Main

Molecular breakthroughs in the 1990s have elevated the inherited LQTS from its previous clinical descriptions as the autosomal recessive Jervell and Lange-Nielsen syndrome associated with deafness and the autosomal dominant Romano-Ward syndrome to a diverse collection of cardiac ion channelopathies(1). Autosomal dominant LQTS is estimated to occur in 1 per 10,000 persons and is speculated to account for nearly half of the 8000 yearly sudden deaths in children. In addition, LQTS, with its perturbed cardiac ion channel genes, provides a molecular model to explore the basis for one form of ventricular arrhythmogenesis that is implicated in a portion of the approximately 300,000 sudden cardiac deaths in the United States each year(2,3).

Marked QT prolongation (QTc ≥ 0.46 s1/2 considered abnormal), symptoms including syncope, seizures, and sudden death, and the polymorphic ventricular tachyarrhythmia known as torsades de pointes comprise the clinical trademarks of inherited LQTS. On a molecular level, over 40 mutations in four cardiac ion channel genes have been identified thus far; KVLQT1 (voltage-gated K channel gene) on chromosome 11p15.5 (LQT1), HERG (human ether-a-go-go related gene) on chromosome 7q35-36 (LQT2), SCN5A on chromosome 3p21-24 (LQT3), and KCNE1 (minK) on chromosome 21q22.1-22.2(LQT5)(4).

Translational research into genotype-phenotype correlations has resulted in the prospect of asymptomatic molecular genetic testing and genotype-targeted therapeutic interventions. For instance, mexiletine may be a useful adjunct to β-blocker therapy in patients having the sodium channel-based LQT3 genotype(5,6). Analyses of ECGs from patients having the LQT1, LQT2, or LQT3 genotypes have been reported to show characteristic patterns for each of the genotypes, namely prolonged T waves (LQT1), small T waves (LQT2), and prolonged QTonset parameters (LQT3)(7). It may be tempting to speculate that prediction of underlying genotype and therefore initiation of genotype-specific therapies may be possible by careful scrutiny of the patient's ECG. For example, if a patient with inherited LQTS has the stereotyped ECG pattern consistent with LQT3, should mexiletine be added to the patient's treatment regimen?

Recently, Ackerman and Porter(8) identified a four-generation kindred with an autosomal dominant form of inherited LQTS. Although several adults were shown subsequently to be affected, the diagnosis of LQTS was established after a near-drowning of a 10-y-old boy who required defibrillation from torsades de pointes ventricular tachyarrhythmia(8). Analysis of the proband's ECG pattern predicted the sodium channel-based LQT3 (Fig. 1). However, we discovered linkage of this family to the chromosome 11p15.5 LQT1 region rather than the chromosome 3p21-24 LQT3 region as predicted by the ECG pattern. Here, we report the characterization of a novel mutation in the KVLQT1 (LQT1) gene as the molecular basis for this family's inherited LQTS.

Figure 1
figure 1

Proband's ECG showing LQT3-like morphology. The proband's lead II ECG recording is shown here with various parameters of repolarization highlighted. The pattern of a normal amplitude, normal duration T wave inscribed after a prolonged ST segment (prolonged QTonset-c) is most consistent with the sodium channel-based LQT3 genotype.

METHODS

Identifying and phenotyping the near-drowning LQTS patient's family. After establishing the diagnosis of LQTS in the proband (lead II QTc = 0.56 s1/2), we obtained ECGs from 29 family members(8). For purposes of linkage analysis, a conservative definition of affected status was used consistent with previous LQTS linkage studies(9,10). A patient was classified phenotypically as affected if they met either of the following two conditions: 1) symptoms plus a QTc ≥ 0.45 s1/2 or 2) QTc ≥ 0.47 s1/2. Unaffected status was given if the patient was asymptomatic and had a QTc ≤ 0.41 s1/2. Finally, a patient was classified as"uncertain" if they either were 1) asymptomatic with a QTc between 0.41 and 0.47 s1/2 or had 2) symptoms but a QTc ≤ 0.44 s1/2. After approval by Mayo Clinic's Department of Pediatrics and Adolescent Medicine's Research Committee and the Mayo Foundation Institutional Review Board, written informed consent was obtained from all participants or their guardians, and blood samples were provided by nearly every available family member (29 of 30).

Genotyping and linkage analyses. Genomic DNA was isolated from the peripheral blood lymphocytes using standard procedures(11). The genomic DNA was genotyped using PCR amplification of 11 polymorphic microsatellite markers that had been linked previously to four LQTS ion channel gene loci: D11S922, D11S4046, and D11S1318 localized to chromosome 11p15.5 (LQT1)(12,13); D7S505, D7S636, and D7S483 localized to chromosome 7q35-36 (LQT2)(10,14); D3S1298, D3S1100, and D3S1767 localized to chromosome 3p21-24 (LQT3)(10); and D21S65 and D21S219 localized to chromosome 21q22.1-22.2 (LQT5)(15). We did not screen for the LQT4 locus on chromosome 4q25-27 as that locus has been identified in only one family to date and a candidate gene has not been established(16).

Amplification of these microsatellite markers used PCR. Foward primers were labeled with phosphoramidite dyes. Each 15-µL reaction contained 50 ng of genomic DNA, 200 µM dNTPs, 8 µM each primer, 0.5 U of Amplitaq Gold (Perkin-Elmer), and 2.5 mM MgCl2. Reactions were cycled in a Perkin-Elmer GeneAmp PCR System 9600 as follows: 10 min at 95°C; then 35 cycles of 30 s at 95°C, 30 s at 55°C, 30 s at 72°C; followed by an extension step of 10 min at 72°C. PCR reactions were held at 4°C until analysis. The PCR products were resolved on a 5% denaturing polyacrylamide gel and detected using an ABI377 DNA sequencer. Genotypes were analyzed using ABI Genescan 2.1 and ABIGenotyper 2.0.

Pairwise linkage analysis was performed using MLINK in the LINKAGE 5.1 software package(17). Similar to previous LQTS linkage studies(9,10), the disease penetrance was set at 0.9 and the LQTn disease gene frequency was assumed to be 0.001 and equal between male and female subjects. The allele frequencies for each DNA marker were assumed to be equally frequent.

Gene mutation identification. Six exons of the KVLQT1 gene encoding most of the transmembrane spanning segments and channel pore (S2-S6) were amplified for DNA sequence analyses. Oligonucleotide sequences for the six primer pairs have been published previously by Wang et al.(13). Each 25-µL PCR included 50 ng of genomic DNA and 20 µM each primer, and cycling conditions were: 5 min at 94°C; then 5 cycles of 20 s at 94°C, 20 s at 64°C, 30 s at 72°C; then 30 cycles of 20 s at 94°C, 20 s at 62°C, 30 s at 72°C; followed by an extension step of 10 min at 72°C (Q. Wang, personal communication, 1998). PCR products were cleaned using the Wizard system (Promega, Madison, WI). The six PCR products from the proband and his unaffected brother were sequenced in both directions using the same primers as used for the first round PCR amplification. DNA sequencing was performed on the ABI377 using approximately 30-100 ng of PCR template and 3.2 pmol of primer. Sequence analysis of the exon encoding the pore-S6 segment of the channel (forward primer, AGGCTGACCACTGTCCCTCT, and reverse primer, CCCCAGGACCCCAGCTGTCCAA) suggested the presence of a deletion event in the affected proband.

ThermoSequenase sequencing (Amersham Life Science, Cleveland, OH) with33 P-labeled ddNTPs was used to define the proband's deletion-mutation. The exon encoding this pore-S6 segment was again amplified by PCR using the afore-mentioned primer pair with the same annealing conditions. The PCR product was enzymatically treated and sequenced in both directions using the same primers as in the initial amplification. The annealing temperature for the 33P-sequencing reaction was 58°C. Samples were run on a 6% denaturing acrylamide gel.

RESULTS

Near-drowning patient's family is linked to chromosome 11p15.5(LQT1). Of the four established LQTS ion channelopathies, linkage studies excluded linkage to the LQT2 (7q35-36), LQT3 (3p21-24), and LQT5(21q22.1-22.2) regions (data not shown). However, there was evidence of linkage to the chromosome 11p15.5 LQT1 locus (Fig. 2) The maternally inherited haplotype 114-190-132 for the DNA markers, D11S922, D11S4046, and D11S1318, co-segregated with the patient's phenotypic status. Two-point linkage analysis demonstrated linkage to these markers with maximum lod scores of 2.07, 3.36, and 2.96 for the markers D11S922, D11S4046, and D11S1318, respectively (Table 1). Each clinically affected patient (proband IV-1, III-1, III-5, III-9, III-11, II-2, II-3, II-5, and II-7) except a maternal aunt (III-4) shared this particular haplotype. The maternal aunt's (III-4) haplotype (126-190-132) represents a recombination event between the markers D11S922 and D11S4046. The candidate gene, KV-LQT1, resides between the D11S922 and D11S4046 markers. This indicates that the recombination event must have occurred telomeric to the location of the mutant KVLQT1 gene. None of the three living individuals who were assigned an uncertain clinical status (I-2, III-2, and IV-3) shared this disease-associated haplotype. Given this chromosome 11p15.5 linkage, the presence of a mutation in the LQT1 gene (KVLQT1) was sought.

Figure 2
figure 2

Linkage of family to chromosome 11p15.5-LQT1. Pedigree for family is shown (circles, female members; squares, male members) with clinical phenotype assigned as either affected (filled), uncertain (hashed), or unaffected (open). The bold arrow points to the proband(IV-1). Genotypes for the chromosome 11p15.5 DNA markers are indicated beneath each symbol and are shown as haplotypes. The actual numbers reflect the size(in base pairs) for each of the three polymorphic DNA markers amplified. The order of the DNA markers is shown from telomere to centromere (top to bottom) as D11S922, D11S4046, and D11S1318. For individuals I-1, II-4, II-6, and III-10, genomic DNA samples were not available. Their inferred genotypes are shown in parentheses. The haplotype (114-190-132) associated with the disease phenotype is enclosed within a rectangle.

Table 1 Pairwise lod scores for chromosome 11p15.5-LQT1 DNA markers

A novel 3-bp deletion in KVLQT1 in the proband is identified. Six exons from the KVLQT1 gene encoding much of the transmembrane spanning ion channel (from S2 through S6) were amplified and sequenced using previously published primers(13). Evaluation of this region of the KVLQT1 gene was chosen as it encodes important functional elements of the ion channel and the majority of LQT1 mutations previously reported are found in this region. The DNA sequences between the proband(IV-1) and his unaffected brother (IV-2) were compared. A 3-bp nucleotide deletion [(Δaag) as shown from the minus strand in Fig. 3] was discovered in the proband's exon that encodes the latter half of the channel pore and most of the final (S6) transmembrane spanning segment (from V308 to A344) (Fig. 3). This 3-bp deletion results in an in-frame deletion of a single amino acid(phenylalanine), ΔF339, without any further disruption of the KVLQT1 potassium channel.

Figure 3
figure 3

Identification of a novel mutation in the candidate gene, KVLQT1. A portion of the DNA sequence from the exon encoding the channel pore through the S6 transmembrane domain is shown here for the proband (IV-1) and his unaffected brother (IV-2).Dashed arrow identifies a mutant allele in the proband's sequence. The nucleotide sequence shown on the right (N, normal allele;M, mutant allele) details the region between the two arrows revealing the 3-bp deletion. The 3-bp deletion is highlighted by a rectangle on the normal sequence (N).

Presence of ΔF339-KVLQT1 mutation confirmed in family members predicted to have inherited LQTS by clinical phenotype and linkage haplotype. Finally, a PCR-based assay was developed to determine the presence or absence of the deletion-containing mutant in the rest of the family (Fig. 4). Using the primers specific for amplifying the exon encoding amino acids 308 through 344, each family member's amplicon was electrophoresed on an 8% acrylamide gel and stained with ethidium bromide. As shown in Figure 4, the abnormal banding pattern (heteroduplex) was identical in the proband, and each of the clinically affected family members who also shared the disease-associated haplotype. This defining band pattern was indeed present in the maternal aunt(III-4, Fig. 2) with the presumed recombination event but was absent in the three individuals (I-2, III-2, and IV-3, Fig. 2) whose clinical status was uncertain. Direct sequence analysis from the nine individuals who shared the proband's heteroduplex pattern confirmed the identical 3-bp deletion (data not shown).

Figure 4
figure 4

Confirmation of ΔF339 mutation in affected individuals. Results of a PCR-based assay amplifying the exon containing the proband's 3-bp deletion (pore-S6) is shown. The pedigree is displayed such that each individual is above the lane containing his/her amplified PCR product. The heteroduplex banding pattern seen on this 8% acrylamide gel for the proband co-segregated with each family member having the affected clinical phenotype and the disease-associated haplotype(filled circles/squares).

DISCUSSION

Ackerman and Porter(8) previously reported the identification of this autosomal dominant inherited LQTS family after a near-drowning of a 10-y-old white boy in a public pool while racing his younger brother(8). This index patient was defibrillated successfully from a torsades de pointes polymorphic ventricular tachyarrhythmia at the poolside. Careful evaluation of the family (history and screening ECG) suggested the inherited LQTS. Evaluations of first degree relatives revealed four additional symptomatic individuals each with an abnormal QTc (≥0.46 s1/2) and eight family members with no previous symptoms compatible with LQTS but a screening ECG with either a borderline QTc (between 0.42 and 0.46 s1/2) or an abnormal QTc. The proband's ECG pattern (Fig. 1) suggested a LQT3-sodium channel-based genotype(7).

We provide molecular genetic evidence that a novel mutation in KVLQT1 is the molecular basis for this family's inherited LQTS (LQT1, Fig. 5). First, polymorphic DNA markers for chromosome 11p15.5-linked LQT1 segregated with clinical disease status with a significant lod score. Furthermore, the 3-bp deletion responsible for theΔF339 mutant in the proband segregated completely with the clinical disease status as well. KVLQT1 is the channel pore-forming subunit that in combination with its β-auxiliary subunit (minK) provides a critical phase 3 repolarizing potassium current, IKs(1820). The novel mutation reported here adds to the already widespread mutational heterogeneity found within the KVLQT1 gene (reviewed by Ackerman in 1998)(4). The majority of previous mutations(18 of 21) have each been single nucleotide missense mutations(13,2124). The three exceptions to the missense mutations have included a 3-bp deletion causing a F167W/G168Δ mutation resulting in a severely truncated channel subunit(13); a 3-bp exon-intron boundary deletion and a splicing mutation both leading to premature termination and truncated proteins(24).

Figure 5
figure 5

Summary of near-drowning patient's family's LQTS genotype. A schematic of the cardiac action potential from a ventricular myocyte is displayed on the left with a prolonged action potential duration (APD) indicated (y axis = membrane potential in mV, x axis = time in milliseconds). The linear topology and chromosomal location of the cardiac potassium channel encoded by the KVLQT1 gene responsible for LQT1 is shown. This novel mutation,ΔF339, occurs in the final transmembrane-spanning segment (S6) adjacent to the amino acids responsible for the channel pore and near the putative A341V "hot-spot."

The mutation described here is unique to KVLQT1 mutations, an amino acid deletion mutation rather than a missense point mutation. This ΔF339 mutant is in close proximity to the putative LQT1 "hot-spot," A341V/E, accounting for nearly one-third of the LQT1 genetic perturbations in previously described LQT1 families(22,24). This suggests that the sixth transmembrane spanning segment (S6) likely contributes key functional characteristics to the adjacent channel pore. Presently, only three KVLQT1 mutations have been functionally characterized: A178P in the S2-S3 cytoplasmic loop, L273F in S5, and T312I in the channel pore(25). Thus far, the basis for action potential prolongation in LQT1 is through a dominant-negative effect with mutant subunits causing a diminution in the number of fully functional KVLQT1-based IKs channels. It should be noted that the A341V mutation is also denoted in the literature as A212V(13) and as A246V(24) with the initial amino acid positions based upon truncated clones at the N-terminus rather than the now known full-length KVLQT1 clone containing 676 amino acids (129 and 95 additional N-terminal residues respectively)(18).

This study also points to the clinical prospects and need for asymptomatic testing. Importantly, two family members (a maternal aunt III-2 and her son IV-3) were believed to have LQTS based upon their screening ECG (lead II QTc= 0.46 s1/2 for both), despite their asymptomatic clinical status. Although QTc determinations from three large LQT1 families would suggest that this degree of QT prolongation predicts having inherited LQTS with a positive predictive value exceeding 90%(26), this ΔF339 mutation was not present in either individual. Without such molecular testing, asymptomatic individuals like these will be presumed to have inherited this potentially fatal arrhythmogenic condition and may experience anxiety as well as unnecessary and expensive adjustments in their insurability. On the other hand, molecular genetic testing will be important to identify carriers of a gene mutation in those individuals who may otherwise be asymptomatic. Predictive testing for these individuals should help to provide adequate surveillance and proper treatment over time.

Finally, this LQT1 mutation raises an important genotype-phenotype inquiry. Characterization of this novel KVLQT1 mutation underscores the importance for cardiologists to resist the temptation of surmising a particular molecular diagnosis and initiating genotype-specific therapies based upon examination of a patient's ECG. A meticulous study by Moss et al.(7) offered the tantalizing prospect that careful evaluation of the ECG could allow direct diagnosis of genetically distinct forms of the inherited LQTS. In analyzing the ECG from six families with LQTS, two families each for LQT1 (n = 76), LQT2 (n = 30), and LQT3 (n = 47), they found some distinctive ECG repolarization features. Namely, LQT1 individuals had prolonged T wave durations (average = 0.262 s1/2, compared with 0.191 and 0.187 for LQT2 and LQT3, respectively). Small T waves appeared unique for the LQT2 genotype (LQT2 = 0.13 mV compared with 0.37 mV and 0.36 mV on average for LQT1 and LQT3). Individuals with the LQT3 genotype displayed essentially normal T waves that were inscribed after a prolonged ST segment reflected by a QTonset-c averaging around 0.341 s1/2 compared with 0.242(LQT1) and 0.290 (LQT2). Given these distinguishing peculiarities, the proband's ECG (Fig. 1) is consistent with the prototypic LQT3 ECG pattern despite the ΔF339 mutation arising from the KVLQT1 potassium channel-based LQT1.

Given some early promising genotype-specific therapies using mexiletine in patients with LQT3(5,6,27), it may be tempting to diagnose those with LQT3 based upon their ECG pattern and initiate gene-targeted mexiletine therapy. Although the significant molecular breakthroughs of the 1990s (Fig. 6) and their potential clinical impact is cause for tremendous excitement, this family's genetic mutation underscores the need for ongoing genotype-phenotype investigations(4,28).

Figure 6
figure 6

Summary of the cardiac channelopathies causing LQTS. This highlights the molecular basis of the inherited long QT syndromes (LQT1-5). The action potential duration of a cardiac cell is prolonged secondary to mutations in critical ion channel proteins that orchestrate the electrical activity of the heart. Displayed are the LQTS genotypes, chromosomal locations, and linear topologies for the genes encoding perturbed LQTS-causing ion channels. The circles reflect the genetic heterogeneity in this syndrome, indicating the approximate positions of the over 40 mutations identified thus far.

We conclude that a novel mutation, ΔF339, in the cardiac potassium channel, KVLQT1, is the molecular basis (LQT1) for this near drowning family's autosomally dominant inherited LQTS. The ΔF339-KVLQT1 mutation segregated with the clinical disease status as predicted phenotypically and by the disease-linked haplotype. The function of this unique mutation has not been elucidated but is believed to disrupt important channel properties such as channel conductance or gating kinetics. Importantly, this potassium channel LQT1 mutation manifests an ECG pattern previously ascribed to the sodium channel-based LQT3 genotype. Although explorations in genotype-phenotype relationships are quite appealing, identification of this novel mutation sounds a cautionary note. Namely, inspection of the LQTS patient's ECG may be insufficient for determining a patient's genotype and therefore, genotype-specific interventions based upon inspection of the ECG should not be made. Ultimately, developments to provide routine clinical molecular diagnostic testing for the inherited long QT syndrome are needed to move the field of LQTS research into the next millennium.