Locus Heterogeneity of Autosomal Dominant Long QT Syndrome
Mark Curran, **5Donald Atkinson, * * Katherine Timothy,11 G. Michael Vincent,*11
Arthur J. Moss, * * Mark Leppert,:" and Mark Keating*§
*Division ofCardiology, *Department ofHuman Genetics, 1Eccles Program in Human MolecularBiologyand Genetics,
'Howard Hughes Medical Institute, University ofUtah Health Sciences Center, Salt Lake City, Utah 84112;
Department ofMedicine, LDS Hospital, Salt Lake City, Utah 84143; and **Department ofMedicine,
University ofRochester Medical Center, Rochester, New York 14642
Autosomal dominant longQT syndrome (LQT) is an inherited
disorder that causes syncope and sudden death from cardiac
arrhythmias. In genetic linkage studies ofseven unrelated fami-
lies we mapped a gene forLQT to the short arm ofchromosome
11 (lp15.5), near the Harvey ras-1 gene (H ras-1). To deter-
mine if the same locus was responsible for LQT in additional
families, we performed linkage studies with DNA markers
from this region (H ras-1 and MUC2). Pairwise linkage analy-
ses resulted in logarithm ofodds scores of -2.64 and -5.54 for
kindreds 1977 and 1756, respectively. To exclude the possibil-
ity that rare recombination events might account for these re-
sults, we performed multipoint linkage analyses using addi-
tional markers from chromosome 1'p'5.5 (tyrosine hydroxy-
lase and D11S860). Multipoint analyses excluded
centiMorgans ofchromosome 11pl5.5 in K1756 and- 13 cen-
tiMorgans in K1977. These data demonstrate that the LQT
gene in these kindreds is not linked to H ras-1 and suggest that
mutations in at least two genes can cause LQT. While the iden-
tification oflocus heterogeneity ofLQT will complicate genetic
diagnosis, characterization ofadditional LQT loci will enhance
our understanding of this disorder. (J. Clin. Invest. 1993.
92:799-803.) Key words: genetic linkage-cardiac arrhythmias
* sudden death * Romano-Ward syndrome * QT prolongation
In the long QT syndromes (LQT),' individuals suffer from
syncope and sudden death due to cardiac arrhythmias, specifi-
cally torsade de pointes and ventricular fibrillation. Many of
these individuals also have prolongation oftheQT interval and
other repolarization abnormalities on electrocardiograms.
LQT can be broadly classified on clinical grounds as acquired
Address correspondence to Mark Keating, Department ofHuman Ge-
netics, Building 533, Room 2100, University ofUtah, Salt Lake City,
Receivedfor publication 19 January 1993 and in revisedform 15
1. Abbreviations used in this paper: H ras-1, Harvey ras-l; LOD, loga-
rithm ofodds; LQT, long QT; TH, tyrosine hydroxylase.
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
Volume 92, August 1993, 799-803
or inherited. The most common acquired form ofLQT is drug-
induced(1), butmetabolic, neurologic, andcardiac abnormali-
ties can also cause this disorder.
LQT can be caused by the inheritance ofa single gene (2).
Two inherited forms ofLQT are clearly defined on phenotypic
and genetic grounds. One is inherited as an autosomal recessive
trait and is associated with congenital neural deafness (3). This
form of LQT, often referred to as the Jervell Lange-Nielsen
syndrome, is rare. The second, more common, heritable form
ofLQT is an autosomal dominant trait and is often called the
Romano-Ward syndrome (4, 5). Individuals with autosomal
dominant LQT have normal hearing and no other obvious
Presymptomatic diagnosis ofinheritedLQT hasbeen based
on family history and electrocardiographic findings. As the QT
interval is affected by age, sex, autonomic tone, and heart rate,
and varies widely among both affected and unaffected individ-
uals, the diagnosis ofLQT in an asymptomatic individual may
be difficult (6). This is a matter ofconsiderable importance as
treatment for LQT exists.
To improve our understanding ofthe mechanisms underly-
ing LQT and to facilitate presymptomatic diagnosis ofthis dis-
order, we have begun to study families with autosomal domi-
nant LQT. In 1991, we reported tight linkage between the LQT
phenotype and the Harvey ras- 1 (H ras-1) gene in several fami-
lies ofnorthern European descent (2,7). These data confirmed
the autosomal dominant inheritance ofan LQT gene in seven
families and mapped that gene to the short arm of chromo-
some 11 (1 lp15.5). The discovery of linked genetic markers
also made genetic testing for LQT feasible for some families.
To determine ifH ras- 1 was useful for genetic testing in addi-
tional patients, we examined other families with autosomal
dominant LQT. In this report, we characterized two new LQT
families, one of Polish descent and another of Italian descent,
and showed that the gene for LQT in these families was not
linked to H ras- 1.
Phenotypic determination. Kindreds 1977 and 1756 were not previ-
ously described. Informed consent was obtained from all study partici-
pants or their guardians, in accordance with standards established by
local institutional review boards. To determine iffamily members and
spouses had evidence ofLQT, we obtained historical data and electro-
cardiograms from each individual before starting drug treatment. One
individual (II-1, K1756) was evaluated while being treated with a beta
blocker for hypertension. We evaluated the presence of syncope, the
number of syncopal episodes, the age at onset of symptoms, and the
Locus Heterogeneity ofLong QTSyndrome
DI 1 S860
Figure 1. LQT pedigrees for K1756
(A) and K1977 (B) showing haplo-
types for H ras-i, MUC2, TH, and
Dl 1S860. Affected individuals are
represented by filled circles (females)
or squares (males). Unaffected indi-
viduals are shown as empty circles or
squares and individuals with an un-
certain phenotype are stippled. DNA
samples for individuals III-4, III-6,
and III-9 were unavailable for geno-
typic analysis. The pedigrees have
been altered to protect confidentiality.
Informed consent was obtained for
each family member before inclusion
in the study. Haplotype recombinants
described in the text occur in individ-
uals III-3 and III-l 1.
Curran et al.
Table I. Clinical Symptoms ofAffected Individuals in KJ 756 and
A II 1
A II 5
A III 5
A III 7
A III 8
A III 10
A III 11
B II 2
B II 4
B III 2
B III 3
B III 5
B III 7
B III 9
QTC, corrected QT interval, defined in references 2, 5, 6, and 7 cited
occurrence ofsudden death. Symptoms such as palpitations andlight-
headedness were not included in the study.
To determine the phenotype of each family member and avoid
misclassifications, we took a conservative approach to phenotypic as-
signment. Each individual was assigned an impression score based on
the presence and extent of symptoms and QTc prolongation. Impres-
sion scores were used to classify family members as affected, unaf-
fected, or uncertain. All phenotypic data were interpreted without
knowledge of genotype. Phenotypic characterization was completed
using the same criteria as in our previous linkage studies (2, 7, 8).
Typing ofRFLP markers. Two DNA markers, known to define
polymorphisms on chromosome 1 lp15.5 were used: pTBB-2 at the H
ras- I locus (9) and pSMUC41 at the MUC2 locus (10). Genotyping of
RFLP markers was performed as previously described (7).
Typing of PCR-based polymorphic markers. Typing of PCR
markers was performed as described by Weber and May (11) with the
following modifications. PCR was carried out on 200 ng ofDNA in a
final vol of 25;d using a DNA thermocycler (Perkin-Elmer Cetus In-
struments, Norwalk, CT). Amplification conditions were 94°C/10
min followed by 30 cycles of60°C/60 s, 72°C/60 s, and 94°C/60 s.25
Alofloading buffer was added to each reaction, samples denatured for
10 min at 94°C and held on ice. A portion (2.5gl)ofeach sample was
separated by electrophoresis on 6% denaturing polyacrylamide gels
which were dried and exposed to x-ray film overnight at -70°C. The
following simple sequence repeat markers were used in this study:
Dl lS860 (12) and tyrosine hydroxylase (TH) (13).
Linkage analysis. To avoid bias all polymorphisms were scored
without knowledge ofan individual's phenotypic status. The LINKAGE
software package was used to perform both pairwise and multipoint
linkage analysis. Penetrance was estimated at 0.90, based on previous
estimates (2). We assumed a disease gene frequency of0.001 and that
female and male recombination frequencies were equal. Marker order
and distance between markers was determined from published genetic
maps and our unpublished data (14).
Table II. Pairwise Lod Scoresfor KJ 756 andK1977
LOD scores (pairwise) between LQT and polymorphic markers used
in this study. LOD scores were calculated assuming a penetrance
value of0.90 and the frequency ofLQT was 0.001 as in our previous
studies. When penetrance was varied from 1.0 to 0.6 LOD scores
ranged from -9.42 to -5.08 (0 = 0.001) at MUC2 in K1756 and
from -4.68 to -2.01 (0 = 0.001) at H ras-l in K1977. Male and fe-
male recombination rates were assumed to be equal. Allele frequen-
cies for D 1 S860 were obtained from published reports (12). Allele
frequencies for pTBB2 were calculated from 196 independent chro-
mosomes, TH allele frequencies were calculated from 200 chromo-
somes and pSMUC41 from 78 chromosomes.
Phenotypic analyses. We studied two multigenerational fami-
lies with autosomal dominant LQT (Fig. 1). Kindred 1756 was
ofItalian descent and included 16 family members. Three indi-
viduals had a history ofsyncope; in all cases, syncope occurred
immediately after awakening (Table I). One member of this
family, individual III-7, died suddenly upon awakening at age
15. This individual also had a history of recurrent syncope
upon awakening. Electrocardiogram data were not available
for this patient. Autopsy results were unremarkable.
Kindred 1977 was ofPolish descent and included 15 family
members. Two members ofthis family had syncopal episodes
and two died suddenly (Table I). Members of this kindred
experienced syncope while at rest and both deaths occurred
during sleep, one at age 20 (individual 111-5) and the second at
age 31 (individual III-9).
All living members of kindreds 1756 and 1977, with the
exception of individual II-1 in K1977, were evaluated for
symptomsand signsofLQT. Weexamined atotal of 15 individ-
uals from K1756 and thirteen individuals from K1977 (Table
I). As a result, six living members from K1756 were classified
as affected, five as unaffected and three as uncertain (Fig. 1). In
K1977, six individuals were classified as affected and three
members as unaffected (Fig. 1).
Locus Heterogeneity ofLong QTSyndrome
Map Position (cM)
0Multipoint LOD Score
Figure 2. M
order and di
the H ras- I
linked to t
ness in thi
H ras-1, tl
score of -
ble for L(
possibility, we performed linkage analyses with polymorphic
markers from different regions of 1 lpl 5.5. Pairwise analysis in
kindred 1756 produced LOD scores of -3.37 (6
TH and -2.83 at (6 = 0.001 ) forDl 1S860 (Table II). Pairwise
LOD scores for kindred 1977 were 0.35 (6 =0.001 ) forTH and
-0.44 (6 = 0.001 ) for Dl S860. Haplotype analyses showed
two marker-marker recombination events, both occurring in
K1756 between the TH and Dl S860 markers. As shown in
II-2 and affected child III-3. The second event occurred be-
tween individual 11-2 and affected child III- 1 1. As both ofthese
recombination events involve transmission ofalleles from the
unaffected, married-in parent, there is no significant impact on
thelinkage analysis. Inspection ofFig. 1 demonstrated the pres-
ence ofseveral obligate recombinants within each family. Mul-
tipoint analyses excluded, at the -2 LOD level, greater than 6
centiMorgans on either side ofH ras-I for both kindreds (Fig.
2, A and B). As the one LOD confidence interval for previous
LQT families which are linked to H ras-
(7), these data indicate that the gene that causes LQT in
kindreds 1756 and 1977 is not linked to H ras-l.
= 0.001) for
1 the first event occurred between the unaffected parent
We conclude that the gene that causes LQT in the two families
described here is not linked to theH ras- 1 gene on chromosome
1 Ip15.5. Since our initial discovery oflinkage in 1991, several
groups have suggested the existence oflocus heterogeneity for
autosomal dominant LQT. In the
ported preliminary data suggesting that the LQT phenotype in
alarge familywith autosomal dominantLQTwas not linked to
Hrash (15). In the sprng of 1992, Benhorn and his col-
leagues described a large Israeli family with an absence oflink-
age between LQT and H ras-( 16). Jeffery et al. reported that
two siblings with autosomal recessive LQT and congenital
neural deafness (Jervell Lange-Nielsen syndrome) had distinct
H ras- 1 genotypes ( 17). Although this kindred was much too
small for statistical analyses, these datasuggestthat autosomal
dominant and autosomal recessive LQT are mechanistically
distinct. In this studywe characterized two families with auto-
somal dominant LQT using consistent and accurate pheno-
typic criteria (2, 7, 8)andmultiplemarkers from chromosome
1 lp15.5. Our study demonstrates statistical evidence of locus
heterogeneityfor autosomal dominantLQT, suggestingthat at
least two different mechanisms for this disorder exist.
It is notyetclear whatpercentageof familialLQTwill be
caused by mutations in agene on chromosome
have examined 13 families. Ofthose, LQTisclearlylinked to
Harvey ras-I (or other chromosome
clearly unlinked in 2. Theremainingfamilies were too small
for significant statistical evaluation. Further work will be re-
quired on additional families before the relative importance of
the 1 lplS.5 genewill be apparent.
As in previous studies, we used conservative criteria for
phenotypic classification of LQT family members (2, 7, 8).
Individuals were assigned an impression score based on the
presence and extent of symptoms and electrocardiographic
signsofLQT. Recentlywe demonstrated that many LQT gene
carriers areasymptomaticand can have a widerangeofresting
QTcvalues;the lowestQTcfor aLQT genecarrier in that study
was 0.41 s whereas the highest QTc for a noncarrier was 0.47 s
fall of 1991, Towbin
Map Position (cM)
[ultipoint linkage analysis of LQT and chromosomeH
irkers for K1756 (A) and K1977 (B). The location map
3osite LOD scores for LQT at different positions in a fixed
). H ras-I is arbitrarily placed at 0 centiMorgans. Marker
istances were based on published maps ofchromosome
geanalyses. To determine ifthe gene causingLQTin
lies was located on chromosome
kage analyses using highly polymorphic markers at
I locus. We obtained a significant negative logarithm
OD) score (LOD score of-2 or lower) of-2.64 for
977 at a recombination fraction (6) of 0.001 (Table
data suggested that the LQT phenotype was not
the H ras-l locus in this kindred. Markers at H ras-l
Lively uninformative for linkage in kindred 1756
re of -0.30 at
s kindred, we used a second DNA marker very near
he MUC2 locus (SMUC 41 ) and obtained a LOD
linked(10), these data suggest that the gene responsi-
?T in both kindreds are not linked to the H ras-I
1 1p15.5, we per-
1 Ip15.5. We
11 markers) in 7 and
= 0.001). To improve informative-
= 0.001 (Table II). As H ras-l and SMUC
onal markers and multipoint analyses. The precise
)f the LQT gene within
therefore, that recombination events between the
s and DNA markers near H ras- 1 might account for
to identify linkage in these families. To exclude this
1 lpl 5.5 is not known. It is
Curran et al.
Multipoint LOD Score
(8). Misclassification of a few family members could lead to Download full-text
false conclusions about the presence or absence oflinkage in a
family, so continued conservative phenotypic classification is
The phenotype of affected members of the families de-
scribed here appears different from that of previously studied
families. The incidence of sudden death was
kindred 1756 and 18% for kindred 1977. By contrast, the inci-
dence of sudden death in three large chromosome 1l-linked
families was lower at
rhythmias in these families also differed from families linked to
1 1p15.5. In kindred 1756 symptoms occurred upon awakening
and in kindred 1977 symptoms occurred at rest. Two individ-
uals in this family died during sleep. By contrast, in chromo-
some 11-linked families arrhythmias were frequently precipi-
tated by exercise or anxiety (50%). In future studies it will be of
interest to determine the phenotypic consequences ofdifferent
The identification of locus heterogeneity for autosomal
dominant LQT will complicate genetic testing for this disorder.
Genetic testing in a family is feasible only after LQT has been
linked to chromosome 11 markers in that family. It will be
difficult, therefore, to use genetic diagnosis in small families.
The mechanism ofQT prolongation in LQT is unknown
but may involveprolonged action potential duration in individ-
ual cardiac myocytes and dispersion ofcardiac repolarization,
or afterdepolarizations which contribute to the expression of
repolarization potentials on the surface ECG. In theory, any-
thing that affects myocellular depolarization and repolariza-
tion could cause LQT. As the autonomic nervous system is an
important mediator of cardiac repolarization, it has also been
hypothesized that genes involved in the development and regu-
lation ofautonomic innervation could be involved in this dis-
order ( 18, 19, 20, 21). Given the multiple possible mecha-
nisms that could cause LQT, it should not be surprising that
two or more different genetic mechanisms may account for this
-4%. The factors that precipitated ar-
We thank J. Robinson, A. Ewart, V. Turner, M. Woodward, P.
Cartwright, and R. White for their help and advice.
This work was supported in part by National Institutes of Health
grants RO IHL-4807 and ROIHL-33843, an American Heart Associa-
tion Established Investigator Award, Public Health Services research
grant No. MOl-RR00064 from the National Center for Research Re-
sources, a Syntex Scholars Award, the Technology Access Section of
the Utah Genome Center, and the LDSHospital Deseret Foundation.
1. Zipes, D. P. 1987. Proarrhythmic effects ofantiarrhythmic drugs. Am. J.
2.Keating,M. T., C. Dunn, D.Atkinson, K.Timothy,G. M. Vincent, and M.
Leppert. 1991.Linkageofa cardiacarrhythmia, thelong QT syndrome, and the
Harveyras-Igene. Science(Wash. DC). 252:704-706.
3. Jervell, A., and F. Lange-Nielsen. 1957. Congenital deaf mutism, func-
tional heart disease withprolongation oftheQT interval, and sudden death. Am.
Heart J. 54:59-78.
4.Ward, 0. C. 1964. A new familial cardiacsyndromein children. J. Ir. Med.
5. Romano, C. 1965. Congenital cardiac arrhythmia. Lancet. i:658-659.
6. Vincent, G. M., D.Jaiswal, and K.Timothy. 1991. The effect ofexercise on
heartrate, QT, QTcandQT/QS2 in the Romano-Ward inherited long QT syn-
drome. Am. J. Cardiol. 68:498-503.
7.Keating, M.T., D. Atkinson, C. Dunn, K. Timothy, G. M. Vincent, and M.
Leppert. 1991. Linkage of thelong QT syndrome to the Harvey ras-1 locus on
chromosome 11. Am. J. Hum. Genet. 49:1335-1339.
8. Vincent, G. M., K. W. Timothy, M. Leppert, and M. Keating. 1992. The
spectrum ofsymptoms and QT intervals in carriers ofthe gene for the long QT
syndrome. N. Engl. J. Med. 327:846-852.
9. Goldfarb, M., K. Shimizu, M. Perucho, and M. Wigler. 1982. Isolation and
preliminary characterization of a human transforming gene from T24 bladder
carcinoma cells. Nature(Lond.). 296:404-409.
10. Griffiths, B., D. J. Matthews, L. West, J. Attwood, S. Povey, D. M. Swal-
low, J. R.Gum, and Y. S. Kim. 1990. Assignment ofthe polymorphic intestinal
mucingene (MUC2) to chromosome 1 lpl5. Ann. Hum. Genet. 54:277-285.
1 1.Weber,J. L., and P. E. May. 1989. Abundant class ofhuman DNApoly-
morphismswhich can betyped using the polymerase chain reaction.Am. J. Hum.
12. McNoe, L. A., M. R. Eccles, and A. E. Reeve. 1991. Dinucleotiderepeat
polymorphism at the Dl 1S860 locus. Nucleic Acids. Res. 20:1161.
13. Polymeropoulos, M. H., H. Xiao, D. S. Rath, and C. R. Merril. 1991.
Tetranucleotide repeat polymorphism at the human tyrosine hydroxylase gene
(TH). Nucleic Acids. Res. 19:3753.
14. Kramer, P., W. Becker, P. Heutink, M. James, C. Julier, M. Lathrop, J. A.
Luty, Z. Wang, J. L. Weber, P. Wilkie, and M. Litt. 1992. A comprehensive
genetic linkage map ofthe Human Genome/NIH/CEPH collaborativeMapping
Group. Science (Wash. DC). 258:67-86.
15. Towbin, J. A., L. Pagotto, B. Sju, J. Robinson, A. Moss, E. R. B.McCabe,
and J. F. Heitmancik. 1992. Romano-Ward long QT syndrome (RWLQTS):
evidence ofgenetic heterogeneity. Pediatr. Res. 31:125A.(Abstr.)
16. Kerem, B., J. Benhorin, Y. M. Kalman, A. Medina, T. D. Dyer, J. Blan-
gero, and J. W. MacCluer. 1992. Evidence for genetic heterogeneity in the long
QT syndrome. Am. J. Hum. Genet. 51:192A. (Abstr.)
17. Jeffery, S., R. Jamieson, M. A. Patton, and J. Till. 1992. Long QT and
Harvey-ras. Letter to the Editor. Lancet. 339:255.
18. Yanowitz, F., J. Preston, and A. Abildskov. 1966. Functionaldistribution
ofright and left stellate inervation to the ventricals. Circ. Res. 18:416-428.
19. Moss, A. J., and J. McDonald. 1971. Unilateral cervical ganglionectomy
for the treatment ofthe long QT syndrome. N. Engl. J. Med. 285:903-904.
20. Schwartz, P., and E. Locati. 1985. The idiopathic long QT syndrome:
pathogenic mechanisms and therapy. Eur. Heart J. 6:103-114.
21. Schwartz, P. J., E. Locati, S. G. Priori, and A. Zaza. 1990. The long QT
syndrome. In Cardiac Electrophysiology from Cell to Bedside. D. P. Zipes andJ.
Jalife, editors. W.B. Saunders Co., Philadelphia. 589-605.
LocusHeterogeneity ofLong QTSyndrome