Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 235

_____________________________ _____________________________

Teratogenicity as a Consequence of Drug-Induced Embryonic

Cardiac Arrhythmia

Common Mechanism for Almokalant, Sotalol, Cisapride, and Phenytoin via Inhibition of IKr

BY

ANNA-CARIN SKÖLD

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2000

2

Dissertation for the Degree of Doctor of Philosophy (Faculty of Pharmacy) in Toxicologypresented at Uppsala University in 2000

ABSTRACT

Sköld, A.-C. 2000. Teratogenicity as a consequence of drug-induced embryonic cardiacarrhythmia. Common mechanism for almokalant, sotalol, cisapride, and phenytoin viainhibition of IKr. Acta Universitatis Upsaliensis. Comprehensive Summaries of UppsalaDissertations from the Faculty of Pharmacy 235. 51 pp. Uppsala. ISBN 91-554-4828-3.

Drugs prolonging the repolarisation phase of the myocardial action potential, via inhibitionof the rapid component of the delayed-rectifying potassium channel (IKr), may cause deathdue to arrhythmia. Selective IKr-blockers are also embryotoxic in rats. The aim of this thesiswas to determine if drugs with the ability to block IKr cause embryotoxicity across species bya common mechanism: pharmacologically induced episodes of embryonic arrhythmia leadingto hypoxia-related damage in the embryo.

The selective IKr -blocker almokalant (ALM) induced, in both mice and rats, concentration-dependent embryonic dysrhythmia in vitro, and embryonic death, growth retardation andmalformations in vivo. These results agree with data indicating that IKr is expressed andfunctional during a restricted period in rodent embryonic life (gestational days 10-14 in rats),but not thereafter. Single doses during this period caused stage-specific external, visceral andskeletal defects in rats. The later the treatment was during this period, the more caudal the siteof the skeletal defect. Overall, the pattern and stage-specificity of defects were very similar tothat reported after transient hypoxia. Pre-treatment with the spin-trapping agent PBN reducedALM teratogenicity, suggesting involvement of reactive species generated during thereoxygenation phase.

In rabbits, which depend on IKr in adult life, the non-selective IKr -blocker sotalol inducedembryonic death at the same developmental stage as ALM treatment in rats. Embryos seem tobe more susceptible to arrhythmia-induced death than the adult. Cisapride, a gastrointestinalstimulating agent with IKr-blockage as a side effect, induced embryonic arrhythmia andidentical stage-specific digital defects as found with ALM in rats. Phenytoin, a non-selectiveIKr-blocking anticonvulsant, induced embryonic arrhythmia and malformations in rats. Theconcentrations required to induce embryonic arrhythmia correlated well with maternal plasmaconcentrations causing malformation (orofacial cleft).

In conclusion, the results support the hypothesis of a common hypoxia-related mechanismresulting in developmental toxicity across species for both selective and non-selectiveIKr-blockers. The use of an in vitro embryo culture model may be a tool for detection of drugswith arrhythmogenic potential.

Key words: IKr, delayed rectifying K+ channel, embryonic cardiac arrhythmia, teratogenicity.

Anna-Carin Sköld, Department of Pharmaceutical Biosciences, Division of Toxicology,Uppsala Universtiy, Box 594, SE-751 24 Uppsala

© Anna-Carin Sköld 2000

ISSN 0282-7484ISBN 91-554-4828-3

Printed in Sweden by Lindbergs Grafiska HB, Uppsala 2000

3

Till minne avfarmor och morfar

4

The thesis is based on the following papers, which will be referred to in the text byRoman numerals:

I. Anna-Carin Sköld, Katrin Wellfelt, and Bengt Danielsson. Stage specificskeletal and visceral defects by the IKr blocker almokalant: Further evidencefor teratogenicity via a hypoxia-related mechanism. Manuscript submitted.

II. Katrin Wellfelt, Anna-Carin Sköld, Alf Wallin, and Bengt Danielsson (1999)Teratogenicity of the class III-antiarrhythmic drug almokalant - role ofhypoxia and reactive oxygen species. Reproductive Toxicology 13, 93-101.

III. Anna-Carin Sköld and Bengt Danielsson (2000). Developmental toxicity ofthe class III antiarrhythmic agent almokalant in mice: adverse effects mediatedvia induction of embryonic heart rhythm abnormalities. ArzneimittelForschung/ Drug Research 50, 520-525.

IV. Anna-Carin Sköld and Bengt Danielsson. The IKr blocking drug sotalol causesdevelopmental toxicity in the pregnant rabbit. Manuscript accepted forpublication in Pharmacology and Toxicology.

V. Anna-Carin Sköld and Bengt Danielsson. Teratogenicity of cisapride:Evidence of relation to its ability to block IKr channels and cause embryoniccardiac arrhythmia. Manuscript submitted.

VI. Bengt Danielsson, Anna-Carin Sköld, Faranak Azarbayjani, Inger Öhman, andWilliam Webster (2000). Pharmacokinetic data support pharmacologicallyinduced embryonic dysrhythmia as explanation to fetal hydantoin syndrome inrats. Toxicology and Applied Pharmacology 163, 164-75.

Reprints were made by kind permission from the publishers: Elsevier Science(Paper II), Editio Cantor Verlag (Paper III), Pharmacology & Toxicology (Paper IV),and Academic Press (Paper VI).

5

CONTENTS

INTRODUCTION...................................................................................................................................7

THE HEART............................................................................................................................................7CARDIAC ARRHYTHMIA.........................................................................................................................8

Definitions........................................................................................................................................8Causes ..............................................................................................................................................9Treatment .........................................................................................................................................9

PHARMACOLOGICALLY INDUCED ARRHYTHMIA..................................................................................10Arrhythmia in humans....................................................................................................................10Arrhythmia in animals....................................................................................................................11

EMBRYONIC ARRHYTHMIA AND ITS RELATION TO HYPOXIA AND DEVELOPMENTAL TOXICITY ...........12COMPOUNDS STUDIED .........................................................................................................................14

Almokalant .....................................................................................................................................14Sotalol ............................................................................................................................................15Cisapride and Mosapride...............................................................................................................15Phenytoin .......................................................................................................................................16

MATERIALS AND METHODS .........................................................................................................17

DEVELOPMENTAL TOXICITY STUDIES ..................................................................................................17EFFECTS ON THE EMBRYONIC HEART...................................................................................................18DETERMINATION OF TEST COMPOUND IN MATERNAL PLASMA.............................................................19

RESULTS AND DISCUSSION ...........................................................................................................20

EMBRYONIC ARRHYTHMIA AND HYPOXIA IN DEVELOPMENTAL TOXICITY (PAPERS I AND II) ..............20Almokalant as model compound for IKr-related developmental toxicity ........................................20Similarities in teratogenicity between hypoxia and IKr-blockers....................................................22Hypoxia versus reoxygenation damage..........................................................................................26

IKR-BLOCKERS, EMBRYONIC ARRHYTHMIA AND DEVELOPMENTAL TOXICITY IN OTHER SPECIESTHAN THE RAT (PAPERS III AND IV)....................................................................................................28

Almokalant .....................................................................................................................................28Developmental toxicity in the mouse.........................................................................................................28Effects on the mouse embryonic heart .......................................................................................................29Correlation between effects on the embryonic heart and developmental toxicity ......................................29

Sotalol ............................................................................................................................................30Developmental toxicity in the rabbit..........................................................................................................30Correlation between the effects on the embryonic heart and developmental toxicity ................................30Comments on developmental toxicity........................................................................................................31

DRUGS THAT BLOCK IKR AS A SIDE EFFECT (PAPERS V AND VI) ..........................................................31Cisapride and mosapride ...............................................................................................................32

Effects on the embryonic heart ..................................................................................................................32Developmental toxicity ..............................................................................................................................32Correlation between the effects on the embryonic heart and developmental toxicity ................................33Comments on developmental toxicity........................................................................................................34

Phenytoin .......................................................................................................................................34Developmental toxicity related to maternal plasma concentrations ...........................................................35Effects on the embryonic heart ..................................................................................................................35Correlation between effects on the embryonic heart and developmental toxicity ......................................36Comments on developmental toxicity........................................................................................................36

WEC: AN INDICATOR FOR DEVELOPMENTAL TOXICITY AND FOR PROARRHYTHMIC POTENTIAL IN ADULTS? ........................................................................................................................................36HUMAN RELEVANCE AND FUTURE STUDIES .........................................................................................39

CONCLUDING REMARKS ...............................................................................................................40

ACKNOWLEDGEMENTS..................................................................................................................41

REFERENCES......................................................................................................................................43

6

ABBREVIATIONS

ALM almokalantAV node atrioventricular nodebpm beats per minuteECG electrocardiogramFDA Food and Drug Administration (US regulatory authority)GD gestation dayHERG human ether-a-gogo related geneIK the delayed rectifying potassium currentIKr rapid component of IK

ip intraperitonealiv intravenousOTC (-)-2-oxo-4-thiazolidine carboxylic acidPBN α-phenyl-N-t-butylnitronepo per osROS reactive oxygen speciesWEC whole embryo culture

7

INTRODUCTION

Cardiac arrhythmia in connection with cardiovascular disease has long been a well-known cause of death. During the last five-six years, increased interest has beenfocussed on pharmacologically induced death due to arrhythmia. This thesis will dealwith arrhythmias induced pharmacologically by a special type of drug, whichprolongs the repolarisation phase of the myocardial action potential, due to inhibitionof the rapid component of the delayed-rectifying potassium channel (IKr). In additionto possessing arrhythmogenic potential, selective IKr-blockers also have embryotoxicand teratogenic properties in animals. A hypothesis has been presented proposing thatthese developmental toxic effects are a result of pharmacologically induced episodesof embryonic cardiac arrhythmias leading to hypoxia-related damage in the embryo(Danielsson, 1993).

The aim of this thesis is to investigate the proposed mechanism, and to study if alsodrugs with primarily another pharmacological action, for which the IKr-blockingproperty is a side effect, cause similar developmental toxicity. An introduction to theelectrophysiology of the heart follows as a brief background together with a briefintroduction of the classification, origin, and treatment of arrhythmias in general.More detailed information is then given on pharmacologically induced arrhythmias,together with the data available at the start of this work on embryonic arrhythmia inrelation to hypoxia and developmental toxicity.

The heart

The heart rate is regulated by electrical impulses. Pacemaker cells in the sinus nodegenerate action potentials that propagate to the atrium, to the atrioventricular node(AV node) and to the rest of the ventricular muscle via Purkinje fibres. The changes inaction potential are mediated via the in- and outflow of different ions through ionchannels. In figure 1 a schematic illustration of an action potential and an ECG areshown (Abrahamsson, 1996; Crumb and Cavero, 2000).

One of the topics in this thesis is disturbance of the heart rhythm - arrhythmia ordysrhythmia. Semantics dictate that the term “arrhythmia” strictly means an absenceof rhythm, while “dysrhythmia” refers to a distortion of cardiac rhythm. In this thesisthe two terms are used interchangeably, with arrhythmia being used in its colloquialand popular sense.

8

Figure 1: A) a schematic action potential of a human cardiac conduction cell correlated withelectrolyte shifts. Phase 0: depolarisation, inward sodium current (INa). Phase 1: rapidrepolarisation, outward potassium current (Ito). Phase 2: plateau of the action potential,L-type Ca2+-channels, inward calcium current (ICa). Phase 3: repolarisation, outwardpotassium currents (IKr, Isus, and possibly IKs) and decay of L-type Ca2+ current. Phase 4:maintaining resting potential, inward rectifier current (IK1). 5: refractory period (duringwhich a new action potential may not be elicited). B) a schematic picture of a surfaceECG.

Cardiac arrhythmia

DefinitionsDisturbances in the heart rhythm are very common. If an ECG is recorded for24 hours, most people show single extra beats. This is normal and usuallyasymptomatic and has no negative prognostic implications. Symptoms due to changesin the heart rhythm are a frequent reason for patients to contact doctors. Thesymptoms are usually palpitations, fatigue, and vertigo. Although many arrhythmiasare benign, others can lead to more serious symptoms needing treatment, especiallydue to negative effects on the circulation. Cardiac arrhythmias are the result ofdisturbed electrical activity of the heart and are classified according to their site oforigin, supraventricular (the atrium or AV node) and ventricular, and according towhether the heart rate is increased (tachycardia) or decreased (bradycardia). It is oftenunclear which of the various mechanisms mentioned below are responsible for thearrhythmia. The most common mechanism causing arrhythmias is re-entry, which isan abnormal conduction circuit. This can occur in both micro- and macro-circuits indifferent parts of the heart. The micro-circuits involve only a very small part of thecardiac muscle, whereas macro-circuits involve larger structures of the heart. Thereare three other basic mechanisms underlying arrhythmias: delayedafter-depolarisation, abnormal pacemaker activity in parts of the heart other than the

P

Q

R

S

T

QT interval

- 80

0

(mV)

(time)

4

0

12

3

45

A B

9

nodal and conducting tissues, and heart block. The heart block can be complete, inwhich the atria and ventricles beat independently at rates determined by their ownpacemakers, or partial, in which a proportion of the sinus beats from the atria aretransmitted to the ventricle (Noguchi and Gill, 1992; Rang et al., 1999).

CausesThe cause of arrhythmias may be pathological changes, “non-cardiac”, orpharmacological. This thesis will discuss pharmacologically induced arrhythmias of aspecial type, induced by class III antiarrhythmic drugs. The most common cause ofarrhythmia is pathological damage to cardiac tissue due to infarction. The cardiacmuscle or the conduction system becomes scarred and fibrous. The macro-circuitsmentioned above can also originate from congenital extra impulse propagationpathways, also known as accessory pathways. The cause of the arrhythmia may alsobe “non-cardiac”, such as intoxication, metabolic disorder, and thyroid toxicosis, etc.

TreatmentWhen the cause of the arrhythmia is “non-cardiac”, then the therapy should of coursebe aimed at eliminating the underlying cause. The treatment of “cardiac” arrhythmiascan be either pharmacological or non-pharmacological or a combination of both. Thenon-pharmacological treatments are electric or surgical methods such as defibrillation,use of a pacemaker, and radio frequency ablation. Vaughan-Williams (Rang et al.,1999) developed in 1970 a classification system for antiarrhythmic drugs based onpharmacological actions and gross electrophysiological properties. It consists of fourmajor groups.

Class I: drugs that block sodium channels, thus reducing the excitability of the non-nodal regions of the heart where the inward sodium current is important forpropagation of the action potential. This class is further subdivided into C (slowdissociation), B (fast dissociation), and A (speed between B and C). Class II:β-adrenoceptor antagonists. Class III: drugs that prolong the refractory period of themyocardium, thus tending to suppress re-entrant rhythms. Class IV: calciumantagonists, which block calcium channels and thus impair impulse propagation innodal areas and in damaged areas of the myocardium. Some drugs, including digitalisand adenosin, do not fit neatly into any of the above categories, but are widely used.

10

Pharmacologically induced arrhythmia

Arrhythmia in humansTreatment of cardiac arrhythmias with antiarrhythmic drugs may lead to the inductionof new arrhythmias. This may sound like a paradox, but considering the complexity ofcardiac electrophysiology, it is not surprising that interfering with any particularmechanism may lead to unpredictable consequences (Abrahamsson, 1996).

In the early part of the 1980s, there was a great deal of interest in developing class ICsodium channel blocking drugs because of their profound effect on myocardialconduction and potent suppression of ventricular ectopic beats. These drugs have alesser effect on myocardial refractoriness. Their use- and rate-dependent propertiesmade them attractive agents for the management of both supraventricular andventricular tachyarrhythmias. The popularity of these drugs was profoundly lessened,however, by the findings of the Cardiac Arrhythmia Suppression Trial (CAST)(Anonymous, 1989; Anonymous, 1992). The study showed, quite unexpectedly, thatthe long-term mortality was increased in patients who were treated with class Iantiarrhythmics. It was thought that the increased mortality was due to proarrhythmicproperties of these drugs. The interest turned instead to agents with class IIIproperties. The target for most of these agents was found to be the rapid component ofthe delayed rectifying potassium channel (Colatsky and Argentieri, 1994).By definition, class III antiarrhythmic drugs prolong the duration of the actionpotential. The result is a lengthening of the effective refractory period with associatedantiarrhythmic activity. The antiarrhythmic efficacy was verified in clinical trials ofthese new compounds (Bashir et al., 1995; Darpö and Edvardsson, 1995; Echt et al.,1995; Sager et al., 1993). The effects of these drugs on myocardial cells manifestsitself as a prolongation of the QT interval (corrected QT interval) on the surface ECG.The consequences of this QT prolongation can be therapeutic or toxic. Therefore itwas disappointing to find that class III antiarrhythmic drugs are alsoproarrhythmogenic. Treatment with these class III antiarrhythmics was associatedwith the risk for a special kind of arrhythmia called Torsade de Pointes (Roden,1994). This is a characteristic type of polymorphous ventricular tachycardia in whichthe QRS axis of each successive beat differs slightly from the preceding one andseems to rotate around the isoelectric line: Torsade de Pointes - twisting of the points(Viskin, 1999).

In the development of Torsade de Pointes, there is a complex interaction of prolongedrepolarisation, the development of early afterdepolarisations with the capacity topropagate and capture the whole heart, and bradycardia (Doig, 1997). The rapidcomponent of the delayed rectifying potassium channel, IKr, is the primary ionchannel involved in repolarisation. The gene HERG (human ether-a-gogo relatedgene) seems to represent the molecular basis of IKr because the K+ currents encodedby HERG display biophysical and pharmacological characteristics similar to those ofIKr. In addition, HERG mutations are responsible for one form of thelong QT syndrome. Thus HERG is a primary target for both congenital

11

(mutation-induced) and acquired (drug-induced) action potential prolongation,delayed repolarisation, and cardiac arrhythmias (Taglialatela et al., 1998).There has been a great deal of interest regarding the safety assessment ofproarrhythmic potential in antiarrhythmic agents, both from the authorities and fromthe pharmaceutical companies. During research in this area it has also been found thatseveral classes of non-cardiovascular drugs, such as psychotropic, prokinetic,antimalarial, antifungal and antihistamine drugs, can also induce prolongation of theQT interval and Torsade de Pointes. This side effect may be even more importantsince the indications for many of these non-cardiovascular drugs are rather mild(Crumb and Cavero, 2000). The European Agency for Evaluation of MedicinalProducts (CPMP) wanted to establish a European consensus concerning how drugswith effects on the QT-interval should be assessed (preclinically and clinically). Theirwork resulted in a “Points to consider”-document: “The assessment of the potentialfor QT-interval prolongation by non-cardiovascular medicinal products” (CPMP986/96/1997). The document proposes practical approaches for preclinical in vitro andin vivo testing of drug candidates to detect the potential for QT prolongation andarrhythmias, and clinical trial design to determine this potential. In the US, the FDAhas focussed on this issue and withdrawn approval for some drugs with relatively mildindications. Even though the concentrations at which the proarrythmia occurs areusually higher than normal therapeutic concentrations, and the number of patientsaffected has been small, the withdrawel is motivated because of interaction effects,and the increased susceptibility of some patients to develop arrhythmias due todifferences in genetic background.

Arrhythmia in animalsThe proarrhythmogenicity of class III antiarrhythmics has also been shown inanimals. The rabbit heart seems to be especially vulnerable towards the induction ofTorsade de Pointes, and the rabbit has been used as a good animal model for testingthis property of new drugs (Abrahamsson, 1996; Carlsson et al., 1993). The delayedrectifier K+ current (IK) represents the sum of two currents: A slow (IKs) and arelatively rapid activating delayed rectifier (IKr) (Carmeliet, 1992). Class III agentsspecifically block the latter current. Not all mammalian hearts have both kineticallydistinct components. For example, both IKr and IKs appear to exist in guinea pigventricular and atrial myocytes, sheep and calf Purkinje fibres, embryonic chick atrialcells, and human atrial cells, whereas only IKr appears to exist in rabbit heart and catventricle (Wiesfeld et al., 1996). In contrast, the pure IKr-blockers almokalant anddofetilide have no effect on the action potential duration in rats (Abrahamsson et al.,1994; Wang et al., 1996), which is why the rat has been considered as an unsuitablespecies for testing proarrhythmic potential. The dominant repolarising outwardcurrent in the rat is Ito and it appears that the action potential is too short to allow foractivation of the IK to any considerable extent (Abrahamsson et al., 1994).

12

Embryonic arrhythmia and its relation to hypoxia anddevelopmental toxicity

The observed species differences regarding IKr in the adult seem to be non-existing inthe embryo. In contrast to adult rodents, experimental studies indicate that the IKr

channel is expressed and functional in the early mouse and rat embryo as soon as theheart starts beating (Abrahamsson et al., 1994; Ban et al., 1994; Davies et al., 1996;Wang et al., 1996; Webster et al., 1996). In the mouse, IKr is expressed in the atriumat an early stage of embryonic development, when it is the dominating repolarisationchannel (Davies et al., 1996). In view of this finding, it has been proposed that IKr

plays an important role in the early development of the mammalian cardiac electricalsystem (Abrahamsson et al., 1994). The presence of IKr in the early embryo in specieswhere this ion channel is suppressed in favour of other ion channels later indevelopment, suggests a universal functional role for the IKr channel across allmammalian species in early embryonic development (Davies et al., 1996; Wang et al.,1996).

It is an interesting but not unexpected finding, in view of the presence andfunctionality of IKr in the early embryo, that the embryonic rat heart reacts withdysrhythmia when exposed to selective IKr-blockers. Thus, embryonic bradycardia hasbeen described for the selective IKr-blockers almokalant, dofetilide, and L-691,121(Abrahamsson et al., 1994; Ban et al., 1994; Spence et al., 1994; Webster et al., 1996)in rat embryos cultured in vitro during the period when the IKr-channel seems to playan important role. The severity of the bradycardia depends on concentration,arrhythmia and cardiac arrest occur at sufficiently high concentrations. The embryonicheart during this stage of development is clearly visible and the heart beatsprominently, making it possible to examine the adverse embryonic cardiac effects byvisual inspection using a stereomicroscope with low magnification.

The effects of almokalant on the rodent embryonic heart in our in vitro model agreewith the results obtained with more traditional electrophysiological methods. Figure 2shows transmembrane recordings of action potentials from the atrium of thespontaneously beating rat embryo hearts treated with different doses of almokalantduring the same stage as almokalant was tested in whole embryo culture (WEC)(Abrahamsson et al., 1994 versus Webster et al., 1996). As seen in figure 2,almokalant induced concentration-dependent bradycardia and arrhythmia in the ratembryo. This may be the result of increased duration of the action potential, leading todecreased heart rate. After extensive prolongation of the action potential, almokalantinduced spontaneous deflection of the membrane potential consistent with earlyafterdepolarisations, which in our in vitro model is manifested as arrhythmia. In thiscontext, it should be mentioned that the transmembrane recordings seen in figure 2,are very similar to those obtained in a sensitive adult rabbit model that mimics thecourse of events in sensitive patients who developed almokalant induced arrhythmia(Abrahamsson et al., 1994). This may also indicate that our in vitro model may beused as a sensitive indicator for the detection of arrhythmogenic potential of IKr-blocking compounds.

13

Figure 2: Reprinted from Cardiovascular Research, 28, Abrahamsson et al. Induction ofrhythm abnormalities in the fetal rat heart. A tentative mechanism for the embryotoxic effectof the class III antiarrhythmic agent almokalant, Copyright (1994), with permission fromElsevier Science.

a): Action potential duration at the 75% repolarisation level (APD75, triangles) and intrinsicheart rate (HR, circles) over time in foetal rat myocardium on gestational day 13. Ordinate:APD75 expressed in ms and HR as beats min-1. Values are means, bars=SD, n=4.

b): The effect of increasing concentrations of almokalant on the spontaneously beating foetalrat heart on gestational day 13. Circles = frequency in beats per minute (beats min-1).Tringles=action potential duration at the 75% repolarisation level (APD75, ms). Values aremeans, bars=SD. The number of experiments decreases due to the occurrence of earlyafterdepolarisations. *p<0.05, †p<0.01, ‡p<0.005 v the control value.

c): Transmembrane recordings of action potentials from the atrium of the spontaneouslybeating foetal rat heart at gestational day 13. (A) In the control situation; (B) After 0.1 µMalmokalant; (C) After 1 µM almokalant; (D) After 10 µM almokalant.

a)

b) c)

14

Several investigators have proposed that the severe bradycardia and arrhythmia thatare induced explain the observed high incidence of embryonic/foetal death due tohypoxia following administration of selective IKr-blockers. The mean foetal weightsof the few surviving foetuses were decreased and there was increased incidence ofmalformations, showing a teratogenic potential of almokalant and of L-691,121(Abrahamsson et al., 1994; Ban et al., 1994; Danielsson, 1993; Spence et al., 1994).It has been suggested that the malformations reported following administration ofselective IKr-blockers are related to episodes of hypoxia as a consequence ofembryonic arrhythmia (Danielsson, 1993). This theory was based on the observationthat similar malformations as those produced by selective IKr-blockers, as well asincreased mortality and decreased foetal weights, are caused by embryonic hypoxia,induced by temporary clamping of the uterine vessels in rats (Brent and Franklin,1960; Franklin and Brent, 1964; Leist and Grauwiler, 1974; Webster et al., 1987).

In the studies with whole embryo culture and L-691,121 and dofetilide (Ban et al.,1994; Spence et al., 1994), wash-out experiments were performed. In theseexperiments, the heart rates were examined during exposure to the drug and afterrinsing and transfer to drug-free medium. The heart rates returned almost to baselinelevels, which indicates a reversible pharmacological effect. The in vivo situation fordrugs with a short half-life resembles the conditions with the wash-out experiments.During peak concentrations of the compound there might be a short period ofarrhythmia, then the heart rhythm might return to normal when the concentrationsdecline. That would mean that a short period of severe hypoxia occurred due toembryonic cardiac arrhythmia, followed by reoxygenation when the heart starts tobeat normally again. This may be what occurs when uterine vessels are clamped for ashort period.

Compounds studied

Almokalant

Almokalant was developed by Astra Hässle in the 1980s as an antiarrhythmic drug. Itis pharmacologically classified as a class III antiarrhythmic agent and it is a selectiveinhibitor of the rapid component of the delayed rectifying potassium channel (IKr)(Abrahamsson, 1996). Almokalant also has proarrhythmic potential and thedevelopment of almokalant was discontinued due to induction of Torsade de Pointes,

15

which occurred in some susceptible patients during the clinical trials (data supplied bythe manufacturer).

Almokalant is a known animal teratogen, but there is no information regarding humanrelevance (since no pregnant women have been treated with almokalant).

Sotalol

Sotalol is a non-selective β-adrenoceptor antagonist and a class III antiarrhythmicagent. It has been available as a β-blocker since the early 1960s and in the early 1970sit was also characterised as a class III agent via inhibition of IKr. The indications forits use include hypertension, angina, and cardiac arrhythmias (Kowey et al., 1997).The β-blocking properties are primarily due to the l-isomer (Beyer et al., 1993), whilethe class III activity is present in both the l- and d-isomers (Sanguinetti andJurkiewicz, 1990). In clinical practice the racemate of sotalol is used.

The incidence of Torsade de Pointes has been reported to be 2-5% for patients treatedwith sotalol (MacNeil, 1997). Despite the wide clinical use of sotalol, there are noexperimental embryo-foetal developmental toxicity studies available in the literature.There is also very limited information on human pregnancy outcome data.

Cisapride and Mosapride

Cisapride and mosapride are two structurally related benzamide derivates whichfacilitate or restore motility in the gastrointestinal tract, most likely via action onseretonin-4 receptors (5-HT4 receptor agonist) (Wiseman and Faulds, 1994; Yoshidaet al., 1993). Cisapride has been on the market for several years, whereas mosapride is

Mosapride Cisapride

16

presently undergoing extensive preclinical and clinical evaluation as a novelprokinetic agent. Recently, it was reported that treatment with cisapride mayoccasionally give rise to arrhyhtmia (1996a; 1996b; Bran et al., 1995; Lewin et al.,1996; Wysowski and Bacsanyi, 1996). It has been shown that cisapride, but notmosapride, inhibits IKr at clinically relevant concentrations (Carlsson et al., 1997). Inview of the discovered proarrhythmic potential, the use of cisapride has beenrestricted in many countries and the drug was removed from the US market on July14, 2000 (Vincent 2000).

There are no published experimental teratology studies on cisapride or mosapride inthe literature, and only limited human cisapride data on pregnancy outcome.

Phenytoin

Phenytoin (PHT) has been used for treating convulsive disorders for over six decades(Merritt and Putnam, 1938) for a review, see Rogawski and Porter (1990). It has alsobeen used as an antiarrhythmic drug in a more limited manner, primarily for treatingdigitalis intoxication (Bigger and Hoffman, 1980). Phenytoin is pharmacologicallyactive by stabilising the membranes of excitable cells, such as neurones and cardiaccells, via interference with voltage-dependent sodium, calcium, and potassium ionchannels, although the exact mechanism is not known. (Kuo and Bean, 1994;Macdonald, 1989; Ragsdale et al., 1991; Yaari et al., 1987). It has recently beenshown that phenytoin inhibits the delayed rectified potassium channel IK, whichmakes the substance particularly interesting for my thesis (Nobile and Lagostena,1998; Nobile and Vercellino, 1997). We also have new data (Danielsson et al.unpublished) that shows that it is the rapid component of IK that is inhibited byphenytoin. Previous work has suggested that PHT has proarrhythmic potential, andeight cases of PHT-induced arrhythmia evolving to asystole and death have beenrecorded, as has marked sinus bradycardia leading to syncope in four of 15 youngvolunteers given PHT by intravenous injection, for a review see Tomson andKennebäck (1997).

Phenytoin is an established human teratogen and foetal adverse effects have beenreported in mice, rats, rabbits and cats. The pattern of foetal adverse effects, describedas “Foetal Hydantoin Syndrome” in humans (Hanson and Smith, 1975), is verysimilar across species, for a review see Finnell and Dansky (1991). It consists ofembryonic death, intra-uterine growth retardation, mild CNS dysfunction, craniofacialabnormalities (including orofacial clefts) and limb reduction defects.

17

AIMS OF THE STUDY

• To elucidate the role of hypoxia and hypoxia/ reoxygenation damage indevelopmental toxicity of selective IKr-blockers mediated via embryonicarrhythmia, using almokalant as a model compound.

• To investigate if two established IKr-blockers, almokalant and sotalol, inducedevelopmental toxicity/ embryonic arrhythmia in other species than the rat.

• To investigate if cisapride and phenytoin (gastrointestinal motility stimulating-,and antiepileptic drug, respectively), which have recently been shown to haveIKr-blocking properties as a side effect, induce embryonic arrhythmia, embryonicdeath and hypoxia-related malformations.

MATERIALS AND METHODS

All animal work was approved by the ethical committees in Uppsala and Stockholm.In all studies performed (both in vivo and in vitro) day 0 of pregnancy was determinedby observing the day of occurrence of vaginal plug and/or sperms in the vaginal smearin the rat and mouse or observance of mating in the rabbit. Details of how the studieswere performed are given in each paper. Below is a summary of the main proceduresused.

Developmental toxicity studies

It is generally accepted in developmental toxicology that there is a higher probabilitythat the observed adverse effects occur in humans if they have been recreated in morethan one animal species. This is the reason why new drugs must be tested in onerodent (usually rat) and one non-rodent species (usually rabbit) according tointernational guidelines. The animals are kept under specified conditions regardingtemperature, humidity and light/dark cycle, usually having free access to feed andwater. All animals are inspected daily for clinical signs. The treatment period shouldcover organogenesis, which means that it should cover the period from implantationto closure of the hard palate. In the mouse, the dosing starts on gestation day (GD) 6and ends on GD 15, in the rat GD 6-16 and in the rabbit GD 6-18. Generally threedose levels are used: a low dose near the expected therapeutic concentration/exposure,a high dose where minimal toxicity is observed, usually as a decrease in maternalbody weight gain, and an intermediate dose. The control group is often given thevehicle or water. Near time of delivery (mouse GD 18, rat GD 21, and rabbit GD 29)a caesarean section is performed. The uterine contents are examined with respect to

18

number of corpora lutea, number of implantation sites, sex and number of viablefoetuses, number of dead foetuses, early and late intrauterine deaths, foetal weights,and external abnormalities. Approximately half of the foetuses in each litter arepreserved in Bouin’s solution for subsequent freehand sectioning in order to detectvisceral anomalies (Barrow and Taylor, 1969). The remaining foetuses are preservedin 95% alcohol for clearing and staining of the skeleton with alizarin red in order todetect skeletal anomalies. Abnormalities can be classified into the following threecategories: 1) major; rare and/or probably lethal, 2) minor; minor differences from'normal' that are detected relatively frequently and 3) variants; alternative structuresthat also occur regularly in the control population.

In this thesis the conventional design has been complemented with more specificinvestigations, including examination of foetal adverse effects after administration onsingle days of pregnancy.

Effects on the embryonic heart

Whole embryo culture of rats and mice has been used in this thesis as a tool forstudying effects on the embryonic heart rhythm. As described in the section“Embryonic arrhythmia and its relation to hypoxia and developmental toxicity” theobserved effects, bradycardia and arrhythmia, agree with results from more traditionalelectrophysiological methods. Explanted embryos were in all experimental designscultured using modified techniques of New (1978). The examination of effects on theheart in NMRI mouse embryos were performed on GD 10 and in Sprague Dawley ratembryos on GD 11. The dams were sacrificed by cervical dislocation (the rats firstbeing lightly anaesthetised by enflurane or CO2), the abdomen was then opened, theuterus excised and the decidual swellings removed under aseptic conditions. Thedecidua and Reichert’s membrane were removed from each swelling and the embryos,with the yolk sac and ectoplacental cone intact, were placed in sterile culture bottleswith one embryo per culture flask. The culture medium consisted of either 20% or~75% immediately centrifuged heat-inactivated rat serum and either 80% or ~25%Tyrode’s salt solution. The culture bottles were gassed with 55% N2, 40% O2 and5% CO2 (vol./vol./vol.) for 5 minutes, and capped tightly and cultured in a rotatingculture system at 37°C and 40 rounds per minute. The embryos were allowed torecover for 60 minutes until an initial heart rate reading was taken under a stereomicroscope. Embryos were removed from the culture medium for approximately25 seconds per examination. The embryos were kept on a heated stage to maintain thebottles at 37°C. Embryos were selected for each experiment based on the followingcriteria:

1) Morphologically normal appearance2) Good yolk sac circulation3) A heart rate of at least 120 beats per minute (bpm) for mice and 140 bpm for rats.

19

Embryos not fulfilling all three criteria were discarded. After the selection ofembryos, the test substance was added to each culture. Controls were treated witheither water or vehicle in corresponding volumes. The embryos were then kept inculture up to two hours, during which the embryos were examined for heart rate,incidence of arrhythmias and cardiac arrest, one to three times, each 15 seconds. Theembryos were video-taped during all observation periods. Their heart rhythms wereevaluated from these videotapes, which increases the accuracy in detecting arrhythmiaand bradycardia.

Determination of test compound in maternal plasma

Blood samples were taken for analysis of the test compounds in maternal plasma.Almokalant and cisapride were analysed at Astra Hässle by reversed phase liquidchromatography and fluorescence detection.Sotalol was analysed at the Department of Medicine and Care, Division ofPharmacology at the University of Linköping. The plasma was analysed with a liquidchromatographic method using fluorescence detection (Gustavsson et al., 1997).Phenytoin was analysed at the Department of Clinical Pharmacology, KarolinskaInstitute at Karolinska Hospital. Concentrations of total and free phenytoin weredetermined with FPIA (fluorescence polarization immunoassay) technology (AxsymPhenytoin package insert 1997 and TDx/TDxFLx Free phenytoin package insert 1997,Abbott Laboratories) (Jolley, 1981; Thomas et al., 1991). Ultrafiltration was used toisolate the free fraction (Gurley et al., 1995).

20

RESULTS AND DISCUSSION

Embryonic arrhythmia and hypoxia in developmental toxicity (PapersI and II)

Papers I and II elucidate the role of hypoxia in the teratogenicity of IKr-blockingagents. We investigated whether IKr-blockers induced the same type of stage-specificdefects as episodic hypoxia induces. Almokalant was used as a model substance.Another aim was to study if pre-treatment with PBN, a radical spin-trapping agent, orOTC, an antioxidant, could reduce the incidence of almokalant-induced defects. Suchpre-treatment may indicate if the defects induced are mainly caused by “purehypoxia” or by reoxygenation damage after an episode of severe embryonicdysrhythmia.

Almokalant as model compound for IKr-related developmental toxicityAlmokalant was administered as single doses during the period when the ratembryonic heart was extremely sensitive to selective IKr-blockers (Ban et al., 1994;Spence et al., 1994; Webster et al., 1996). Almokalant was given orally on GDs 10,11, 12, 13 or 14 and the doses were different on different days, adjusted in order toallow the majority of the foetuses to survive until term. No adverse clinical signs wereobserved in the dams that could be related to the treatment with almokalant.Almokalant increased the incidence of embryonic/foetal death, decreased the meanfoetal weights and induced external malformations in similarity to what has beenshown with other IKr blockers in conventionally designed teratology studies (Ban etal., 1994; Marks and Terry, 1996; Spence et al., 1994). The main results aresummarised in table 1.

We also showed that these external defects were phase-specific, with orofacial cleftsafter administration on GD 11 (figure 3) and digital reduction defects on the forepawafter administration on GD 13 (figure 4) and on the hindpaw after administrationGD 14. Our results regarding almokalant-induced external defects confirm results ofWebster et al. (1996). The similarities in developmental toxicity between the differentIKr-blockers strongly support the idea that the teratogenicity is a class effect forIKr-blockers that is mediated by a common mechanism.

21

Figure 3: A rat foetus on GD 21with cleft lip and palate afteradministration of 125 mg/kgALM on GD 11.

Figure 4: A rat foetus withbrachydactyly of the left forepawafter administration of 25 mg/kgALM on GD 13.

Almokalant, dofetilide, and L-691,121 all induced bradycardia and cardiac arrest inwhole embryo culture. It is generally accepted that the observed dose-dependent highincidences of foetal death for these substances are a consequence of observed adverseeffects on the embryonic heart (Abrahamsson et al., 1994; Ban et al., 1994; Spence etal., 1994; Webster et al., 1996). The similarity in the pattern of malformations mayalso indicate a common hypoxia-related mechanism as explanation to theirteratogenicity. The induced embryonic bradycardia and arrhythmia/cardiac arrestcould lead to hypoxia and the generation of free radicals at reoxygenation in theembryo, which in turn could cause the malformations. This mechanism will bediscussed further in the following sections.

22

Table 1: A summary of the main effects induced by almokalant in Papers I and II,together with reported effects for other selective IKr-blockers.

Abnormality Almokalant Dofetilide Ibutilide L-691,121

Rodent whole embryo culture

Bradycardia Yes Yes Yes

Arrhythmia Yes Yes Yes

Cardiac arrest Yes Yes Yes

Reproductive data

Embryonic/ foetal death Yes Yes Yes Yes

Decreased foetal weights Yes Yes Yes Yes

Defects observed externally

Cleft lip/ palate GD 11 GD 11 GD 6-15 GD 6-17

Short/ kinked/ threadlike/ absent tail GD 11-12 GD 11 GD 6-15 GD 6-17

Anal atresia GD 11 GD 6-15

Digital defects: forepaw GD 13 GD 13 GD 6-15a GD 6-17a

Digital defects: hindpaw GD (13) 14 GD (13) 14 GD 6-15a GD 6-17a

Oedema GD 15 GD 6-15 bGD 10, 12aNo distinction was made between the fore- and the hindpaw.bIn whole embryo culture

Similarities in teratogenicity between hypoxia and IKr-blockersHypoxia was one of the first teratogenic agents to be identified and studied, for reviewsee Grabowski (1970). Embryonic hypoxia can be obtained by for example, reducingthe oxygen tension in the air or by mechanically clamping the uterine blood vesselsleading to impaired uteroplacental blood flow. It is also secondary to pharmacologicaleffects by various drugs. The early postimplantion rat embryo, during the periodGD 6- GD 11, resists the induced hypoxia relatively well, and many embryos survive90-120 min of uterine vessel clamping (Franklin and Brent, 1964; Leist andGrauwiler, 1973). A low incidence of congenital malformations is seen in thesurviving foetuses including CNS defects such as dilated ventricles (hydrocephalus),orofacial clefts, and heart defects. By GD 12 the embryos become more sensitive, andduring GDs 13 - 16 most embryos are unable to survive 60 min of hypoxia. Shorterperiods of hypoxia on these days, 30-45 min, induce a range of malformations, themost common of which are distal digital defects. Other less frequently seen defectsinclude tail, face, and urogenital abnormalities (Leist and Grauwiler, 1974; Webster etal., 1987). Pathological examination (2-24 h after the clamping procedure) showedthat malformations such as cleft palate and digital defects are preceded by oedema,dilated blood vessels, haemorrhage, blisters and the subsequent development of

23

necrosis (Leist and Grauwiler, 1974; Webster et al., 1987). After GD 16, the foetusagain becomes less sensitive to hypoxia induced by uterine vessel clamping (Leist andGrauwiler, 1973). Hypoxia could also induce a variety of skeletal defects. The typeand localisation depended on the timing of the hypoxia episode relative toorganogenesis: a clear-cut head to tail shift of a variety of skeletal defects wasobserved as the developmental stage of the induced oxygen deficiency advanced(Degenhardt, 1958; Ingalls and Curley, 1957; Leist and Grauwiler, 1973; Morawa andHan, 1968; Murakami and Kameyama, 1963).

A summary of phase-specific external, visceral, and skeletal malformations induced byalmokalant in our studies (in the rat) is presented in table 2, together with defectsreported to be induced by hypoxia. Almokalant induced a relatively high incidence ofcardiac interventricular septum defects (~10%) and vascular malformations (~12%)consisting of the absence (for example lack of inominate), abnormal vessels (forexample enlarged aorta with reduced pulmonary trunk) or transposition of vessels(such as the azygos vein, and retroesophageal right subclavian) after dosing on GD 10and 11 and to a lesser degree on GD 12 and 13. In view of this, it is interesting thatepisodic hypoxia also causes phase-specific interventricular septal defects andabnormalities of the aortic arch (Leist and Grauwiler, 1973).

24

Table 2: Stage specific abnormalities induced by almokalant (Papers I and II) or hypoxia.

Abnormality Almokalant Hypoxia

Reproductive data

Embryonic/ foetal death Yes Yes

Decreased foetal weights Yes Yes

External/visceral defects

Cleft lip/palate GD 11 GD ~7-15

Digital defects of the forepaw GD 13 GD 13 (14)

Digital defects of the hindpaw GD 14 (13) GD 14 (13)

Short/ kinked/ absent tail GD 11-12 GD ~7-12

CNS/ Brain defects GD 11 GD ~ 8-12

Cardiac ventricular septum GD 10-11 (12)

Cardiovascular abnormalitiese.g. absence or transposition ofvessels

GD 10-11 (12-13) GD 8-10

Absence/ malposition ofpostcaval lung lobe

GD 10 (12)

Attenuated diaphragm GD 10-11 (12)

Dilated urethras GD 10-11 (12)

Abnormalities of the kidney GD 10-11 (12) GD ~ 7-12

Undescended testis GD 11-12 GD ~ 12

Skeletal defectsSkull bones GD 10-14

Thoracic GD 10-11 (lower12-14)

GD ~ 10

Lumbar GD 10-11 GD ~ 10

Sacral GD 11 GD ~ 11

Caudal GD 11-12 GD ~ 11

Sternebrae GD 10-14 GD 14-15

27th presacral GD 10-11

Ribs/extra rib GD 10 GD ~ 10

Almokalant induces lung abnormalities and imperforation of the anus, resembling theeffects of ibutilide. The latter drug leads to portions of lung being absent and analatresia after dosing during GD 6 - 16 (Marks and Terry, 1996). For almokalant thelung defects were most accentuated after dosing on GD 10, and ~20 % of the foetusesand litters had malposition/absence of the post caval lobe of the lung, while analatresia was induced most effectively by dosing on GD 11. Some CNS defects, such asdilated ventricles, and zigzag-shaped pia mater were also induced after dosing onGD 11. New malformations, not previously described for IKr-channel blocking drugs,were found in this study. These consisted of urogenital defects such as increased renal

25

pelvic cavitation, or absence of the renal papilla and undescended testes. Thesedefects were mainly observed after treatment on GD 10 and 11.

Figure 5: A rat foetusskeleton with an example ofalmokalant-induced defectson the thoracic level afteradministration on GD 10.

In this respect it is worth noticing that undescended testis (Franklin and Brent, 1964;Ingalls et al., 1952) and renal agenesis / hypoplasia (Brent and Franklin, 1960;Franklin and Brent, 1964) and hydronephrosis (Leist and Grauwiler, 1973) have beeninduced after episodic hypoxia. The same is true for attenuated diaphragm /abdominal wall defects and other occasionally observed malformations in this study(abnormal structure of endocrine glands) and in the study with ibutilide(aplebharia/agnathia) (Marks and Terry, 1996). In all cases similar defects have beenfound after embryonic hypoxia due to clamping of uterine vessels during GDs 7-12 inthe rat (Brent and Franklin, 1960; Leist and Grauwiler, 1973; Leist and Grauwiler,1974).

Almokalant caused skeletal defects in the present study (in the rat) on alladministration days between GD 10 and GD 14. Skeletal examination showed a highincidence of vertebral abnormalities on the thoracic level on GD 10 (an example isshown in figure 6) and on lower thoracic- to caudal level on GD 11, compared tocontrols. Both vertebral centra and arches were affected, as were the ribs, with defectsthat involved misalignment, incomplete ossification of varied type and extent, andfusion. On GD 10 a very high incidence of an extra (27th) presacral vertebrae wasnoticed compared to controls. Increased incidences of sternebral defects, compared tocontrols were obtained on all days, while almost no vertebral defects were observedon GD 13 or 14. On GD 13 and 14 digital defects such as brachydactyly, syndactyly,and oligodactyly were induced on the forepaw and hindpaw, respectively. Skeletaldefects showed a clear trend; the later the treatment the more caudal was the site ofthe defect. The phase-specificity for induction of skeletal defects by almokalant is

26

thus almost completely in accordance with that which has been reported for skeletalmalformations caused by hypoxia in rodents (Ingalls and Curley, 1957; Leist andGrauwiler, 1973; Morawa and Han, 1968) and rabbits (Degenhardt, 1958).

Hypoxia versus reoxygenation damageThe exact mechanism that gives rise to hypoxia-related malformations is not fullyunderstood. It has long been known that a reduced oxidation rate will primarily affectorgans that are at a rapid proliferation phase at the time, causing discontinuation indevelopment. The hypoxia may result either in organs that consist of sensitive cellsdying, becoming retarded or having their differentiation capacity affected in somemanner (Naujoks, 1953). This could result in an overall embryonic growth retardationthat would be more pronounced in the areas with most cell proliferation, such as facialprominences, leading for example to failure of closing the palate (Bronsky et al.,1986). A new paper has studied how hypoxia plays a global role in developmentalphysiology and morphogenesis in rat embryos on GD 9 – 11 (Chen et al., 1999). Thehypoxia may act as a physiological signal co-ordinating some group of genesimportant in development

However, several studies indicate that damage after hypoxic episodes is related to freeradicals that are generated during the reoxygenation phase. In adult tissues, generationof free radicals is a well established mechanism for tissue damage duringreoxygenation of the ischemic heart after myocardial infarction, and the centralnervous system (CNS) after stroke (Gutteridge, 1993; McCord, 1993). The majorcause of damage is readmission of oxygen and not the re-establishment of flow(Kalyanaraman et al., 1995). Reactive oxygen species (ROS) have also beenidentified in embryonic tissues after episodes of hypoxia followed by normoxia. In astudy by Fantel et al. ROS (superoxide anion radicals generated in mitochondria)were detected in the embryonic tissues corresponding to the distal parts of the digitsafter two transient 30-minute episodes of hypoxia separated by normoxia in GD 14 ratembryos cultured in vitro (Fantel et al., 1995).

To investigate the role of free radicals in almokalant-induced developmental toxicity,the effects of pre-treatment with the spin-trapping agent PBN, (α-phenyl-N-t-butylnitrone) with capacity to capture free radicals, and the antioxidant OTC (2-oxo-4-thiazolidine carboxylic acid) -a promotor of gluthathione synthesis, were studied.Pre-treatment with PBN resulted in a decrease in the incidence of almokalant-induceddefects both on GD 11 and 13. No foetuses with orofacial clefts, ventricular septumdefects, tail defects or distal digital defects were observed in almokalant-treated damsafter pre-treatment with PBN. The only observed visceral malformations in the PBNpre-treated dams was a vessel defect (malposition of the aortic arch) in one foetus, andone foetus with dilated urethras treated on GD 11. Pre-treatment with OTC on GD 11reduced the incidence of septal defects. On GD 13 there was a reduction in digitaldefects, compared to dams given almokalant only. However, pre-treatment with OTCdid not cause a reduction in almokalant-induced orofacial clefts or tail defects. The

27

statistical evaluation showed that OTC did not significantly alter the incidence ofalmokalant-induced internal or external malformations on GD 11 or 13.

These results suggest that the almokalant-induced defects are mainly caused byreoxygenation damage after an episode of severe embryonic dysrhythmia, rather than“pure hypoxia”. The mechanism for such embryonic damage is unknown, but mayinvolve generation of free radicals, such as reactive oxygen and nitric oxide speciesduring the reoxygenation phase (Fantel, 1996; Fantel et al., 1999).

The fact that PBN prevented almokalant-induced malformations to a much higherextent than OTC, may have several explanations. PBN is a lipophilic radical spin-trapping agent, which rapidly penetrates cell membranes and is most likely distributedboth inside and outside of cells in embryonic tissues. PBN thus seems to have thecapacity to directly capture (trap) generated reactive oxygen species both within thecell and in the circulation. OTC in contrast, acts indirectly by promoting glutathionesynthesis by functioning as an intracellular delivery system (Williamson et al., 1982).Glutathione is the major intracellular reductant present in developing embryonictissues (Allen and Balin, 1989). Recent studies suggest that glutathione is releasedinto the circulation, presumably as a defence against ROS duringhypoxia/reoxygenation (Liu et al., 1994). However, the synthesis of glutathione (viaOTC) and the release of extracellular glutathione as a response to generated freeradicals are probably slower processes, with a lower capacity for detoxification thanimmediate trapping of free radicals by PBN, available in high concentrations bothintra - and extracellularily. An increased protective effect by PBN, compared to OTC,was also seen in a study by Zimmerman et al. (1994). PBN, but not OTC, completelyprevented cocaine-induced embryonic vascular disruption thought to be mediated byROS.

The only cardiovascular malformation detected in the group pre-treated with PBN wasa defect of the great vessels (malposition of the aortic arch) on GD 11. Even if thisfinding does not exclude hypoxia - reoxygenation damage, it suggests thatabnormalities of the great vessels arise through another mechanism. In view of this, itis of great interest to note that abnormalities of the great vessels have been inducedexperimentally by modifying the normal blood flow pattern (Jaffee, 1967; Sweeney etal., 1979) and by production of episodes of embryonic bradycardia, ventriculararrhythmia and temporary cardiac arrest (Rajala et al., 1984). A wide variation ofvessel defects, including the malposition of the great vessels, can be induced byselectively increasing, decreasing, or misdirecting blood flow in the developingcardiovascular system by mechanical methods or by cardioactive substances (Gilbertet al., 1980; Kirby, 1987; Gilbert et al., 1980). Lack of sufficient blood flow andpressure changes have been related to the induction of abnormally absent vessels.Thus, the results suggest that the observed cardiovascular defects are result of both theinduced embryonic arrhythmia per se, resulting in pressure changes and misdirectionof embryonic blood flow, and of the hypoxia-related damage.

28

As mentioned previously, it is generally accepted that the increased incidence ofembryonic/foetal death by selective IKr-blockers is a consequence of severebradycardia or of cardiac arrest. The results in this study support this hypothesis, sincepre-treatment with PBN or OTC did not significantly decrease the incidence ofembryonic death. Further, the decreased foetal weights in this and in other studieswith class III antiarrhythmics may be directly related to hypoxia, (due to bradycardia)in embryos surviving until term.

IKr-blockers, embryonic arrhythmia and developmental toxicity in otherspecies than the rat (Papers III and IV).

The purpose of this part of the thesis was to investigate if established IKr-blockersinduce embryonic dysrhythmia and developmental toxicity in other species than therat. Almokalant was studied in NMRI-mice in a conventional developmental toxicitystudy and in whole embryo culture. Further, sotalol was studied in New ZealandWhite rabbits in a limited embryo-foetal developmental toxicity study, which wasdivided into three parts, with different dosing regimes.

AlmokalantDevelopmental toxicity in the mouseAlmokalant was administered in the diet during GDs 6-15 in doses of 50, 125, and300 µmol/kg. No signs of dysfunction were observed in the pregnant dams. A dose-dependent embryonic lethality was observed resulting in a 100% embryo mortality atthe highest tested concentration in agrement with previous studies (Abrahamsson etal., 1994; Danielsson, 1993) . The mouse embryos are thought to have died duringGDs 9-12 as indicated by the decreased maternal body weight gain from GDs 9-12and onwards in the group with total litter loss. The plasma concentration ofalmokalant was measured in satellite animals that had been dosed with 300 µmol/kg,which resulted in a mean maternal plasma concentration of 1.3 µmol/l on GD 15.

The surviving foetuses in the control, 50 and 125 µmol/kg groups were examined forexternal, visceral and skeletal defects. There were no significant differences betweenthe treated groups concerning external- and visceral malformations. The mean foetalweight was statistically significantly reduced in the group treated with 125 mg/kg. Inthe 50 and 125 µmol/kg groups, the incidence of some minor skeletal defects wasstatistically significantly increased compared with the control group. These minorskeletal defects were manifested in the intermediate dose group by a higher incidenceof unossified centers of cervical vertebrae (~12% affected foetuses versus ~7% and~6% in the low dose and control groups, respectively), and a decreased degree ofcalcification of the forepaw and hindpaw (~17% affected foetuses versus ~8% and~9% in the low dose and control groups, respectively). The decreased calcificationwas mainly localised to the distal phalanges. These results may reflect slight digitalreduction defects of the same type as those observed in the rat (Webster et al., 1996;

29

Papers I and II). In the low dose group, the increase was due to a higher incidence ofcervical ribs and an extra sternebral ossification center compared with the controlgroup. The percentage of foetuses showing skeletal variants (extra ribs) wassignificantly increased in the low and intermediate dose groups compared withcontrols. Overall, the results show that almokalant is a developmental toxicant inmice, causing dose-dependent increase in embryonic death and skeletal defects, aswell as decreased mean foetal weights. The absence of external, visceral, and someskeletal defects (compared to the rat), may be explained by the wide dose intervalbetween the intermediate and high dose groups. Further, single dose exposure onspecific days of gestation might increase the chance of observing / inducing theseeffects. An example of such a dose regime with cisapride is presented in the nextsection (pages 31-34).

Effects on the mouse embryonic heartAlmokalant was tested in GD 10 mouse embryos cultured in vitro. This daycorresponds in developmental stage to GD 11 in the rat, when almokalant was tested.Almokalant caused bradycardia/arrhythmia at concentrations higher than 325 nmol/l,in a concentration-dependent manner. A statistically significant decrease in heart rate(bradycardia) was obtained compared to control embryos as follows; 650 nmol: 13%,1.3 µmol: 20%, 2.6 µmol: 27% and 5.2 µmol: 42%. At the concentration 2.6 µmol,arrhythmia (irregular heart rhythm and/or episodes of cardiac arrest) was observed inone out of 17 examined embryos, and at 5.2 µmol, arrhythmia was observed in 7 outof 15 examined embryos. The results obtained when mice embryos were cultured invitro with almokalant were similar to those obtained when rats were cultured withalmokalant and other pure IKr-blockers (Ban et al., 1994; Spence et al., 1994; Websteret al., 1996).

Correlation between effects on the embryonic heart and developmental toxicityThe concentration (650 nmol/l) where almokalant reduced the embryonic heart rate invitro correlates well with the observation that a concentration of 1.3 µmol/l in vivo,caused embryonic death in the present study. This strengthens the theory that thepharmacologically induced embryonic bradycardia/arrhythmia that might lead toimpaired embryonic circulation and hypoxia as a secondarily effect mediates thedevelopmental toxicity of IKr-blockers not only in rats, but also in mice. The mainadverse manifestations in this study (increased incidence of embryo lethality in the300 and 125 µmol/kg groups and decreased foetal weight in the 125 µmol/kg group)are well known adverse manifestations produced by hypoxia due to different causes(Grabowski, 1970). In view of the similarities in adverse effects of the embryonicheart in the mouse and in the rat, we conclude that almokalant has teratogenicpotential also in the mouse.

30

SotalolDevelopmental toxicity in the rabbitThis study aimed to investigate if the IKr-blocker sotalol could induce developmentaltoxicity in rabbits during the corresponding part of embryonic development as IKr-blockers had caused developmental toxicity in the rat. Initially, repeated dosing wasperformed on GDs 13-16. Administration of 150-300 mg/kg sotalol during this periodresulted in 100% embryonic death. Thereafter, single doses were given on GD 8, 9,10, 12, 13, 14, 15, 16, or 17. Single doses of 100 mg/kg of sotalol on GD 8, 9, or 10,resulted in a slight increase in the incidence of embryonic death, compared tohistorical controls. Marked increased embryonic death (55-90%) was noticed after asingle dose of 100-150 mg/kg of sotalol on GD 12, 13, 14, 15 or 16. No external orvisceral abnormalities were observed in the surviving foetuses. In a subsequent part ofthe study, different doses of sotalol (50, 60, 75, 85, or 100 mg/kg) were administeredon GD 14, one of the most sensitive days for induction of embryonic/foetal death.There was total litter loss (100% embryonic death) at 85 and 100 mg/kg.Administration of 60 and 75 mg/kg resulted in increased embryonic death, decreasedlitter size and a corresponding increase in mean foetal weight. The dose of 50 mg/kggave a statistically significant elevated level of embryonic/foetal death (28.8%), butdid not severely affect litter size or mean foetal weight. No external or visceralabnormalities were observed in the surviving foetuses.

In contrast to the adult rat, the adult rabbit shows a high susceptibility of developingarrhythmias (Torsade de Pointes type) when exposed to IKr-blocking drugs such asalmokalant and dofetilide (Carlsson et al., 1993). In this study, two animals diedwithout any adverse clinical signs preceding death. It is reasonable to assume that thetwo cases of maternal death (one at 225 mg/kg after three days of dosing and one aftera single dose of 100 mg/kg) were secondary to maternal cardiac arrhythmia. Theobservation that dose levels in the range of 100 – 225 mg/kg induced death in onlytwo out of 22 adult animals (9%), compared to the high incidence of embryonic death(approximately 70%), indicates that the early embryonic heart is more susceptible toreact with arrhythmias than the adult heart when exposed to IKr-blockers.

Correlation between the effects on the embryonic heart and developmental toxicityIt is reasonable to assume that the sotalol-exposed rabbit embryos die as aconsequence of embryonic bradyarrhythmia/ cardiac arrest, in the same way as hasbeen proposed happens after embryonic exposure to IKr-blockers in the rat. Thisassumption is strengthened by results from studies in rat WEC where sotalol causeddose-dependent bradyarrhythmia in the embryonic heart on GDs 11 and 13. Sotalolcaused a statistically significant embryonic cardiac effect at 50 µM (Webster et al.,1996). The magnitude of these cardiac adverse effects has been related to embryonicdeath in vivo in this species. In this study in rabbits, a single dose of sotalol(50 mg/kg) on GD 14 resulted in embryonic death (28% of the embryos) at a Cmax of~46 µM. These results indicates that the rabbit embryonic heart is at least as likely toreact with lethal bradyarrhythmia as the rat embryonic heart.

31

Comments on developmental toxicityThe results clearly showed that the IKr-blocker sotalol causes phase- andconcentration-dependent embryonic death in the rabbit in the same way and duringthe corresponding period as IKr-blockers cause embryo lethality in the rat. It could beargued that it is the β-blocking property of sotalol that is responsible for theembryonic bradycardia and subsequent developmental toxicity, and not the IKr-blockage. However, results from rodents indicate that this is not the case. In the rat,the heart begins to beat around GD 10, but does not receive the extrinsic autonomicinnervation until about GD 16-18 (Adolph, 1965; Gomez, 1958; Hogg, 1957). Thesinus node is innervated on about GD 16. In early embryonic life, there is a slow fixedheart rate, which increases progressively with age. β-adrenergic receptors are presentin the non-innervated embryonic rodent heart (Martin et al., 1973; Robkin et al.,1974). For example, in the foetal heart β1-receptors are already in place in earlydevelopment (GD 12), but receptor coupling to adenylate cyclase (indicating someaspect of physiological function) is low at this stage. In contrast, coupling shows aspike at GD 18 and furthermore, on this day the concentration of β-receptors in foetaltissues is five times higher than on GD 12 (Slotkin et al., 1994). These data suggest anassociation between β-receptor expression and cell differentiation in late foetal stagesand may contribute to explaining the relative absence of effects of β-antagonists(despite the occurrence of beta-receptors) on the embryonic/foetal heart. Altogether,these results suggest that the ability of β-blockers to cause embryonic/foetalbradycardia by direct pharmacological action is low during the susceptible stages forinduction of hypoxia-related malformations. This is also supported by the fact thatβ-blockers as a group have not shown any developmental toxic potential in teratologystudies (Danielsson and Webster, 1997).

Drugs that block IKr as a side effect (Papers V and VI)

In this part of the thesis the purpose was to investigate if two compounds, cisaprideand phenytoin, whose primary action is not IKr-blockage, also have the potential tocause embryonic arrhythmia, embryonic death and hypoxia-related malformations inthe same way as selective IKr-blockers. Both compounds have recently been shown tohave IKr-blocking properties as a side effect. Cisapride – (Carlsson et al., 1997),Phenytoin – IK (Nobile and Vercellino, 1997), phenytoin – IKr (Danielsson et al.unpublished data). Mosapride was used as a reference compound and included in thewhole embryo culture. Mosapride is structurally related to cisapride and has the samepharmacological activity (gastrointestinal motility stimulation) but it does not blockIKr at therapeutic concentrations (Carlsson et al., 1997). Phenytoin is an antiepilepticdrug (Rogawski and Porter, 1990).

The background data available for cisapride and phenytoin was completely different.The proarrhythmic potential of cisapride in adults and its IKr-blocking properties wereknown, while the information regarding developmental toxicity was very scarce.However, if the theory about IKr blockage being the cause of developmental toxicityvia induction of embryonic cardiac arrhythmia and hypoxia/reoyxgenation damage is

32

correct, then cisapride ought to induce similar effects as pure IKr-blockers such asalmokalant. With phenytoin it was the other way round: the potential to causedevelopmental toxicity both in animals and humans was well established, but not themechanism. Different theories regarding the mechanism have been proposed, forexample the formation of reactive metabolites (Dansky and Finnell, 1991; Finnell andDansky, 1991). In view of the similarities reported in the malformation patterns andearly changes preceding the defects, including haemorrhage and necrosis of distalparts of digits between PHT (Danielsson et al., 1992) and IKr-blockers (Webster et al.,1996) in animals studies, we decided to investigate if PHT would affect theembryonic heart at concentrations inducing developmental toxicity.

Cisapride and mosaprideEffects on the embryonic heartIn rat whole embryo culture, the control group heart rate decreased slightly (~6%) atthe end of the study period (at 2 hours), compared to the original heart rate. Mosapridedid not statistically significantly affect the embryonic heart rhythm at any of the testedconcentrations; 100, 1,000 and 2,000 ng/ml. The heart rate was approximately 90% ofthe heart rate recorded before administration of mosapride at the end of the studyperiod. Cisapride (100 ng/ml) caused slight bradycardia (~10 %) but no irregular heartrhythm, similar to what had been obtained with mosapride. In contrast, exposure to1,000 ng/ml of cisapride resulted in severe bradycardia manifested as a mean decreasein heart rate of 34 % and an irregular heart rhythm observed in almost all embryos (15out of 16). The arrhythmia was of the same kind as been described for the selectiveIKr-blocker almokalant in previous studies (rat – Webster et al., 1996, mouse –Paper III). Cisapride caused decreased heart rate and arrhythmia in rat embryos duringthe same period as that in which selective IKr-blockers induce identical embryoniccardiac adverse effects (Abrahamsson et al., 1994; Ban et al., 1994; Spence et al.,1994; Webster et al., 1996).

Developmental toxicitySince mosapride did not induce embryonic cardiac arrhythmia in vitro, only cisapridewas investigated for hypoxia-related malformations in vivo in the rat. Data suppliedby the manufacturer (unpublished) showed no evidence of any embryotoxic orteratogenic effects of mosapride in the rat and in the rabbit in conventional teratologystudies even at very high doses (300 mg/kg - rat, 125 mg/kg - rabbit). This data agreewith the observed absence of arrhythmogenic potential in cultured embryos.

33

Figure 6: Defects of the left forepaw after administration of cisapride to pregnant rats on GD 13.

A single dose of cisapride on GD 13 caused deteriorated condition such as piloerection and decreased motor activity in this study. The signs of maternal toxicitywere dose-dependent, with the most severe signs of the longest duration (up to twodays) after 200 mg/kg. Dosing with 75 mg/kg and 50 mg/kg did not seem to affect thedams. Cisapride was shown to have both embryotoxic and teratogenic potential. Ashas been observed in studies with different selective IKr-blockers (Danielsson,1993),(Ban et al., 1994; Spence et al., 1994; Webster et al., 1996) high incidences(55-82%) of embryonic death and a decrease in mean foetal weight in survivingfoetuses were seen after administration of high doses (100, 160 and 200 mg/kg) ofcisapride. When the dose was lowered to 75 mg/kg, the incidence of embryonic deathdecreased considerably (15%) and at 50 mg/kg there was no statistically significantdifference in comparison to controls. In both these dose groups (50 and 75 mg/kg) themajority of foetuses could be examined for external malformations. Identical digitalreduction defects (brachydactyly) and fusion of digits (syndactyly) as those seen aftersingle dose administration of the selective IKr-blockers almokalant and dofetilide aswell as hypoxia (Papers I and II; Leist and Grauwiler, 1973; Webster et al., 1996), ofthe same stage of development (GD 13) were obtained with cisapride (figure 6).

Correlation between the effects on the embryonic heart and developmental toxicityBlood samples taken one hour after oral administration showed that a maternal plasmaconcentration of 1,100 ng/ml resulted in adverse effects on the embryonic/foetaldevelopment (embryonic death and malformations). Corresponding concentrations ofcisapride, 1,000 ng/ml in the whole embryo culture system, caused severe bradycardiaand arrhythmia.

34

Comments on developmental toxicityThese results indicate that drugs with IKr-blocking properties as a side effect alsocause embryonic death and phase-specific defects due to induced embryonicarrhythmia, in the same way as selective IKr-blockers do. The only known majordifference between cisapride and mosapride is the ability of cisapride to block the IKr.The IKr-blocking potential of cisapride is thus most likely of major importance toexplain why almost all rat embryos exposed to cisapride had arrhythmia and severebradycardia at 1,000 ng/m in this study, and while no embryos exposed to 2,000 ng/mlof mosapride showed such effects. These results further strengthen the idea thatcisapride’s potential to cause embryonic arrhythmia and malformations are aconsequence of its ability to exert pharmacological action via inhibition of IKr. In vitrostudies show that the potency of cisapride to block IKr is similar to the selectiveIKr-blocker dofetilide (Carlsson et al., 1997). The differences in developmentaltoxicity between cisapride and mosapride could not be attributed to differences inpharmacokinetics, since both drugs are readily absorbed, with a similar half-life inrats (1-2 hours) and transfer to the foetus in this species (cispride- (Michiels et al.,1987), mosapride- unpublished data). Neither can the differences be attributed todifferences in their chemical structures or primary pharmacological effect (Carlsson etal., 1997).

There are no published experimental teratology studies on cisapride in the openliterature. The FDA pregnancy labelling, which summarises non-clinical studiesconducted by the manufacturer, mentions embryotoxicity/foetotoxicity at160 mg/kg/day and reduced birth weight of pups and adversely affected pup survivalat 40 and 160 mg/kg/day in rats. In rabbits, cisapride was embryotoxic and foetotoxicat 20 mg/kg/day or higher. The pregnancy text mentions that there was no evidence ofa teratogenic potential of cisapride. In this study, we have observed a sharp doseresponse curve, with no adverse effects at 50 mg/kg, pronounced embryonic death andteratogenicity at 75 mg/kg, and almost total litter loss at 160 mg. A likely reason formissing the teratogenic potential in the original rat teratology study is thus the widedose interval between the two doses used (40mg/kg/day and 160 mg/kg/day).

Phenytoin

Phenytoin has been associated with a variety of developmental toxicity manifestationsincluding embryonic death, intrauterine growth retardation/decreased foetal weights,and a variety of malformations including orofacial clefts, limb defects, skeletalossification defects, heart and urogenital abnormalities, as well as CNS defects suchas dilated ventricles (hydrocephalus) in both animal and human studies (Dansky andFinnell, 1991; Finnell and Dansky, 1991). Both almokalant and hypoxia have beenassociated with a very similar pattern of developmental toxic effects (see table 2). Thepurpose of this study was to investigate if PHT would affect the embryonic heart atconcentrations that induce developmental toxicity. In this study in the rat, embryonicdeath, decreased foetal weights and orofacial clefts were chosen to represent

35

developmental toxicity after a single dose of PHT on GD 11, since they are easilyrecognised. Further orofacial cleft was induced with almokalant on GD 11. Thematernal plasma concentrations, both total and free, of PHT were followed during24 hours after dosing with PHT.

Developmental toxicity related to maternal plasma concentrationsA non-toxic dosing regime of 150 mg/kg po (Rowland et al., 1990) did not induce anydevelopmental toxicity. The plasma concentration of PHT peaked 4 hours after dosing(Cmax total 63 µM and Cmax free 7 µM). It declined slowly over the next 20 hours to41 µM (total) and 4.4 µM (free) at the end of the sampling period.A dosing regime, 100 mg/kg ip, which has been reported to be slightlydevelopmentally toxic (Harbison and Becker, 1972), induced a small decrease inmean foetal weight. This dose resulted in peak concentrations at around 8 hours(Cmax total 150 µM and Cmax free 16 µM) which declined to 47 µM (total) and 6.3 µM(free) by 24 hours. During the 8 hours after dosing the free plasma levels after100 mg/kg ip were 2-7 times higher than those seen after administration of 150 mg/kgpo. The AUC for this dose was more than double that of the 150 mg/kg oral dose. Anestablished developmentally toxic dosing regime, 150 mg/kg ip (Harbison andBecker, 1972), induced a relatively high incidence of early embryonic death (24%)with almost all litters (9/11) affected. Foetal weight was statistically significantlyreduced compared to control. As predicted, this single dose of PHT on GD 11 inducedteratogenicity, specifically orofacial clefts, which occurred in 12 out of 105 foetuses.There was persistently high plasma concentrations (Cmax total ∼240 µM and Cmax free33 µM) for the entire 24 hours following dosing giving an AUC (total or free) doublethat of the 100 mg/kg ip dose.

Effects on the embryonic heartPhenytoin caused a dose-dependent decrease in heart rate in rat WEC. When ratembryos were cultured in 75% rat serum, the embryonic heart rates were statisticallysignificantly decreased after the addition of 200 µM (-32%) or 100 µM (-16%) ofphenytoin, when compared to the heart rates before addition of phenytoin. This agreeswith earlier studies by our group, where the heart rates were compared to controls(Danielsson et al., 1997).

A “bridging” experiment between in vivo and in vitro conditions was also performed.The dams were administered phenytoin in vivo, and the embryos were removed fromthe uterus after four hours, placed in culture medium in vitro, and the heart rates wereexamined after 30 minutes in culture. The doses and routes of administration werechosen to represent both developmentally toxic and non-toxic levels of phenytoin asobserved in the in vivo part of this study. The embryos from the dams treated with thedevelopmentally toxic dose of 150 mg/kg phenytoin ip had a statistically significantlydecreased heart rate, 25% lower than those from the control dams. After treatmentwith the developmentally non-toxic dose of 150 mg/kg phenytoin po, there was aslight, but not statistically significant decrease of 7%. In general, the embryos with

36

slower heart rates were paler and the contraction force of the heartbeats seemed lowercompared to the embryos with normal heart rates.

Correlation between effects on the embryonic heart and developmental toxicityA dosing regime of phenytoin that does not cause any foetal adverse effects(150 mg/kg po) resulted in a total plasma concentration that was lower than thatwhich causes embryonic bradyarrhythmia.In contrast, dosing regimes that resulted in total and free plasma concentrations thatcaused slight (100 mg/kg ip) and marked (150 mg/kg ip) bradyarrhythmia caused inthis experiment slight and marked developmental toxicity, respectively.

Comments on developmental toxicityThe close correlation that we have shown between embryonic arrhythmia anddevelopmental toxicity for phenytoin strengthens the idea that non-selective IKr-blocking drugs should be considered as potential developmental toxicants, due to theirability to cause embryonic arrhythmia. In addition to acting on K+ channels,phenytoin acts on both Na+ and Ca2+ channels. In view of this it cannot be excludedthat the actions of Na+ and Ca2+ ions contribute to the cardiac dysrhythmia in theembryo. However, in view of the considerable similarities in developmental toxiceffects between selective IKr-blockers, and our very recent data showing that PHTinhibits IKr (Danielsson et al. unpublished), it is reasonable to assume that action onIKr is the major reason for the observed embryonic dysrhythmia.

WEC: an indicator for developmental toxicity and for proarrhythmicpotential in adults?

Interest in reliable and easy tests for proarrhythmic potential of especially non-cardiovascular drugs has increased dramatically. Various tests have been proposed bythe regulatory agencies, the medicinal industry themselves, and by Contract ResearchOrganisations (CROs), including studies with Purkinje fibres, heart organ baths fromspecies expressing functional IKr, and cell-lines expressing HERG. One of the majordrawbacks has been that the relatively simple mouse and rat are not suitable due to thefact that their hearts lack IKr-dependence. This is true for the adult rodent heart, butnot for the rodent embryonic heart as has been shown in this thesis. We propose thatobservation/recording of heart rhythm in rodent embryos cultured in vitro during arestricted sensitive period of development is a good test for studying the potential ofsubstances to induce developmental toxicity, and arrhythmia in sensitive adults.Table 3 shows a compilation of compounds tested in WEC. The drugs have beentested up to 20 times the clinical concentration. Some of the tested compounds areanti-epileptic drugs such as trimethadione, carbamazepine and phenobarbital that havebeen associated with a similar spectrum of foetal adverse effects as phenytoin(Dansky and Finnell, 1991; Finnell and Dansky, 1991). Their common capacity to

37

cause embryonic dysrhythmia during the sensitive period via interference with ionchannels seems to be a likely unifying mechanism for their foetal adverse effects.Another interesting pair of compounds are the antihistamines terfenadine andfexofenadine. Terfenadine has induced Torsade de Pointes in some susceptiblepatients and it blocks IKr (Taglialatela et al., 1998). Fexofenadine is a metabolite ofterfenadine and is today often used instead of terfenadine since it seems to lackproarrhythmic potential.

Although not all data for all compounds are available the trend is clear. Induction ofsevere embryonic bradycardia (= statistically significant and usually at least ~20%decrease in heart rate) and arrhythmia seems to be a good indicator of bothdevelopmental toxicity and proarrhythmic potential in adults. Using rodent wholeembryo culture may have some of these advantages over other proposed test systems:

• it is a relatively simple method to carry out• it is based on the rat or the mouse, which are common laboratory animals used in

toxicology testing• the effects are tested on a whole organ, not only a specific cell type, which could

increase the precision of the test• since it is an intact, although immature, heart it may react with arrhythmia also for

compounds acting on other channels than IKr, such as the L-type Ca2+ and Na+

channels, which are expressed at least from GD 11 in the mouse embryo, andother types of arrhythmia than those induced by IKr may also be of importance.

• it is probably possible to further develop the method by including ECG recordingsof the embryonic heart (as in the study by (Abrahamsson et al., 1994)).

Further studies testing more compounds, however, are needed before the wholeembryo culture can be established as a good test system for developmental toxic andproarrhythmic potential.

38Tabl

e 3:

Sum

mar

y of

com

poun

ds te

sted

in w

hole

em

bryo

cul

ture

.

Subs

tanc

e

WEC

Seve

rebr

adyc

ardi

a

WEC

Arrh

ythm

ia,

Car

diac

arre

stI K

r-blo

ckag

e

Anim

al m

odel

sEm

bryo

toxi

city

/hy

poxi

a-re

late

dte

rato

geni

city

Clin

ical

ly

Tera

toge

nici

ty

Clin

ical

lyAr

rhyt

hmia

,C

ardi

ac a

rrest

,To

rsad

e de

Poi

ntes

Alm

okal

ant

Yes

Yes

Yes

Yes

No

info

rmat

ion

Yes

Dof

etili

deYe

sYe

sYe

sYe

sN

o in

form

atio

nYe

sL

691,

121

Yes

Yes

Yes

Yes

No

info

rmat

ion

Yes

Sota

lol

Yes

Yes

Yes

Yes

No

info

rmat

ion

Yes

Phen

ytoi

nYe

sYe

sYe

sYe

sYe

sYe

sPh

enob

arbi

tal

Yes

No

? U

nder

inve

stig

atio

nYe

sYe

sN

o in

form

atio

n

Car

bam

azep

ine

Yes

Yes

Yes,

pre

l dat

aYe

sYe

sYe

sTr

imet

hadi

one/

dim

etha

dion

eYe

sYe

s?

Und

erin

vest

igat

ion

Yes

Yes

No

info

rmat

ion

Valp

roic

aci

dN

oN

o?

Und

erin

vest

igat

ion

No

(yes

, but

not

hypo

xia

rel.)

No

(yes

, but

not

hypo

xia

rel.)

No

info

rmat

ion

Viga

batr

inN

oN

o?

Und

erin

vest

igat

ion

No

info

rmat

ion

No

info

rmat

ion

No

info

rmat

ion

Cis

aprid

eYe

sYe

sYe

sYe

sN

o in

form

atio

nYe

sM

osap

ride

No

No

No

No

info

rmat

ion

No

info

rmat

ion

Terf

enad

ine

Yes,

pre

l dat

aN

o, p

rel d

ata

Yes

No

info

rmat

ion

No

info

rmat

ion

Yes

Fexo

fena

dine

No,

pre

l dat

aN

o, p

rel d

ata

No

info

rmat

ion

No

info

rmat

ion

No

info

rmat

ion

No

39

Human relevance and future studies

The human relevance of the embryonic arrhythmia related mechanism by IKr-blockersteratogenicity is unknown. However, indirect evidence indicates that the proposedmechanism may be relevant to humans. Experimental data indicates that the earlyembryo expresses IKr in mice, rats, and rabbits, and that it reacts with bradycardia andarrhythmia when exposed to IKr-blockers. These data suggest a universal functionalrole across all mammalian species (including humans) for IKr channels in thedevelopment of the mammalian cardiac electrical system.

Phenytoin is an established human teratogen and has recently been shown to inhibitIKr (Danielsson et al. unpublished). Phenytoin causes very similar patterns ofdevelopmental toxic effects in a variety of species (including humans, rabbits, rats,and mice), which strengthens the idea of a common mechanism across species. Thefoetal hydantoin syndrome in humans includes both major malformations such asorofacial clefts and minor malformations such as distal digital defects, and growthretardation. It has been possible to recreate these manifestations in animal models attherapeutic concentrations. Both phenytoin and selective IKr-blockers induce orofacialclefts and distal digital defects in animals preceded by identical damage with respectto vascular disruption and haemorrhage. In clinical practice, it is well known thatepisodes of severe hypoxia during labour may result in foetal intracranialhaemorrhage and permanent tissue damage. Little information is available, however,regarding the consequences of hypoxia in the human embryo in early pregnancy. It isnot unreasonable to assume that the human embryo may respond to hypoxic episodeswith stage-specific damage, in the same way as other mammalian species.

It would be interesting to investigate if human embryos express functionalIKr-channels, and if so, during which stage of development. Another interesting studywould be to determine whether children born with PHT-induced malformations sufferfrom long QT syndrome, being more susceptible towards drug-induced prolongationof the QT interval. Regarding the promising method of monitoring rodent embryonicheart rhythms with WEC, further studies will tell if it can be used as one reliable testfor proarrhtymic potential in adults as well as teratogenicity of non-cardiovasculardrugs – a “cardiac Ames test”.

40

CONCLUDING REMARKS

During the work with this thesis, general interest in compounds acting on IKr, also fordrugs not being developed as class III antiarrhythmics, increased dramatically, due totheir proarrhythmic potential. The interest has been focussed on cardiac arrhythmia inadults and not on their potential for causing developmental toxicity. But, as has beenshown in this thesis, these two characteristics are strongly linked.

Referring to the aims with this study:

• The different studies (Papers I-III) with the selective IKr-blocker almokalantstrongly support a hypoxia-related mechanism due to embryonic arrhythmiaacross species (rat, mouse).

• The decrease in almokalant-induced malformations by the spin-trapping agentPBN (Papers I-II) suggests that the defects are mainly caused by reoxygenationdamage after episodes of severe embryonic dysrhythmia, rather than “purehypoxia”.

• Sotalol induced embryonic death in the rabbit (Paper IV). The adult rabbit heartexpresses functional IKr and is a sensitive species for IKr-induced arrhythmia. Theembryonic heart seems to be more sensitive than the adult heart, reacting witharrhythmia at lower concentrations than the adult heart, as indicated in this study.

• The observed developmental toxicity has been linked to the IKr-inhibition not onlyfor the selective IKr-blocker almokalant, but also for non-selective IKr-blockerssuch as sotalol, cisapride, and phenytoin. This leads to the hypothesis that if acompound exerts a blocking effect on IKr as the therapeutic mechanism or as a sideeffect, it has the potential to cause embryonic/foetal toxicity and teratogenicity.This might occur irrespective of the original indications that the compound wasdeveloped for.

• Since the heart in rodent embryos, in contrast to adult hearts in rodents, expressesa functional IKr-channel, WEC (whole embryo culture) with observation of theheart rhythm may become a new, relatively simple screening method. Thus, theembryonic heart may during a restricted period serve as an indicator forproarrhythmic potential of different compounds, and of their potential to inducedevelopmental toxicity.

41

ACKNOWLEDGEMENTS

This work was performed at AstraZeneca R&D, Södertälje, Safety Assessment, and atthe Department of Pharmaceutical Biosciences, Division of Toxicology, UppsalaUniversity. I wish to express my sincere gratitude to everyone who made this workpossible, and especially:

To Bengt Danielsson, my supervisor, who offered me the opportunity to become aPh.D. student (and a study director at the same time). Thank you for sharing yourknowledge in reproductive toxicology and for always being enthusiastic, even whenthe results did not turn out as expected.

To Peter Moldéus, Head of Safety Assessment, for his generosity in giving me thetime and opportunity to do this work.

To Lennart Dencker, Head of the Division of Toxicology, for taking me in as aninternal/external Ph.D. student and for your kind support and interest in my work.

To my co-authors Alf Wallin, Faranak Azarbayjani, Inger Öhman, and WilliamWebster for a pleasant, fruitful collaboration and valuable scientific discussions.

To Katrin Wellfelt collaborator and co-author, for “skilful technical contribution”,friendship, laughs and talk about any- or everything, and especially for the workduring your maternity leave.

To Arvid Wellfelt for accompanying your mother to Safety Assessment (no choiceI suppose), even though I wasn’t too happy about your throwing up on me.

To all present and past friends and colleagues at the Department of PharmaceuticalBiosciences, Division of Toxicology.

To all present and past friends and colleagues at the Department of ReproductiveToxicology, Safety Assessment, especially those involved in my studies, and thatmeans more or less everyone.

To my liaison officer Anita Annas for friendship and helping me to organise things inUppsala after I left for Gärtuna.

To Raili Engdahl and Lena Norgren for all help and pleasant company in the lab inUppsala, especially for helping me to manage to watchmaker forceps under themicroscope in the beginning of this work.

To Terri Mitchard and Elaine Gordon for taking urgent measures in revisions of theEnglish language now and then.

42

To Karin Cederbrant my “friend-PhD-student” at Safety Assessment forencouragement in every aspect of work and life, and nice company during late nightsof the writing process. Fuskigt att du blev doktor före mig, hm.

To my parents Gerd and Erling, who have always believed in me, encouraged me anddone everything in their power to support me, remember all the crumpled uppapers.....

To my husband Johan for endless love, support, encouragement, patience, anddelicious meals. I’m really looking forward to spending more time with you.

To all my other (BMC-) friends, who I’ve spent many joyful hours with, and whohaven’t tired on repeated phone-calls when I needed encouragement or just a break.

To MFR, CFN, and AstraZeneca for financial support.

43

REFERENCES

1996a). ADRAC on cisapride, acyclovir, SSRIs. Scrip. 2158, 27-27.

1996b). Warning on cisapride interactions. Scrip. 2123, 21-21.

Abrahamsson, C. Proarrhythmias related to delayed repolarisation. Importance ofdispersion of repolarisation and early afterdepolarisations. 1996. Göteborg,Vasastadens Bokbinderi AB.

Abrahamsson, C., Palmer, M., Ljung, B., Duker, G., Bäärnhielm, C., Carlsson, L., andDanielsson, B. (1994). Induction of rhythm abnormalities in the fetal rat heart.A tentative mechanism for the embryotoxic effect of the class IIIantiarrhythmic agent almokalant. Cardiovascular Research. 28, 337-344.

Adolph, E. F. (1965). Capacities for regulation of heart rate in fetal, infant, and adultrats. American Journal of Physiology. 209, 1095-1105.

Allen, R. G. and Balin, A. K. (1989). Oxidative influence on development anddifferentiation: an overview of a free radical theory of development. FreeRadical Biology & Medicine. 6, 631-661.

Anonymous (1989). Preliminary report: effect of encainide and flecainide onmortality in a randomized trial of arrhythmia suppression after myocardialinfarction. The Cardiac Arrhythmia Suppression Trial (CAST) Investigators[see comments]. New England Journal of Medicine. 321, 406-412.

Anonymous (1992). Effect of the antiarrhythmic agent moricizine on survival aftermyocardial infarction. The Cardiac Arrhythmia Suppression Trial IIInvestigators. New England Journal of Medicine. 327, 227-233.

Azarbayjani, F. and Danielsson, B. R. (1998). Pharmacologically induced embryonicdysrhythmia and episodes of hypoxia followed by reoxygenation: a commonteratogenic mechanism for antiepileptic drugs? Teratology. 57, 117-126.

Ban, Y., Konishi, R., Kawana, K., Nakatsuka, T., Fujii, T., and Manson, J. M. (1994).Embryotoxic effects of L-691,121, a class III antiarrhythmic agent, in rats.Archives of Toxicology. 69, 65-71.

Barrow, M. V. and Taylor, W. J. (1969). A rapid method for detecting malformationsin rat fetuses. Journal of Morphology. 127, 291-305.

Bashir, Y., Thomsen, P.-E. B., Kingma, J. H., Moller, M., Wong, C., Cobbe, S. M.,Jordaens, L., Campbell, R. F., Rasmussen, H. S., and Camm, A. J. (1995).Electrophysiologic profile and efficacy intravenous dofetilide (UK- 68,798), anew class III antiarrhythmic drug, in patients with sustained monomorphicventricular tachycardia. American Journal of Cardiology. 76, 1040-1044.

44

Beyer, T., Brachmann, J., and Kubler, W. (1993). Comparative effects of d-sotaloland l-sotalol on the atrioventricular node of the rabbit heart. Journal ofCardiovascular Pharmacology. 22, 240-246.

Bigger, J.T. and Hoffman, B. F. (1980). The Pharmacological Basis of Therapeutics.MacMillan Publishing Co, New York.Bigger, J.T. and Hoffman, B. F.(Antiarrhyhtmics Drugs. In A. G. Gilman, L. S. Goodman, and A. Gilman,Eds.), pp. 761-792. MacMillan Publishing Co, New York.

Bran, S., Murray, W. A., Hirsch, I. B., and Palmer, J. P. (1995). Long QT syndromeduring high-dose cisapride. Archives of Internal Medicine. 155, 765-768.

Brent, R. L. and Franklin, J. B. (1960). Uterine Vascular Clamping: New Procedurefor the Study of Congenital Malformations. Science. 132, 89-91.

Bronsky, P. T., Johnston, M. C., and Sulik, K. K. (1986). Morphogenesis of hypoxia-induced cleft lip in CL/Fr mice. Journal of Craniofacial Genetics &Developmental Biology - Supplement. 2, 113-128.

Carlsson, L., Abrahamsson, C., Andersson, B., Duker, G., and Schiller-Linhardt, G.(1993). Proarrhythmic effects of the class III agent almokalant: importance ofinfusion rate, QT dispersion, and early afterdepolarisations [see comments].Cardiovascular Research. 27, 2186-2193.

Carlsson, L., Amos, G. J., Andersson, B., Drews, L., Duker, G., and Wadstedt, G.(1997). Electrophysiological characterization of the prokinetic agents cisaprideand mosapride in vivo and in vitro: implications for proarrhythmic potential?Journal of Pharmacology & Experimental Therapeutics. 282, 220-227.

Carmeliet, E. (1992). Voltage- and time-dependent block of the delayed K+ current incardiac myocytes by dofetilide. Journal of Pharmacology & ExperimentalTherapeutics. 262, 809-817.

Chen, E. Y., Fujinaga, M., and Giaccia, A. J. (1999). Hypoxic MicroenvironmentWithin an Embryo Induces Apoptosis and Is Essential for ProperMorphological Development. Teratology. 60, 215-225.

Colatsky, T. J. and Argentieri, T. M. (1994). Potassium channel blockers asantiarrhythmic drugs. Drug Development Research. 33, 235-249.

Crumb, W. J. J. and Cavero, I. (2000). Perspectives on the Potential of Non-Cardiovascular Drugs to Adversely Alter Cardiac Repolarisation. EuropeanPharmaceutical Contractor. 63-72.

45

Danielsson, B.R. (1993). Developmental Toxicology Preclinical and Clinical Data inRetrospect. Tryckgruppen, Stockholm.Danielsson, B.R. (Malformations andFetal Hypoxia Induced by Exaggerated Pharmacological Effects. In A.Sundwall, B. R. Danielsson, O. Hagberg, E. Lindgren, P. Sjöberg, and A.Viktor, Eds.), pp. 45-52. Tryckgruppen, Stockholm.

Danielsson, B. R., Azarbayjani, F., Sköld, A. C., and Webster, W. S. (1997). Initiationof phenytoin teratogenesis: pharmacologically induced embryonic bradycardiaand arrhythmia resulting in hypoxia and possible free radical damage atreoxygenation. Teratology. 56, 271-281.

Danielsson, B. R., Danielson, M., Rundqvist, E., and Reiland, S. (1992). Identicalphalangeal defects induced by phenytoin and nifedipine suggest fetal hypoxiaand vascular disruption behind phenytoin teratogenicity. Teratology. 45, 247-258.

Danielsson, B.R. and Webster, W. S. (1997). Drug Toxicity in EmbryonicDevelopment II; Advances in Understanding Mechanisms of Birth Defects:Mechanistic Understanding of Human Developmental Toxicants. Springer,Berlin Heidelberg.Danielsson, B.R. and Webster, W. S. (CardiovascularActive Drugs. In R. J. Kavlock and G. P. Daston, Eds.), pp. 161-190. Springer,Berlin Heidelberg.

Dansky, L. V. and Finnell, R. H. (1991). Parental epilepsy, anticonvulsant drugs, andreproductive outcome: epidemiologic and experimental findings spanningthree decades; 2: Human studies. Reprod.Toxicol. 5, 301-335.

Darpö, B. and Edvardsson, N. (1995). Effect of almokalant, a selective potassiumchannel blocker, on the termination and inducibility of paroxysmalsupraventricular tachycardias: A study in patients with Wolff-Parkinson-Whitesyndrome and atrioventricular nodal reentrant tachycardia. Journal ofCardiovascular Pharmacology. 26, 198-206.

Davies, M. P., An, R. H., Doevendans, P., Kubalak, S., Chien, K. R., and Kass, R. S.(1996). Developmental changes in ionic channel activity in the embryonicmurine heart. Circulation Research. 78, 15-25.

Degenhardt, K. (1958). Analysis of Intra-Uterine Malformations of the VertebralColumn Induced by Oxygen Deficiency in Rabbits. Proc Xth Int.Congr.ofGenetics. 2, 66-67.

Doig, J. C. (1997). Drug-induced cardiac arrhythmias: incidence, prevention andmanagement. Drug Safety. 17, 265-275.

46

Echt, D. S., Lee, J. T., Murray, K. T., Vorperian, V., Borganelli, S. M., Crawford, D.M., Friedrich, T., and Roden, D. M. (1995). A randomized, double-blind,placebo-controlled, dose-ranging study of dofetilide in patients with induciblesustained ventricular tachyarrhythmias. Journal of CardiovascularElectrophysiology. 6, 687-699.

Fantel, A. G. (1996). Reactive oxygen species in developmental toxicity: review andhypothesis. Teratology. 53, 196-217.

Fantel, A. G., Person, R. E., Tumbic, R. W., Nguyen, T. D., and Mackler, B. (1995).Studies of mitochondria in oxidative embryotoxicity. Teratology. 52, 190-195.

Fantel, A. G., Stamps, L. D., Tran, T. T., Mackler, B., Person, R. E., and Nekahi, N.(1999). Role of free radicals in the limb teratogenicity of L-NAME (N(G)-nitro-(L)-arginine methyl ester): a new mechanistic model of vasculardisruption. Teratology. 60, 151-160.

Finnell, R. H. and Dansky, L. V. (1991). Parental epilepsy. anticonvulsant drugs, andreproductive outcome: epidemiologic and experimental findings spanningthree decades; 1: Animal studies. Reprod.Toxicol. 5, 281-299.

Franklin, J. B. and Brent, R. L. (1964). The effect of uterine vascular clamping on thedevelopment of rat embryos three to fourteen days old. Journal ofMorphology. 115, 273-290.

Gilbert, E. F., Bruyere, H. J., Jr., Ishikawa, S., Foulke, L. M., and Heimann, S. R.(1980). Role of decreased cardiac output in isoproterenol-inducedcardiovascular teratogenesis in chick embryos. Teratology. 21, 299-307.

Gomez, H. (1958). The development of the innervation of the heart in the rat embryo.Anat.Rec. 130, 53-71.

Grabowski, C.T. (1970). Advances in Teratology. Logos Press Ltd.,London.Grabowski, C.T. (Embryonic oxygen deficiency - a physiologicalapproach to analysis of teratological mechanisms. In D. H. M. Woollam, Ed.),pp. 125-169. Logos Press Ltd., London.

Gurley, B. J., Marx, M., and Olsen, K. (1995). Phenytoin free fraction determination:comparison of an improved direct serum injection high-performance liquidchromatographic method to ultrafiltration coupled with fluorescencepolarization immunoassay. Journal of Chromatography B: BiomedicalApplications. 670, 358-364.

Gustavsson, C. G., Vinge, E., Norlander, B. O., and Pantev, E. (1997).Pharmacokinetic evaluation of a case of massive sotalol intoxication. Annalsof Pharmacotherapy. 31, 856-859.

47

Gutteridge, J. M. (1993). Free radicals in disease processes: a compilation of causeand consequence. Free Radical Research Communications. 19, 141-158.

Hanson, J. W. and Smith, D. W. (1975). The fetal hydantoin syndrome. J.Pediatr. 87,285-290.

Harbison, R. D. and Becker, B. A. (1972). Diphenylhydantoin teratogenicity in rats.Toxicol.Appl.Pharmacol. 22, 193-200.

Hogg, I. (1957). Developing innervation of the heart in the albino rat. (abstract).Anat.Rec. 127, 470-470.

Ingalls, T. H. and Curley, F. J. (1957). Principles governing the genesis of congenitalmalformations induced in mice by hypoxia. New England Journal ofMedicine. 257, 1121-1127.

Ingalls, T. H., Curley, F. J., and Prindle, R. A. (1952). Experimental production ofcongenital anomlies, Timing and degree of anoxia as factors causing fetaldeaths and congenital anomlies in the mouse. New England Journal ofMedicine. 247, 758-768.

Jaffee, O. C. (1967). The development of the arterial outflow tract in the chickembryo heart. Anat.Rec. 158, 35-42.

Jolley, M. E. (1981). Fluorescence polarization immunoassay for the determination oftherapeutic drug levels in human plasma. Journal of Analytical Toxicology. 5,236-240.

Kalyanaraman, B., Konorev, E. A., Joseph, J., and Baker, J. E. (1995). Is free radicalgeneration an important component of ischemia/reperfusion injury? RedoxReport. 1, 181-184.

Kirby, M. L. (1987). Cardiac morphogenesis--recent research advances. Pediatr.Res.21, 219-224.

Kowey, P. R., Marinchak, R. A., Rials, S. J., and Bharucha, D. (1997). Pharmacologicand pharmacokinetic profile of class III antiarrhythmic drugs. AmericanJournal of Cardiology. 80, 16G-23G.

Kuo, C. C. and Bean, B. P. (1994). Slow binding of phenytoin to inactivated sodiumchannels in rat hippocampal neurons. Molecular Pharmacology. 46, 716-725.

Leist, K. H. and Grauwiler, J. (1973). Influence of the developmental stage onembryotoxicity following uterine vessel clamping in the rat. EuropeanTeratology Society: 2nd Conference. 173-176.

Leist, K. H. and Grauwiler, J. (1974). Fetal pathology in rats following uterine-vesselclamping on day 14 of gestation. Teratology. 10, 55-67.

48

Lewin, M. B., Bryant, R. M., Fenrich, A. L., and Grifka, R. G. (1996). Cisapride-induced long QT interval [see comments]. Journal of Pediatrics. 128, 279-281.

Liu, P., Fisher, M. A., Farhood, A., Smith, C. W., and Jaeschke, H. (1994). Beneficialeffects of extracellular glutathione against endotoxin-induced liver injuryduring ischemia and reperfusion. Circulatory Shock. 43, 64-70.

Macdonald, R. L. (1989). Antiepileptic drug actions. Epilepsia. 30 Suppl 1, S19-S28.

MacNeil, D. J. (1997). The side effect profile of class III antiarrhythmic drugs: focuson d,l-sotalol. American Journal of Cardiology. 80, 90G-98G.

Marks, T. A. and Terry, R. D. (1996). Developmental toxicity of ibutilide fumarate inrats after oral administration. Teratology. 54, 157-164.

Martin, S., Levey, B. A., and Levey, G. S. (1973). Development of the cardiac betaadrenergic receptor in fetal rat heart. Biochem.Biophys.Res.Commun. 54, 949-954.

McCord, J. M. (1993). Human disease, free radicals, and the oxidant/antioxidantbalance. Clinical Biochemistry. 26, 351-357.

Merritt, H. H. and Putnam, T. J. (1938). A new series of anticonvulsant drugs testedby experiments on animals. Arch.Neurol.Psychiatry. 39, 1003-1015.

Michiels, M., Monbaliu, J., Hendriks, R., Geerts, R., Woestenborghs, R., andHeykants, J. (1987). Pharmacokinetics and tissue distribution of the newgastrokinetic agent cisapride in rat, rabbit and dog. Arzneimittel-Forschung.37, 1159-1167.

Morawa, A. P. and Han, S. S. (1968). Studies on hypoxia--I. Gross and histologicinfluences of maternal anoxia upon the developing rat foetus. Archives of OralBiology. 13, 745-754.

Murakami, M. and Kameyama, Y. (1963). Vertebral Malformation in the FoetusCaused by Maternal Hypoxia During Early Stages of Pregnancy.J.Embryol.exp.Morph. 11, 107-118.

Naujoks, H. (1953). Der Einfluss kurzfristigen Sauerstoffmangles auf dieEntwicklung des Huhnchens in den ersten funf Bruttagen. Beitr.Path.Anat.113, 221-252.

New, D. A. (1978). Whole-embryo culture and the study of mammalian embryosduring organogenesis. Biological Reviews of the Cambridge PhilosophicalSociety. 53, 81-122.

49

Nobile, M. and Lagostena, L. (1998). A discriminant block among K+ channel typesby phenytoin in neuroblastoma cells. British Journal of Pharmacology. 124,1698-1702.

Nobile, M. and Vercellino, P. (1997). Inhibition of delayed rectifier K+ channels byphenytoin in rat neuroblastoma cells. British Journal of Pharmacology. 120,647-652.

Noguchi, J.K. and Gill, M. A. (1992). Clinical Pharmacy and Therapeutics. Williams& Wilkins, Baltimore.Noguchi, J.K. and Gill, M. A. (Cardiac arrhythmias. InE. T. Herfindal, D. R. Gourley, and L. Lloyd Hart, Eds.), pp. 635-676.Williams & Wilkins, Baltimore.

Ragsdale, D. S., Scheuer, T., and Catterall, W. A. (1991). Frequency and voltage-dependent inhibition of type IIA Na+ channels, expressed in a mammalian cellline, by local anesthetic, antiarrhythmic, and anticonvulsant drugs. MolecularPharmacology. 40, 756-765.

Rajala, G. M., Kuhlmann, R. S., and Kolesari, G. L. (1984). The role of dysrhythmicheart function during cardiovascular teratogenesis in epinephrine-treated chickembryos. Teratology. 30, 385-392.

Rang, H.P., Dale, M. M., and Ritter, J. M. (1999). Pharmacology. ChurchillLivingstone, Edinburgh.Rang, H.P., Dale, M. M., and Ritter, J. M.(The heart. ) pp. 277-250. Churchill Livingstone, Edinburgh.

Robkin, M., Shepard, T. H., and Baum, D. (1974). Autonomic drug effects on theheart rate of early rat embryos. Teratology. 9, 35-44.

Roden, D. M. (1994). Risks and benefits of antiarrhythmic therapy. New EnglandJournal of Medicine. 331, 785-791.

Rogawski, M. A. and Porter, R. J. (1990). Antiepileptic drugs: pharmacologicalmechanisms and clinical efficacy with consideration of promisingdevelopmental stage compounds. Pharmacological Reviews. 42, 223-286.

Rowland, J. R., Binkerd, P. E., and Hendrickx, A. G. (1990). Developmental toxicityand pharmacokinetics of oral and intravenous phenytoin in the rat.Reprod.Toxicol. 4, 191-202.

Sager, P. T., Nademanee, K., Antimisiaris, M., Pacifico, A., Pruitt, C., Godfrey, R.,and Singh, B. N. (1993). Antiarrhythmic effects of selective prolongation ofrefractoriness: Electrophysiologic actions of sematilide HCl in humans.Circulation. 88, 1072-1082.

Sanguinetti, M. C. and Jurkiewicz, N. K. (1990). Two components of cardiac delayedrectifier K+ current. Differential sensitivity to block by class III antiarrhythmicagents. Journal of General Physiology. 96, 195-215.

50

Slotkin, T. A., Lau, C., and Seidler, F. J. (1994). Beta-adrenergic receptoroverexpression in the fetal rat: distribution, receptor subtypes, and coupling toadenylate cyclase activity via G-proteins. Toxicology & AppliedPharmacology. 129, 223-234.

Spence, S. G., Vetter, C., and Hoe, C. M. (1994). Effects of the class IIIantiarrhythmic, dofetilide (UK-68,798) on the heart rate of midgestation ratembryos, in vitro. Teratology. 49, 282-292.

Sweeney, L. J., Clark, E. B., and Rosenquist, G. C. (1979). Aortic arch anomalies inexperimentally prodiced left heart hypoplasia. Pediatric Research. 13, 487-487. (abstract).

Taglialatela, M., Pannaccione, A., Castaldo, P., Giorgio, G., Zhou, Z., January, C. T.,Genovese, A., Marone, G., and Annunziato, L. (1998). Molecular basis for thelack of HERG K+ channel block-related cardiotoxicity by the H1 receptorblocker cetirizine compared with other second-generation antihistamines.Molecular Pharmacology. 54, 113-121.

Thomas, K. C., Hullin, D. A., and Davis, S. J. (1991). Comparison of enzymeimmunoassay and fluorescence polarization immunoassay as techniques formeasuring anticonvulsant drugs on the same analytical instrument.Therapeutic Drug Monitoring. 13, 172-176.

Tomson, T. and Kennebäck, G. (1997). Arrhythmia, heart rate variability, andantiepileptic drugs. Epilepsia. 38, S48-S51.

Vincent, G. M. (2000-08-17). Homepage of “The Sudden Arrhythmia DeathSyndrome (SADS) Foundation”. http://www.sads.org/

Viskin, S. (1999). Long QT syndromes and torsade de pointes. Lancet. 354, 1625-1633.

Wang, L., Feng, Z. P., Kondo, C. S., Sheldon, R. S., and Duff, H. J. (1996).Developmental changes in the delayed rectifier K+ channels in mouse heart.Circulation Research. 79, 79-85.

Webster, W. S., Brown-Woodman, P. D., Snow, M. D., and Danielsson, B. R. (1996).Teratogenic potential of almokalant, dofetilide, and d-sotalol: drugs withpotassium channel blocking activity. Teratology. 53, 168-175.

Webster, W. S., Lipson, A. H., and Brown-Woodman, P. D. (1987). Uterine traumaand limb defects. Teratology. 35, 253-260.

Wiesfeld, A. C., De Langen, C. D., Crijns, H. J., Bel, K. J., Hillege, HL, Wesseling,H., and Lie, K. I. (1996). Rate-dependent effects of the class III antiarrhythmicdrug almokalant on refractoriness in the pig. Journal of CardiovascularPharmacology. 27, 594-600.

51

Williamson, J. M., Boettcher, B., and Meister, A. (1982). Intracellular cysteinedelivery system that protects against toxicity by promoting glutathionesynthesis. Proceedings of the National Academy of Sciences of the UnitedStates of America. 79, 6246-6249.

Wiseman, L. R. and Faulds, D. (1994). Cisapride. An updated review of itspharmacology and therapeutic efficacy as a prokinetic agent in gastrointestinalmotility disorders. Drugs. 47, 116-152.

Wysowski, D. K. and Bacsanyi, J. (1996). Cisapride and fatal arrhythmia [letter]. NewEngland Journal of Medicine. 335, 290-291.

Yaari, Y., Hamon, B., and Lux, H. D. (1987). Development of two types of calciumchannels in cultured mammalian hippocampal neurons. Science. 235, 680-682.

Yoshida, N., Omoya, H., Kato, S., and Ito, T. (1993). Pharmacological effects of thenew gastroprokinetic agent mosapride citrate and its metabolites inexperimental animals. Arzneimittel-Forschung. 43, 1078-1083.

Zimmerman, E. F., Potturi, R. B., Resnick, E., and Fisher, J. E. (1994). Role ofoxygen free radicals in cocaine-induced vascular disruption in mice.Teratology. 49, 192-201.