Genetic and molecular basis of cardiac arrhythmias. Impact on clinical management

22 Pages·1999·0.64 MB·English

EuropeanHeartJournal(1999)20,174–195
ArticleNo.euhj.1998.1220,availableonlineathttp://www.idealibrary.comon
WorkingGroupReport
Genetic and molecular basis of cardiac arrhythmias
Impact on clinical management
S. G. Priori, J. Barhanin, R. N. W. Hauer, W. Haverkamp, H. J. Jongsma,
A. G. Kleber, W. J. McKenna, D. M. Roden, Y. Rudy, K. Schwartz, P. J. Schwartz,
J. A. Towbin and A. Wilde*
Introduction ology relevant to arrhythmias. Rather, they try to put
into a very practical perspective the ways in which
Clinical cardiologists who manage arrhythmias are ongoing progress in genetics may aVect day-to-day
increasingly faced with new complexities in manage- clinical management.
ment decisions. The once obscure science and jargon of Therecognitionthatdiversityincardiacelectro-
medical genetics is assuming a much more prominent physiology and indeed in many aspects of cardiac func-
position in the mainstream medical literature, with tion can be attributed to variable expression of specific
seemingly weekly reports of new mutations to explain genes or variability in the function of their protein
what once seemed very obscure diseases. This rapidly products has the potential to alter the way in which we
expanding knowledge base places the clinician—who think about normal and abnormal electrical heart func-
usually trained when the concepts were not a major tion. The third part of the article reviews the potential
component of the medical school or fellowship training for a genetic approach to understanding diversity in
curricula—at a disadvantage in making day-to-day cardiac function, focusing in particular on ion channels
decisionswithrespecttomanagingcommonsymptoms, and gap junction proteins as the central players in
such as unexplained syncope or heart failure. Even normal and abnormal electrophysiology. Moreover,
entertaining a diagnosis such as the congenital long QT integration of molecular function into a single cell and
syndrome or hypertrophic cardiomyopathy used to be of single cells into cellular networks reveals a multitude
a medical curiosity. Now, with increased public and of interactions eventually determining the generation
physician awareness of these and even more esoteric and conduction of the cardiac action potential and,
conditions, the questions on patient management have therefore, arrhythmogenesis.
becomemorecommon,andmorecomplex.Theyinclude Thistextistheoutcomeofaworkshopconvened
not only broad questions like ‘How can I establish (or bytheStudyGrouponMolecularBasisofArrhythmias
better yet, rule out) a diagnosis?’ but also more specific oftheWorkingGrouponArrhythmiasoftheEuropean
issues such as ‘Should this patient undergo genetic Society of Cardiology.
testing? Where? How? And how can I interpret the
results?’
Thefirstandsecondpartsofthisarticleattempt Inherited arrhythmogenic disorders
toanswerthesequestions.Theyneitherteachmolecular
geneticsnordotheyprovideanexhaustivereviewofthe Long QT syndrome
current state-of-the-art of molecular and genetic cardi-
The long QT syndrome is a familial disease[1,2]
Key words: Sudden cardiac death, genetics, arrhythmia, characterized by abnormally prolonged ventricular re-
moleculardiagnosis,basicelectrophysiology. polarization and a high risk of malignant ventricular
tachyarrhythmias, occurring often, but not always, in
*See Appendix for details of Workshop on which this report is thesettingofhighadrenergicactivity,thatis,physicalor
based.
emotional stress. Two major clinical syndromes have
Manuscriptsubmitted14July1998,andaccepted15July1998. beencharacterizedbasedonthepatternoftransmission
of the disease: a more common autosomal dominant
This article is also published in Circulation, Volume 99: 2 and
9February1999. form with a pure cardiac phenotype (Romano–Ward[3])
andarareautosomalrecessiveformcharacterizedbythe
Correspondence:SilviaG.Priori,MD,PhD,MolecularCardiology
coexistence of cardiac abnormalities and congenital
and Electrophysiology Laboratory, Fondazione ‘S. Maugeri’
IRCSS,ViaFerrata,8,27100Pavia,Italy. deafness (Jervell and Lange-Nielsen[4]).
0195-668X/99/030174+22$18.00/0
Cardiac arrhythmias 175
Table 1 Mutations in families with familial dilated cardiomyopathy
Disease Locus Gene Reference
XLCM Xp.21.2 Dystropin Muntonietal.[70]
Muntonietal.[71]
Milasinetal.[72]
Ortiz-Lopezetal.[68]
Barth Xq28 G.4.5 Bioneetal.[69]
ADDCM 1q32 ? Durandetal.[73]
2p31 ? Siuetal.[74]
9q13-q21 ? Krajinovicetal.[75]
10q21-q23 ? Bowlesetal.[76]
3p22-p25 ? Olsenetal.[78]
15q14 Actin Olsenetal.[79]
CDDC 1p1-1q1 ? Kassetal.[77]
FHCM 1q3 cTnT Thierfelderetal.[39]
Watkinsetal.[54]
3p MELC Poetteretal.[38]
7q3 ? MacRaeetal.[43]
11p11.2 MyPBC Bonneetal.[40]
Watkinsetal.[41]
12q23-q24.3 MRLC Poetteretal.[38]
14q11-q12 BetaMHC Geisterferetal.[37]
15q2 AlfaTM Thierfelderetal.[39]
Watkinsetal.[54]
19p13.2-q13.2 ? Kimuraetal.[42]
FA.Fib 10q22-q24 ? Brugadaetal.[89]
PFHB-I 19q13.2-q13.3 ? Brinketal.[90]
DeMeeusetal.[92]
LQTS(R–W) 3p21-p23 SCN5A Wangetal.[9]
4q25-q27 ? Schottetal.[8]
7q35-q36 HERG Curranetal.[12]
11p15.5 KvLQT1 Wangetal.[5]
21q22.1-p22 minK Splawskietal.[15]
LQTS(JLN) 11p15.5 KvLQT1 Neyroudetal.[19]
Splawskietal.[20]
21q22.1-q22 minK Schultze-Bahretal.[16]
ARVD 1q42-q43 ? Rampazzoetal.[61]
14q12-q22 ? Severinietal.[60]
14q23-q24 ? Rampazzoetal.[59]
2q32.1-q32.2 ? Rampazzoetal.[62]
NAXOS 17q21 ? Coonaretal.[58]
F-IVF 3p21-p23 SCN5A Chenetal.[82]
XLDCM=X-linked dilated cardiomyopathy; ADDCM=autosomal dominant dilated cardio-
myopathy; Barth=Barth syndrome; CDDCC=conduction defect and dilated cardiomyopathy;
FHCM=familial hypertrophic cardiomyopathy; FA.Fib=familial atrial fibrillation; PFHB-
I=progressivefamilialheartblocktypeI;LQTS(R–W)=longQTsyndromeRomano–Wardtype;
LQTS(JLN)=longQTsyndromeJervellandLange-Nielsentype;ARVD=arrhythmogenicright
ventricular cardiomyopathy; NAXOS=Naxos disease; F-IVF=familial idiopathic ventricular
fibrillation.
Long QT syndrome genes ofthesefivelocihavebeendescribed,sothereareother
Five loci[5–8] have been associated with the Romano– disease genes. The recognition that the long QT syn-
Ward long QT syndrome and they are located on dromeisactuallyagroupofionchanneldiseaseswitha
chromosomes3,4,7,11and21(Table1).Asillustrated similar phenotype has led to the new terminology for
in Fig. 1, four long QT syndrome disease genes, each mutations:(i)LQT1onKvLQT1,(ii)LQT2onHERG,
encoding an ion channel protein, have been identified: (iii) LQT3 on SCN5A, (iv) LQT5 on minK. Although
SCN5A,encodingthecardiacsodiumchannel(chromo- theprevalenceofeachvariantofthelongQTsyndrome
some3)[6,9–11];HERG,encodingtheI potassiumchan- has not been precisely defined, LQT1 is the most fre-
Kr
nelprotein(chromosome7)[6,12];KvLQT1,encodingthe quentlyencounteredform,whereasLQT3andLQT5are
alpha subunit of the I potassium channel protein rare.
Ks
(chromosome 11)[5,13,14] and KCNE1, encoding minK,
an ancillary subunit for the I channel complex (chro- Mutations in long QT syndrome genes
Ks
mosome 21)[15,16]. The gene at the chromosome 4 locus Most of the mutations identified to date in the long
(LQT4)hasnotbeenidentified.Familieslinkedtonone QT syndrome genes are missense mutations. These
EurHeartJ,Vol.20,issue3,February1999
176 Working Group Report
HERG: 7q35–36
KvLQT1: 11p15·5 N
P
P
minK: 21q22·1-22·2
C
N C N C
SCN5A: 3p21–23
N C
Figure1 LQTSgenes Chromosomallocationsofthegenesandpredictedtopologyoftheionchannel
proteins associated with the genetic variants of the LQTS.
mutations are not confined to a single location but are the gene (for example, close to the regions encoding
found at various positions within each gene in diVerent specific structures such as the pore, the voltage sensor,
families. Thus, in most aVected families, the long the S1–S6 region, or the N or C-terminal portions) or
QT syndrome is due to a distinctive, or ‘private’ mu- thetypeofmutation(thenatureoftheaminoacidsub-
tation. This remarkable genetic heterogeneity probably stitution, missense mutation vs deletions or insertions)
contributestothevariabilityintheclinicalpresentation. may also play a role.
A few mutational ‘hot-spots’ (such as specific
positions within a gene mutated in multiple families) Functional consequences of mutations
have been identified in KvLQT1[17] and HERG[18]. Un- ThechannelscarryingI andI aremultimeric;thatis,
Kr Ks
relatedkindredsworldwidewiththesamemutationcan alleles from both parents are thought to contribute to
therefore be studied to test the logical hypothesis that the channel complexes. When mutations in KvLQT1,
they share common clinical or epidemiological features. KCNE1orHERGareexpressedaloneorwithwild-type
Contrary to expectations, initial studies indicate that alleles in oocytes or in other cell lines, they exhibit ‘loss
substantial phenotypic heterogeneity remains even with offunction’,i.e.thetotalcurrentcarriedbythedefective
an identical long QT syndrome gene abnormality. This, channel complexes is reduced. Some of the mutations
in turn, suggests that variable expression of as-yet- not only reduce current but also modify channel
unidentified ‘modifier genes’ contributes to the clinical kinetics. Many HERG and KvLQT1 mutations have
manifestations of the disease. beenidentifiedas‘dominantnegative’becausewhenthe
TheJervellandLange-Nielsen(autosomalreces- mutant protein is co-expressed with the native pro-
sive)variantofthelongQTsyndrome(inwhichaVected tein[13,14,26,27]theresultingdefectincurrentexceeds50%.
subjects have especially long QT intervals) arises in One explanation for this phenomenon is that incorpor-
individuals who inherit abnormal KvLQT1 or minK ation of a single abnormal protein subunit into the
allelesfrombothparents.Theabnormalallelecanbethe tetrameric channel structure is suYcient to alter the
same(usuallyinconsanguineousfamilies)[19,20]ordiVer- overall behaviour of the current.
ent (‘compound heterozygosity’)[16]. Thus, parents of By contrast, mutations in the SCN5A channels
subjects with Jervell and Lange-Nielsen carry long QT causea‘gainoffunction’[10,11].Thesemutationsproduce
syndrome mutations, although most (but not all) are a persistent late sodium current which is not present
aysmptomatic. Recently, a family with apparent auto- physiologically and which is due to defective inactiva-
somalrecessivelongQTsyndromewithoutdeafnesshas tion. In all described mutations, the sodium current is
also been identified[21]. These findings all suggest that increased because of late reopenings of the channels,
‘gene dosage’ determines the phenotype (two abnormal while in the three amino-acid deletion (˜KPQ) long
alleles appear worse than one), and also highlights the lastingburstsofchannelactivityarealsopresent.These
extraordinary variability in the long QT syndrome mutationsalsodiVerinseverity,withthe˜KPQdeletion
phenotype[22–25]. The location of the mutations within being associated with a quantitatively larger increase in
EurHeartJ,Vol.20,issue3,February1999
Cardiac arrhythmias 177
latesodiuminwardcurrent[11].ItisgenerallydiYcultto shouldshortenQTinthemaswell.Theputativeroleof
develop specific therapies for loss of function (for I in cardiac physiology suggests an especially favour-
Ks
example, the K+ channel defects described above). By ableeVectofbeta-blockadeandtheavoidanceofvigor-
contrast, the gain in abnormal function exhibited by ous increase in heart rate (i.e. competitive sports) in
mutantSCN5Ageneproductsraisesthepossibilitythat LQT1 and LQT5. These examples demonstrate that
acurecouldbeaccomplishedbypharmacologicalagents gene-specific therapy may be feasible in the long QT
that inhibit the ‘gained’ function, i.e. block the late syndrome.However,itshouldbeemphasizedthatlong-
I . Indeed, some data suggest that these currents termtrialsarenotyetavailable,andthat,atthepresent
Na
are especially sensitive to block by mexiletine or time, beta-blockers remain the first choice therapy.
lidocaine[10,11].
Drug-induced long QT syndrome
Genotype–phenotype correlations It has long been postulated that drug-induced long QT
The diVerent time and voltage dependence of the ionic syndromemightrepresentagenetically-mediated‘forme
currents involved in the long QT syndrome may help fruste’ of the long QT syndrome[32]. Recent studies
explain some aspects of the variable phenotype and have identified relatively large numbers of individuals
raise the possibility of gene-specific treatment. Indeed, who carry ‘silent’ mutations on long QT syndrome
available data on several hundred genotyped patients genes[22–25]. Thus, these persons, whose long QT syn-
indicate the existence of gene-specific diVerences in the dromemutationsbythemselvesproduceanalterationin
triggers for cardiac events[28]. Exercise-related events repolarizing currents that is insuYcient to prolong the
dominate the clinical picture in I -related long QT QT interval at rest, may be especially sensitive to any
Ks
syndrome(LQT1)[28].I isthepredominantK+ current drugsthataVectsK+ currents.Thecombinationofeven
Ks
inconditionsofhighsympatheticactivity,particularlyat amodestdegreeofI blockade,inducedbyavarietyof
Kr
shortercyclelengths.Thus,reducedI willbepredicted drugsusedformultiplepurposes[33]andthesilentmuta-
Ks
to lead to inadequate action potential shortening with tions could produce a major prolongation in action
adrenergic stress and thereby account for the high potential that triggers the onset of Torsades de pointes.
prevalenceofarrhythmiceventsinthesepatientsduring Indeed, occasional patients with typical drug-induced
exercise. By contrast, most LQT3 patients experience long QT syndrome and underlying mutations on long
events during sleep or at rest; they are also able to QTsyndromegeneshavenowbeenidentified.However,
markedlyshortentheirQTintervalduringexercise[29].In therarityofthisphenomenonmeansthatgenetictesting
these cases, it seems likely that the presence of normal in patients with drug-induced long QT syndrome is not
K+ currents produces normal action potential shorten- yet warranted in the absence of other indications (for
ing during exercise; however, at rest, defective inactiva- example, family history, long baseline QT)[34,35].
tionofI willresultinanincreaseintheplateauinward
Na
Na+-current. This apparently ‘nice’ distinction between
LQT1andLQT3is,however,complicatedbythereality Familial hypertrophic cardiomyopathy
that LQT2 patients also tend to display events both at
rest and during exercise, thus pointing to the persistent Hypertrophic cardiomyopathy[36] is transmitted as an
limitations in current understanding. autosomal dominant disease. Its clinical phenotype is
There is an emerging sense that gene-specific characterized by unexplained and inappropriate clinical
therapy may be feasible for some forms of long QT left and/or right ventricular hypertrophy, which may be
syndrome.Thisrelatedbothtopharmacologicaltherapy severe4to5cm),mildorevenabsent.Characterization
aswellastoadviceregardinglifestyle.Adisorderbased of the distribution of left ventricular hypertrophy is
ondisturbedinactivationkineticsofthesodiumchannel arbitrary, but by convention hypertrophy is considered
(LQT3) seems likely to respond to a sodium channel to be either asymmetric septal hypertrophy, concentric
blocker. Indeed, in LQT3 patients, the QT-interval or predominantly distal ventricular. Any pattern of
seems to shorten more than in LQT1 and in LQT2 hypertrophy, however, may be seen including hyper-
patients in response to mexiletine, but individual trophyconfinedtotheposteriororfreewall.Character-
exceptions do exist[29] and significant shortening of QT istic histological features included myocyte disarray
intervals by sodium channel blockers has been reported surrounding areas of increased loose connective tissue.
in some LQT2 patients[30]. It is also possible that Clinically, there is marked haemodynamic
while mexiletine or similar drugs shorten QT in LQT3, heterogeneityamongpatientswithfamilialhypertrophic
beta-blockade might still be required to suppress cardiomyopathy. Systolic function may be hyper-
arrhythmias. As the amplitude of I increases when dynamic (with or without obstruction), ‘normal’ or
Kr
extracellular potassium concentration is increased, at- impaired (10–15%). Diastolic dysfunction is the usual
tempts have been undertaken to increase K+ levels in physiological abnormality, although the precise abnor-
longQTsyndromepatients.Todate,theQTintervalhas malityofventricularfillingandcomplianceisextremely
beenshowntoshortensignificantlyinLQT2patients[31], variable.
but neither LQT1 nor LQT3 patients have been tested Familial hypertrophic cardiomyopathy-related
withthisapproach.BecauseI functionisnormalinthe arrhythmias occur both at the ventricular and at the
Kr
latter subjects, elevating potassium to increase I atrial level. Importantly, sudden cardiac death in
Kr
EurHeartJ,Vol.20,issue3,February1999
178 Working Group Report
1q3: cTnT ACTIN
19p13·2–19q13·2: cTnl
15q2:alpha TM
11p11·2:MYBPC3
3p: MELC
MYOSIN ?
14q11: beta MHC
12q23–24·3: MRLC
7q3
Figure 2 FHCM genes Chromosomal locations of the genes and locations of the sarcomeric proteins
associated with the genetic variants of FHCM.
familial hypertrophic cardiomyopathy is not necessarily troponin I on chromosome 19[42]. An additional locus
caused by ventricular arrhythmias. Atrial fibrillation in has been identified on chromosome 7 in a large family
thepresenceofanaccessorypathway,bradyarrhythmias with both familial hypertrophic cardiomyopathy and
andischaemiamayallleadtosuddendeath.Inb-myosin cardiac preexcitation (WolV–Parkinson–White)[43].
heavy chain related patients (probably the majority), The prevalence of the diVerent gene abnormali-
hypertrophy itself does not seem to be the main deter- ties in familial hypertrophic cardiomyopathy is being
minantofmalignantventriculararrhythmia.Onecaveat delineated. To date, information on less than 100 geno-
ininterpretingelectrophysiologicalchangesintheseset- typed families suggests that mutations in beta-myosin
tings is that a common secondary response to injury heavy chains and myosin binding protein C are more
(such as pressure overload or coronary occlusion) is common than the others. In addition to this locus
cardiac hypertrophy, which in diseased hearts produces heterogeneity, there is, as in the long QT syndrome,
furtherfunctionalchanges,notablyincalciumhandling. marked allelic heterogeneity for all the recognized dis-
Thus, the extent to which any of the observed electro- easegenes,andtodatemorethan85diVerentmutations
physiological alterations are primary or secondary to have been reported (for reviews see[34,44,45]). The
the response to the disease process requires further majorityofmutationsaremissensemutations,although,
study. for the cardiac myosin binding protein C gene, most of
the mutations lead to an early stop codon, resulting
Familial hypertrophic cardiomyopathy genes in truncated mutant proteins[46]. Functional studies of
As illustrated in Fig. 2, there is considerable genetic mutant myosin indicates that sarcomeric contractile
heterogeneity in familial hypertrophic cardiomyopathy. performance is depressed[47–49]. This in turn suggests
Mutations in seven sarcomeric protein genes have been that myocyte hypertrophy characteristic of familial
identified in families with familial hypertrophic cardio- hypertrophic cardiomyopathy reflects a compensatory
myopathy (Table 1). These are: (1) beta-myosin heavy response. The molecular (or other) determinants of
chainonchromosome14[37];(2)cardiacessentialmyosin myocyte disarray and myocardial fibrosis (interstitial
light chain on chromosome 3[38]; (3) cardiac regulatory and replacement) remain unclear. It may well be that
myosin light chain on chromosome 12[38]; (4) cardiac these latter responses relate to the type of mutation
troponinTonchromosome1[39];(5)alphatropomyosin (e.g. greater with troponin-related disease) and that
on chromosome 15[39]; (6) cardiac myosin binding sudden death and clinical arrhythmia are the clinical
protein-C on chromosome 11[40,41]; and (7) cardiac consequences of extensive disarray and fibrosis.
EurHeartJ,Vol.20,issue3,February1999
Cardiac arrhythmias 179
Genotype–phenotype correlations (14q23-q24 and 14q12- q22)[59,60]. A third locus was
Information on the genotype–phenotype relation in locatedonchromosome1(1q42-q43)[61],andthefourth
familial hypertrophic cardiomyopathy is still prelimi- on[62]chromosome2(2q32.1-q32.2)(Table1).Theauto-
nary,asthepublisheddataongenotypedpatientsrelates somal recessive syndrome variant of arrhythmogenic
toonlyafewhundredindividualsfromcentresthatmay rightventriculardysplasiahasbeenlinkedtoalocuson
reflect diVerent referral biases. It is nevertheless clear chromosome 17 (17q21), within the gene encoding
that the phenotype varies not only with the type of a keratin, a reasonable candidate for the entity[58].
mutation, but also within individuals bearing the same Further advances will facilitate recognition of the
mutation. The 403 codon in beta-myosin is a hot spot non-arrhythmic clinical presentations and the broader
for mutations; the arginine to glutamine mutation is phenotype of arrhythmogenic right ventricular
associated with a poor prognosis, whereas the arginine dysplasia/Naxos disease (Table 1).
to tryptophan mutation appears more benign[50–52].
Current practice suggests that if the ECG and two-
dimensionalechocardiogramarenormalbyage25,then Dilated cardiomyopathy
thepatientcanbesafelyreassuredthatheorshewillnot
develop clinical familial hypertrophic cardiomyopathy. Dilated cardiomyopathy is a genetically heterogeneous
However,myosinbindingproteinCmutationsappearto andclinicallyheterogeneousdisease[63],whichcanaVect
be associated with age-related penetrance during adult newborns, children, adolescents, adults and the elderly.
life[40,41,53]. Further information confirming the impres- The disease may be associated with other organ or
sion that adult onset of disease is an important feature muscle abnormalities or present as a pure disorder.
seen with myosin binding protein C mutations would Malignant life-threatening ventricular arrhythmia as
thus have a significant impact on management and well as atrial arrhythmia with a serious impact on
counselling.ThediseasecausedbytroponinTmutations cardiac function are frequently associated with the dis-
appears associated with mild or absent hypertrophy, a order. As in familial hypertrophic cardiomyopathy,
20 to 25% incidence of non-penetrance and a high sudden death in dilated cardiomyopathy may also be
incidenceofprematuresuddendeath(possiblygreaterin caused not only by ventricular arrhythmias but also by
young men although the numbers are small) which can bradyarrythmias. Whenever spontaneous ventricular
occur even in the absence of significant clinical left arrhythmia have been clinically documented, the induc-
ventricular hypertrophy[54–56]. ibility and reproducibility of the arrhythmia in electro-
physiological studies is usually low, favouring the
possibility of a predominant role for non-reentrant
Arrhythmogenic right ventricular dysplasia mechanisms[64,65]. At least 30% of cases of dilated
cardiomyopathy are inherited (that is, familial dilated
Arrhythmogenicrightventriculardysplasiaisarecently cardiomyopathy, familial dilated cardiomyopathy) with
recognised familial cardiomyopathy[57]. The disease is a significant percentage of the remaining cases being
characterized by fibro-fatty replacement of the right acquired (that is, myocarditis, ischaemic heart disease
ventricularmyocardiumandlife-threateningventricular etc.). Inherited dilated cardiomyopathy may have
tachyarrhythmias originating from the right ventricle. autosomal dominant, autosomal recessive, X-linked or
Occasionally, the left ventricular myocardium is in- mitochondrial transmission (Table 1).
volved as well. Disease progression is associated with
left ventricular involvement (50%), atrial dilatation and Molecular basis of dilated cardiomyopathy
arrhythmias with embolic risk. Malignant ventricular To date, genes for X-linked and autosomal dominant
arrhythmiasareacommonmanifestationofthedisease. dilatedcardiomyopathyhavebeenmapped,demonstrat-
Inducibility and reproducibility in the clinical electro- inggeneticheterogeneity[66].ThegenesfortwoX-linked
physiological laboratory is high, suggesting that re- cardiomyopathies have been identified: the dystrophin
entrant mechanisms related to the distinctive structural genewhichisalsoresponsibleforDuchenneandBecker
changes are likely. The disease appears especially com- muscular dystrophy[67,68], and G4.5 in Barth Syndrome
moninNorthEasternItaly(prevalence1:1000)withan (X-linked cardioskeletal myopathy with neutropenia,
autosomal dominant inheritance (30%). An autosomal abnormal mitochondria and 3-methylglutaconic
recessive variant of arrhythmogenic right ventricular aciduria)[69].Multiplemutationsinbothgeneshavealso
dysplasia which is associated with a distinctive extra- been reported[68–72].
cardiac phenotype (woolly hair and palmoplantar kera- Dystrophin is a large cytoskeletal protein which
toderma)hasbeenreportedfromtheislandofNaxosin is found on the inner face of the sarcolemma and
Greece[58]. attaches at its N-terminal domain to F-actin in the
matrix and to the dystrophin-associated glycoprotein
Molecular basis of arrhythmogenic right ventricular complex (an oligomeric transmembrane protein) at its
dysplasia C-terminal domain. The protein encoded by the G4.5
To date, four loci for autosomal dominant arrhyth- gene is called ‘tafazzin’ but its function is unknown.
mogenicrightventriculardysplasiahavebeenidentified, Genes for autosomal dominant dilated cardio-
two of which are in close proximity on chromosome 14 myopathy have been mapped to six diVerent loci thus
EurHeartJ,Vol.20,issue3,February1999
180 Working Group Report
far.‘Pure’dilatedcardiomyopathyhasbeenlocalizedto channels[88].Laterduringthe‘remodelling’thatappears
1q32, 2p31, 9q13, and 10q21-q23[73–76], while dilated to accompany chronic atrial fibrillation, changes in
cardiomyopathy with conduction defects has been expression and/or distribution of connexin proteins
mapped to 1p1-1q1[77] and 3p22-3p25[78]. Recently, and/or other ion channel proteins, as well as changes in
mutations in cardiac actin[79] located on chromosome cellular ultrastructure, may play a role.
15q14 have been identified, therefore so far actin is the Inherited atrial fibrillation is considered uncom-
only known gene for autosomal dominant dilated car- mon and has been reported with autosomal dominant
diomyopathy.Basedonthisfinding,Olsenetal.[79]have transmission. Recently, familial atrial fibrillation has
now proposed that dilated cardiomyopathy results as a been mapped to 10q22-q24 (a region of approximately
consequenceofdefectivetransmissionofforceincardiac 11cM)inthreefamilies[89].Expansionofthepreviously
myocytes leading to heart failure. identified kindreds has allowed further refining of the
map position and limitation of the gene critical region.
A fascinating issue concerning atrial fibrillation
is its association with other disorders, such as dilated
Idiopathic ventricular fibrillation and the
cardiomyopathy, familial hypertrophic cardiomyopathy
Brugada syndrome
andLQT4andthepossibilitythatamutationinagene
responsible for one of these associated disorders could
Anotherinterestinggroupofpatientswhichhasbecome
causefamilialatrialfibrillation.Forinstance,isitsimply
a target for genetic studies is represented by individuals
circumstantial that a familial dilated cardiomyopathy
with so-called idiopathic ventricular fibrillation (that is,
locus[76] and the mapped atrial fibrillation locus are
patients with a normal heart that experience cardiac
withinthesamerelativelysmallregionof10q21-q24?Is
arrest with documented ventricular fibrillation)[80]. A there something diVerent about the clinical course, and
subgroup of these patients experience sudden death
thusthecausativegeneresponsibleforLQT4[8]inwhich
which may occur in families, apparently have no struc-
prolonged QTc appears to be associated with a high
turalheartdiseaseandhaverightprecordialSTsegment
incidence of atrial fibrillation and slower heart rates
elevation, sometimes with right bundle branch block
thantypicallyseeninthelongQTsyndrome?Couldthis
(Brugada syndrome[81]). These electrocardiographic beadiVerenttypeofgene(i.e.notanionchannel)ora
characteristics may depend on exaggerated transmural
new channel disorder?
diVerences in action potential configuration, especially
in the right ventricular outflow tract. This could arise
from dysfunction of a number of ion currents, such as
Progressive familial heart block
I , L-type Ca2+ current (I ) and I .
to Ca(L) Na
AtleastonevariantoftheBrugadasyndromeis
Two forms of progressive familial heart block[90] which
causedbydefectsinthesodiumchannelgene(SCN5A), diVer in their ECG characteristics, have been reported.
i.e.thesamegeneimplicatedinLQT3[82].IntheBrugada
Thefirst,progressivefamilialheartblock-I,isdefinedon
syndrome, the mutations identified apparently lead to a
the ECG by evidence of bundle branch disease such as
loss of function while in LQT3 all cause a gain of
right bundle branch block, left anterior hemiblock, left
function.Thus,thelongQTandtheBrugadasyndromes
posteriorhemiblock,orcompleteheartblockwithbroad
appear to be separate allelic disorders.
QRS complexes. Progression of disease occurs with
Evidence that not all patients with the Brugada
changes in the ECG, from a normal ECG to right
syndrome have defects on the cardiac sodium channel
bundle branch block to complete heart block. Typical
(Priori et al., 1998 personal communication) suggest
manifestationsofthediseasearesyncope,suddendeath,
that, in analogy with the other inherited cardiac dis-
orStokes–Adamsattacks.Thesecondformofprogress-
eases, genetic heterogeneity is also present in Brugada
ive familial heart block, known as progressive familial
syndrome.
heart block-II, presents with complete heart block and
narrow QRS complexes and is believed to occur due
to atrioventricular nodal disease with atrioventricular
Atrial fibrillation block and an idionodal escape rhythm. Typically
these patients present with sinus bradycardia and left
Perhaps the commonest arrhythmia requiring interven- posterior hemiblock, and develop syncope and Stokes–
tionisatrialfibrillation.Dataarenowemergingfroma Adams attacks.
numberoflaboratoriesonthepotentialmolecularbasis Genetically, progressive familial heart block-I is
of electrophysiological changes observed in atria that better studied than progressive familial heart block-II
have been fibrillating for hours to days and those that and appears to be inherited in an autosomal dominant
have been fibrillating for weeks to months[83–87]. They fashion. Brink et al.[91] studied three South African
all share a marked shortening of refractoriness, prob- families with progressive familial heart block-I, includ-
ably reflecting decreased action potential duration ing one nine generation kindred, for linkage analysis.
early during atrial fibrillation. Available data suggest Using 86 family members (39 aVected), linkage was
that a major mechanism is decreased inward current identified on chromosome 19 at 19q13.2-q13.3 and the
through L-type calcium channels and possibly sodium gene was localized to within 10cM of the kallikrein
EurHeartJ,Vol.20,issue3,February1999
Cardiac arrhythmias 181
locus. Confirmation of this localization was subse- necessarythatallthegenesandallthemutationswithin
quentlyreportedbyBouvagnetetal.inalargeLebanese these genes causing a given disease be identified. This is
family[92]. Other candidate genes within the mapped not yet even close to reality for any of the inherited
region include Apolipoprotein C2 (ApoC2), creatine arrhythmogenic diseases discussed here. As a conse-
kinase MM isoform (CK-MM), myotonic dystrophy, quence,physiciansstillgenerallyhavetorelyonclinical
troponin T and the histidine rich Ca2+ binding protein criteria to establish these diagnoses.
(a luminal sarcoplasmic reticulum protein). Myotonic For some diseases, not even the specific aVected
dystrophy,CK-MM,andApoC2havebeenexcludedas gene(s) are known. In these cases (for example, familial
the causative genes. atrial fibrillation, progressive familial atrioventricular
block),theavailablegeneticinformationisderivedfrom
linkage studies and provides only data on which chro-
Familial WolV–Parkinson–White syndrome mosomalregionthedisease-geneislocated.Ifthisregion
islarge,itmaytakeyearsbeforethegeneresponsiblefor
Familial WolV–Parkinson–White has been rarely re- the disease is located. Thus, at this stage of knowledge,
portedbutaninheritedformofthesyndromeassociated molecular screening for these entities is limited to re-
with familial hypertrophic cardiomyopathy has been searchactivities;itisnotpossibletoconsidergenotype–
described and its locus mapped to chromosome 7q3[43]. phenotype correlations and, most importantly, the
It is unknown whether a single defect is responsible for nature of the defect underlying the disease remains
bothaspectsofthesyndromeoriftwogenesarelocated undefined. Linkage studies can, nonetheless, provide
in close proximity (i.e. contiguous gene syndrome) and important information on whether only one gene is
thus frequently co-segregate. In the latter case, familial associated with the disease or if genetic heterogeneity
WolV–Parkinson–White could be caused by a single exists (i.e. several genes accounting for a disease).
gene-defect on chromosome 7. However, other associ- When a gene responsible for a disease is ident-
ations of familial hypertrophic cardiomyopathy and ified, it then becomes possible to search for specific
WolV–Parkinson–White have also been identified. For mutations. The organization and the sequence of the
exampleKimuraetal.[42]foundmutationsinthecardiac disease genes are often not entirely known, and thus
troponin I gene (on chromosome 19) in patients with mutations are usually searched for only (at least
familial hypertrophic cardiomyopathy and WolV– initially) in portions of the gene. As a consequence, a
Parkinson–White. Furthermore, some children with positive finding (i.e. the identification of a mutation) is
mitochondrial abnormalities and metabolic disease diagnostic, while a negative finding in a linked gene
(Pompe disease) associated with familial hypertrophic suggeststhatmutationsmaybepresentinanunexplored
cardiomyopathy also have been noted to have WolV– regionofthegene(orthatthelinkageisincorrect).The
Parkinson–White. Therefore, it currently appears that diagnostic power of molecular screening is further
WolV–Parkinson–White may have multiple diVerent limited(forallarrhythmogenicdisordersdiscussedhere)
genetic aetiologies. bythepresenceofgeneticheterogeneityandthelackof
identification of all of the genes responsible for the
disease.
Molecular diagnosis of inherited
arrhythmogenic disorders Implications of molecular diagnosis on
patient management
Role of DNA screening in diagnosis of
inherited arrhythmogenic diseases When the genetic bases of familial hypertrophic cardio-
myopathy and the long QT syndrome were elucidated,
Thepossibilityofageneticdiagnosismeansthatgenetic thehopeofmolecularbiologistsandcliniciansalikewas
testing in routine clinical practice is also a possibility. that it would become possible to reach, in a relatively
Applications could include pre-clinical diagnosis and shorttime,someimportantgoalstoestablishgenotype–
identification of patients who might benefit from pro- phenotype correlations. In this respect, valuable infor-
phylactictreatmentforsuddendeath.Forthisapproach mation would be the ability to categorize mutations as
tobecomeareality,however,severalconditionsmustbe ‘mild’ vs ‘severe’, in order to guide the therapeutic
met: the development of routine clinical DNA diag- approachonthebasisofthepredictedrisk.Forthetime
nostic testing facilities; a suYciently large database to being, this goal has not been achieved, and we are still
determine risk in relation to genotype, as well as the far from being able to predict adverse or favourable
recognized heterogeneity in phenotype; an estimate of prognoses based on the genetic defect.
theeYcacyofavailabletreatments;andaconsideration A major goal in the long QT syndrome and
of the cost implications. familial hypertrophic cardiomyopathy remains, to have
Molecular diagnosis has the potential to define suYcient genotyped patients to understand the diag-
with 100% sensitivity and 100% specificity the genetic nostic, functional and prognostic implications of the
status of any member of an aVected family. However, diVerent mutations. A problem in genetic testing
for this potential to become fully expressed, it is in the long QT syndrome and familial hypertrophic
EurHeartJ,Vol.20,issue3,February1999
182 Working Group Report
cardiomyopathy is that the disease-associated gene and weightand/ormildmorphologicalfeaturesintheyoung
specificmutationsarestillbeingidentified.Thisresearch should lead to testing for mutations in the cardiac
information is not yet widely implemented in commer- troponinTgene.Prematuresuddendeathinassociation
cial laboratories, and the resource demands for such an with obvious morphological features in the young have
eVort on a routine (or ‘service’) basis are generally been associated with the Arg403Glu and Arg453Cys
beyond those available to the research laboratories mutation in the beta-myosin heavy chain gene, and
engaged in the problem. these mutations could be tested in this clinical context.
Identification in the proband of troponin or myosin
Genetic testing for the long QT syndrome heavy chain mutations, which are associated with poor
When should genetic testing be considered in dealing prognosis, would permit an early or even a pre-clinical
withlongQTsyndromepatients[93].Thecardiologistwill diagnosis in family members with the potential for
confront three clinical scenarios: lifestyle modifications (avoidance of competitive exer-
The first situation is the patient who has a cise) and prophylactic treatment (amiodarone or im-
definite diagnosis based on established clinical diag- plantable cardioverter defibrillator) to prevent sudden
nostic criteria. Here, genetic testing is not absolutely death.
necessary because the cardiologist has most elements
necessarytomakeadecisionaboutinitiationoftherapy.
Genetic testing for autosomal dominant dilated
However, genetic testing could be useful because, de-
cardiomyopathy, arrhythmogenic right ventricular
pending on the gene (and ultimately even the specific
dysplasia, familial atrial fibrillation, progressive familial
mutation) identified as responsible for the disease,
atrioventricular block
modifications in management[29,94] may be suggested.
Until specific genes are discovered and characterized,
Examplesdiscussedaboveincludetheadditionofmexi-
moleculardiagnosisshouldbeconsideredaresearchtool
letine in LQT3 or lifestyle modifications such as limi-
onlyinlargesizefamiliesinwhichlinkageanalysismay
tation of strenuous or competitive exercise in LQT1. It
be performed.
should be pointed out, however, that in symptomatic
patients with an established diagnosis of long QT syn-
Genetic testing for dilated cardiomyopathy
drome, implementation of therapy with beta-blockers
In both X-linked dilated cardiomyopathy and Barth
shouldnotbedelayedwhilewaitingforgeneticscreening
syndrome, definite diagnosis at the molecular level may
results.
beusefulclinicallysincebotharerapidlyprogressiveand
A second scenario occurs when the diagnosis of
severe disorders. In the case of X-linked dilated cardio-
longQTsyndromeisonlysuspectedorthepatienthasa
myopathy where anticongestive and antiarrhythmic
borderline diagnosis based on clinical criteria. Under
management initially, and cardiac transplantation
thesecircumstances,genetictestingcouldbeveryuseful
shortly thereafter, is life-saving, determination of a
in establishing the diagnosis because identification of a
mutation could help diagnose pre-symptomatic male
mutated long QT syndrome gene would convert a sus-
gene-carriers. In Barth syndrome, therapeutic options
pecteddiagnosistoacertainoneandwouldremovethe
are less clear-cut but a definitive diagnosis in family
cardiologist’s hesitation in making therapeutic choices.
members and potentially in fetuses could be similarly
However,failuretoidentifyamutationdoesnotruleout
useful.
the diagnosis (since only a minority of mutations have
The recent identification of mutations in the
beenidentifiedtodate).Althoughgenetictestinginthis
actin gene opens the opportunity to perform family
situationisnotyetwidelyavailable,techniquestoauto-
screeningformutations;however,untiltheprevalenceof
mate screening for the hundreds of possible known
actin-related autosomal dominant dilated cardiomy-
mutationsarenowbeingdevelopedandwillprobablybe
opathy is defined, the cost/benefit ratio of actin gene
available in the next 5–10 years. A third scenario is an
screening cannot be defined.
apparently asymptomatic relative of a patient with the
long QT syndrome. Here, genetic testing can be es-
pecially useful if the disease-causing mutation has pre- Genetic testing for Brugada syndrome
viously been identified in the proband. Otherwise the The identification of mutations in the cardiac sodium
issues are the same as those for evaluating the ‘border- channel in families with Brugada syndrome opens the
line’ long QT syndrome diagnosis. possibility to screen patients with the disease. The im-
portance of identifying the defect obviously consists of
Genetic testing for hypertrophic cardiomyopathy the ability to identify the carriers before they become
Similar considerations apply in familial hypertrophic symptomatic.Thisisparticularlyimportantforadisease
cardiomyopathy. Comprehensive screening of the in which the first manifestation is often cardiac arrest.
disease-causing genes would, from the clinical perspec- However, since not all patients with Brugada syndrome
tive,bebothinappropriateandimpracticalatthistime. have mutations in the sodium channel (Priori et al.,
SpecificclinicalsituationsexistwhereDNAdiagnosisis personal communication), the cost/benefit ratio of mu-
likelytohaveanimportantimpactonmanagement.For tation screening in the sodium channel gene cannot be
example, sudden death/resuscitated ventricular fibril- defineduntiltheprevalenceofthegeneticvariantofthe
lation in association with normal or near normal heart form associated to sodium channel defects is defined.
EurHeartJ,Vol.20,issue3,February1999
Cardiac arrhythmias 183
Ethical aspects of molecular screening by an inborn alteration and, for the most part, are
characterized by a single alteration in one or more
Importantethicalaspectsareinvolvedwhenconsidering disease genes. This has allowed the use of ‘paradigms’;
DNA screening in families aVected by a congenital namely,diseasessuchasthelongQTsyndromeinwhich
disease. Discussion with families about the information it has been possible to trace specific mutations on ion
that could be provided by genetic testing is a most channelgenestotheirelectrophysiologicalconsequences
important first step. The experience of the team (which inthepatient.Unfortunatelyforthepractisingcardiolo-
should include an appropriately trained genetic coun- gist, these ‘simple’ diseases constitute only a small part
sellor) in caring for patients with similar disorders is of the clinical conditions associated with cardiac ar-
an important component for patient acceptance. One rhythmias.ThemajorityofcasesaVectpatientsinwhom
specific objective of counselling in arrhythmogenic dis- the arrhythmogenic substrate is complex. Indeed, the
orders is to help the patient decide whether he/she expression of the molecular systems responsible for
should undergo genetic screening at all. These are the normal and abnormal electrical activity vary signifi-
considerations involved when an individual is deciding cantly, depending on a variety of factors including age,
whether or not to be tested: regional factors (type of cells, myocardial perfusion),
(a) The patient should be given information on the (1) and underlying chronic diseases such as cardiac
samplerequired;(2)useofthesample;(3)resultsof hypertrophy, myocardial infarction, and heart failure.
the test performed; (4) implications of these results To approach this complex system of interacting
for management of the patient and their family; (5) molecular functions requires a somewhat diVerent ap-
who will have access to the results. proach from that required to consider monogenic dis-
(b) The patient should use this information to decide ease. Accordingly, in this section we discuss broader
whether to give or withhold consent. themes that are essential to understand the integration
(c) There should be no coercion by anyone (healthcare of gene expression, of ion channel function, and cell
team, family members, insurance companies). coupling in multicellular networks. This must be con-
(d) The consent form that patients sign should include sidered a first step toward the comprehension of more
statements that (1) all blood samples are coded to frequentandmorecomplexarrhythmogenicconditions.
prevent identification; and (2) the results of screen-
ing will be communicated only to the patient and
thatnodisclosurewillbemadetothirdparties(not Diversity of gene expression in the heart
evenfamilymembers)withouttheirwrittenconsent.
Asymptomatic patients should have the option of Understanding cell–cell variability in the cardiac action
providing samples (e.g. for family study) without potentialshapeandthemechanismsunderlyingimpulse
being informed of the results. propagation is the key to understanding normal and
abnormalcardiacelectrophysiology.Muchofthisvaria-
bility can be attributed to variability in the character-
Reimbursement and costs issues
istics of individual ion currents whose integrated
behaviour determines the shape and duration of action
Currently, DNA screening for arrythmogenic disorders
potentials in individual cardiac cells, as well as to
is not considered a routine test, and therefore costs are
variability in cell–cell communications. Ion currents are
notusuallycoveredbyinsurance.Linkagetoisolatethe
now recognized to flow through specific pore-forming
disease gene can be performed in large families. When
membraneproteins,termedionchannels.Thefirstgene
small pedigrees or single patients, in whom linkage
encodinganionchannelwasclonedin1984[95],andthe
cannot be applied are studied, the only approach for
succeeding decade and a half has seen the cloning of
DNA screening is the systematic search for known
genesencodingmostionchannelsexpressedintheheart,
mutations in any disease-linked gene. As discussed
and in many other tissues[96–98]. Many of the proteins
above, clinical evaluation may help in selecting gene(s)
these genes encode share common structures and can
to be screened first. Depending on the size of the gene
be viewed as members of the same superfamily. For
and on the number of genes to be screened, costs may
example, Fig. 3 shows the tremendous diversity of
be substantial (>$1000 US per gene screened in each
mammaliangenesthatmakeupthefamilyofpotassium
family). Currently, costs are covered almost exclusively
channel genes. Since potassium channels are made up
by research funding of the laboratories involved in the
of four ion channel alpha subunit proteins, which are
field. The development of automated screens and iden-
not necessarily identical, the potential for diversity in
tificationofmoremutationsmaychangethisinthenear
potassium currents is even greater than shown. This is
future.
further compounded by the identification of ancillary
subunits (the products of diVerent genes) that can as-
Molecular basis of cardiac
semble with potassium channel tetramers to modulate
electrophysiology and arrhythmias their function[99–101]. Figure 4 illustrates the major ion
currents in the heart and the genes whose protein
In the first two parts we have discussed monogenic productsarethoughttoformtheirstructuralbasis.The
arrhythmicdisorders.Thesearedeterminedorfavoured dramatic increase in molecular genetic information
EurHeartJ,Vol.20,issue3,February1999