J

Origin of the Swedish long QT syndrome Y111C/KCNQ1 foundermutation

Annika Winbo, MD,* Ulla-Britt Diamant, MSc,† Annika Rydberg, MD, PhD,*ohan Persson, Consultant in Mathematics and Computer Science,* Steen M. Jensen, MD, PhD,†

Eva-Lena Stattin, MD, PhD‡

From the *Department of Clinical Sciences, Pediatrics, †Department of Public Health and Clinical Medicine, HeartCentre, and ‡Department of Medical Biosciences, Medical and Clinical Genetics, Umeå University, Umeå, Sweden.

tymvl

Csptgna

Km

A(

BACKGROUND The Y111C/KCNQ1 mutation causes a dominant-negative effect in vitro but a benign clinical phenotype in aSwedish long QT syndrome population.

OBJECTIVE The purpose of this study was to investigate theorigin (genealogic, geographic, genetic, and age) of the Y111C/KCNQ1 mutation in Sweden.

METHODS We identified 170 carriers of the Y111C/KCNQ1 muta-tion in 37 Swedish proband families. Genealogic investigation wasperformed for all families. Haplotype analysis was performed in 26probands, 21 family members, and 84 healthy Swedish controls,using 15 satellite markers flanking the KCNQ1 gene. Mutation agewas estimated using ESTIAGE and DMLE computer software andregional population demographic data.

RESULTS All probands were traced back to a northern river valleyregion. A founder couple born in 1605/1614 connected 26 of 37families. Haplotyped probands shared 2–14 (median 10) uncom-mon alleles, with frequencies ranging between 0.01 and 0.41(median 0.16) in the controls. The age of the mutation was

November 26, 2010.)

1547-5271/$ -see front matter © 2011 Heart Rhythm Society. All rights reserved

hat is, 600 years (95% confidence interval 450; 850) assuming 25ears per generation. The number of now living Swedish Y111Cutation carriers was estimated to approximately 200–400 indi-iduals for the mutation age span 22–24 generations and popu-ation growth rates 25%–27%.

ONCLUSION The Y111C/KCNQ1 mutation is a Swedish long QTyndrome founder mutation that was introduced in the northernopulation approximately 600 years ago. Enrichment of the mu-ation was enabled by a mild clinical phenotype and strong re-ional founder effects during population development of theorthern inland. The Y111C/KCNQ1 founder population constitutesn important asset for future genetic and clinical studies.

EYWORDS Dominant-negative mutation; Founder mutation; Geneutation; Ion channel; Long QT syndrome

BBREVIATIONS CI � confidence interval; cM � centi-Morgan106 base pairs); LQTS � long QT syndrome

(Heart Rhythm 2011;8:541–547) © 2011 Heart Rhythm Society. All

estimated to 24 generations (95% confidence interval 18; 34), rights reserved.

IntroductionLong QT syndrome (LQTS) is an autosomal dominantinherited arrhythmic disorder that can be caused by sev-eral hundred different mutations in at least 12 separategenes, most of which affect the function of cardiac ionchannels.1–3 LQTS disease phenotype spans from asymp-tomatic carriership, with or without QT prolongation onthe ECG, to syncope and sudden cardiac death fromventricular arrhythmia. Characteristically the LQTS-causing mutation is family specific, but founder muta-tions have been identified.2,4 Founder mutations are mu-tations that are identical by descent and are enriched in a

This research was supported by grants from the Swedish Heart-LungFoundation, the Heart Foundation of Northern Sweden, and the NorthernCounty Councils Cooperation Committee. Address reprint requests andcorrespondence: Dr. Annika Winbo, Department of Clinical Sciences,Pediatrics, Umeå University, 90185 Umeå, Sweden. E-mail address:annika.winbo@pediatri.umu.se. (Received November 1, 2010; accepted

population derived from a limited gene pool. The na-scence and development of a founder population is in-fluenced by conditions such as environmental factors andsocioethnic constructs as well as the specific properties ofthe mutation itself. The population development of Swe-den encompassed a relative isolation of the river valleysspanning the northern part of the country from northwestto southeast, resulting in strong regional founder effects.5

We previously described the clinical phenotype of 80Swedish carriers of the Y111C mutation in the KCNQ1gene.6 The Y111C mutation, first reported in a NorthAmerican female in the year 2000, has a strong domi-nant-negative electrophysiologic effect in vitro whilepresenting with a surprisingly mild clinical phenotype inSwedish carriers.6 –9

The aim of this study was to explore the possibilitythat the high occurrence of the Y111C/KCNQ1 mutationin the Swedish population is caused by a founder effectby investigating the origin (genealogic, geographic, ge-

netic, and age) of the mutation.

. doi:10.1016/j.hrthm.2010.11.043

eIfMflrytDNrfivmrcwalinoc

542 Heart Rhythm, Vol 8, No 4, April 2011

MethodsPatients and familiesSwedish Y111C/KCNQ1 index families were recruited fromthe regional LQTS Family Clinic in Umeå, a national in-ventory of LQTS patients, and national referrals to theaccredited laboratory of the Department of Clinical Genet-ics, Umeå University Hospital. Probands (index cases) weredefined as the first identified mutation-carrier in an indexfamily. Index families were defined as the proband plus anymutation-carrier in the extended family identified throughthe cascade-screening process, encompassing a voluntaryand consecutive direct mutation analysis of first-degree rel-atives.10,11

Genealogy and geographyGenealogic investigation in Sweden is enabled by compre-hensive population records initiated in 1686 that annuallydescribe the births, deaths, marriages, and migration ofinhabitants in Swedish parishes. Data were collected fromlocal parish registers and catechetical examination recordsat the Umeå University research archive. Digital genealogicand census archives were accessed from the Swedish ar-chive information homepage (www.svar.ra.se) and fromprivate genealogic databases. The genealogic investigationwas performed by tracing all maternal and paternal ancestrallines of the Y111C/KCNQ1 probands throughout as manygenerations as possible. The most likely founder couple wasdefined as the couple that connected the highest number ofprobands in the least generations. The geographic clusteringand migration of ancestors were analyzed by noting birth-places, throughout the 17th to the 20th centuries, on aregional map. The origin of ancestors’ spouses, defined asfrom within or without the region, was noted in order toinvestigate prevailing marriage patterns. DISGEN computersoftware (Lakewood, CO, USA) was used for organizingand storing genealogic data. Pedigrees and geographic mapswere constructed using Open Source software (Inkscapevector graphics editor) and GIMP (GNU image manipula-tion program).

Molecular genetics analysesMolecular diagnosis of mutation-carriers was obtained bybidirectional sequencing of the KCNQ1 gene using the CEQ8000 (Beckman Coulter, Fullerton, CA, USA) or by directmutation analysis using MGB probes by TaqMan 7000

Figure 1 Overview of the 15 chosen microsatellite markers, 6 upstream11 over a physical distance of 8 cM (106 base pairs). Location of microsafollowed by the number under the arrows. Locations of mutation (Y111Crectangle, respectively.

(Applied Biosystems, Foster City, CA, USA). Analyses

were performed on DNA extracted from whole blood usingstandard methods.

Haplotype analysis was performed using 15 microsatel-lite markers, 6 upstream and 9 downstream of the KCNQ1gene located on the short arm of chromosome 11, withmarkers spanning a total distance of approximately 8 centi-Morgan (8 � 106 base pairs; Figure 1). Microsatellite mark-rs were chosen from the National Centre for Biotechnologynformation Entrez Gene database with respect to distancerom the KCNQ1 gene using deCODE, Généthon, and

arshfield genetic maps. Forward and reverse primersanking the microsatellite region were used for each sepa-ate marker (Sigma-Aldrich, St Louis, MO). Fragment anal-sis was performed according to manufacturers’ instruc-ions, using an automated capillary electrophoresis-basedNA Sequencer (Wave 3500 HT, Transgenomic, Omaha,E, USA). Solutions and material for polymerase chain

eaction mix were manufactured by GE Healthcare (Fair-eld, CT) and Applied Biosystems. GeneMapper softwareersion 3.7 (Applied Biosystems) was used to analyze theicrosatellite data. All data were manually analyzed. Mic-

osatellite marker analysis was performed on two mutation-arriers in separate generations from each index family,henever possible, to aid in resolving the disease-associ-

ted alleles. Based on the pattern of shared alleles, the mostikely ancestral haplotype was reconstructed. In order tonvestigate the frequency of disease-associated alleles inoncarriers, microsatellite marker analysis was performedn 168 control chromosomes from 84 healthy military re-ruits matched for origin (northern Sweden).

Estimation of mutation age and prevalenceThe age of a mutation can be inferred assessing the decay ofa shared haplotype in mutation-carriers of a common de-scent due to recombination and/or mutation.12 Due to theinherent degree of uncertainty when attempting to infermutation age from present-day observations, we used twoseparate approaches to estimate the age and associated prev-alence of the Y111C/KCNQ1 mutation. First, we usedESTIAGE computer software in order to obtain an estimateof mutation age, including computations of the 95% confi-dence interval (CI), where mutation age was defined as theage of the most recent common ancestor of the sampledprobands.13 The ESTIAGE input file included the extent ofshared alleles among probands, the frequency of disease-

ownstream of the KCNQ1 gene, located on the short arm of chromosomearkers are indicated by short vertical arrows. Markers are named D11S-

ene (KCNQ1) are indicated by the vertical bold line and horizontal gray

and 9 dtellite m) and g

associated alleles in healthy controls, and the recombination

DtpDagtewgirdtwiatDtio

B

543Winbo et al Origin of Swedish LQTS Y111C/KCNQ1 Founder Mutation

frequencies for the 15 microsatellite markers. Recombina-tion frequencies were calculated using the physical dis-tances between the Y111C mutation and the separate mark-ers, and the standard correspondence 1 cM � 106 base pairs.Separate estimates were performed, exploring the impact onthe estimate by the factors mutation model (stepwise orequal) and mutation rate (10–6 to 10–3). Second, we used

MLE computer software (www.dmle.org) in order to es-imate the prevalence of the mutation from a range ofossible values for mutation age, including 95% CI. TheMLE input file included the full haplotype of probands

nd controls for the 15 markers, an estimate of populationrowth rate per generation, and an estimate of the propor-ion sampled. The average population growth rate per gen-ration, evaluated as a discrete variable [population gro-th � e^(ln [end population/start population]/number ofenerations) –1] was calculated using regional demograph-cs data from the Jämtland and Västernorrland Countiesegarding the time period 1570–1950 from Statistics Swe-en (www.scb.se). Proportion sampled (i.e., proportion ofhe total Y111C population constituted by the probands)as viewed as the unknown variable, and iterations over the

nterval 0.0001–0.5 were performed. The upper limit forcceptable values of proportion sampled was corrected forhe number of identified now living mutation-carriers.MLE analysis results were interpreted assuming that the

rue value for mutation age occurred where the confidencentervals computed by the DMLE and ESTIAGE softwareverlapped.

Ethical considerationsAll individuals, including the legal guardian of children,gave written informed consent for haplotype analysis. Thestudy was approved by the regional ethical review board inUmeå.

ResultsWe identified 170 mutation-carriers of the LQTS mutationY111C/KCNQ1 in 37 apparently unrelated Swedish indexfamilies. Genealogic investigation, including geographictracing of earliest known ancestors, was performed for allfamilies. Haplotype analysis was performed for 26 of the 37families. In 21 of these 26 families, two generations ofmutation-carriers were available to aid in the identificationof the disease-associated haplotype.

Genealogy, geographic origin, migration, andmarriage patternsBy tracing the birthplaces of all ancestors of the now livingY111C probands, a common geographic origin was identi-fied. Both maternal and paternal ancestors born in the 19thcentury were found to be clustered in the geographic regionof the Ångerman River valley with tributaries, spanningacross mid-northern Sweden (Figure 2). Tracing the ances-tral lines of the Y111C probands further back led to theidentification of a common ancestor couple, born in 1605/

1614. The identified couple connected 26 of the 37 probands

over a distance of 13 generations and approximately 400years (Figure 3). The successive migration and spread of theY111C population along the riverbeds of the ÅngermanRiver throughout the 17th to 20th centuries was recon-structed by noting the birth places of the descendants of thefounder couple (obligate mutation-carriers) on regionalmaps (Figure 4).

Analysis of marriage patterns in the Y111C populationperformed by investigating the origin of spouses chosen bythe obligate-mutation carriers revealed a marked preferencefor marrying individuals originating from the same rivervalley. All spouses of obligate mutation-carriers born dur-ing the 17th and 18th centuries were born within the sameriver valley. During the 19th and 20th centuries, the corre-sponding numbers were 93% and 62%, the latter corre-sponding to increased migration in the late 20th century.

Haplotype analysisIn order to evaluate whether the probands shared a commongenetic origin, fragment analysis of 15 microsatellitemarker loci flanking the Y111C mutation was performed in26 probands, 21 family members, and 168 control chromo-somes. A shared haploblock extending over 2–14 (median

Figure 2 Map of Sweden/Scandinavia illustrating the geographic cluster-ing of ancestors of all 37 long QT syndrome Y111C/KCNQ1 index familiesduring the 19th century. The rivers and tributaries flowing from northwest tosoutheast across northern Sweden are depicted as lines. The river valley regionof interest, shown in magnification in the upper left corner, is shaded in gray.

irthplaces of Y111C ancestors are indicated by dots.

10) marker loci was identified (Table 1). The frequencies for

544 Heart Rhythm, Vol 8, No 4, April 2011

the shared alleles ranged between 0.01 and 0.41 (median0.16) in the controls. An alternative upstream haploblockspanning at least five markers (no upper boundary identi-fied) was seen in a subgroup of eight probands genealogi-cally connected to a more recent common ancestor born inthe early 18th century. The occurrence of the alternativehaploblock, a result of a probable recombination event inthat sub-branch of the pedigree, constitutes a link betweenthe genealogic and the haplotype data (Figure 5). The ex-tent of shared alleles between probands with (19) andwithout (7) genealogic connection to the Y111C pedigreewas comparable (2–14 alleles, median 10 vs 3–14 alleles,median 12).

Mutation age and prevalenceThe estimated age of the Y111C mutation computed by theESTIAGE software (assuming a mutation rate of 10–6 permarker and per generation) was 24 generations with a 95%CI ranging from 18 to 34 generations. Assuming that onegeneration is 25 years, the mutation was estimated to be 600years old (95% CI 450; 850). Estimates were not affected bythe mutation model used (stepwise or equal) and were onlyslightly affected by the mutation rate. The estimated age ofthe Y111C mutation computed by the DMLE software (as-suming a population growth rate per generation of 25% anda proportion of population sampled of 0.1) was 23 genera-tions (95% CI 16; 36) with a corresponding mutation age of575 years (95% CI 400; 900). The average growth rate pergeneration in the region of interest (Jämtland and Väster-norrland counties) between the years 1570 and 1950 wascalculated as between 25% and 27% using regional popu-lation demographic data. Mutation age estimates were onlyslightly affected by the variable proportion of populationsampled. The value for proportion of population sampledvaried between 0.07 and 0.12 for the mutation age span

Figure 3 Long QT syndrome Y111C/KCNQ1 population pedigree con-necting 26 probands through a line of obligate mutation-carriers to afounder couple born in 1605/1614, spanning 13 generations. Probands aredepicted as diamonds. H indicates probands with available haplotype data.

22–24 generations and population growth rates 25%–27%.

Assuming that the sampled probands constitute the calcu-lated 7%–12% of the total Y111C population, the number ofnow living Y111C mutation-carriers would be somewherebetween 200 and 400 individuals (217–371). This in turnwould translate into an approximate estimate of Y111Cmutation prevalence in mid-northern Sweden (Jämtland andVästernorrland counties plus the neighboring Västerbottencounty) of 1:1,500–3,000 (population �6 � 105) comparedto �1:25,000–35,000 for the entire Swedish population(�9.3 � 106).

DiscussionWe studied the origin of the Y111C/KCNQ1 founder mu-tation, an important cause of LQTS in Sweden. By inves-tigating the genealogy and haplotype data of 37 SwedishY111C probands, a shared geographic origin in the Ånger-man River valley area was found, a founder couple born in1605/1614 connecting 26 of 37 probands was identified, andtraces of the original ancestral haploblock was seen in theDNA of all 26 haplotyped probands. Mutation dating placedthe convergence of the Y111C bloodlines in the 15th cen-tury, at the beginning of colonization of the northern Swed-ish inland. Documented migration and marriage patterns of

Figure 4 Regional maps (A-D) of the Ångerman River valley area inmid-northern Sweden during four centuries (17th–20th centuries). Thebirth places of the founder couple and their descendants, indicated by dots,demonstrate the migration and spread of the long QT syndrome Y111C/KCNQ1 population over time. The Ångerman River is depicted by forkinglines. The region is delimited by the Norwegian border on the left and the

Gulf of Bothnia/Baltic Sea on the right.

PPPPPPPPPPPPPPPPPPPPPP2PPPf

adgL set of p

545Winbo et al Origin of Swedish LQTS Y111C/KCNQ1 Founder Mutation

the Y111C population during the 17th through 20th centu-ries were consistent with the development of a foundereffect. The separate methods used to investigate the occur-rence of the Y111C mutation in Sweden showed excellentagreement and mutual support.

Founding of the Swedish Y111C populationThe earliest ancestors of the Swedish Y111C populationprobably were among the pioneers who settled in the north-ern inland and formed part of the local founding gene pool.Colonization started in the coastal areas in the early 14thcentury, and the settlers used the rivers as travel routes andsuccessively colonized their shores. The natural boundariesof the river valleys, with mountains and wild forests sepa-rating them, constituted effective geographic isolation.Within these regions, immigration was limited and marriagepatterns included a preference for spouses from the samearea and a certain degree of consanguineous marriages,effectively forming genetic subisolates specific to each rivervalley.14

Mutation age, distribution, and prevalenceAccording to the mutation dating, introduction of the

Table 1 Reconstruction of the ancestral haplotype in 26 long Qmicrosatellite loci flanking the KCNQ1 gene

M6 M5 M4 M3 M2 M1 M

1 3 7 5 1 9 2 22 3 7 5 1 9 2 23 3 7 5 1 9 2 24 3 7 5 1 9 2 25 3 7 5 1 9 2 26 3 7 5 1 9 2 27 3 7 5 1 9 2 28 3 7 5 1 9 2 29 3 7 5 1 10 2 210 4 7 5 3 2 8 211 4 1 9 5 10 2 212 3 1 5 1 10 2 213 4 6 1 4 9 2 214 4 1 9 5 10 2 215 4 1 9 5 10 2 216 4 1 9 5 10 2 217 4 1 9 5 10 2 218 4 1 9 5 10 2 219 4 1 9 5 10 2 220 4 1 9 5 10 2 221 4 5 4 18 10 2 622 4 1 9 5 10 2 23 4 1 9 5 10 2 224 4 1 9 5 10 2 225 4 1 9 5 10 2 226 4 1 9 7 10 2 2rq 0.34 0.26 0.05 0.01 0.08 0.11 0.

On the top x-axis, the 15 microsatellite markers (M), 6 upstream and 9nd outward. The alleles are sequentially numbered according to size, incisease-associated alleles in the controls are given. On the y-axis, theenealogic connection. Probands with a family member with haplotype daight gray fields indicate a separate upstream haploblock shared by a sub

Y111C mutation into the Swedish population was concur-

rent with colonization of the northern inland, which is inagreement with the finding that all probands shared a com-mon origin in the Ångerman River valley. Because mostdescendants of the founder couple pertained within the sameregion well into the 20th century, the geographic spread ofY111C mutation-carriers appears to have been limited. Ac-cordingly, few Y111C probands have been identified insouthern Sweden, except in the major urban regions ofStockholm and Gothenburg. This corresponds well to thehigh regional prevalence of the Y111C mutation suggestedby our data as well as experience from our regional LQTSFamily Clinic, where the Y111C mutation explains approx-imately one third of LQTS probands.

The Y111C mutation is a common cause of LQTS inSweden but to our knowledge has not, apart from six casesin North America,7,15 been further reported in the mutationinventories of neighboring countries or elsewhere.16,17 As acuriosity, genealogic records revealed that seven individualspertaining to the Y111C population emigrated to NorthAmerica in the 19th century, as did 1.5 million Swedes.18 Itis plausible that the North American Y111C carriers couldbe of Swedish descent, possibly representing cuttings of the

rome probands carrying the Y111C/KCNQ1 mutation at 15

M2 M3 M4 M5 M6 M7 M8 M9

8 5 7 2 5 4 3 28 5 7 2 5 4 3 28 5 7 2 5 4 3 28 5 7 2 5 4 5 28 5 7 2 5 4 5 28 5 7 2 5 4 3 28 5 7 2 5 4 3 28 5 3 1 5 8 3 28 5 7 2 5 4 3 18 5 7 7 4 4 3 38 5 7 2 5 4 3 58 5 7 2 5 4 3 38 5 7 2 5 5 3 48 5 7 2 5 4 2 38 5 7 2 5 4 3 38 5 7 2 5 5 2 38 5 7 2 5 5 3 48 5 7 2 5 5 2 48 5 7 2 5 4 4 28 5 7 2 5 4 2 22 2 6 6 1 4 3 38 5 6 2 1 5 3 38 5 7 2 5 4 2 28 5 7 2 5 4 2 38 5 7 2 5 5 5 35 5 3 2 4 4 2 10.15 0.41 0.11 0.16 0.2 0.24 0.41 0.11

ream of the KCNQ1 gene (vertical black line), are numbered from the genecontrol alleles (n � 186). On the bottom x-axis. frequencies (frq) of theds (P) are ordered according to association in pedigree. P20–P26 lackindicated in bold. Dark gray fields indicate assumed ancestral haplotype.robands.

T synd

1

36

downstludingprobanta are

founder tree. Curiosities aside, our study indicates that al-

s

nprgcmflrhW

546 Heart Rhythm, Vol 8, No 4, April 2011

though the Y111C/KCNQ1 mutation has been of greatestclinical impact in mid-northern Sweden, ancestral roots inthe Ångerman River valley constitute an augmented prob-ability for Y111C carriership irrespective of current local-ity.

LQTS founder populationsLarge LQTS founder populations have been described inother relatively isolated populations, such as the Boer im-migrants in South Africa and the Finnish population.19,20

Due to their relative genetic homogeneity, LQTS founderpopulations constitute excellent human research models inwhich elusive concepts such as genotype–phenotype corre-lations and the impact of modifying factors on diseasephenotype can be studied in the “in vivo” context.1,4 TheSouth African A341V/KCNQ1 founder population, charac-terized by an unusually malignant clinical phenotype de-spite a relatively mild functional effect of the mutation invitro, has provided useful insights into risk modifiers forLQTS and arrhythmia.21–23

Possible implications of the Swedish Y111Cfounder populationThe 0.05% annual incidence of life-threatening cardiacevents before beta-blocker therapy previously documentedin a subset of the Swedish Y111C founder population likelyfacilitated the propagation and enrichment of the mutationin the population.6 However, the finding of such a lowincidence of life-threatening cardiac events despite the�75% mutation-associated reduction of Ks channel func-

Figure 5 Combined genealogic and haplotype data for the long QTyndrome Y111C/KCNQ1 population presented within a temporal context.

Horizontally forking branches of the Y111C pedigree on the right, span-ing 13 generations from the year 1600 to present day, lead up to therobands in the end branches on the left. Probands (index families) areepresented on the Y-axis from top to bottom as the 19 index families withenealogic connection, followed by the 7 families that lack an establishedonnection to the Y111C pedigree. The alleles found at the microsatellitearkers for each proband are depicted as 15 boxes on the X-axis (6/9),anking the KCNQ1 gene, represented by the vertical line. Black boxesepresent the ancestral allele. Gray boxes represent the separate upstreamaploblock shared by probands pertaining to a sub-branch of the pedigree.hite boxes represent all other alleles.

tion demonstrated in vitro in itself raises questions of

whether other as yet unknown factors affect the clinicalphenotype of this founder population. Clearly, functionaleffects do not translate directly into clinical reality in LQTS,a concept that was first proposed based on the clinicalobservations of the large A341V population and that hasnow been corroborated.6,19

The discrepancy between the expected and the resultantphenotype found in the Y111C founder population indicatesits potential role as an “in vivo” research model for identi-fying factors that influence clinical severity in LQTS. Thecommon origin of the Y111C probands allows for the pos-sibility of genetic material with potentially modifying prop-erties to co-segregate with the deleterious mutation in thepopulation. In this study, we present haplotype data dem-onstrating a wide haploblock shared by the majority ofsampled probands containing promoter regions, genes, andsingle nucleotide polymorphisms of potential interest aswell as a separate upstream haploblock in a subset of moreclosely related probands. Further studies are needed to ver-ify or dismiss the existence of a genetic modifying influenceon the clinical phenotype of the Y111C/KCNQ1 founderpopulation.

Study limitationsComplete data were not available for all 37 probands. Ge-nealogic connection to the pedigree was not found for 11 of37 probands; however, it is well known that all bloodlinesare not conceived within wedlock. The mutation age esti-mates placed the convergence of the probands’ bloodlinesbefore the 16th century, which further explains why allprobands could not be connected using available genealogicdata. DNA for haplotype analysis was available for only 26of 37 probands. However, the extent of shared alleles (me-dians) did not significantly differ between probands withand without genealogic connection, suggesting that the an-alyzed proband sample is representative. There are signifi-cant inherent limitations to inferring mutation age and prev-alence from empiric observations, and estimates should beinterpreted with care. However, the separate methods usedwere in excellent agreement, and the mutation age estimateswere in accordance with conclusions drawn from the ge-nealogic data and regional population history.

ConclusionThe Y111C/KCNQ1 founder mutation probably was intro-duced in the inland of northern Sweden by early settlers inthe 15th century. Subsequent population developmentwithin the relatively isolated river valleys caused strongfounder effects that in combination with the mutations’ mildphenotype probably enabled enrichment of this LQTS mu-tation in the northern Swedish population. This conclusionis supported by the presented clinical, epidemiologic, ge-nealogic, and haplotype data. This large LQTS founderpopulation constitutes an important asset for future genetic

and clinical studies.

1

1

1

1

1

1

1

1

1

1

2

2

2

547Winbo et al Origin of Swedish LQTS Y111C/KCNQ1 Founder Mutation

AcknowledgmentsWe thank all of the families that participated in this study.We thank Dr. Emmanuelle Genin (Hopital Paul Brousse,Villejuif Cedex, France) for graciously allowing us to usethe ESTIAGE computer software. We thank SusannHaraldsson, Medical Laboratory Assistant at the Depart-ment of Medical and Clinical Genetics, Umeå UniversityHospital, Umeå, for expert technical assistance. Illustrationsfor maps and figures were provided by illustrator ErikWinbo (erikwinbo.artworkfolio.com).

References1. Schwartz PJ. The congenital long QT syndromes from genotype to phenotype:

clinical implications. J Intern Med 2006;259:39–47.2. Hedley PL, Jorgensen P, Schlamowitz S, et al. The genetic basis of long QT and

short QT syndromes: a mutation update. Hum Mutat 2009;30:1486–1511.3. Lu JT, Kass RS. Recent progress in congenital long QT syndrome. Curr Opin

Cardiol 2010;Mar 10[Epub ahead of print].4. Brink PA, Schwartz PJ. Of founder populations, long QT syndrome, and destiny.

Heart Rhythm 2009;6(11 Suppl):S25–S33.5. Einarsdottir E, Egerbladh I, Beckman L, Holmberg D, Escher SA. The genetic

population structure of northern Sweden and its implications for mapping ge-netic diseases. Hereditas 2007;144:171–180.

6. Winbo A, Diamant UB, Stattin EL, Jensen SM, Rydberg A. Low incidence ofsudden cardiac death in a Swedish Y111C type 1 long-QT syndrome population.Circ Cardiovasc Genet 2009;2:558–564.

7. Splawski I, Shen J, Timothy KW, et al. Spectrum of mutations in long-QTsyndrome genes. KVLQT1, HERG, SCN5A, KCNE1, and KCNE2. Circulation2000;102:1178–1185.

8. Jespersen T, Rasmussen HB, Grunnet M, et al. Basolateral localisation ofKCNQ1 potassium channels in MDCK cells: molecular identification of anN-terminal targeting motif. J Cell Sci 2004;117(Pt 19):4517–4526.

9. Dahimene S, Alcolea S, Naud P, et al. The N-terminal juxtamembranous domain

of KCNQ1 is critical for channel surface expression: implications in the Romano-Ward LQT1 syndrome. Circ Res 2006;99:1076–1083.

2

0. Hofman N, Tan HL, Alders M, van Langen IM, Wilde AA. Active cascadescreening in primary inherited arrhythmia syndromes. J Am Coll Cardiol 2010;55:2570–2576.

1. Schwartz PJ. Cascades or waterfalls, the cataracts of genetic screening are beingopened on clinical cardiology. J Am Coll Cardiol 2010;55:2577–2579.

2. Slatkin M, Rannala B. Estimating allele age. Annu Rev Genomics Hum Genet2000;1:225–249.

3. Genin E, Tullio-Pelet A, Begeot F, Lyonnet S, Abel L. Estimating the age of raredisease mutations: the example of triple-A syndrome. J Med Genet 2004;41:445–449.

4. Bittles AH, Egerbladh I. The influence of past endogamy and consanguinity ongenetic disorders in northern Sweden. Ann Hum Genet 2005;69(Pt 5):549–558.

5. Kapplinger JD, Tester DJ, Salisbury BA, et al. Spectrum and prevalence ofmutations from the first 2500 consecutive unrelated patients referred for theFAMILION long QT syndrome genetic test. Heart Rhythm 2009;6:1297–1303.

6. Fodstad H, Swan H, Laitinen P, et al. Four potassium channel mutations accountfor 73% of the genetic spectrum underlying long-QT syndrome (LQTS) andprovide evidence for a strong founder effect in Finland. Ann Med 2004;36:53–63.

7. Berge KE, Haugaa KH, Fruh A, et al. Molecular genetic analysis of long QTsyndrome in Norway indicating a high prevalence of heterozygous mutationcarriers. Scand J Clin Lab Invest 2008;68:362–368.

8. Statistical Yearbook of Sweden 2004. Stockholm, Sweden: Statistics Sweden,2004.

9. Crotti L, Spazzolini C, Schwartz PJ, et al. The common long-QT syndromemutation KCNQ1/A341V causes unusually severe clinical manifestations inpatients with different ethnic backgrounds: toward a mutation-specific riskstratification. Circulation 2007;116:2366–2375.

0. Marjamaa A, Salomaa V, Newton-Cheh C, et al. High prevalence of four longQT syndrome founder mutations in the Finnish population. Ann Med 2009;41:234–240.

1. Brink PA, Crotti L, Corfield V, et al. Phenotypic variability and unusual clinicalseverity of congenital long-QT syndrome in a founder population. Circulation2005;112:2602–2610.

2. Schwartz PJ, Vanoli E, Crotti L, et al. Neural control of heart rate is anarrhythmia risk modifier in long QT syndrome. J Am Coll Cardiol 2008;51:920–929.

3. Crotti L, Monti MC, Insolia R, et al. NOS1AP is a genetic modifier of thelong-QT syndrome. Circulation 2009;120:1657–1663.