Latent Genetic Backgrounds and Molecular Pathogenesis in Drug-Induced Long-QT SyndromeCLINICAL PERSPECTIVE
Background— Drugs with IKr-blocking action cause secondary long-QT syndrome. Several cases have been associated with mutations of genes coding cardiac ion channels, but their frequency among patients affected by drug-induced long-QT syndrome (dLQTS) and the resultant molecular effects remain unknown.
Methods and Results— Genetic testing was carried out for long-QT syndrome–related genes in 20 subjects with dLQTS and 176 subjects with congenital long-QT syndrome (cLQTS); electrophysiological characteristics of dLQTS-associated mutations were analyzed using a heterologous expression system with Chinese hamster ovary cells together with a computer simulation model. The positive mutation rate in dLQTS was similar to cLQTS (dLQTS versus cLQTS, 8 of 20 [40%] versus 91 of 176 [52%] subjects, P=0.32). The incidence of mutations was higher in patients with torsades de pointes induced by nonantiarrhythmic drugs than by antiarrhythmic drugs (antiarrhythmic versus others, 3 of 14 [21%] versus 5 of 6 [83%] subjects, P<0.05). When reconstituted in Chinese hamster ovary cells, KCNQ1 and KCNH2 mutant channels showed complex gating defects without dominant negative effects or a relatively mild decreased current density. Drug sensitivity for mutant channels was similar to that of the wild-type channel. With the Luo-Rudy simulation model of action potentials, action potential durations of most mutant channels were between those of wild-type and cLQTS.
Conclusions— dLQTS had a similar positive mutation rate compared with cLQTS, whereas the functional changes of these mutations identified in dLQTS were mild. When IKr-blocking agents produce excessive QT prolongation (dLQTS), the underlying genetic background of the dLQTS subject should also be taken into consideration, as would be the case with cLQTS; dLQTS can be regarded as a latent form of long-QT syndrome.
Received February 29, 2008; accepted July 6, 2009.
Congenital long-QT syndrome (cLQTS) is characterized by abnormally prolonged ventricular repolarization and familial inheritance, leading to polymorphic ventricular tachycardia (torsades de pointes [TdP]), causing sudden cardiac death.1,2 In contrast, secondary long-QT syndrome can be induced by a variety of commercially available drugs, including antiarrhythmic drugs, antihistamines, antibiotics, and major tranquilizers.3 In patients with drug-induced long-QT syndrome (dLQTS), after a washout period of the culprit drugs, the QT interval usually returns to within normal range. Genetic factors may underlie the susceptibility to drug-induced serious adverse reactions such as a long QT interval and TdP. Sesti et al4 demonstrated that a polymorphism of the KCNE2 gene (T8A) is present in 1.6% of the population and is associated with drug-induced TdP related to quinidine and to sulfamethoxazole/trimethoprim administration. We have also previously reported that a mutant SCN5A channel (L1825P), found in an elderly woman with cisapride-induced TdP, appeared to have unique electrophysiological characteristics with both loss and gain of functions for the cardiac sodium current.5 In addition, there have been several case reports with long-QT syndrome–associated gene mutations in dLQTS.6–10 The accurate prevalence of long-QT syndrome–related gene mutations in a larger dLQTS cohort, however, remains unknown, and the relationship between genotypes and cellular electrophysiology has not been fully examined. The present study therefore aimed to survey mutations in long-QT syndrome–related genes responsible for dLQTS in 20 patients who had been referred to our institutes consecutively over the past 11 years and analyze the functional effects induced by these mutations.
Clinical Perspective on p 511
Blood samples of 305 subjects with long-QT syndrome, comprising 196 long-QT syndrome probands and 109 family members were referred to Kyoto University Graduate School of Medicine and Shiga University of Medical Sciences from March 1996 to January 2006 for genetic analysis. A diagnosis of dLQTS was made in those subjects who had not previously been diagnosed with long-QT syndrome and who only developed typical ECG features of QT prolongation after administration of culprit drugs. A diagnosis of cLQTS was made in those subjects with clinical phenotypes of long-QT syndrome, but without the involvement of secondary factors (eg, drugs, hypokalemia, or bradycardia). Among the subjects, 20 probands had drug-induced cardiac events (10.2% of long-QT syndrome probands). Their clinical information was collected, including family history of sudden death age 30 years or younger and long-QT syndrome members, previous syncope, ECGs, and serum electrolyte levels at the time of cardiac events. TdP was defined as either nonsustained or sustained ventricular tachycardia showing variation in the electronic polarity of the QRS complex and a “short-long-short” initiating sequence.11 Written informed consent was obtained from all subjects in accordance with the guidelines approved by our institutional review board. QT intervals were measured in lead II or V5, using the Bazett formula,12 before and after allowing sufficient time for the complete washout of drugs.
Schwartz scores13 were calculated in all probands. A Schwartz score ≧4 points indicates that long-QT syndrome is definitely present, a score of 2 or 3 points indicates there is a strong possibility that long-QT syndrome is present, whereas a score ≦1 point indicates a low probability of long-QT syndrome, respectively.13
The protocol for genetic analysis was approved by and performed under the guidelines of the Institutional Ethics Committee at Shiga University of Medical Science. Genomic DNA was isolated from peripheral white blood cells using conventional methods. Genetic screening was performed for KCNQ1, KCNH2, SCN5A, KCNE1, KCNE2, and KCNJ2, using polymerase chain reaction/single-strand conformational polymorphism (PCR-DHPLC, WAVE system, Transgenomic Inc, Omaha, Neb) analysis.14,15 For the abnormal DHPLC patterns, we determined the DNA sequences on both strands with an automated sequencer (PRISM 3130 Sequencer, Perkin Elmer, Calif).
The expression plasmids pIRES2-EGFP/KCNQ1(wild type[WT]/KCNQ1) and pRc-CMV/KCNH2 (wild-type [WT]/KCNH2) were kindly provided by Dr J. Barhanin (UMR 6097 CNRS and Université de Nice Sophia Antipolis, Valbonne, France) and Dr M. Sanguinetti (University of Utah, Salt Lake City, Utah), respectively. The mutations were introduced using overlap PCR. The mutant plasmids were constructed by substituting the 857-bp XhoI-BglII for R231C and R243H mutants, 287-bp EagI-BstEII for the D342V mutant, 794-bp BstEII-BglII for the H492Y mutant, or 392-bp BglII-SphI fragments for S706F and M756V mutants, respectively, for the corresponding fragments of WT/KCNQ1 or WT/KCNH2.
Functional potassium channels were expressed transiently in Chinese hamster ovary (CHO) cells by transfecting the same amount of α subunit plasmids (1 μg/mL KCNQ1 cDNA or 2 μg/mL KCNH2 cDNA). For the analysis of IKs currents, the same amount of pIRES/CD8-KCNE1 was coexpressed. Cells were trypsinized, diluted with Dulbecco Modified Eagle’s Medium (DMEM, Nacalai Tesque, Kyoto, Japan) supplemented with 10% fetal bovine serum, 30 U/mL penicillin, and 30 μg/mL streptomycin. The DMEM used for cell culture dishes was changed to OptiMEM (Invitrogen, Carlsbad, Calif) for transfection, and, after the addition of 10 μL lipofectamine (Invitrogen) and cDNA, the cells were incubated at 37°C for 3 hours, unless otherwise described. OptiMEM was then replaced by DMEM and the cells were subjected to electrophysiological measurements 48 to 72 hours after transfection. Cells expressing the potassium channels were selected through detection of green fluorescent protein and by decoration with anti-CD8 antibody-coated beads.
Whole-cell patch-clamp recordings were made at 37°C, using an EPC-8 patch-clamp amplifier (HEKA, Lambrecht, Germany) with pipettes filled with (in mM): 70 aspartate, 70 KOH, 40 KCl, 10 KH2PO4, 1 MgSO4, 3 Na2-ATP, 0.1 Li2-GTP, 5 HEPES, and 5 EGTA (pH 7.3 with 1N KOH), with a resistance of 2.0 to 4.0 mol/LΩ. The external superfusate contained (in mM): NaCl 140, KCl 5.4, MgCl2 0.5, CaCl2 1.8, NaH2PO4 0.33, glucose 5.5, and HEPES 5 (pH 7.4 with NaOH). Data were filtered at 2 kHz. Data acquisition was performed using PatchMaster acquisition software (HEKA). The holding potential was set at −80 mV. Current densities (pA/pF) were calculated for each cell studied, by normalizing peak tail current amplitude to cell capacitance. Current-voltage relations were fitted with the Boltzmann function: equation
where V1/2 indicates the potential at which the activation or inactivation is half-maximal, Vm the test potential, and k the slope factor.
Drug sensitivities were examined by various concentrations for erythromycin, disopyramide, and pirmenol (gift from Pfizer Inc, Groton, Conn). Depolarizing pulses were applied every 15 seconds and peak tail currents at −60 mV after +20 mV test potential were recorded in the absence or presence of various concentrations of agent. Percent inhibition was calculated by dividing the peak amplitude in the presence of drug by control. Drug concentration-inhibition relations were fitted to the Hill equation; equation
where IC50 is the amount of drug necessary to produce the half-maximal inhibition of IKr tail currents, and n is the Hill coefficient for the fit.
The dynamic Luo-Rudy model (Clancy and Rudy 2001 model) of a ventricular cell was used, with recent modifications and action potentials were simulated using a previously reported model.16 The ratio of IKr and IKs conductance of M cell layer was set at 23:7. Based on the experimental data of voltage-clamp recordings of KCNQ1 and KCNH2 channels heterologously expressed in CHO cells, we constructed Markov or Hodgkin-Huxley models for simulated mutant channels as compared with mutants associated with cLQTS (see supplemental data 1). To make the mutant channel models, we decreased the conductance of each channel as appropriate for the decreased current density and looked for adequate changes for mutant channels by changing each coefficient value, in turn, for gating states associated with impaired gating defects. The simulation for voltage-clamp experiments was calculated using the 4th-order Runge-Kutta method with a fixed-time step of 0.020 ms. The simulation programs (see supplemental data 1) were coded in C++ and implemented for personal computers.
Experimental data are expressed as mean±SE and other clinical data as mean±SD, and the statistical comparisons were made using the unpaired Student t test. Differences in the positive mutation rate between 2 groups were analyzed by χ2 and Fisher exact probability test. Statistical significance was considered as P<0.05.
Molecular Genetics of dLQTS
Table 1 summarizes the clinical characteristics of 20 subjects with dLQTS (14 women and 6 men; mean age, 65±16 years). Nineteen subjects had TdP with marked QT prolongation, and 1 had syncope without documented TdP after taking 1 of the drugs listed in Table 1. The average QTc interval before taking drugs, available for 15 subjects, was 446±29 ms. The QTc interval was significantly prolonged to 616±91 ms after taking 1 of the culprit drugs (versus QTc interval before taking drugs, P<0.001) and significantly shortened to 441±33 ms after washout of drugs (versus QTc interval after taking drugs, P<0.001). However, in 3 patients (cases 8, 13, and 18 in Table 1) a prolonged QT interval of over 480 ms was maintained after washout. The average R-R interval immediately before TdP was 1.2±0.4 seconds, and the average serum potassium level after TdP was 3.9±0.6 mEq/mL. In the majority of subjects (n=14, 70%), dLQTS was induced by antiarrhythmic drugs (disopyramide, pirmenol, cibenzoline, procainamide, and aprindine in 8, 3, 1, 1, and 1 subjects, respectively; cases 1 to 14); the remaining cases were induced by antihistamines, antibiotics, psychiatric, or miscellaneous drugs (cases 15 to 20 in Table 1). None had a family history of long-QT syndrome, whereas 3 subjects (cases 4, 14, and 20) had unexplained syncope and another subject (case 17) had a family member with sudden cardiac death. During a mean follow-up period of 52±44 months after discontinuing drugs, 1 subject without any gene mutations (case 9) had recurrent ventricular fibrillation. Compared with cLQTS, the age at first cardiac event in subjects diagnosed with dLQTS was significantly older (dLQTS versus cLQTS; 65±16 versus 19±18 years, P<0.001). No subject in the dLQTS group had a family member with long-QT syndrome (dLQTS versus cLQTS; 0% versus 24% subjects, P<0.01). The QTc interval in dLQTS subjects was significantly shorter than in cLQTS subjects (dLQTS versus cLQTS; 446±29 versus 507±71 ms, P<0.001); the Schwartz score was significantly lower in dLQTS than in cLQTS (dLQTS versus cLQTS; 0.9±1.4 versus 4.1±2.1, P<0.001).
In 8 (40%) dLQTS subjects, the genetic analysis identified 8 mutations in LQTS-related genes; 2 KCNQ1, 5 KCNH2, and 1 SCN5A (Table 1). These variants were not observed in the control subjects (220 chromosomes from non-cLQTS and non-dLQTS subjects), suggesting that they represented disease-related mutations. The 91 cLQTS patients also had gene mutations such as 33 KCNQ1, 36 KCNH2, 12 SCN5A, 2 KCNE2, and 8 compound mutations, and the positive mutation rate was similar between dLQTS and cLQTS subjects (8 of 20 [40%] versus 91 of 176 [52%] subjects, respectively, P=0.32, Figure 1A). These mutations were found in only 3 of 14 (21%) subjects with TdP induced by antiarrhythmic drugs. In contrast, 5 of 6 subjects with TdP induced by nonantiarrhythmic drugs had gene mutations (83%; versus patients with TdP induced by antiarrhythmic drugs, P<0.05) (Figure 1B). Seven of the 8 mutations were located in the nonpore regions (red circles in Figure 1C), except for the A614V-KCNH2 mutation in a case of hydroxyzine-induced TdP.17 The green circles in Figure 1C indicate the location of 8 mutations previously reported in dLQTS.5–10
Clinical Characteristics of Genotyped dLQTS Subjects
Detailed subject characteristics are presented in Table 1 and the supplemental data 2. Case 2 (M756V/KCNH2) was a 63-year-old man who was admitted with syncope after taking pirmenol (300 mg/d). Case 4 (H492Y/KCNH2) was a 52-year-old woman who had syncope while taking disopyramide (300 mg/d). She had previously had 1 episode of unexplained syncope. Case 14 (R243H/KCNQ1) was a 52-year-old woman who had repetitive syncope after taking aprindine (60 mg/d). Case 15 (S706F/KCNH2) was a 21-year-old woman who complained of sudden onset of palpitations and dyspnea after taking amphetamine and methamphetamine. An SCN5A mutation (L1825P/SCN5A) (case 16), has been reported previously5; this concerned a 70-year-old woman who had TdP and prolongation of the QT interval after taking cisapride (5 mg/d) and pirmenol (200 mg/d). Case 17 (D342V/KCNH2) was a 70-year-old woman who had repetitive syncope due to erythromycin intake (1200 mg/d). This subject had a 24-year-old sister who had died suddenly, but it is not unknown if her sister had suffered from LQTS. Case 18 (R231C/KCNQ1) was a 72-year-old woman who had presyncope while taking probucol (250 mg/d). Case 20 concerned a 34-year-old woman (A614V/KCNH2 according to Table 1) who had syncope and QT prolongation induced by hydroxyzine (3 mg/d).17 None of the 20 subjects in the study had structural heart disease or a family history of documented LQTS. In genotyped families, a genetic test for 5 members revealed 1 R243H/KCNQ1 mutation carrier. This carrier had no syncope and a normal QT interval (QTc, 400 ms). The Schwartz scores in the 8 subjects with mutations were 1.0±1.5 points (range, 0 to 4 points). Thus, among the genotyped dLQTS subjects in this current study, there was only a low or moderate possibility of LQTS being present, both before taking the drugs or after their withdrawal.
Electrophysiological Characteristics of Mutations Associated With dLQTS
The biophysical effects of the respective KCNQ1 and KCNH2 mutations were analyzed using a heterologous expression system with the CHO cell line. The upper panel of Figure 2A shows representative current traces for WT and 2 mutant KCNQ1 channels; the lower panel shows those recorded from cells cotransfected with WT and each mutant. On their own, both mutants displayed smaller currents compared with WT, whereas the R231C channel was a slightly open K+ leak channel. Under heterozygous conditions, however, they displayed currents comparable to those of WT without dominant negative effects.
At the end of the depolarizing pulse to +40 mV, current densities were smaller than those of WT channels (77.0±10.6 pA/pF) in R231C/WT (39.6±11.9 pA/pF, P<0.05 versus WT) but not in R243H/WT (61.9±10.7 pA/pF) (Figure 2B). On the other hand, the R243H channels showed a significant positive shift of steady-state activation curve (Figure 2C). Half activation voltages (V1/2) and k were −8.2±3.1 mV and 12.6±0.6 for WT, −12.8±3.6 mV and 13.6±1.3 for R231C/WT, 1.7±3.1 mV and 12.9±0.6 for R243H/WT, respectively (V1/2: WT versus R243H/WT, P<0.05). Figure 2D shows representative families of current traces (left panel) and time constants (right panel) of deactivation in each channel. The time course of deactivating kinetics could be fitted by a single-exponential function. The R243H/WT channel had faster deactivation process over −60 mV, whereas this process in the R231C/WT channel was slower than WT under −90 mV.
Figure 3A shows representative families of current traces recorded during depolarizing pulses from CHO cells transfected with KCNH2 cDNAs as indicated in the graph. The left column depicts current traces from cells transfected with each construct alone and the right column those recorded under heterozygous conditions except for the uppermost traces (WT 1 μg). When expressed alone, D342V was nonfunctional and 3 other mutations displayed functional channels. When coexpressed with WT, D342V showed weak dominant negative effects, whereas the other 3 channels showed nondominant negative effects. The functional outcome of the remaining mutation, A614V-KCNH2, 17 has recently been reported. Using an oocyte expression system, Nakajima et al18 reported that the A614V channel showed loss of function in a dominant negative manner, and the results from the present study were almost identical (18% current density of WT).
Current densities at the end of a 2-second depolarization pulse were calculated in multiple cells and plotted as a function of test potential in Figure 3B. At the end of the depolarizing pulse to +20 mV, the densities were smaller than those of WT channels (36.2±6.8 pA/pF) in D342V/WT (18.7±4.5 pA/pF, P<0.05 versus WT) and S706F/WT (21.9±2.7 pA/pF, P=0.055 versus WT). Those in H492Y/WT and M756V/WT were also smaller than in the WT, but the difference did not reach statistical significance.
The mean peak amplitudes of tail currents at −60 mV on repolarization from various test potentials (2-second duration) plotted as a function of test potentials are displayed in Figure 3C. H492Y and M756V channels displayed amplitudes of peak tail currents similar to WT. For example, the peak tail current densities after a test pulse to +20 mV were 66.2±10.5 pA/pF for WT, 62.1±13.0 pA/pF for H492Y/WT, and 58.4±7.7 pA/pF for M756V/WT. On the other hand, the current densities of D342V/WT (1.5±0.9 pA/pF, P<0.001 versus WT) and S706F/WT (38.9±7.3 pA/pF, P<0.05 versus WT) channels were significantly smaller than WT. D342V/WT channel displayed currents smaller than WT 1 μg (indicated by dotted line), whereas S706F/WT currents were similar to WT 1 μg in size. Thus, D342V channels had weakly dominant negative suppression effects on reconstituted IKr-like currents, whereas S706F had no dominant negative suppression effect. To examine the voltage dependence for activation, Boltzmann function curves were fitted to the relationship between peak tail currents and test voltages under respective conditions and are represented by solid lines in Figure 3C. Half inactivation voltages (V1/2) were −23.2±1.6 mV for WT, −19.8±3.0 mV for D342V/WT, −24.6±1.4 mV for H492Y/WT, −26.1±1.5 mV for S706F/WT, and −19.8±1.7 mV for M756V/WT, respectively (Figure 3D). Therefore, the mutations did not affect the voltage-dependent activation of the reconstituted IKr-like channel.
Whether the mutations affected the inactivation kinetics of KCNH2 channels was then assessed. Figure 4A shows the voltage dependence of availability of WT and mutant channels measured by a brief repolarization method (inset). With a 1-second depolarizing pulse, peak tail current amplitudes (arrow in the inset) after short preconditioning voltage pulses (5 ms) are plotted against the voltage of conditioning pulse. Mutant channels caused significant voltage shift of channel availability to the hyperpolarizing direction compared with WT (V1/2 of −58.3±4.7 mV for WT, V1/2 of −61.4±6.5 mV for D342V/WT, −77.8±4.7 mV for H492Y/WT, −70.1±3.2 mV for S706F/WT, and −71.1±4.6 mV for M756V/WT, respectively). The time course to recovery from or the development of the inactivation (recovery from inactivation at hyperpolarized potentials and development of inactivation at >−70 mV) was analyzed by double or triple pulse protocols. The time course of inactivating kinetics could be fitted by a single-exponential function. Time constants thus calculated were significantly smaller than WT and mutant channels over a wide range of voltage (between −50 and +40 mV for D342V/WT and H492Y/WT; between −30 and +20 mV for S706F/WT; between −40 and 0 mV for M756V/WT, Figure 4B). These results demonstrate that drug-induced LQTS mutants have accelerated inactivation kinetics. Figure 4C shows representative families of current traces (left panel) and time constants (right panel) of deactivation in each channel. When the time course of deactivating kinetics was fitted by a double-exponential function, the S706F/WT channel slightly accelerated the deactivation process (Figure 4C).
Most of the drugs that induced LQTS in the study subjects have been known to block IKr in a concentration-dependent manner,8,17,19–22 whereas IKr channel with KCNH2 mutations may have different drug sensitivities compared with the WT channel. Figure 5 shows 3 sets of drug concentration-current inhibition relationships associated with KCNH2 mutations with respect to erythromycin (5A), disopyramide (5B), and pirmenol (5C). In each case of drug-induced LQTS, an electrophysiological assay of current inhibition by the respective culprit drug was performed, for example, hydroxyzine,17 erythromycin, disopyramide, and pirmenol. IC50s were not significantly different between WT and the respective mutant channels, suggesting that a change in drug sensitivity was not involved in causing the drug-induced TdP in the study subjects.
Computer Simulation of Ventricular Action Potentials
To compare how the functional changes caused by mutants affect ventricular action potentials, a simulation study was conducted using the Luo-Rudy computer model that incorporated the Markov16 or Hodgkin-Huxley23 process gating for the mutant channels (Figure 6A). Table 2 shows parameters of simulation that have been changed to fit to experimental results. First, the IKr or IKs conductance was reduced to the level observed in the D342V/WT, A614V/WT, and R231C/WT channels. The deactivation time course for R231C/WT was also fitted by modifying a parameter. Second, the transition rate was changed accordingly from inactivation to open states (αi) and fitted with the experimental data seen in H492Y/WT and M756V/WT channels. Acceleration of inactivation induced by modifying the transition rate αi reproduced smaller amplitudes for I-V relationships and negative shift of steady-state inactivation curve (Figure 4A); these findings were compatible with the experimental results (Figure 3, A through C). Third, some parameters were altered to simulate the S706F/WT model, by reducing the IKr conductance, modifying αi, and increasing the transition rate from open to deactivation states. This was followed by change to parameters associated with the activation and deactivation rates for the R243H/WT model. Finally, not only was INa conductance decreased in the subjects, but the burst mode was also added to simulate the sustained current for the L1825P model.24 The L1825P/WT channel was heterogeneously simulated to equally mix WT and L1825P models.
In the simulated M cells using the Luo-Rudy model, the order of increase in magnitude of action potential duration (APD) was A614V/WT>R231C/WT>D342V/WT>S706S/WT>R243H/WT=L1825P/WT >H492Y/WT>M756V/WT for dLQTS mutations (Figure 6B, middle panel). Typically, for cLQTS mutations, the simulated APD was longer than for WT or drug-induced models, whereas APDs in drug-induced models were intermediate between those in WT and those in cLQTS (clinical information for simulated cLQTS are presented in supplemental Table 1).
Finally, an effort was made to reproduce action potentials in the presence of IKr-blocking drugs (Figure 6C). Early afterdepolarizations appeared in all mutants where there were smaller reductions in the IKr conductance compared with the corresponding WT. Because drug sensitivities for WT and mutant channels were not different (Figure 5), the same inhibition rate was used in both the WT and mutant models and the IKr conductance was gradually reduced at the cycle length of 1200 ms. As shown in Figure 6C, typically, when the IKr conductance was theoretically decreased to 89% of the basal conductance for each channel, the D342V/WT model began to develop early afterdepolarizations, whereas the WT model produced only a 2.9% increase in APD.
There were 3 major findings.1 In 8 of 20 consecutive dLQTS subjects, 5 KCNH2, 2 KCNQ1, and 1 SCN5A heterozygous missense mutations were identified; there was a similar positive mutation rate with dLQTS compared with cLQTS.2 Both KCNQ1 and KCNH2 mutants possessed loss of function effects on reconstituted IKs- or IKr-like channels.3 The functional changes in mutant channels reconstituted by the computer simulation resulted in a mildly prolonged APD, suggesting that the dLQTS may partially have a genetic background, especially mild or latent long-QT syndrome–associated mutations.
Mutations in dLQTS
Potential torsadegenic drugs are used in the clinical setting and include antiarrhythmic drugs, antibiotics, antihistamines, psychiatric drugs, and cholinergic antagonists. These might induce TdP, which may lead to the sudden cardiac death of individuals whose QT intervals were within normal range before taking the drug. Several drugs such as cisapride and terfenadine have been withdrawn from the market because of these possible side effects.25,26 The incidence of dLQTS is not high, and the drugs lead to TdP in only a small percentage of individuals, suggesting that there may be an underlying genetic background that predisposes these individuals to the risk.
In the medical literature, 13 KCNQ1 or KCNH2 mutations (13 of 15 mutations, 87%) associated with dLQTS, including 6 of the mutations identified in this present study, have been located in nonpore regions (Figure 1B). Mutations in nontransmembrane regions have been shown to cause either mild long-QT syndrome or benign clinical phenotypes.27 Mutation sites may influence the clinical and basic electrophysiological characteristics of patients with dLQTS. It is of interest that the functional assay of our 7 mutations resulted in various levels of loss of function, but most of them showed no dominant negative suppression, which is usually observed in the classic cLQTS. Clinical characteristics during the drug intake were not significantly distinct from those of cLQTS, which shares a similar genetic background with the dLQTS. Several polymorphisms have been shown to be associated with dLQTS.4,28 Abbott et al4 identified a polymorphism (T8A) of the KCNE2 gene encoding MiRP, a β-subunit for the IKr channel, which is present in 1.6% of the population and is associated with TdP induced by quinidine or sulfamethoxazole/trimethoprim administration. Splawski et al28 also found a heterozygous polymorphism involving substitution of serine with tyrosine (S1102Y) in the sodium channel gene SCN5A among blacks that increased the risk for drug-induced TdP. The polymorphism was present in 57% of 23 patients with proarrhythmic episodes but in only 13% of control subjects. These findings suggested that common genetic variations may increase the risk for development of drug-related arrhythmias.
Roden et al29 reported that cisapride could rescue trafficking of L1825P channel with a potentially sustained current and revealed a new mechanism of dLQTS with the Luo-Rudy model. However, in the genotyped subjects in the present study, dLQTS mainly resulted from the IKr-blocking effect of culprit drugs in the presence of latent genetic backgrounds. Cardiac repolarization reserve may protect subjects against the drug-induced QT prolongation by IKr-blocking drugs.30,31 In the presence of latent genetic backgrounds, however, reduction in the repolarization reserve unveils the presence of so-called “concealed” long-QT syndrome when drugs with IKr-blocking effects are administered. The presence of borderline prolongation of the QT interval, together with personal information such as unexplained previous syncope and family history of premature sudden death, may help to prevent drug-induced arrhythmia even if a subject’s Schwartz score is low, because they could have a potential risk of TdP. Special attention should be paid to family members of the index subject with drug-induced QT prolongation because ≈30% of family members were found to have a predisposing genetic background in the present study. Indeed, they may have inherited the risk for being susceptible to dLQTS.
Limitations of the Study
This study has some limitations. Bacause of the small cohort of dLQTS subjects, this study was not powered to quantify the overall prevalence of ion channel mutations in the group of subjects with drug-induced TdP; there was also a possible selection bias in the population. As for the causative agents, we were unable to test the action of amphetamine and methamphetamine because it was impossible to obtain these illegal drugs for clinical study. However, a previous report has shown that 3,4-methylenedioxy methamphetamine (ecstasy, NMDA) prolongs the APD of hippocampal neurons by blocking the conductance of a resting K+ channel.32 It is quite possible that these drugs also suppressed cardiac K+ currents and induced QT prolongation and TdP in 1 subject in our study, based on her medical records. Regarding protein trafficking, 2 mutations in this study, A614V in KCNH217 and L1825P in SCN5A,29 had been reported to be trafficking-deficient mutations. Though protein trafficking of other mutants remains unclear, especially R243H in KCNQ1, and H492Y, S706F, and M756V in KCNH2, would be not trafficking-deficient because, under heterozygous conditions, these mutants showed adequate current density compared with WT. In the simulation study, the parameter settings could mimic mutant channels. In addition, the setting of the parameters might have innumerable patterns, and it therefore remains possible that other combinations of patterns could also simulate mutant channels.
We thank Arisa Ikeda for excellent technical assistance.
Sources of Funding
This work was supported by the Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science and the Biosimulation and Health Sciences Research Grant (H18-Research on human Genome-002) from the Ministry of Health, Labor, and Welfare of Japan (M.H.) and a Grant-in-Aid for Young Scientists from the Ministry of Education, Culture and Technology of Japan (H.I.).
Moss AJ, Schwartz PJ, Crampton RS, Tzivoni D, Locati EH, MacCluer J, Hall WJ, Weitkamp L, Vincent GM, Garson A Jr. The long-QT syndrome: prospective longitudinal study of 328 families. Circulation. 1991; 84: 1136–1144.
Haverkamp W, Breithardt G, Camm AJ, Janse MJ, Rosen MR, Antzelevitch C, Escande D, Franz M, Malik M, Moss A, Shah R. The potential for QT prolongation and proarrhythmia by non-antiarrhythmic drugs: clinical and regulatory implications: report on a policy conference of the European Society of Cardiology. Eur Heart J. 2000; 21: 1216–1231.
Sesti F, Abbott GW, Wei J, Murray KT, Saksena S, Schwartz PJ, Priori SG, Roden DM, George AL Jr, Goldstein SA. A common polymorphism associated with antibiotic-induced cardiac arrhythmia. Proc Natl Acad Sci U S A. 2000; 97: 10613–10618.
Makita N, Horie M, Nakamura T, Ai T, Sasaki K, Yokoi H, Sakurai M, Sakuma I, Otani H, Sawa H, Kitabatake A. Drug-induced long-QT syndrome associated with a subclinical SCN5A mutation. Circulation. 2002; 106: 1269–1274.
Yang P, Kanki H, Drolet B, Yang T, Wei J, Viswanathan PC, Hohnloser SH, Shimizu W, Schwartz PJ, Stanton M, Murray KT, Norris K, George AL, Roden DM. Allelic variants in long-QT disease genes in patients with drug-associated torsades de pointes. Circulation. 2002; 105: 1943–1948.
Bellocq C, Wilders R, Schott JJ, Louérat-Oriou B, Boisseau P, Le Marec H, Escande D, Baró I. A common antitussive drug, clobutinol, precipitates the long QT syndrome. Mol Pharmacol. 2004; 66: 1093–1102.
Haverkamp W, Shenasa M, Borggrefe M, Breithardt G. Cardiac Electrophysiology. 4th ed. Philadelphia: Saunders; 1999: 886–899.
Bazett HC. An analysis of the time relations of electrocardiograms. Heart. 1920; 7: 353–370.
Schwartz PJ, Moss AJ, Vincent GM, Crampton RS. Diagnostic criteria for the long QT syndrome: an update. Circulation. 1993; 88: 782–784.
Clancy CE, Rudy Y. Cellular consequences of HERG mutations in the long QT syndrome: precursors to sudden cardiac death. Cardiovasc Res. 2001; 50: 301–313.
Sakaguchi T, Itoh H, Ding WG, Tsuji K, Nagaoka I, Oka Y, Ashihara T, Ito M, Yumoto Y, Kubota T, Zenda N, Higashi Y, Takeyama Y, Matsuura H, Horie M. Hydroxyzine, an H1 receptor antagonist of first generation, inhibits reconstituted IKr currents. J Pharmacol Sci. 2008; 108: 462–471.
Nakajima T, Furukawa T, Tanaka T, Katayama Y, Nagai R, Nakamurea Y, Hiraoka M. Novel mechanism of HERG current suppression in LQT2: shift in voltage dependence of HERG inactivation. Circ Res. 1998; 83: 415–422.
Daleau P, Lessard E, Groleau MF, Turgeon J. Erythromycin blocks the rapid component of the delayed rectifier potassium current and lengthens repolarization of guinea pig ventricular myocytes. Circulation. 1995; 91: 3010–3016.
Clancy CE, Rudy Y. Na+ channel mutation that causes both Brugada and long-QT syndrome phenotypes: a simulation study of mechanism. Circulation. 2002; 105: 1208–1213.
Donger C, Denjoy I, Berthet M, Neyroud N, Cruaud C, Bennaceur M, Chivoret G, Schwartz K, Coumel P, Guicheney P. KVLQT1 C-terminal missense mutation causes a forme fruste long-QT syndrome. Circulation. 1997; 96: 2778–2781.
Splawski I, Timothy KW, Tateyama M, Clancy CE, Malhotra A, Beggs AH, Cappuccio FP, Sagnella GA, Kass RS, Keating MT. Variants of SCN5A sodium channel implicated in risk of cardiac arrhythmia. Science. 2002; 297: 1333–1336.
Liu K, Yang T, Viswanathan PC, Roden DM. New mechanism contributing to drug-induced arrhythmia: rescue of a misprocessed LQT3 mutant. Circulation. 2005; 112: 3239–3246.
Jost N, Virág L, Bitay M, Takács J, Lengyel C, Biliczki P, Nagy Z, Bogáts G, Lathrop DA, Papp JG, Varró A. Restricting excessive cardiac action potential and QT prolongation: a vital role for IKs in human ventricular muscle. Circulation. 2005; 112: 1392–1399.
Roden DM, Yang T. Protecting the heart against arrhythmias: potassium current physiology and repolarization reserve. Circulation. 2005; 112: 1376–1378.
Premkumar L, Ahern G. Blockade of a resting potassium channel and modulation of synaptic transmission by ecstasy in the hippocampus. J Pharmacol Exp Ther. 1995; 274: 718–722.
Drug-induced long-QT syndrome (dLQTS) is a disease associated with the appearance of a prolonged QT interval and torsades de pointes after taking a culprit drug or drugs; the QT interval usually returns to within normal range after a washout period of these drugs. The clinical phenotype of dLQTS that appears during administration of these drugs resembles that of congenital long-QT syndrome (cLQTS), and “latent” genetic factors may underlie the susceptibility of a subject to drug-induced serious adverse reactions, such as a long QT interval and torsades de pointes. In the analysis of cLQTS-associated genes encoding cardiac ion channel-composing proteins, this study revealed that dLQTS had a similar positive mutation rate compared with cLQTS. When reconstituted in Chinese hamster ovary cells, KCNQ1 and KCNH2 mutant channels showed complex gating defects without dominant negative effects or a relatively mild decreased current density. With the Luo-Rudy simulation model of action potentials, action potential durations of most mutant channels were between those of wild-type and cLQTS. In conclusion, although the dLQTS subjects had genetic backgrounds that were similar to cLQTS subjects, the functional changes associated with these mutations identified in dLQTS were different from those in cLQTS. Thus, we believe that dLQTS can be regarded as a latent form of long-QT syndrome. When IKr-blocking agents produce excessive QT prolongation, the underlying genetic background of the dLQTS subject should be taken into consideration, as would be the case with cLQTS.
The online-only Data Supplement is available at http://circep.ahajournals.org/cgi/content/full/CIRCEP.109.862649/DC1.