 patients and comparison with 25 control subjects. Am J Cardiol 44 : 39-45.

Ruberman, W, E Weinblatt, JD Goldberg, CW Frank, and S Shapiro. 1977. Ventricular premature beats and mortality after myocardial infarction. N Eng J Med 297 : 750-757.

Rushmer, RF, N Watson, D Harding, and D Baker. 1963. Effects of acute coronary occlusion on performance of right and left ventricles in intact unanaesthetized dog. Am Heart J 66 : 522-531. 153

Russell, DC, MF Oliver, and J Wojtczak. 1977. Combined electrophysiological technique for assessment of the cellular basis of early ventricular arrhythmias. Lancet ii : 686-688.

Sagawa, K. 1978. The ventricular pressure-volume diagram revisited. Circ Res 43 : 677-687.

Sagawa, K. 1981. Editorial: The pressure-volume relation of the ventricle: Definition, modifications and clinical use. Circulation 63 : 1223-1227.

Samson, WE, and AM Sher. 1960. Mechanism of ST segment alteration during acute myocardial injury. Circ Res 8 : 780-787.

Sato, R, A Noma, Y Kurachi, and H Irisawa. 1985. Effects of intracellular acidification on membrane currents in ventricular cells of the guinea pig. Circ Res 57 : 553-561.

Schmidt, FO, and J Erlanger. 1928. Directional differences in the conduction of the impulse through the heart muscle and their possible relation to extrasystolic and fibrillatory contractions. Am J Physiol 87 : 326-347.

Seyama, I. 1977. The effect of Na, K and Cl ions on the resting membrane potential of sinoatrial node cell of the rabbit. Jap J Physiol. 27 : 577-588.

Seyama, I. 1979. Characteristics of the anion channel in the sino-atrial node cell of the rabbit. J Physiol 294 : 447-460.

Sheridan, DJ, PA Penkoske, BE Sobel, and PB Corr. 1980. Alpha-adrenergic contribution to dysrhythmias during myocardial ischaemia and reperfusion in cats. J Clin Invest 65 :161 -171.

Sherlag, BJ, RH Helfant, Jl Haft, and AN Damato. 1970. Electrophysiology underlying ventricular arrhythmias due to coronary ligation. Am J Physiol 219 : 1665-1671. 154

Shiki, K, and DJ Hearse. 1987. Preconditioning of ischaemic myocardium: reperfusion-induced arrhythmias. Am J Physiol 253 : H1470-H1476.

Singh, BN. 1983. Amiodarone: Historical development and pharmacological profile. Am Heart J 106 : 788-797.

Singh, BN, LH Opie, DC Harrison, and FI Marcus. 1987. Antiarrhythmic agents. In Drugs for the heart. Edited by L. Opie. 54-90. Orlando, Florida, U.S.A.: Grune and Stratton, Inc.

Sobel, BE, PB Corr, AK Robinson, RA Goldstein, FX Witkowski, and MS Klein. 1978. Accumulation of lysophosphoglycerides with arrhythmogenic properties in ischaemic myocardium. J Clin Invest 61 : 546-553.

Steenbergen, C, G Deleeuw, CH Barlow, B Chance, and JR Williamson. 1977. Heterogeneity of the hypoxic state in perfused rat heart. Circ Res 41 : 606-615.

Stockman, R, and WB Charteris. 1901. On the action of iodine and iodides chiefly with regard to the muscle of the frog. J Phvsiol 26 : 277-281.

Tchou, PJ, N Kadri, J Anderson, JA Caceres, M Jazayeri, and M Akhtar. 1988. Automatic implantable defibrillators and survival of patients with left ventricular dysfunction and malignant ventricular arrhythmias. Ann Intern Med 109 : 529- 534.

Temesy-Armos, PN, S Vanderbrug Medendropp, S Goldstein, JR Landis, RF Leighton, G Ritter, CM Vasu, RA Wolfe, and A Acheson. 1988. Predictive value of ventricular arrhythmias in resuscitated out-of-hospital cardiac arrest victims. Eur Heart J 9 : 625-633.

Tennant, R, and CJ Wiggers. 1935. The effect of coronary occlusion on myocardial contraction. Am J Phvsiol 112 : 351 -361.

The working party on nitrate and related compounds. 1987. Nitrate, nitrite and N- nitroso compounds in food: the twentieth report of the steering group on food 155 surveillance. Ministry of Agriculture, Fisheries and Food. NTIS, Food surveillance paper 20, H.M.S.O., London.

Tosaki, A, and DJ Hearse. 1987. Protective effect of transient calcium reduction against reperfusion-induced arrhythmias in the rat hearts. Am J Physiol 253 : H225-H233.

Trolese-Mongheal, Y, P Duchene-Marullaz, JF Trolese, M Leinot, JC Lamar, and P. Lacroix. 1985. Sudden death and experimental acute myocardial infarction. Am J Cardiol 56 : 677-681.

Tsushihashi, K, and MJ Curtis. 1990. Tedisamil possesses direct defibrillatory activity during myocardial ischaemia and during reperfusion. Br J Pharmacol 99 : 231 P.

Tzivoni, D, A Keren, H Granot, S Gottlieb, J Benhorin, and S Stern. 1983. Ventricular fibrillation caused by myocardial reperfusion in Prinzmetal's angina. Am Heart J 105 : 323-325.

Vaughan Williams, EM. 1970. Classification of anti-arrhythmic drugs. In Symposium on Cardiac Arrhythmias. Edited by E. Sandoe, E. Flensted-Jansen and K. Olesen. 449-472. Sodetalje, Sweden: AB Astra.

Vaughan Williams, EM. 1984. A classification of antiarrhythmic actions after a decade of new drugs. J Clin Pharmacol 24 :129-147.

Vaughan Williams, EM. 1985. Delayed ventricular repolarization as an anti arrhythmic principle. Eur Heart J 6 (Suppl D) : 145-149.

Vaughan-Jones, RD. 1979a. Non-passive chloride distribution in mammalian heart muscle: microelectrode measurement of the intracellular chloride activity. J. Phvsiol 295 : 83-109.

Vaughan-Jones, RD. 1979b. Regulation of chloride in quiescent sheep heart Purkinje fibres studied using intracellular chloride and pH-sensitive micro- 156 electrodes. J Physiol 295 :111-137.

Walker, MJA, MJ Curtis, DJ Hearse, RWF Campbell, MJ Janse, DM Yellon, SM Cobbe, SJ Coker, JB Harness, DWG Harron, AJ Higgins, DG Julian, MJ Lab, AS Mannining, BJ Northover, JR Parratt, RA Riemersma, E Riva, DC Russell, DJ Sheridan, E Winslow, and B Woodward. 1988. The Lambeth Conventions: guidelines for the study of arrhythmias in ischaemia, infarction and reperfusion. Cardiovasc Res 22 : 447-455.

Wallenstein, S, CL Zuker, and JL Fleiss. 1980. Some statistical methods useful in circulation research. Circ Res 47 :1-9.

Wexler, LF, BH Lorell, S Momomura, EO Weinberg, JS Ingwall, and CS Apstein. 1988. Enhanced sensitivity to hypoxia-induced diastolic dysfunction in pressure- overload left ventricular hypertrophy in the rat: Role of high-energy phosphate depletion. Circ Res 62 : 766-775.

Williams, DO, BJ Scherlag, RR Hope, N El-Sherrif, and R. Lazzara. 1974. The pathophysiology of malignant ventricular arrhythmias during acute myocardial ischaemia. Circulation 50 : 1163-1172.

Winkler, B, S Sasson, K Binz, and W Schaper. 1984. Myocardial blood flow and myocardial infarction in rats, guinea pigs and rabbits. J Mol Cell Cardiol 16 (Suppl II): 48.

Winslow, E. 1984. Methods for the detection and assessment of antiarrhythmic activity. Pharmac Ther 24 : 401-403.

Witkowski, FX, and PB Corr. 1984. Mechanisms responsible for arrhythmias associated with reperfusion of ischaemic myocardium. Ann N Y Acad Sci 427 : 187-208.

Wollenberger, A, E-G Krausse, and G Heier. 1969. Stimulation of 3’,5’-cyclic AMP formation in dog myocardium following arrest of blood flow. Biochem Biophys Res Com 36 : 664-670. 157

Woodward, B, and MNM Zakaria. 1985. Effect of some free radical scavengers on repefusion-induced arrhythmias in the isolated rat heart. J Mol Cell Cardiol 17 : 485-493.

Yusuf, S, R Collins, R Peto, C Furberg, MJ Stampfer, SZ Goldhaber, and CH Hennekens. 1985. Intravenous and intracoronary fibrinolytic therapy in acute myocardial infarction: overview of results on mortality, reinfarction and side- effects from 33 randomized controlled trials. Europ Heart J 6 : 556-585.

Ziegelhoffer, A, EG Krause, M Fedelesova, J Styk, R Kvetnansky, I Blasig, and A Wollenberger. 1976. Changes in cyclic AMP levels in the non-ischaemic and ischaemic myocardium after coronary artery ligation. Seventh European Congress of Cardiology : 691.