Long QT 1 Mutation KCNQ1^A344VIncreases Local Anesthetic Sensitivity of the Slowly Activating Delayed Rectifier Potassium Current

The mutant KCNQ1 A344V was created by site directed mutagenesis from human KCNQ1. All constructs were cloned in the pcDNA3 expression vector for expression in Chinese hamster ovary (CHO) cells and in the pGEM expression vector for complementary RNA (cRNA) synthesis. The cRNA was synthesized in vitro  with the T7 mMESSAGE mMACHINE Kit (Ambion, Austin, TX) according to the manufacturer's protocol. The concentration was determined with the RiboGreen method (RiboGreen RNA Quantification Reagent; Molecular Probes, Eugene, OR). The total amount of cRNA was 2 ng per oocyte for injection in Xenopus laevis  oocytes. The cRNA of KCNQ1 and KCNE1 were injected in a 1:1 ratio. Two-electrode voltage clamp experiments were performed 2–7 days after the cRNA injection. Preparation of Xenopus  oocytes was performed as described previously.25Oocytes were incubated at 17°C in gentamicin containing oocyte Ringer's solution (75 mm NaCl, 2 mm KCl, 2 mm CaCl², 1 mm MgCl², 5 mm Na-pyruvate, 5 mm HEPES, pH adjusted to 7.5 with NaOH, 50 μg/ml gentamicin; all from Sigma-Aldrich, Taufkirchen, Germany). Cell culture of CHO cells was performed as described previously.14 

Electrophysiologic experiments were performed as described previously.14,25Different pulse protocols were used for characterization of the channels and to establish their pharmacologic sensitivities. The holding potential was −100 mV for Xenopus  oocytes and −80 mV for CHO cells. To analyze the activation, depolarizing pulses were applied from −80 to +100 mV in 10-mV steps. The duration of the depolarization was 3 s for KCNQ1 alone and 5 s for KCNQ1/KCNE1; tail currents were recorded at −60 mV. For pharmacologic experiments, single square pulses to +60 or +80 mV for 3 and 5 s were used. Repetitive pulses were applied to determine that steady state inhibition was reached. For whole cell recordings in CHO cells, series resistance was 2.5–5.0 MΩ and was actively compensated for by at least 85%. A leak subtraction protocol was used. The recorded signal was filtered at 2 kHz and stored with a sampling rate of 5 kHz for analysis.

Bupivacaine (Sigma-Aldrich, Taufkirchen, Germany), S  (−)-ropivacaine, (AstraZeneca, Södertalje, Sweden), and mepivacaine (Sigma-Aldrich) were dissolved in the respective extracellular recording solution. Terfenadine (Sigma-Aldrich) was prepared as 5 mm stock in dimethyl sulfoxide and diluted to 1 μm in extracellular recording solution. A hydrostatically driven perfusion system was used to apply the drugs onto the cells and to exchange the extracellular solutions. All experiments were performed at room temperature.

Data were analyzed with Pulse Fit software (HEKA Elektronik, Lambrecht, Germany) and with KaleidaGraph software (Synergy Software, Reading, PA). The normalized tail currents during the activation protocol were fitted by a Boltzmann equation: I = I^max/[1 + exp((V^0.5− V^m)/k)], where V^0.5is the voltage of half-maximal activation, V^mis the membrane potential, and k is the slope factor. The inhibition of currents by local anesthetics was quantified by the reduction of the maximal current during the test pulse: f = 1 − (I^{max, drug}/I^{max, control}). Concentration–response curves were fitted by a Hill function: f = 1/[1 + (IC⁵⁰/c)^h], where IC⁵⁰is the concentration of half-maximal inhibition, c is the concentration of the local anesthetic, and h is the Hill coefficient. Statistical significance was tested using a two-sided Student t  test or F test to compare concentration–response curves26(Excel; Microsoft, Redmond, WA). P  values of 0.05 or less were regarded as significantly different. Data are presented as mean ± SD unless stated otherwise; n values indicate the number of experiments.

Computer simulations were conducted using a modified version of the Luo-Rudy dynamic (LRd) model (C++ code)27,28downloaded from the Internet.#The program was translated into C#, compiled with Visual Studio 2003, Net Framework 1.1 (Microsoft), and executed on personal computers. The program was used to model action potentials for epicardial, endocardial, and midmyocardial layers of the heart.29The effect of the mutation KCNQ1^A344Von the cardiac action potential was simulated as a 15-mV shift in the voltage dependence of I^Ksactivation. Parameters for I^Ksactivation were altered to fit the experimental data. The influence of bupivacaine on the cardiac action potential was simulated by inhibition of sodium channels,30L-type calcium channels,31and repolarizing potassium channels14,32present in human heart by a clinically relevant concentration of 3 μm (inhibition of I^Na: 55%, I^Kr: 15%; I^L-Ca: 5%; I^towas not included in the LRd model) and by a toxicologically relevant concentration of 30 μm bupivacaine33(inhibition of I^Na: 90%, I^Kr: 60%, I^L-Ca: 20%). Effects of the mutation A344V were simulated as an additional 15-mV shift in I^Ksactivation and inhibition of I^Ks(5% for 3 μm, 10% for 30 μm). Action potential duration (APD) was automatically calculated as time difference between begin of depolarization and 90% repolarization. Numerical results were written in ASCII formatted text files and visualized with KaleidaGraph software.

KCNQ1 wild-type and mutant channels were heterologously expressed in Xenopus laevis  oocytes. Expression of KCNQ1^wtresulted in a slowly activating, noninactivating voltage-dependent current. Expression of KCNQ1^A344Vyielded functional channels displaying a voltage-dependent inactivation of the macroscopic current not seen in wild-type channels (figs. 1A and B, upper panel). Coexpression of the pore-forming subunit together with the accessory subunit KCNE1 slowed activation of the currents and increased current amplitudes for both wild-type and mutant. Notably, the macroscopic voltage-dependent inactivation of KCNQ1^A344Vwas abolished by coexpression with KCNE1 (figs. 1A and B, lower panel). The voltage dependence of activation was marginally changed by the mutation when KCNQ1 was expressed alone (fig. 1Cand table 1). Coexpression with KCNE1 caused a shift in the voltage dependence of activation compared with KCNQ1 without KCNE1 by +53 mV for KCNQ1^wt/KCNE1 and by +80 mV for KCNQ1^A344V/KCNE1, resulting in a 30-mV difference between wild-type and mutant channel complexes (fig. 1Cand table 1). However, the maximal current amplitudes were not significantly different between KCNQ1^wtand KCNQ1^A344Vor between KCNQ1^wt/KCNE1 and KCNQ1^A344V/KCNE1 (P > 0.05; table 1).

The pharmacologic effects of different amino-amide local anesthetics on KCNQ1^wtand KCNQ1^A344Valone as well as on complexes formed by KCNQ1^wtor KCNQ1^A344Vand KCNE1 were analyzed in Xenopus  oocytes. Because of the differences in activation behavior, different protocols were used for the pharmacologic experiments. Channels were activated by 3-s pulses to +60 mV for KCNQ1^wtand KCNQ1^A344V, 5-s pulses to +60 mV for KCNQ1^wt/KCNE1, and 5-s pulses to +80 mV for KCNQ1^A344V/KCNE1. Both wild-type and mutant channels were inhibited by bupivacaine in a concentration-dependent manner (figs. 2A and B). The concentration–response data were mathematically described by Hill functions (fig. 2Cand table 2). The mutant channel KCNQ1^A344Vwas 17-fold more sensitive to the inhibition by bupivacaine than KCNQ1^wt. Coexpression with KCNE1 reduced the sensitivity by a factor of 0.4 for wild-type and by a factor of 0.6 for mutant channels. The ratio of sensitivity between wild-type and mutant channels (22-fold increase by the mutation A344V) was only slightly changed. Bupivacaine (100 μm) did not significantly change the voltage dependence of activation either for KCNQ1^A344V(V^0.5=−26.4 ± 3.2 mV, n = 7; P > 0.05) or for KCNQ1^A344V/KCNE1 (V^0.5=+54.8 ± 5.0 mV, n = 5; P > 0.05). The inhibition of KCNQ1^A344Vchannels or of KCNQ1^A344V/KCNE1 channel complexes was not voltage dependent for depolarizations to potentials between −40 and +100 mV.

To establish whether the lipophilicity of amino-amide local anesthetics influences the pharmacologic potency on KCNQ1 inhibition, the effects of ropivacaine and mepivacaine were analyzed next. Complete concentration–response curves were determined for inhibition of KCNQ1^A344Vand for inhibition of KCNQ1^A344V/KCNE1 (fig. 2Dand table 3). The rank order of inhibitory potency was bupivacaine > ropivacaine > mepivacaine and thus correlated with the length of the N-substituent (fig. 2E). Inhibition of KCNQ1^wtand KCNQ1^wt/KCNE1 by ropivacaine and mepivacaine was analyzed at a concentration of 1 mm (fig. 2F). The rank order of inhibitory potency was the same for KCNQ1^wtand KCNQ1^wt/KCNE1 and KCNQ1^A344Vand KCNQ1^A344V/KCNE1, whereas the sensitivity was significantly increased by the mutation A344V for all three local anesthetics.

The effect of the mutation A344V on the biophysical and pharmacologic properties of KCNQ1/KCNE1 channel complexes was independent of the expression system because we obtained the same results in mammalian CHO cells (tables 1 and 2and fig. 3). Although the absolute values for the half-maximal activation differ between CHO cells and Xenopus  oocytes, the relative difference between KCNQ1^wt/KCNE1 and KCNQ1^A344V/KCNE1 is similar. Because of the small currents of KCNQ1 channels without KCNE1 in CHO cells, pharmacologic experiments in CHO cells were only performed with KCNQ1/KCNE1 channel complexes. The sensitivity of KCNQ1/KCNE1 channels in CHO cells was increased by a factor of 3.6 for wild-type and by a factor of 2.9 for mutant channel complexes when compared with Xenopus  oocytes. However, the increase in bupivacaine sensitivity of KCNQ1^A344V/KCNE1 compared with KCNQ1^wt/KCNE1 was similar (18-fold in CHO cells, 22-fold in Xenopus  oocytes).

Simulating the heterozygous state in a patient, KCNQ1^wtand KCNQ1^A344Vwere coexpressed in Xenopus  oocytes in a 1:1 ratio. Figures 4A and Bshow the current response of KCNQ1^wt/KCNQ1^A344Vexpressed without and with KCNE1. In the absence of KCNE1, the channels displayed voltage-dependent inactivation, similar to KCNQ1^A344Vwhen expressed alone. The voltage of half maximal activation of KCNQ1^wt/KCNQ1^A344Vwas not significantly different from the half-maximal activation of KCNQ1^wt(fig. 4C; V^0.5=−26.8 ± 4.0 mV, slope = 10.9 ± 1.2 mV, n = 5; P > 0.05). Coexpression of KCNQ1^wt/KCNQ1^A344Vwith KCNE1 caused a shift of the voltage dependence of activation to more depolarizing potentials (fig. 4C). The V^0.5value was 38.6 ± 3.9 mV (slope = 15.2 ± 1.7 mV, n = 5) which was between the V^0.5values of KCNQ1^wt/KCNE1 (P < 0.05) and KCNQ1^A344V/KCNE1 (P < 0.05).

The pharmacologic properties of heteromeric KCNQ1^wt/KCNQ1^A344Vchannels were analyzed at a concentration of 100 μm bupivacaine (fig. 4D). The sensitivity of the channel to the inhibitory action of bupivacaine was also increased in the heteromeric channel compared with the wild-type. The action of 100 μm bupivacaine on the different channel complexes was compared (fig. 4E). The heteromeric channel was less sensitive than the mutant channel KCNQ1^A344Vfor bupivacaine (P < 0.05) but significantly more sensitive than the wild-type channel, regardless of coexpression with KCNE1. The difference between expression without and with KCNE1 was not significant (P > 0.05). Similar to inhibition of KCNQ1^A344V, the block of KCNQ1^wt/KCNQ1^A344Vwas not voltage dependent, and the half-maximal activation was not changed by application of 100 μm bupivacaine.

To establish whether the increased sensitivity of KCNQ1^A344Vwas specific for amino-amide local anesthetics, we investigated the effect of terfenadine on wild-type and mutant KCNQ1 channels. Similar to local anesthetics, terfenadine preferentially blocks HERG channels.34The effect of 1 μm terfenadine, a value close to the IC⁵⁰for the inhibition of HERG potassium channels,34was examined on HERG channels and on KCNQ1 wild-type and mutant channels expressed with and without KCNE1 (fig. 5). In agreement with literature,34terfenadine (1 μm) inhibited HERG channels by 54 ± 11% (n = 6) and only marginally affected KCNQ1^wtchannels (inhibition: 4.5 ± 1.7%, n = 5). The sensitivity of KCNQ1 channels for terfenadine was significantly increased by the mutation A344V in both the homomeric and the heteromeric situations (inhibition of KCNQ1^A344V: 17 ± 6.7%, n = 5, P < 0.05; of KCNQ1^wt/KCNQ1^A344V: 7.0 ± 1.5%, n = 5, P < 0.05). However, the inhibition of KCNQ1 wild-type and mutant channels by 1 μm terfenadine was significantly reduced when they were coexpressed with KCNE1. As a result, there was no significant difference between inhibition of wild-type and mutant or heteromeric channels when they were coexpressed with KCNE1.

We used the LRd model for the guinea pig cardiac action potential27,28to simulate the effect of the mutation on cardiac action potential morphology (fig. 6). The simulation was performed with a basic cycle length of 1,000 ms for 500 cycles for epicardial, endocardial, and midmyocardial layers of the heart. According to the LRd model, the depolarizing shift of 15 mV in the voltage dependence of I^Ksactivation as present in the heterozygous situation resulted in a prolongation of the action potential (fig. 6A). This prolongation was larger in the endocardial layer (41 ms or 19%) than in the epicardial layer (0.2 ms or 0.1%) and largest in the midmyocardial layer. Furthermore, in midmyocardial myocytes, repolarization did not occur within one stimulation interval, and early afterdepolarizations (EADs) developed.

The effects of 3 μm and 30 μm bupivacaine on the cardiac action potential simulated for the homozygous and heterozygous conditions are demonstrated in figures 6B and C(see Materials and Methods for values of inhibition). In wild-type conditions (fig. 6B), 3 μm bupivacaine induced a prolongation of the cardiac action potential only in the midmyocardial layer of the heart by 16%. At 30 μm, bupivacaine prolonged the action potential in the epicardial and endocardial layers by 8% and 12%, respectively. In addition, the higher concentration of bupivacaine caused the development of EADs in the midmyocardial layer. Simulation of the effect of 3 μm bupivacaine for the heterozygous condition (fig. 6C) demonstrated that the local anesthetic prolonged APD in the epicardial layer by 23% and further prolonged APD in the endocardial layer by 10%. The low concentration of bupivacaine furthermore induced EADs and a loss of repolarization in the midmyocardial layer of the myocardium. At 30 μm, bupivacaine prolonged the APD in the epicardial layer by 56% and caused EADs in the midmyocardial and endocardial layers with a loss of repolarization in the midmyocardial layer.

In this work, we describe biophysical and pharmacologic properties of the LQT1 mutant KCNQ1^A344V.21The electrophysiologic features of this mutation in the lower part of the S6 helix are a depolarizing shift of the voltage dependence of activation and induction of voltage-dependent inactivation.

A positive shift in the voltage dependence of activation has been reported for LQT1 mutations in the S4 domain as well as for LQT1 mutations in the C-terminal region of KCNQ1.19,35Similar to our findings, the shift is only observed when the KCNQ1 mutants are coexpressed with KCNE1. It is reasonable to assume that mutations in the voltage sensor containing S4 domain may change the voltage dependence of activation by directly affecting the positive charge of the voltage sensor. However, there are several possible explanations for the effect of a mutation in the S6 region on activation gating. First, the coupling between the activation gate and the voltage sensor may be impaired. Mutagenesis studies in voltage-gated potassium channels demonstrate that the lower part of the S6 region couples with the S4–S5 linker and that mutations in this region affect the voltage dependence of gating.36,37These results were substantiated by the recently resolved crystal structure of Kv1.2 revealing a close vicinity of the S4–S5 linker and the S6 helix.38Second, the shift in the voltage dependence of activation may also be compatible with a destabilization of the open state relative to the closed state due to profound structural changes in the pore region. This destabilized channel would require higher activation energy and therefore a more positive membrane potential to open.39 

The mutant channels KCNQ1^A344Vexhibit macroscopic inactivation that is abolished by coexpression with KCNE1. A similar voltage-dependent inactivation in the absence of KCNE1 is caused by other mutations in both the S6 and S5 regions of KCNQ1.40,41The voltage-dependent inactivation of KCNQ1 channels has been suggested to result from a constriction of the pore, mediated by interaction of the residues F340 and L273 in the S6 and S5 helix with V310 in the pore helix of the channels.42A change in the size of these residues leads to an enhanced inactivation behavior due to a destabilization of the open state of the pore.42The residue A344 is a single turn below F340 in the S6 helix, and an increased side chain at position 344 may disrupt the interactions of the S6 helix with the pore as well. Because KCNE1 eliminates delayed inactivation of KCNQ1^wtchannels by stabilizing the pore by an indirect mechanism,43KCNE1 may prevent inactivation in KCNQ1^A344Vchannels by the same mechanism. Our results indicate that the interaction of KCNE1 with KCNQ1 is not disturbed by the LQT1 mutation A344V because the mutant channels display the biophysical effects attributed to coassembly of KCNQ1 with KCNE1, such as a slower activation, a positive shift in the activation midpoint, and abolishing of inactivation.15,43Taken together, our results allow us to hypothesize that the effects of the mutation A344V on channel gating may result from a destabilization of the open state of the pore.

KCNQ1 channels as well as KCNQ1/KCNE1 channel complexes are insensitive to many pharmacologic agents that block HERG channels, such as local anesthetics.16The LQT1 mutation A344V not only alters the gating behavior of KCNQ1 channels and KCNQ1/KCNE1 channel complexes, but it also changes their pharmacologic sensitivity to local anesthetics. The single amino acid exchange in the S6 region of KCNQ1 leads to a 22-fold increase in the sensitivity of KCNQ1/KCNE1 channel complexes to bupivacaine, rendering KCNQ1^A344V/KCNE1 channel complexes nearly as sensitive as HERG channels.14,32This effect did not depend on the expression system because similar results were obtained in CHO cells and in Xenopus  oocytes. Both wild-type and mutant channels were approximately three times less sensitive to bupivacaine when expressed in Xenopus  oocytes than in CHO cells. A similar reduction in sensitivity caused by the expression system has previously been described for inhibition of HERG channels by bupivacaine as well.14,16,32Although we cannot exclude that the increased sensitivity of KCNQ1^A344Vis related to the changes in the gating behavior, this seems less likely for several reasons. First, local anesthetic sensitivity of KCNQ1 and of KCNQ1/KCNE1 channels is increased to a similar extend by the mutation, although the effects of the mutation A344V on the gating are qualitatively different in KCNQ1 channels and in KCNQ1/KCNE1 channel complexes. Furthermore, the residue A344 mediates binding of benzodiazepines to KCNQ1 channels,40and it has been described as a potential site of interaction in a pharmacophore model of KCNQ1 blockers.44Finally, a direct structural effect of the mutation A344V is strongly supported by the experiments with heterologous expression of KCNQ1^wtand KCNQ1^A344Vchannels in a 1:1 ratio. The increased sensitivity of the mutant channels seems to be a dominant effect, because there was no significant difference between the inhibition of KCNQ1^A344V/KCNE1 and KCNQ1^wt/KCNQ1^A344V/KCNE1 complexes by 100 μm bupivacaine (fig. 4E). In contrast, the changes in the gating seem not to be dominant. The V^0.5value of the KCNQ1^wt/KCNQ1^A344V/KCNE1 complex is between the value for KCNQ1^wt/KCNE1 and KCNQ1^A344V/KCNE1 (fig. 4C). The increased sensitivity is thus rather caused by structural changes in the pore than by changes in channel gating.

Hydrophobic interactions between the drug molecule and the channel pore seem to be important for inhibition of KCNQ1^A344Vand KCNQ1^A344V/KCNE1 by local anesthetics because plotting the log IC⁵⁰against the lipophilicity of the local anesthetics yields a linear correlation with a correlation coefficient close to unity (fig. 2E). Hence, it may be hypothesized that the mutation A344V generates a hydrophobic interaction site. In addition or as an alternative, the mutation A344V might change the pore structure in such a way that the hydrophobic residues F340 and I33744become better accessible for local anesthetics. The observation that the mutation did not alter the sensitivity of KCNQ1/KCNE1 channel complexes to terfenadine suggests that not only the lipophilicity of the drug (terfenadine45: logD^oct= 4.4; bupivacaine46: logD^oct= 3.4) but also the size of the drug molecule (terfenadine: 471.7 g/mol; bupivacaine: 324.9 g/mol) influences the effect of the mutation A344V. In addition, the three-dimensional structure and orientation of the drug molecule may also be a contributing factor. The pharmacogenetic results may thus be specific for hydrophobic pharmacologic agents of a critical size such as local anesthetics.

The depolarizing shift in the voltage dependence of activation of KCNQ1/KCNE1 channel complexes caused by the mutation A344V may explain the pathophysiologic mechanism of this LQT1 mutation. We applied the LRd model for the guinea pig cardiac action potential27,28to simulate the effect of the mutation on cardiac action potential morphology (fig. 6). This model has been used before to simulate the effect of LQT mutations on cardiac action potentials.47,48These simulations demonstrate an increased dispersion and a repolarization deficiency for the mutant resulting in a more heterogeneous excitation pattern compared with the wild type. Because differences in APD and an increased dispersion of the repolarization can increase the risk for severe cardiac arrhythmia,49these simulations confirm the arrhythmogenic potential of the LQT1 mutation A344V.

Simulating the effects of bupivacaine on the cardiac action potential under wild-type conditions showed a prolongation of the cardiac action potential in the epicardial and endocardial layer and development of EADs in the midmyocardial layer. The results of the modeling are therefore consistent with the observation that bupivacaine intoxication causes QT prolongation and an increase in QT dispersion.2Under heterozygous conditions, the local anesthetic prolonged APD in epicardial layer and induced EADs in endocardial and midmyocardial layers of the myocardium. Because of different ion channel densities across the ventricular wall that are incorporated in the LRd model as different I^Krto I^Ksratios,29effects of the mutation and of bupivacaine were more pronounced in endocardial and midmyocardial layers of the heart.

The results of these simulations are in agreement with experimental data on human as well as on canine cardiomyocytes demonstrating that I^Ksblockade causes a more pronounced prolongation of the action potential in the presence of I^Krblockade.24,50Mutations in LQT related genes that per se  do not cause severe clinical effects or show a low penetrance may thus only become apparent when additional factors, such as drug-induced I^Krblock, increase the susceptibility to develop severe cardiac arrhythmia.22,51,52Because mutation carriers of A344V present with a mild phenotype of LQT121that may not become clinically apparent, our results confirm and extend the concept of I^Ksserving as a repolarization reserve.22,23LQT1 patients may be exposed to an increased risk of experiencing severe cardiac arrhythmia during local anesthetic intoxication. This study's results allow us to hypothesize that genotype-directed perioperative treatment of patients with congenital LQT syndrome is warranted.

In summary, the results of our study demonstrate that the LQT1 mutation A344V profoundly changes the gating of KCNQ1 channels and of KCNQ1/KCNE1 channel complexes. The mutation induces a voltage-dependent inactivation in KCNQ1 channels and a depolarizing shift in activation of KCNQ1/KCNE1 channel complexes. Furthermore, the mutation leads to a 22-fold increase in local anesthetic sensitivity. The increased sensitivity results from structural changes in the three-dimensional structure induced by the mutation A344V or by the de novo  formation of an interaction site. The results of our study suggest that certain forms of the LQT syndrome may constitute a specific rather than a nonspecific risk factor for the development of severe cardiac arrhythmia during anesthesia.

The authors thank James P. Dilger, Ph.D. (Professor, Department of Anesthesiology, State University of New York, Stony Brook, New York), for critically reading the manuscript and Andrea Zaisser (Technician, Institute for Neural Signal Transduction, University Medical Center Hamburg-Eppendorf, Hamburg, Germany) for technical assistance. The authors thank Olaf Pongs, Ph.D. (Professor and Director of the Institute of Neural Signal Transduction, University Medical Center Hamburg-Eppendorf), for his generous support.