Anesthesiology 2005; 103:102–12 © 2005 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc. Local Anesthetic Interaction with Human Ether-a-go-go–related Gene (HERG) Channels Role of Aromatic Amino Acids Y652 and F656 Cornelia C. Siebrands, M.Sc.,* Nicole Schmitt, Ph.D.,† Patrick Friederich, M.D.‡

Background: Human ether-a-go-go–related gene (HERG) po- and local anesthetics.10–12 Pharmacologic inhibition of tassium channels constitute a potential target involved in car- I may cause drug-induced long QT syndrome, severe diotoxic side effects of amino-amide local anesthetics. The mo- Kr lecular interaction site of these low-affinity blockers with HERG ventricular arrhythmia, and sudden cardiac death. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021 channels is currently unknown. The aim of this study was to HERG channels lack the Pro-Val-Pro motif conserved in determine the effect of the mutations Y652A and F656A in the other Kv channels that is supposed to produce a kink in putative drug binding region of HERG on the inhibition by the S6 helix.13 This results in a larger pore-forming cavity bupivacaine, ropivacaine, and mepivacaine. of HERG channels, allowing preferential trapping of Methods: The authors examined the inhibition of wild-type 14 and mutant HERG channels, transiently expressed in Chinese many structurally unrelated drugs. A scanning analysis hamster ovary cells by bupivacaine, ropivacaine, and mepiva- of the S6 transmembrane domain of the channel identi- caine. Whole cell patch clamp recordings were performed at fied two aromatic amino acids that are important for room temperature. high-affinity drug binding to HERG: tyrosine 652 and Results: Inhibition of HERG wild-type and mutant channels phenylalanine 656.2 Mutating these residues to alanine by the different local anesthetics was concentration dependent, stereoselective, and reversible. The sensitivity decreased in the increases the IC50 value of MK-499, terfenadine, and 2 order bupivacaine > ropivacaine > mepivacaine for wild-type cisapride between 100- and 650-fold. All high-affinity and mutant channels. The mutant channels were approxi- blockers (IC50 values within the nanomolar range) tested mately 4–30 times less sensitive to the inhibitory action of the are suggested to bind to this region of the channel2,15 by different local anesthetics than the wild-type channel. The con- ␲ centration–response data were described by Hill functions (bu- hydrophobic interaction with Phe656 and -cation inter- 16 .action or ␲-stacking with Tyr652 ؍ ؍ ␮ ؎ ؍ pivacaine: wild-type IC50 22 2 M,n 38; Y652A IC50 ؍ ␮ ؎ 95 5 M,n 31). The mutations resulted in a change of the The situation is different for low-affinity blockers (IC50 stereoselectivity of HERG channel block by ropivacaine. The values within the micromolar range) of HERG channels. potency of the local anesthetics to inhibit wild-type and mutant Mutating either of the aromatic amino acids to alanine channels correlated with the lipophilicity of the drug (r > 0.9). Conclusions: These results indicate that local anesthetics has severe effects on the inhibition by some compounds, 17–19 specifically but not exclusively interact with the aromatic resi- such as vesnarinone and quinidine, but not by oth- dues Y652 and F656 in S6 of HERG channels. ers, such as the selective serotonin reuptake inhibitor fluvoxamine20 and the antiparkinsonian drug budip- HUMAN ether-a-go-go–related gene (HERG) codes for ine.21 Different structural requirements have therefore the pore-forming component of the rapid delayed recti- been suggested to mediate low-affinity block of HERG 1 22 fier channel (IKr) in the heart. HERG channels constitute channels. toxicologically relevant targets for many structurally and Based on their IC50 values, amino-amide local anesthet- functionally unrelated substances, such as antiarrhyth- ics have to be regarded as low-affinity blockers of HERG mic drugs,2,3 antihistamines,4,5 psychoactive drugs,6 gas- channels10–12 with a so far unknown molecular site of trointestinal prokinetic agents,7,8 macrolide antibiotics,9 interaction. These local anesthetics differ by the length of their N-substituent, which is a butyl group (bupiva- caine), a propyl group (ropivacaine), or a methyl group * Ph.D. Student, Department of Anesthesiology and Institute for Neural Signal (mepivacaine). The length of the substituent determines Transduction, ‡ Privatdozent, Department of Anesthesiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany. † Postdoctoral Researcher, the lipophilicity and may thus be a structural require- Department of Medical Physiology, The Panum Institute, University of Copenha- ment for hydrophobic interactions between the drug gen, Copenhagen, Denmark. and the channel protein. It was shown before that the Received from the Department of Anesthesiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany. Submitted for publication September length of the N-substituent influences the potency of 2, 2004. Accepted for publication April 20, 2005. Supported by grant No. FR local anesthetics to block Kv1.5 channels23 and HERG 1625/1-1 from the Deutsche Forschungsgemeinschaft, Bonn, Germany; the De- 11 partment of Anesthesiology, University Hospital Hamburg-Eppendorf, Hamburg, channels. If amino-amide local anesthetics were to Germany; and the Institute for Neural Signal Transduction, University Hospital interact with the aromatic amino acids in the S6 region, Hamburg-Eppendorf, Hamburg, Germany. Levobupivacaine, ropivacaine, and mepivacaine were gifts from AstraZeneca, So¨dertalje, Sweden. Dr. Friederich has the potency to inhibit HERG channels would be ex- received lecture/travel fees from Abbott, Wiesbaden, Germany. Presented in part pected to correlate with the lipophilic properties of the at the Annual Meeting of the German Society of Physiology, Leipzig, Germany, March 14–17, 2004. drugs. Point mutations of these aromatic amino acids Address reprint requests to Dr. Friederich: Zentrum fu¨r Ana¨sthesiologie, Uni- may furthermore be expected to alter the relation be- versita¨tsklinik Eppendorf, Martinistrasse 52, 20251 Hamburg, Germany. Address electronic mail to: [email protected]. Individual article re- tween inhibitory potency and lipophilic drug properties. prints may be purchased through the Journal Web site, www.anesthesiology.org. Inhibition of HERG channels by amino-amide local anes-

Anesthesiology, V 103, No 1, Jul 2005 102 LOCAL ANESTHETIC INTERACTION WITH HERG 103

11,12 thetics is also stereoselective. If local anesthetics HEPES, 10 mM sucrose, and 0.1 mg/ml phenol red (all were to interact with the aromatic amino acids Y652 and from Sigma), adjusted to a pH of 7.4 with NaOH. To F656, mutating these amino acids might also alter stereo- record inward tail currents of HERG channels, an extra- selectivity of local anesthetic inhibition. cellular solution with high [Kϩ] was used, containing 40

Therefore, the aim of this study was to clarify whether mM NaCl, 100 mM KCl, 2 mM CaCl2,2mM MgCl2,5mM inhibition of HERG wild-type (wt) and mutant channels HEPES, 10 mM sucrose, and 0.1 mg/ml phenol red, ad- is related to the lipophilic properties of bupivacaine, justed to a pH of 7.4 with NaOH. ropivacaine, and mepivacaine. It was furthermore in- Series resistance was 2.5–6.0 M⍀ and was actively tended to establish whether and to what extent the compensated for by 85%. A leak subtraction protocol aromatic amino acids Y652 and F656 in the S6 region was used except for recordings with high extracellular ϩ ϩ

influence drug affinity and stereoselective local anes- [K ] ([K ]o). The recorded signal was filtered at 2 kHz Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021 thetic inhibition. As a prerequisite to these experiments, and stored with a sampling rate of 5 kHz for analysis. the gating changes induced by mutating these residues2 Bupivacaine (Sigma), levobupivacaine, S(Ϫ)-ropivacaine, needed to be characterized first. The results of this study R(ϩ)-ropivacaine, and mepivacaine (all from AstraZen- may help to identify structural requirements for low- eca, So¨dertalje, Sweden) were dissolved in the extracel- affinity block of HERG channels by amino-amide local lular solution. A hydrostatically driven perfusion system anesthetics. was used to apply the drugs onto the cells and to ex- change the extracellular solutions. All experiments were performed at room temperature. Materials and Methods Different pulse protocols were used for characteriza- tion of the channels and to establish their pharmacologic Cell Culture sensitivities. The holding potential was Ϫ80 mV for all Chinese hamster ovary cells were cultured in 50 ml- experiments. For the activation protocol, cells were de- flasks (NUNC, Roskilde, Denmark) at 37°C in MEM Alpha polarized from Ϫ80 to ϩ60 mV in 10-mV steps for 1 s, medium (GIBCO; Invitrogen, Carlsbad, CA) with 10% and tail currents were recorded at Ϫ40 mV. In high fetal calf serum, 100 U/ml penicillin, and 100 mg/ml ϩ [K ] , the tail potential was changed to Ϫ120 mV. To streptomycin in a humidified atmosphere (5% CO ). o 2 analyze the deactivation, channels were activated by a Cells were subcultured in 35-mm diameter monodishes 2-s pulse to ϩ60 mV, and deactivating tail currents were (NUNC) at least 1 day before transfection. recorded at potentials from Ϫ120 to ϩ40 mV in 10-mV ϩ steps. The same protocol was used in high [K ]o.To Molecular Biology and Transfection of Cells compare the three channels, currents were normalized The mutants HERG Y652A and F656A were created by to the minimum. The deactivation was fitted with an site-directed mutagenesis. All channels were cloned in exponential function that yielded two time constants. the pcDNA3 expression vector. Chinese hamster ovary For the steady state inactivation, a three-step protocol ␮ cells were transiently transfected with 1 g HERG wt or was used: First, the cells were depolarized for 2 s from ␮ ␮ mutant cDNA, 0.5 g EFGP cDNA, and 3 l lipo- Ϫ80 to ϩ40 mV; then, a pulse from Ϫ100 to ϩ60 mV in fectamine reagent (Invitrogen) per dish according to the 10-mV steps was applied for 30 ms; and finally, the manufacturer’s protocol after 1 day. Cells were cotrans- membrane potential was held at ϩ40 mV for 0.5 s. The fected with an EGFP pcDNA3 construct to verify suc- inactivation was corrected for the closing during cessful transfection. Only green fluorescing cells were the hyperpolarizing step at very negative potentials as used for patch clamp experiments. Patch clamp experi- described previously.25 The corrected curves were fitted ments were performed 1 or 2 days after transfection. by a Boltzmann function. During the instantaneous acti- vation protocol channels were activated and inactivated Electrophysiology by depolarization from a holding potential of Ϫ80 to Whole cell currents were recorded using the patch ϩ20 mV for 2 s. After recovery from inactivation by clamp technique24 with an EPC-9 amplifier and Pulse repolarization to Ϫ80 mV for 30 ms, the instantaneous software version 8.50 (HEKA Electronik, Lambrecht, current was recorded at potentials from ϩ40 mV to Germany). Patch electrodes were pulled from borosili- Ϫ120 mV in 10-mV steps. The time constants of inacti- cate glass capillary tube (World Precision Instruments, vation were derived from a monoexponential fit of the Saratoga, FL) on a horizontal puller (P-97; Sutter Instru- decay. This protocol was also used for pharmacologic ment Co., Novato, CA) and had a pipette resistance of experiments. For this purpose, the total length of the 1.5–3.5 M⍀. The internal solution contained 160 mM protocol was reduced, and the potential was increased Ϫ ϩ KCl, 0.5 mM MgCl2,10mM HEPES, and 2 mM Na-ATP (all in 30-mV steps from 120 to 60 mV. from Sigma, Deissenhofen, Germany), adjusted to a pH For the pharmacologic experiments with HERG wt and of 7.2 with KOH. The external solution contained 135 Y652A, a ramp protocol was used.12,26 Cells were depo- Ϫ ϩ mM NaCl, 5 mM KCl, 2 mM CaCl2,2mM MgCl2,5mM larized from a holding potential of 80 mV to 60 mV

Anesthesiology, V 103, No 1, Jul 2005 104 SIEBRANDS ET AL.

Fig. 1. Activation and deactivation of HERG wild-type (wt) and HERG Y652A currents in Chinese hamster ovary cells. (A) Original current traces evoked by the activation protocol shown in the inset. (B) Normalized tail current amplitudes

(Itail/Imax) were fitted by a Boltzmann function. The voltage of half-maximal ac- Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021 tivation V0.5 and the slope factor were not significantly different (P > 0.5). (C) Rep- resentative current traces of HERG wt and HERG Y652A evoked by the deactiva- tion protocol (see inset). (D) Voltage de- pendence of deactivation. Normalized

current amplitudes (I/Imax) are shown. (.for each 6 ؍ n)

Ϫ ϩ Ϯ and repolarized to 80 mV within 1 s. In high [K ]o,a are presented as mean SD unless stated otherwise; n single pulse to ϩ20 mV for 2 s and a tail potential of values indicate the number of experiments. Ϫ120 mV was used for the pharmacologic tests. Repet- itive pulses were applied to determine that steady state Results inhibition was reached. Properties of HERG wt, HERG Y652A, and HERG Data Analysis F656A Data were analyzed with Pulse Fit software (HEKA Representative current traces of HERG wt and HERG Electronik) and with Kaleidagraph software (Synergy Y652A elicited by the activation protocol are shown in Software, Reading, PA). The normalized tail currents figure 1A. The voltage dependence of activation of HERG during the activation protocol were fitted by a modified Y652A channels was not significantly different from wt ϭ ϩ Ϫ Boltzmann equation: I Imax/[1 exp((V0.5 Vm)/k)], channels (fig. 1B; wt: voltage of half-maximal activation ϭ Ϯ ϭ Ϯ ϭ where V0.5 is the voltage of half-maximal activation, Vm V0.5 4.85 4.23 mV, slope 8.51 0.83 mV, n 6; ϭ Ϯ ϭ Ϯ is the membrane potential, and k is the slope factor. The Y652A: V0.5 5.35 4.31 mV, slope 8.79 1.27 mV, exponential decay of deactivation or inactivation was n ϭ 12). Also, the time constants of activation did not fitted with one or two time constants according to the differ between wt and mutant channels. However, the ϭ ϩ Ϫ ␶ ϩ following equation: y C Afastexp( t/ fast) decay of HERG Y652A currents was faster than that of ␶ ␶ ␶ Aslowexp(t/ slow), with the time constants fast and slow HERG wt currents (exponential fit of the decay after ␶ ϭ Ϯ ␶ ϭ and the amplitudes Afast and Aslow. The inhibition of cur- activation to 50 mV, wt: 1.42 0.5 s; Y652A: rents was quantified by the reduction of the maximal cur- 0.86 Ϯ 0.3 s; P Ͻ 0.01). Figure 1C shows exemplary rent during the ramp protocol or during a single pulse in current traces obtained by the deactivation protocol. ϩ case of the high-[K ]o experiments. Also, the charge cross- Both HERG wt and HERG Y652A currents exhibited ing the membrane during the ramp protocol (Qramp) and inward rectification with no difference in the voltage the reduction of Qramp were analyzed. The charge crossing dependence of deactivation (fig. 1D). The steady state the membrane is equivalent to the time integrals of current inactivation was analyzed next (fig. 2A). The voltage traces and was determined using Pulse Fit software. The dependence of HERG Y652A current inactivation was fractional block f was calculated by the following formula: shifted to more positive potentials compared with HERG ϭ Ϫ ϫ ϩ ϭϪ Ϯ ϭ f 1 [2 Imax, drug/(Imax, control Imax, washout)]. wt (fig. 2B; V0.5 36.2 10.6 mV, n 6 for Y652A ϭϪ Ϯ ϭ Ͻ Concentration–response curves were fitted by a Hill func- versus V0.5 69.6 7.8 mV, n 7 for wt; P 0.05). ϭ ϩ h tion: f 1/[1 (IC50/c) ], where IC50 is the concentration Instantaneous currents inactivated in a voltage-depen- of half-maximal inhibition, c is the concentration of the dent manner (fig. 2C) with time constants that were local anesthetic, and h is the Hill coefficient. Statistical smaller for HERG Y652A than for wt channels (fig. 2D; significance was tested using a two-sided Student t test or n ϭ 4 each, P Ͻ 0.05). analysis of variance (Excel; Microsoft, Redmond, WA). Data HERG F656A channels only conducted very small cur-

Anesthesiology, V 103, No 1, Jul 2005 LOCAL ANESTHETIC INTERACTION WITH HERG 105

Fig. 2. Inactivation of HERG wild-type (wt) and HERG Y652A currents. (A) Rep- resentative current traces of HERG wt and HERG Y652A showing the steady state inactivation. Note that the current for HERG Y652A does not approach the steady state level after hyperpolarization. (B) Uncorrected curves for the voltage

dependence of steady state inactivation. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021 Tail currents were normalized to the

maximum (Itail/Imax). (C) An instanta- neous activation protocol was used to de- termine the inactivation time constants at different test potentials. (D) Voltage dependence of inactivation time con- ␶ stants ( inact). The differences between HERG wt and HERG Y652A were highly significant (P < 0.01, tested by analysis of variance).

rents under low extracellular [Kϩ].2 Inward currents of 615 Ϯ 12 ms for HERG Y652A (n ϭ 91). These values this mutation were therefore recorded using high extra- correspond to voltages of Ϫ40 Ϯ 1 mV (HERG wt) and cellular [Kϩ].20 To compare all three channels, HERG wt Ϫ26 Ϯ 2 mV (HERG Y652A), respectively. All local and HERG Y652A currents were recorded under these anesthetics reduced HERG wt and HERG Y652A currents conditions as well (fig. 3). Tail current densities at Ϫ120 in a concentration-dependent and reversible manner ϩ mV under high extracellular [K ] were 968 Ϯ 475 pA/pF (figs. 4B–F). The inhibition was quantified as the de- for HERG wt, 571 Ϯ 272 pA/pF for HERG Y652A, and crease of the maximum current as well as the reduction 129 Ϯ 49 pA/pF for HERG F656A channels, respectively of the charge. The concentration–response data were (n ϭ 5 for each). Only the difference between HERG mathematically described by Hill functions (figs. 5A and F656A and HERG wt was significant. Exemplary current B and table 1). The concentration of half-maximal inhi- ϩ traces evoked by the activation protocol in high [K ]o bition (IC50) by the local anesthetics was between 4 and are shown in figure 3A. Voltage dependence of activa- 10 times higher for HERG Y652A compared with the wt tion did not differ between HERG wt and HERG Y652A channel. The Hill coefficients were close to unity for currents, whereas HERG F656A channels activated with inhibition of both channels by all local anesthetics. Ap- aV0.5 shifted by 10 mV to more negative potentials (fig. plication of local anesthetics furthermore caused a con- 3B). Figure 3C shows deactivating currents of all three centration-dependent and reversible rightward shift of channels. HERG F656A hyperpolarized the voltage de- the peak HERG wt current and therefore increased the pendence of deactivation (fig. 3D) and slowed the deac- time to peak current response (figs. 4B–F). In contrast, tivation kinetics, whereas deactivation kinetics were ac- the application of the local anesthetics caused a concen- celerated by HERG Y652A (fig. 3D, inset). The tration-dependent and reversible leftward shift of the differences between HERG wt and the mutant channels peak HERG Y652A current and therefore decreased the were significantly different (confirmed by analysis of time to peak current response. Because of the reduced variance). sensitivity of HERG Y652A, the shift is also more pro- nounced at higher concentrations of local anesthetics. Local Anesthetic Sensitivity of HERG wt and HERG For example, 30 ␮M bupivacaine caused a shift of the Y652A time to peak HERG wt current response of ϩ62 Ϯ 10 ms The inhibitory effects of the local anesthetics bupiva- (n ϭ 9), and 100 ␮M bupivacaine caused a shift of caine, levobupivacaine, ropivacaine, and mepivacaine ϩ90 Ϯ 27 ms (n ϭ 5), corresponding to shifts in voltage on HERG wt and HERG Y652A currents were investi- of ϩ9 Ϯ 1 and ϩ13 Ϯ 4 mV, respectively. For HERG gated using the ramp protocol. This protocol evoked Y652A, 100 ␮M bupivacaine caused a shift of the time to bell-shaped currents (fig. 4A). The mean time to peak peak current of Ϫ37 Ϯ 17 ms (n ϭ 9), and 300 ␮M current was 723 Ϯ 10 ms for HERG wt (n ϭ 116) and bupivacaine caused a shift of Ϫ71 Ϯ 18 ms (n ϭ 9).

Anesthesiology, V 103, No 1, Jul 2005 106 SIEBRANDS ET AL. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021

Fig. 3. Activation and deactivation of HERG wild-type (wt), HERG Y652A, and HERG F656 currents with 100 mM extracellular potassium. (A) Original currents elicited by the activation protocol shown in the inset. Inward tail currents were recorded at ؊120 ؎ ؍ ؎ ؍ mV. (B) Normalized tail current amplitudes (Itail/Imax) were fitted by a Boltzmann function (wt: V0.5 1.40 4.68 mV, slope 10.43 ؎ ؍ ؎ ؊؍ ؍ ؎ ؍ ؎ ؊؍ ؍ 1.93 mV, n 6; Y652A: V0.5 1.68 6.78 mV, slope 10.79 1.29 mV, n 6; F656A: V0.5 8.86 4.69 mV, slope 10.69 C) Representative current traces of HERG wt, HERG Y652A, and HERG Y652A evoked by the deactivation protocol) .(7 ؍ mV, n 1.10 ؍ ؉ with high extracellular [K ]. (D) Voltage dependence of deactivation (n 6 for each). Currents were normalized (I/Imax). The inset ␶ ␶ shows the voltage dependence of the fast deactivation time constant fast. The values were significantly different between HERG wt and HERG Y652A and between HERG wt and HERG F656A (P < 0.01, tested by analysis of variance).

These values correspond to shifts in voltage of Ϫ5 Ϯ 2 R(ϩ)-ropivacaine was a more potent inhibitor of HERG and Ϫ10 Ϯ 2 mV, respectively. For both HERG wt and Y652A currents than S(Ϫ)-ropivacaine (83 ␮M:41Ϯ 3%, ϭ Ϯ ϭ Ͻ ␮ Ϯ HERG Y652A, the IC50 values of the local anesthetics n 7 vs. 27 2%, n 7; P 0.05; 250 M:63 3%, increased in the order levobupivacaine Ͻ bupivacaine Ͻ n ϭ 7 vs. 49 Ϯ 3%, n ϭ 7; P Ͻ 0.05). ropivacaine Ͻ mepivacaine. The inhibition of HERG wt and HERG Y652A channels The influence of the mutation Y652A on stereoselec- by bupivacaine was further analyzed by applying an tive inhibition of HERG currents was further analyzed. instantaneous current protocol (fig. 6; see Materials and

The wt channel was twice as sensitive to levobupiva- Methods section). Bupivacaine inhibited Imax of both Ϯ Ϯ caine than to bupivacaine (IC50 12.7 1.4 vs. 21.9 1.6 channels (fig. 7) in a voltage-independent manner (fig. ␮ M; table 1), whereas the ratio of IC50 values of both 7A). The local anesthetic significantly altered the voltage drugs approximated unity for the mutant channel HERG dependence and the size of the inactivation time constants Ϯ Ϯ ␮ Y652A (IC50 82.9 3.2 vs. 95.2 5.0 M; table 1). For of HERG wt but not of HERG Y652A channels (fig. 7B; ropivacaine, the influence of HERG Y652A on stereose- tested by analysis of variance). The time courses of deacti- lective inhibition was even more pronounced (compare vating currents (Ϫ120 mV) of both HERG wt and HERG Ϫ ␶ ϭ figs. 4D and E). S( )-ropivacaine was a more potent Y652A were slowed by bupivacaine (HERG wt: deact, fast ϩ Ϯ ␶ ϭ Ϯ inhibitor of HERG wt currents than R( )-ropivacaine (25 18.98 4.31 ms, deact, slow 179.3 44.9 ms under ␮ Ϯ ϭ Ϯ ϭ Ͻ ␮ ␶ ϭ Ϯ M:53 5%, n 8 vs. 28 2%, n 5; P 0.05; 83 M: control and washout conditions; deact, fast 27.14 5.89 Ϯ ϭ Ϯ ϭ Ͻ ␶ ϭ Ϯ 79 4%, n 8 vs. 60 4%, n 5; P 0.05), whereas ms, deact, slow 259.6 63.3 ms under 20 mM bupiva-

Anesthesiology, V 103, No 1, Jul 2005 LOCAL ANESTHETIC INTERACTION WITH HERG 107 Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021

Fig. 4. (A) Normalized and superimposed traces of HERG wild-type (wt; solid line) and HERG Y652A (dotted line) currents evoked by the ramp protocol (see inset). Note that the peak current is reached later for the wt channel than for the mutant HERG Y652A. (B–F) Examples of current traces and effect of bupivacaine (B), levobupivacaine (C), S(؊)-ropivacaine (D), R(؉)-ropivacaine (E), and mepivacaine (F) on HERG wt and HERG Y652A currents evoked by the ramp protocol. caine; n ϭ 5, paired experiments, P Ͻ 0.05; HERG Y652A: HERG wt and HERG Y652A channels were in addition ␶ ϭ Ϯ ␶ ϭ Ϯ ϩ deact, fast 12.13 2.50 ms, deact, slow 59.57 21.37 established under high extracellular [K ]. Bupivacaine ␶ ϭ ␮ Ϯ ϭ ms under control and washout conditions; deact, fast (100 M) inhibited wt currents by 66 5% (n 5). Both Ϯ ␶ ϭ Ϯ 13.28 2.12 ms, deact, slow 90.71 27.66 ms under 90 mutants reduced the sensitivity (figs. 8A and B), with ␮M bupivacaine; n ϭ 5, paired experiments, P Ͻ 0.01). HERG F656A currents being less sensitive than HERG Y652A (32 Ϯ 4% for Y652A, n ϭ 5; 17 Ϯ 6% for F656A, Local Anesthetic Sensitivity of HERG F656A n ϭ 5; P Ͻ 0.01). Because of the low sensitivity of HERG To meaningfully compare the pharmacologic sensitiv- F656A channels, it was not possible to determine whole ities between all three channels, bupivacaine effects on concentration–response curves for the different local

Fig. 5. Concentration dependence of HERG wild-type (wt) and HERG Y652A channel block induced by bupivacaine (b), mepivacaine (m) (A), levobupiva- .(caine (lb), and S(؊)-ropivacaine (S-r) (B Inhibition was measured as the reduction of the maximum current during the ramp protocol. Curves were fitted by Hill func- tions (parameters are shown in table 1). Each point represents 3–10 experiments.

Anesthesiology, V 103, No 1, Jul 2005 108 SIEBRANDS ET AL.

Table 1. Parameters Derived from the Fit of the Concentration–Response Data by Hill Functions

Substance wt, Mean Ϯ SEM Y652A, Mean Ϯ SEM

␮ ␮ IC50, M Hill n IC50, M Hill n

Data obtained from inhibition of maximum current Bupivacaine 21.9 Ϯ 1.6 0.91 Ϯ 0.06 31 95.2 Ϯ 5.0 1.21 Ϯ 0.07 38 Levobupivacaine 12.7 Ϯ 1.4 0.85 Ϯ 0.07 41 82.9 Ϯ 3.2 1.11 Ϯ 0.05 25 S-ropivacaine 23.9 Ϯ 1.6 1.00 Ϯ 0.06 50 240.3 Ϯ 9.5 0.97 Ϯ 0.04 27 Mepivacaine 156.2 Ϯ 11.1 0.89 Ϯ 0.06 28 475.6 Ϯ 50.3 1.03 Ϯ 0.10 27 Data obtained from

reduction of charge Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021 Bupivacaine 21.5 Ϯ 3.4 0.88 Ϯ 0.12 31 83.9 Ϯ 5.9 1.27 Ϯ 0.11 38 Levobupivacaine 10.8 Ϯ 1.2 1.01 Ϯ 0.12 41 71.8 Ϯ 5.9 1.16 Ϯ 0.11 26 S-ropivacaine 21.0 Ϯ 0.8 1.18 Ϯ 0.05 48 212.3 Ϯ 12.7 1.00 Ϯ 0.06 27 Mepivacaine 127.0 Ϯ 9.9 0.96 Ϯ 0.07 28 407.9 Ϯ 43.0 1.04 Ϯ 0.11 27

anesthetics. Instead, local anesthetic sensitivity was com- for HERG F656A) compared with drug-free conditions ϩ pared at a concentration of 1 mM. HERG F656A channels under high extracellular [K ]. exhibited the same order of decreasing pharmacologic sensitivity as HERG wt (fig. 8C). However, HERG wt currents were inhibited more potently by S(Ϫ)-ropiva- Discussion caine than by R(ϩ)-ropivacaine, whereas HERG F656A currents were inhibited more potently by R(ϩ)-ropiva- In the current work, the effects of the mutations caine than by S(Ϫ)-ropivacaine (R-ropivacaine vs. S-ropi- Y652A and F656A in the S6 region of HERG potassium vacaine: 74 Ϯ 2% vs. 84 Ϯ 1%, n ϭ 5, paired experi- channels2 on the inhibition by amino-amide local anes- ments, P Ͻ 0.01 for wt; 43 Ϯ 5% vs. 31 Ϯ 5%, n ϭ 4, thetics have been established. Mutating either of the paired experiments, P Ͻ 0.05 for F656A). The time aromatic residues diminished the block by these local constants of deactivation in high extracellular [Kϩ] and anesthetics 4- to 10-fold (Y652A) and 20- to 30-fold the effect of bupivacaine on deactivation (figs. 8D and E) (F656A), respectively. In accord with previous stud- revealed that high extracellular [Kϩ] slows the deactiva- ies,2,16 the voltage dependence of activation and deacti- tion kinetics of both HERG wt and HERG Y652A. How- vation remained unchanged by the mutation Y652A. ever, deactivation of HERG channels is slowed by bupiv- However, the voltage dependence of steady state inacti- acaine (100 ␮M) under normal as well as high vation was shifted to more positive potentials, and the extracellular [Kϩ], whereas deactivation of HERG Y652A inactivation and deactivation time constants were signif- and HERG F656A follows a faster time course under the icantly faster. This may explain the observation that the influence of bupivacaine (100 ␮M for HERG Y652A, 1 mM time to peak current was shorter for HERG Y652A than

Fig. 6. Effect of bupivacaine on the instan- taneous activation of HERG wild-type (wt; A) and HERG Y652A (B). Concentrations

near the respective IC50 values were cho- sen to investigate the effect of the local anesthetic on inactivation and deactiva- tion time constants. Shown are original current traces under control conditions, after application of bupivacaine (20 ␮M for HERG wt, 90 ␮M for HERG Y652A) and after washout of the drug. The inhibition of current and the changes in kinetics are fully reversible upon washout. The pulse protocol is shown in the inset.

Anesthesiology, V 103, No 1, Jul 2005 LOCAL ANESTHETIC INTERACTION WITH HERG 109

Fig. 7. Analysis of bupivacaine effects on HERG wild-type (wt; 20 ␮M bupivacaine) and Y652A (90 ␮M bupivacaine) currents during the instantaneous protocol (fig. 6). (A) Inhibition of peak current is not voltage dependent. (B) Inactivation is faster in the mutant channels HERG Y652A than in HERG wt (compare with fig. 2D). Application of bupivacaine Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021 (bupi) accelerates inactivation kinetics in HERG wt but not in HERG Y652A chan- ؍ ␶ nels ( inact inactivation time con- stants). (C and D) Deactivation at ؊120 mV was fitted with two time constants ␶ ␶ ( deact, fast and deact, slow), which were both smaller for HERG Y652A than for HERG wt. Application of bupivacaine in- creases both time constants in HERG wt and in HERG Y652A channels.

for HERG wt during the ramp protocol. The mutation current decline of HERG wt channels under drug condi- F656A reduced the current density by approximately tions, the time constants of current decline of HERG 90% compared with the wt channel. Furthermore, the Y652A under drug conditions likely reflect channel in- voltage dependence of activation and deactivation were activation rather than drug–channel interaction. The al- shifted to more negative potentials, and the deactivation ready accelerated inactivation time constant of HERG kinetics were slowed. Y652A channels under control conditions may therefore The effects of bupivacaine, levobupivacaine, and ropi- have precluded resolving the time course of the bupiv- vacaine but not of mepivacaine on HERG wt channels acaine effect. The extent of inhibition of the instanta- have been analyzed before.11 However, instead of con- neous peak current response of HERG wt and HERG ventional stimulation protocols, we used a ramp proto- Y652A was identical to the extent of inhibition predicted col26 to establish drug sensitivity of wt and mutant chan- from the respective concentration–response curves ob- nels. Such a protocol has repeatedly been used to tained with the ramp protocol. Taken together, these measure drug sensitivity of HERG channels.12,27,28 This results may suggest that ramp current inhibition mainly

protocol yielded IC50 values and Hill coefficients of bu- reflects open channel inhibition. This idea is supported pivacaine, levobupivacaine, and ropivacaine that are by the experiments with high extracellular [Kϩ]. Inhibi- nearly identical to those obtained with conventional tion of HERG wt as well as HERG Y652 is reduced by the pulse protocols.11 Furthermore, the different current presence of high extracellular [Kϩ]. This has previously responses of HERG wt and HERG Y652A to the ramp been described for the open channel blocker E-4031.29 It protocol suggest that differences in channel gating be- may therefore be hypothesized that interaction with the tween wt and mutant channels are elicited by the ramp open channel pore constitutes an important mechanism protocol. The ramp protocol therefore allows compar- contributing to local anesthetic inhibition of wt11 as well ing local anesthetic sensitivity of wt and mutated HERG as mutated HERG channels. channels. The effect of bupivacaine on these channels The results of our study demonstrate that the muta- was in addition assessed with an instantaneous current tions in the S6 segment are not conservative with regard protocol.25 By this protocol, the effect of bupivacaine on to ion channel gating. Therefore, it cannot entirely be channels in the open state, on channel deactivation, and ruled out that differences in channel gating induced by on channel inactivation was studied. Bupivacaine in- the mutations Y652A and F656A may have influenced creased the fast and slow time constants of deactivation local anesthetic affinity. Mutations that disable inactiva- in both HERG wt and HERG Y652A channels compatible tion (G628C-S631C) or selective inhibition of HERG with open channel block. Bupivacaine accelerated mac- channel inactivation by addition of Cd2ϩ ions or removal roscopic current decline of wt but not of HERG Y652A of Naϩ ions reduce block by D-sotalol.30 Removal of channels. Because inactivation of HERG Y652A under C-type inactivation reduces bupivacaine sensitivity and drug-free conditions is already faster than macroscopic confers voltage dependence to the inhibition of HERG

Anesthesiology, V 103, No 1, Jul 2005 110 SIEBRANDS ET AL. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021

Fig. 8. (A) Representative current traces of HERG wild-type (wt), HERG Y652A, and HERG F656A channels under control conditions ؉ and after application of 100 ␮M bupivacaine. Experiments were performed with high extracellular [K ], and inward currents were (recorded at ؊120 mV (see inset). (B) Effect of 100 ␮M bupivacaine on HERG wt, HERG Y652A, and HERG F656A currents. (C ,F656A ;5 ؍ S(؊)-ropivacaine (wt, n ,(6 ؍ F656A, n ;3 ؍ Inhibition of HERG wt and HERG F656A channels by 1 mM bupivacaine (wt, n Highly significant difference ** .(7 ؍ F656A, n ;3 ؍ and mepivacaine (wt, n ,(4 ؍ F656A, n ;5 ؍ R(؉)-ropivacaine (wt, n ,(4 ؍ n ؊ ␶ ␶ between wt and mutant channels (P < 0.01). Deactivation at 120 mV was fitted with two time constants, deact, fast (D) and deact, slow (E), for HERG wt and the two mutants. Mutation Y652A accelerates deactivation; mutation F656A slows deactivation. Application of ␮ ␶ bupivacaine (100 M for HERG wt and Y652A, 1 mM for HERG F656A) increases deact for HERG wt but decreases it for the mutant channels.

channels.10 However, as judged from the results ob- the concentration of extracellular [Kϩ]. Both results tained with the instantaneous current protocol, voltage point to a common inhibitory mechanism. In addition, dependence of inhibition does not differ between HERG the inhibitory potency of the local anesthetics signifi-

wt and HERG Y652A. Also, inhibition of both channels cantly correlated with the number of the CH2 groups of HERG wt and HERG Y652A was reduced by increasing the local anesthetics for HERG wt (figs. 9A and B), HERG

Fig. 9. (A) Correlation between the log

IC50 for HERG wild-type (wt) and Y652A and the length of the N-substituent of the homolog series of local anesthetics used in our experiments. Bupivacaine has four

CH2 groups, ropivacaine has three, and mepivacaine has one (linear fit for wt: ؍ for Y652A: y ;0.91 ؍ ؊ 0.29x, r 2.47 ؍ y B). Correlation of) .(0.92 ؍ ؊ 0.25x, r 2.96 the inhibitory effect of 1 mM local anes- thetic on HERG wt and F656A channels and the length of the N-substituent (lin- ;0.97 ؍ ؉ 0.15x, r 0.33 ؍ ear fit for wt: y .(0.91 ؍ ؉ 0.13x, r 0.032 ؍ for F656A: y

Anesthesiology, V 103, No 1, Jul 2005 LOCAL ANESTHETIC INTERACTION WITH HERG 111

Y652A (fig. 9A), and HERG F656A (fig. 9B). The correla- rhythmic action at lower concentrations and cardiotoxic tions had similar slope factors. If the quantitatively and side effects at higher concentrations. It may be worth qualitatively different changes in channel gating were to considering that because of the specific structure of play a fundamental role in reducing local anesthetic HERG channels, both antiarrhythmic as well as proar- affinity, a conserved response to extracellular [Kϩ]as rhythmic action may increase with the lipophilicity of well as a conserved relation between physicochemical amino-amide local anesthetics and may be more pro- properties of the drugs and blocking affinity would be nounced for the S(Ϫ)-enantiomers. unlikely. The conserved relation between physicochem- In summary, our results suggest that local anesthetics ical properties of the drugs and blocking affinity would interact with the inner cavity of the HERG channel pore. furthermore point to a nonspecific membrane-mediated Interaction may involve the aromatic residues Tyr652 31,32

effect underlying inhibition by the local anesthetics. and Phe656. However, the correlation between the li- Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021 As an alternative, it may be hypothesized that hydropho- pophilic properties of the drugs and inhibitory potency bic interactions occurring at a site different from the remains unaltered by the mutations Y652A and F656A. inner S6 cavity22 are crucial for local anesthetic interac- Additional hydrophobic domains of the channel contrib- tion. Both alternatives seem less likely. uting to local anesthetic interaction with HERG channels HERG wt channels were inhibited more potently by thus remain to be identified. the S(Ϫ)-enantiomer than by the R(ϩ)-enantiomer of ropivacaine. Mutating the aromatic amino acids Tyr652 The authors thank Olaf Pongs, Ph.D. (Director of the Institute of Neural Signal Transduction, University Medical Center Hamburg-Eppendorf, Germany), for his or Phe656 to alanine reversed the stereoselectivity of generous support; Andrea Zaisser (Medicinal Technical Assistant, Institute for ropivacaine block. This change in the stereoselectivity of Neural Signal Transduction, University Medical Center Hamburg-Eppendorf, Ger- many) for technical assistance; and Anna Solth, M.Sc., and Axel Neu, M.D. block suggests a direct interaction of the drug molecules (Research Fellows, Institute for Neural Signal Transduction, University Medical with the residues Tyr652 and Phe656 rather than a Center Hamburg-Eppendorf, Germany), for critically reading the manuscript. membrane-mediated effect. The aromatic residues in the S6 region of HERG channels are therefore likely to be involved in the interaction with the local anesthetics References investigated. Kv channels exhibit a reversed local anes- 1. Sanguinetti MC, Jiang C, Curran ME, Keating MT: A mechanistic link be- 23,33 tween an inherited and an acquired cardiac arrhythmia:HERG encodes the IKr thetic stereoselectivity compared with HERG chan- potassium channel. Cell 1995; 81:299–307 nels, with the R(ϩ)-enantiomer being more potent than 2. Mitcheson JS, Chen J, Lin M, Culberson C, Sanguinetti MC: A structural basis Ϫ for drug-induced long QT syndrome. Proc Natl Acad SciUSA2000; 97:12329–33 the S( )-enantiomer. Exchanging either of the aromatic 3. Spector PS, Curran ME, Keating MT, Sanguinetti MC: Class III antiarrhyth- residues in the S6 of HERG for alanine results in a pattern mic drugs block HERG, a human cardiac delayed rectifier Kϩ channel: Open- channel block by methanesulfonanilides. Circ Res 1996; 78:499–503 of stereoselective inhibition resembling that of Kv chan- 4. Suessbrich H, Waldegger S, Lang F, Busch AE: Blockade of HERG channels nels. These results indicate that both aromatic amino expressed in Xenopus oocytes by the histamine receptor antagonists terfenadine and astemizole. FEBS Lett 1996; 385:77–80 acids are necessary for the pattern of stereoselective 5. Roy M, Dumaine R, Brown AM: HERG, a primary human ventricular target inhibition by local anesthetics specific for HERG chan- of the nonsedating antihistamine terfenadine. Circulation 1996; 94:817–23 6. Rampe D, Murawsky MK, Grau J, Lewis EW: The antipsychotic agent nels. The conserved correlation of inhibitory potency sertindole is a high affinity antagonist of the human cardiac potassium channel and lipophilic properties of the drugs, however, suggests HERG. J Pharmacol Exp Ther 1998; 286:788–93 7. Rampe D, Roy ML, Dennis A, Brown AM: A mechanism for the proarrhyth- that additional hydrophobic amino acids residing in the mic effects of cisapride (Propulsid): High affinity blockade of the human cardiac 16 inner cavity of HERG channels may be involved in local potassium channel HERG. FEBS Lett 1997; 417:28–32 8. Walker BD, Singleton CB, Bursill JA, Wyse KR, Valenzuela SM, Qiu MR, Breit anesthetic action. SN, Campbell TJ: Inhibition of the human ether-a-go-go-related gene (HERG) Our results demonstrate that local anesthetic affinity of potassium channel by cisapride: Affinity for open and inactivated states. Br J Pharmacol 1999; 128:444–50 HERG channels is determined by the lipophilic proper- 9. Stanat SJ, Carlton CG, Crumb Jr, WJ Agrawal, KC, Clarkson CW: Character- ties of the drugs. The longer the alkyl side chain is, the ization of the inhibitory effects of erythromycin and clarithromycin on the HERG potassium channel. Mol Cell Biochem 2003; 254:1–7 more potent the drug is. Also, the orientation of the local 10. Lipka LJ, Jiang M, Tseng GN: Differential effects of bupivacaine on cardiac anesthetic molecule influences drug affinity of HERG K channels: Role of channel inactivation and subunit composition in drug- Ϫ channel interaction. J Cardiovasc Electrophysiol 1998; 9:727–42 channels, with the S( )-isomer being more potent than 11. Gonzalez T, Arias C, Caballero R, Moreno I, Delpon E, Tamargo J, Valen- the R(ϩ)-isomer. HERG channels constitute well-estab- zuela C: Effects of levobupivacaine, ropivacaine and bupivacaine on HERG channels: Stereoselective bupivacaine block. Br J Pharmacol 2002; 137:1269–79 lished, toxicologically relevant targets for many structur- 12. Friederich P, Solth A, Schillemeit S, Isbrandt D: Local anaesthetic sensitiv- ally and functionally unrelated substances, including lo- ities of cloned HERG channels from human heart: comparison with HERG/MiRP1 2,10–12 and HERG/MiRP1 T8A. Br J Anaesth 2004; 92:93–101 cal anesthetics. Inadvertent intravascular injection 13. del Camino D, Holmgren M, Liu Y, Yellen G: Blocker protection in the of amino-amide local anesthetics, such as those investi- pore of a voltage-gated Kϩ channel and its structural implications. Nature 2000; 403:321–5 gated in this study, are capable of inducing severe ven- 14. Mitcheson JS, Chen J, Sanguinetti MC: Trapping of a methanesulfonanilide 34,35 tricular arrhythmia and sudden death. On the other by closure of the HERG potassium channel activation gate. J Gen Physiol 2000; 115:229–40 hand, modest inhibition of HERG channels may result in 15. Lees-Miller JP, Duan Y, Teng GQ, Duff HJ: Molecular determinant of 3 class III antiarrhythmic action. Like the class III antiar- high-affinity dofetilide binding to HERG1 expressed in Xenopus oocytes: Involve- 30 ment of S6 sites. Mol Pharmacol 2000; 57:367–74 rhythmic agent sotalol and other proarrhythmic 16. Fernandez D, Ghanta A, Kauffman GW, Sanguinetti MC: Physicochemical 36 drugs, local anesthetics may therefore exhibit antiar- features of the HERG channel drug binding site. J Biol Chem 2004; 279:10120–7

Anesthesiology, V 103, No 1, Jul 2005 112 SIEBRANDS ET AL.

17. Kamiya K, Mitcheson JS, Yasui K, Kodama I, Sanguinetti MC: Open chan- rent during the cardiac ventricular action potential. Pflugers Arch 1998; 436: nel block of HERG K(ϩ) channels by vesnarinone. Mol Pharmacol 2001; 60: 843–53 244–53 27. Zhou Z, Gong Q, Ye B, Fan Z, Makielski JC, Robertson GA, January CT: 18. Sanchez-Chapula JA, Navarro-Polanco RA, Culberson C, Chen J, Sangui- Properties of HERG channels stably expressed in HEK 293 cells studied at netti MC: Molecular determinants of voltage-dependent human ether-a-go-go physiological temperature. Biophys J 1998; 74:230–41 related gene (HERG) Kϩ channel block. J Biol Chem 2002; 277:23587–95 28. Sesti F, Abbott GW, Wei J, Murray KT, Saksena S, Schwartz PJ, Priori SG, 19. Sanchez-Chapula JA, Ferrer T, Navarro-Polanco RA, Sanguinetti MC: Volt- Roden DM, George Jr, AL Goldstein SA: A common polymorphism associated age-dependent profile of human ether-a-go-go-related gene channel block is with antibiotic-induced cardiac arrhythmia. Proc Natl Acad SciUSA2000; influenced by a single residue in the S6 transmembrane domain. Mol Pharmacol 97:10613–8 2003; 63:1051–8 29. Wang S, Morales MJ, Liu S, Strauss HC, Rasmusson RL: Modulation of HERG 20. Milnes JT, Crociani O, Arcangeli A, Hancox JC, Witchel HJ: Blockade of affinity for E-4031 by [Kϩ]o and C-type inactivation. FEBS Lett 1997; 417:43–7 HERG potassium currents by fluvoxamine: Incomplete attenuation by S6 muta- 30. Numaguchi H, Mullins FM, Johnson Jr JP Johns, DC, Po SS, Yang IC, tions at F656 or Y652. Br J Pharmacol 2003; 139:887–98 Tomaselli GF, Balser JR: Probing the interaction between inactivation gating and 21. Scholz EP, Zitron E, Kiesecker C, Lueck S, Kathofer S, Thomas D, Weretka Dd-sotalol block of HERG. Circ Res 2000; 87:1012–8 S, Peth S, Kreye VA, Schoels W, Katus HA, Kiehn J, Karle CA: Drug binding to 31. Friederich P, Urban BW: Interaction of intravenous anesthetics with hu- aromatic residues in the HERG channel pore cavity as possible explanation for man neuronal potassium currents in relation to clinical concentrations. ANESTHE- Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/1/102/358525/0000542-200507000-00017.pdf by guest on 24 September 2021 acquired Long QT syndrome by antiparkinsonian drug budipine. Naunyn SIOLOGY 1999; 91:1853–60 Schmiedebergs Arch Pharmacol 2003; 368:404–14 32. Friederich P, Benzenberg D, Trellakis S, Urban BW: Interaction of volatile 22. Mitcheson JS. Drug binding to HERG channels: Evidence for a “non- anesthetics with human Kv channels in relation to clinical concentrations. ANES- aromatic” binding site for fluvoxamine. Br J Pharmacol 2003; 139:883–4 THESIOLOGY 2001; 95:954–8 23. Valenzuela C, Delpon E, Tamkun MM, Tamargo J, Snyders DJ: Stereoselec- 33. Franqueza L, Longobardo M, Vicente J, Delpon E, Tamkun MM, Tamargo tive block of a human cardiac potassium channel (Kv1.5) by bupivacaine enan- J, Snyders DJ, Valenzuela C: Molecular determinants of stereoselective bupiva- tiomers. Biophys J 1995; 69:418–27 caine block of hKv1.5 channels. Circ Res 1997; 81:1053–64 24. Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ: Improved patch- 34. Albright GA: Cardiac arrest following regional anesthesia with etidocaine clamp techniques for high-resolution current recording from cells and cell-free or bupivacaine. ANESTHESIOLOGY 1979; 51:285–7 membrane patches. Pflugers Arch 1981; 391:85–100 35. Polley LS, Santos AC: Cardiac arrest following regional anesthesia with 25. Smith PL, Baukrowitz T, Yellen G: The inward rectification mechanism of ropivacaine: Here we go again! ANESTHESIOLOGY 2003; 99:1253–4 the HERG cardiac potassium channel. Nature 1996; 379:833–6 36. Webster R, Leishman D, Walker D: Towards a drug concentration effect 26. Hancox JC, Levi AJ, Witchel HJ: Time course and voltage dependence of relationship for QT prolongation and torsades de pointes. Curr Opin Drug Discov expressed HERG current compared with native “rapid” delayed rectifier K cur- Dev 2002; 5:116–26

Anesthesiology, V 103, No 1, Jul 2005