- Email: [email protected]
MiRP1T8A
British Journal of Anaesthesia 92 (1): 93±101 (2004)
DOI: 10.1093/bja/aeh026
LABORATORY INVESTIGATIONS Local anaesthetic sensitivities of cloned HERG channels from human heart: comparison with HERG/MiRP1 and HERG/MiRP1T8A P. Friederich1*, A. Solth1 2, S. Schillemeit2 and D. Isbrandt2 1
Department of Anaesthesiology, University Hospital Hamburg Eppendorf, Martinistr. 52, D-20251 Hamburg, Germany. 2Institute of Neural Signal Transduction, Center for Molecular Neurobiology, University of Hamburg, Germany
Methods. Whole-cell patch-clamp recordings and site-directed mutagenesis were combined to compare local anaesthetic sensitivities of cloned and mutated human potassium channel subunits. The ion channels were activated by a protocol that approximated ventricular action potentials. Results. The amide local anaesthetics bupivacaine, levobupivacaine and ropivacaine inhibited HERG channels at toxicologically relevant concentrations, with IC50 values of 20 (SEM 2) mM (n=29), 10 (1) mM (n=40) and 20 (2) mM (n=49), respectively. Hill coef®cients were close to unity. There were no indications of qualitative differences in channel inhibition between the three anaesthetics. The putative subunit MiRP1 did not alter local anaesthetic sensitivity of HERG channels. The common single nucleotide polymorphism producing MiRP1T8A did not increase local anaesthetic sensitivity of HERG/MiRP1 channels. Conclusions. Amide local anaesthetics target HERG and HERG/MiRP1 channels with identical potency. The effects on these ion currents may signi®cantly contribute to local-anaestheticinduced cardiac arrhythmia. MiRP1T8A does not seem to confer an increased risk of severe cardiac side-effects to carriers of this common polymorphism. Br J Anaesth 2004; 92: 93±101 Keywords: anaesthetics local; anaesthetics, toxicity; complications, arrhythmia Accepted for publication: July 21, 2003
Accidental intravascular injection of long-acting amide local anaesthetics such as bupivacaine may cause prolongation of ventricular action potentials, monitored as prolongation of the QT interval, and severe ventricular arrhythmia known as torsades de pointes.1 2 Prolongation of the QT interval and the induction of torsades de pointes ventricular arrhythmia are well known clinical manifestations of the long QT syndrome.3 The inherited form of this disease may be associated with mutations in several cardiac ion channel
genes such as the HERG (human ether-aÁ-go-go) gene.4 5 HERG channels underlie the repolarizing cardiac potassium current IKr.4 6 They are critical for the maintenance of normal rhythm in the human heart.4 Mutations in the HERG gene may lead to dysfunctional HERG channels and a reduced IKr. This correlates with the prolongation of ventricular action potentials, as well as an increase in susceptibility to ventricular arrhythmia.7 Inhibition of HERG channels has also been associated with drug-induced
Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2004
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
*Corresponding author. E-mail: [email protected] Background. Myocardial potassium channels are complexes formed by different subunits. The subunit composition may in¯uence the cardiotoxic action of local anaesthetics. The effects of amide local anaesthetics on HERG channels co-expressed with the putative subunit MiRP1 have not been established. It is also unclear if the common polymorphism MiRP1T8A that predisposes individuals to drug-induced cardiac arrhythmia increases local-anaesthetic sensitivity of HERG/MiRP1 channels. This may suggest the presence of genetic risk factors for local-anaesthetic-induced cardiac arrhythmia.
Friederich et al.
long QT syndrome and ventricular ®brillation.8 Long-acting amide local anaesthetics may therefore exhibit cardiotoxic effects in part by inhibiting HERG channels.9±11 In human myocardium HERG channels may associate with minKrelated peptide 1 (MiRP1) to form IKr.12 13 This auxillary subunit encoded in KCNE2 has been reported to alter the pharmacological sensitivity of HERG channels.12 Several mutations in hKCNE2, as well as the common T8A polymorphisms of MiRP1, have been shown to predispose individuals to drug-induced cardiac arrhythmia.12 14 KCNE2 may thus also constitute a susceptibility gene for localanaesthetic-induced ventricular arrhythmia. Myocardial potassium channels are complexes formed by different subunits.15 The subunit composition may in¯uence the effects of local anaesthetics on these membrane proteins.12 16 The effects of amide local anaesthetics on complexes formed by HERG and MiRP1 remain to be compared. It is therefore unclear if co-expression of HERG and MiRP1 alters the local anaesthetic sensitivity of HERG channels. If this were the case, MiRP1T8A might confer an increased sensitivity to HERG/MiRP1 channels, which may be indicative of an additional risk for carriers to develop local-anaesthetic-induced cardiac arrhythmia. Furthermore, it may suggest a role for genetic testing in the prevention of local-anaesthetic-induced cardiac arrhythmia. This study was designed to investigate if the local anaesthetic sensitivity of HERG channels is altered by the presence of MiRP1 or MiRP1T8A.
CTACTTTATCCAATTTCGCACAGACGCTGGAAGACGTCTTCCGAAG-3¢) and the pcDNA6-BGH primer (Invitrogen, Groningen, The Netherlands). Successful mutagenesis and the MiRP1T8A cDNA sequence were veri®ed by sequencing. The respective cDNAs were cloned into pcDNA6 expression vectors for transfection of CHO cells.
Transfection of CHO cells
Electrophysiology of potassium channels Whole-cell currents were measured using the patch clamp method18 with an EPC-9 ampli®er and Pulse software version 8.11 (HEKA Elektronik, Lambrecht, Germany). Patch electrodes with a pipette resistance of 2.0±4.0 MW were pulled from borosilicate glass capillary tubes (World Precision Instruments, Saratoga, USA) and ®lled with a solution comprising 160 mM KCl, 0.5 mM MgCl2, 10 mM HEPES, 2 mM Na-ATP (all from Sigma, Deissenhofen, Germany), adjusted to pH 7.2 with KOH. The external solution applied to the cells was 135 mM NaCl, 5 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 5 mM HEPES, 10 mM sucrose, phenol red 0.1 mg ml±1 (all from Sigma, Deissenhofen, Germany) adjusted to pH 7.4 with NaOH. The capacitance of the cells was 18.2 (SD 5.7) pF (n=20). Series resistance was 2.5±7.5 MW and was actively compensated for by at least 80%. A leak subtraction protocol was not used in this study. The data were therefore only recorded from cells in which tight mega-ohm or giga-ohm seals had been formed. As leak currents tend to increase during the course of an experiment, we only analysed data that did not show more than a 6% difference in any tested parameter between control and wash-out conditions. The liquid junction potential was calculated according to the Henderson equation.19 It accounted for 3.86 mV and was not corrected for. Bupivacaine (Sigma, Deissenhofen, Germany), levobupivacaine (AstraZeneca, SoÈdertalje, Sweden) and ropivacaine (AstraZeneca, SoÈdertalje, Sweden) were dissolved in the extracellular solution. The drugs were superfused onto the cells by a hydrostatically driven perfusion system. We studied the activity of cardiac potassium channels using a pulse protocol that approximates a cardiac ventricular action
Methods Cell culture Chinese hamster ovary (CHO) cells stably or transiently transfected with cloned human potassium channels were grown as non-con¯uent monolayers in MEM Eagle's Alpha medium (Life Technologies, Paisley, Scotland) containing 10% fetal calf serum (Biochrom, Berlin, Germany), penicillin 100 IU ml±1 and streptomycin 100 mg ml±1 (Life Technologies, Paisley, Scotland). The cells were cultured in 50 ml ¯asks (NUNC, Roskilde, Denmark) at 37°C in a humidi®ed atmosphere (carbon dioxide 5%). Cells were subcultured in 35 mm diameter monodishes (NUNC, Roskilde, Denmark) for at least 24 h before electrophysiological experiments were performed.
Cloning of ion channel subunits and cDNAcontaining expression vectors and in vitro mutagenesis Fragments encoding the complete open reading frames of the KCNH2 or KCNE2 genes were ampli®ed from human cardiac cDNA by the reverse transcriptase polymerase chain reaction (PCR) employing primers speci®c for the particular gene.17 The amino acid exchange T8A was introduced into KCNE2 by PCR using the primer KCNE2-T8A (5¢-ATGT94
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
CHO cells were seeded in 35 mm diameter monodishes 1 day before transfection. Transfection was performed using 3.5 ml of lipofectamine reagent (Gibco-BRL, Rockville, USA) according to the manufacturer's protocol and 1.0±2.0 mg of the respective ion channel plasmid DNA. Coexpression of HERG channels with MiRP1 and MiRP1T8A within the cell membrane was veri®ed using internal ribosomal entry site eGFP constructs of MiRP1 and MiRP1T8A, respectively. Only those cells transfected with HERG/MiRP1 or HERG/MiRP1T8A that emitted green ¯uorescence when stimulated by ultraviolet light were investigated.
Local anaesthetic sensitivities of cloned HERG channels
potential.20 The potassium channels were activated from a holding potential of ±80 mV by a step depolarization of the cell membrane to +60 mV and subsequent repolarization within 1000 ms back to the holding potential of ±80 mV (ramp protocol). The ramp protocol was followed by a 30 ms depolarization of the membrane to a test potential of +40 mV.17 Repetitive ramps were applied to determine that steady-state inhibition was reached. The voltage dependence of local anaesthetic action on HERG channels was investigated after a 2 s depolarizing test pulse to +60 mV by stepping the membrane potential for 0.5 s to potentials of ±120 mV, ±40 mV and +40 mV. Experiments were performed at room temperature. The recorded signal was ®ltered at 2 kHz and stored with a sampling rate of 5 kHz on a personal computer for later analysis.
Data analysis The current±time relationship observed with the ramp protocol for cardiac potassium channel activation was used to quantify the charge crossing the membrane during the ramp protocol (Qramp). Similarly, the current±time relationship observed during the step protocol for channel activation was used to quantify the charge crossing the membrane during the step protocol (Qstep). Inhibition of currents evoked by the ramp protocol was measured as inhibition of the maximal outward current and as reduction of Qramp. Inhibition of the current during the step protocol was analysed as a reduction of Qstep. The charge crossing the membrane is equivalent to the time integrals of current traces and was determined using Pulse software version 8.11 (HEKA Elektronik, Lambrecht, Germany). Inhibition was quanti®ed by the ratio of time integrals of current traces in the presence of the drug to the mean of the time integrals before application and after wash out of the drug (inhibition=1± Qdrug/[(Qcontrol+Qwash out)/2]) and by the ratio of the size of the current in the presence of the drug to the mean of the size of the current before application and after wash out of the drug (inhibition=1±Imax drug/[(Imax control+Imax wash out)/2]). The concentration±response curves were ®tted to the Hill equation (e/emax=cg/[IC50g+cg], where e=effect, emax=maximal effect, c=anaesthetic concentration, g=Hill coef®cient and IC50=concentration of half-maximal effect) using nonlinear regression. Regression analysis was performed using Kaleidagraph (Syngery Software, Reading, Pennsylvania, USA). Data points are given as mean (SD) unless stated otherwise. Statistical analysis was performed with an unpaired Student's t-test; n is the number of experiments.
Local anaesthetic sensitivity of HERG The current traces in Figure 3A demonstrate the effects of bupivacaine, levobupivacaine and ropivacaine on HERG currents after stimulation by the ramp protocol. The drugs reversibly inhibited HERG currents in a concentrationdependent manner. Inhibition was immediate and resulted from the wash in of the drugs. Wash out occurred within the same time scale as the wash in (5±10 s). The inhibitory effects were quanti®ed as a reduction of Qramp. Concentration-dependent reduction of Qramp was described
Results Properties of HERG, HERG/MiRP1 and HERG/ MiRP1T8A evoked by ramp protocols HERG outward currents were evoked from a holding potential of ±80 mV by a ramp protocol descending over 95
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
1000 ms from a membrane potential of +60 mV back to the holding potential of ±80 mV (Fig. 1A). This protocol was chosen because it approximates cardiac ventricular action potentials.20 Activation of HERG channels by the ramp protocol resulted in a bell-shaped current response, with an Imax of 0.5 (0.2) nA (n=20) at a membrane potential of ±40 mV. The size of Imax corresponded to a current density of 32 (12) pA/pF (Fig. 1B). Qramp was 356 (107) pC (n=20) (Fig. 1C). The time of the ramp protocol necessary to evoke peak HERG current was 711 (41) ms (n=15) (Fig. 1D). The current responses of HERG/MiRP1 and HERG/ MiRP1T8A were also bell-shaped, with Imax values of 0.7 (0.5) nA (n=10) and 0.8 (0.4) (n=10), respectively. The size of Imax corresponded to current densities of 35 (11) pA/pF in the case of HERG/MiRP1 and 33 (12) in the case of HERG/ MiRP1T8A (Fig. 1B). Qramp was 376 (203) pC for HERG/ MiRP1 and 397 (183) for HERG/ MiRP1T8A (Fig. 1C). The time to peak current response did not differ between HERG, HERG/MiRP1 and HERG/MiRP1T8A channels (Fig. 1D). The ramp was followed by an extra 30 ms test pulse to +40 mV (Fig. 2A). This additional depolarizing step assays an important property of HERG channels.21 Deactivation of HERG channels is much slower than their recovery from inactivation. This additional depolarizing step evoked rapidly inactivating outward currents passing through HERG, HERG/MiRP1 and HERG/ MiRP1T8A. The densities of the currents were 120 (46) pA/pF (n=18), 114 (67) pA/pF (n=9) and 114 (36) pA/pF (n=10), respectively (Fig. 2B). The current densities did not differ signi®cantly. The time course of inactivation of HERG, HERG/MiRP1 and HERG/ MiRP1T8A was adequately ®tted by a single exponential function with time constants (t) of 3.0 (0.4) ms (n=14), 3.1 (0.4) ms (n=9) and 3.0 (0.6) ms (n=9), respectively (Fig. 2C). The ratio of the size of the current in response to the step depolarization to the preceding ramp protocol (Imax step/Imax ramp) was 3.4 (0.8) (n=5), 3.3 (1.3) (n=6) and 3.6 (0.7) (n=13) for HERG, HERG/MiRP1 and HERG/ MiRP1T8A channels, respectively. The ratios did not differ between these channels (Fig. 2D). The ramp protocol combined with the step depolarization highlights two important physiological functions of cardiac HERG channels.17 We used this protocol to analyse local anaesthetic action on HERG, HERG/MiRP1 and HERG/MiRP1T8A channels.
Friederich et al.
Under control conditions Imax was reached after 723 (SD 36) ms of the test pulse (mean of control and wash out, n=19). The occurrence of Imax was signi®cantly delayed by bupivacaine 30 mM, to 766 (33) ms (n=7, P<0.05), levobupivacaine 10 mM, to 796 (49) ms (n=6, P<0.05) and ropivacaine 25 mM, to 803 (10) ms (n=6, P<0.05). As a consequence, inhibition of HERG channels was time dependent. Relating the size of Imax under control conditions to the size of the current at the same time of the ramp protocol during the drug effect yielded an inhibition of Imax by bupivacaine, levobupivacaine and ropivacaine of 60 (4)% (n=7), 48 (6)% (n=6) and 53 (4)% (n=6), respectively. Relating the size of Imax during drug action to the size of the current at the same time of the ramp protocol under control conditions yielded a signi®cantly smaller inhibition (P<0.05) by all three local anaesthetics: 55 (3)% by bupivacaine (n=7), 39 (3)% by levobupivacaine (n=6) and
mathematically by ®tting Hill functions to the concentration±response data (Fig. 3B, Table 1). The mean IC50 values of bupivacaine, levobupivacaine and ropivacaine were 20 (SEM 2) mM (n=29), 10 (1) mM (n=40) and 20 (2) mM (n=49), respectively (Fig. 3B, Table 1). The Hill coef®cients were close to unity (Table 1). Inhibition of HERG channels was also quanti®ed as inhibition of Imax. The IC50 values for inhibition of Imax were essentially identical to those for reduction of Qramp: 21 (SEM 1) mM (n=29), 11 (1) mM (n=40) and 22 (1) mM (n=56) for bupivacaine, levobupivacaine and ropivacaine, respectively. The mean Hill coef®cients were again close to unity: 0.95 (SEM 0.03), 0.95 (0.1) and 1.0 (0.02) for bupivacaine, levobupivacaine and ropivacaine, respectively. All three anaesthetics in¯uenced the shape of the current trace in response to the 1000 ms ramp protocol (Fig. 3A).
96
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
Fig 1 (A) HERG currents in response to a ramp protocol. Original current traces and the stimulation protocol are shown. Activation of these ion channels by the ramp protocol resulted in a bell-shaped current response. (B) The current densities did not differ between HERG, HERG/MiRP1 and HERG/MiRP1T8A channels expressed in Chinese hamster ovary cells. (C) The currents passing through the channels were quanti®ed as charge ¯ow (area under the current±time curve). Charge ¯ow did not differ between HERG, HERG/MiRP1 and HERG/MiRP1T8A channels. (D) MiRP1 and MiRP1T8A did not in¯uence the time to peak current response of the HERG channels. Data are mean and SD.
Local anaesthetic sensitivities of cloned HERG channels
43 (7)% by ropivacaine (n=6). Accordingly, the shape of the remaining HERG current trace after drug application was altered. Inhibition of HERG was voltage dependent, as assayed by a deactivating pulse protocol. Bupivacaine 30 mM inhibited the maximal currents by 40 (6)% at ±120 mV (n=7), by 51 (3)% at ±40 mV (n=7) and by 49 (5)% at +40 mV (n=7). Inhibition was signi®cantly different between membrane potentials of ±120 mV and ±40 (P<0.05) and between membrane potentials of ±120 mV and +40 mV (P<0.05) but was not signi®cantly different between membrane potentials of ±40 mV and +40 mV. The inhibitory effects of the local anaesthetics on HERG currents evoked by the step depolarization following the
ramp was also reversible and concentration dependent. The effects were quanti®ed as a reduction of Qstep. Fitting Hill functions to the concentration±response data (Fig. 3C) yielded IC50 values for the reduction of Qstep 2±3 times higher than for the inhibition of Qramp. Hill coef®cients were again close to unity (Table 1).
Local anaesthetic sensitivity of HERG/MiRP1 and HERG/MiRP1T8A The in¯uence of MiRP1 on local anaesthetic sensitivity of HERG channels and the in¯uence of the T8A polymorphism of KCNE2 on local anaesthetic sensitivity of IKr were 97
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
Fig 2 (A) HERG/MiRP1 currents in response to a ramp protocol followed by a rectangle depolarization. Original current traces and the stimulation protocol are shown. Activation of these ion channels by the additional depolarizing step evoked inactivating outward currents. (B) The current densities of the inactivating outward currents did not differ between HERG, HERG/MiRP1 and HERG/MiRP1T8A channels expressed in Chinese hamster ovary cells. (C) The time course of inactivation of HERG, HERG/MiRP1 and HERG/ MiRP1T8A channels was adequately ®tted by a single exponential function. The time constants of channel inactivation did not differ signi®cantly. (D) Current amplitudes evoked by the step depolarization were threefold larger than those evoked by the ramp protocol. MiRP1 and MiRP1T8A did not in¯uence the ratio of peak current responses of the HERG channels. Data are mean and SD.
Friederich et al.
MiRP1T8A (Qstep) accounted for 33 (13)% (n=6), 26 (9)% (n=13) and 29 (5)% (n=9) by bupivacaine, levobupivacaine and ropivacaine, respectively (Fig. 4C). The degree of inhibition of HERG/MiRP1T8A was not signi®cantly different from that seen for HERG/MiRP1 channels.
investigated at concentrations close to the IC50 values for inhibition of HERG channels (Qramp) of all three local anaesthetics. If MiRP1 or MiRP1T8A alter the local anaesthetic sensitivities of HERG channels, this effect can best be detected at the steepest part of the concentration±response curve for inhibition of HERG channels. All three anaesthetics reversibly inhibited both HERG/MiRP1 and HERG/ MiRP1T8A channels (Fig. 4A). Inhibition of HERG/MiRP1 (Qramp, Fig. 4B) accounted for 55 (3)% (n=10) by bupivacaine 30 mM, 48 (6)% (n=7) by levobupivacaine 10 mM and 49 (8)% (n=12) by ropivacaine 25 mM. Inhibition of Qstep (Fig. 4C) accounted for 37 (6)% (n=5), 27 (8)% (n=4) and 29 (9)% (n=8) by bupivacaine, levobupivacaine and ropivacaine, respectively. The extent of inhibition of HERG/MiRP1 was not signi®cantly different from that of HERG channels (P>0.05). Inhibition of HERG/MiRP1T8A (Qramp) accounted for 55 (6)% (n=7), 43 (5)% (n=14), and 50 (6)% (n=9) by bupivacaine, levobupivacaine and ropivacaine (Fig. 4B). The inhibitory effect on HERG/
Discussion
Fig 3 (A) HERG current traces after activation by ramp protocols (shown beneath the current traces). Currents are shown under control conditions and with bupivacaine 30 mM (left), levobupivacaine 10 mM (middle), ropivacaine 25 mM (right), and after wash out. Local anaesthetics reduced the size of the time integrals of current traces (Qramp), inhibited the maximal outward current (Imax) and increased the time needed to reach the peak current. (B) Concentration±response curves for the inhibition of Qramp by bupivacaine, levobupivacaine and ropivacaine. The rank order of potency for the inhibition of Qramp was levobupivacaine > bupivacaine = ropivacaine, with pIC50 values of 4.7 M, 5 M and 4.7 M, respectively. (C) The concentration±response curve for the reduction of Qstep by bupivacaine, levobupivacaine and ropivacaine. Local anaesthetics reduced the size of the time integrals of the current traces (Qstep). pIC50 values were 4.4 M, 4.5 M and 4.3 M, respectively (see Table 1).
98
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
The present work explored the pharmacological sensitivity of cloned repolarizing potassium channel complexes from human cardiac ventricle to the action of three clinically used amide local anaesthetics. The cardiac ion channels were activated by a protocol that approximates ventricular action potentials.20 This protocol has previously been used to study drug action on HERG and HERG/MiRP114 and facilitates interpretation of the clinical signi®cance of our results. The responses of the channels to the ramp protocol were in agreement with predictions derived from their respective roles during ventricular action potentials.7 The ramp
Local anaesthetic sensitivities of cloned HERG channels
protocol was followed by a sudden depolarization, highlighting the capability of these ion channels to counteract Table 1 Hill parameters for bupivacaine, levobupivacine and ropivacaine derived from concentration±response curves for the inhibition of HERG channels. The channels were activated with a ramp protocol (Qramp) followed by a step depolarization (Qstep). IC50=concentration at half-maximal effect. Data are mean (SEM).
Inhibition Qramp Number of experiments Maximum effect (%) Hill coef®cient IC50 (mM) Inhibition Qstep Number of experiments Maximum effect (%) Hill coef®cient IC50 (mM)
Bupivacaine
Levobupivacaine
Ropivacaine
29 94 (2) 0.95 (0.08) 20 (2)
40 91 (2) 0.97 (0.09) 10 (1)
49 93 (2) 0.99 (0.07) 20 (2)
34 90 (5) 1.14 (0.18) 43 (8)
28 94 (7) 0.75 (0.14) 32 (10)
26 93(5) 1.01 (0.12) 50 (8)
Fig 4 (A) HERG/MiRP1T8A current traces after activation by ramp protocols (shown beneath the current traces). Currents are shown under control conditions and with bupivacaine 30 mM (left), levobupivacaine 10 mM (middle), ropivacaine 25 mM (right) and after wash out of the drugs. Local anaesthetics reduced the size of the time integrals of current traces (Qramp), inhibited the maximal outward current (Imax) and increased the time needed to reach peak current. (B) Comparison of inhibition (Qramp) of HERG, HERG/MiRP1 and HERG/MiRP1T8A by bupivacaine 30 mM, levobupivacaine 10 mM and ropivacaine 25 mM. Inhibition of these channel complexes by local anaesthetics was not signi®cantly different between HERG, HERG/MiRP1 and HERG/MiRP1T8A. (C) Comparison of inhibition (Qstep) of HERG, HERG/MiRP1 and HERG/MiRP1T8A by bupivacaine 30 mM, levobupivacaine 10 mM and ropivacaine 25 mM. Inhibition of these channel complexes by all three anaesthetics was not signi®cantly different between HERG, HERG/MiRP1 and HERG/MiRP1T8A.
99
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
sudden depolarizations such as those observed during ventricular extrasystoles.17 The design of our study thus allowed investigation of local anaesthetic action on HERG, HERG/MiRP1, and HERG/MiRP1T8A by a protocol that accounts for the complex biophysical response of these ion channels during ventricular action potentials.20 Establishing drug sensitivity of cardiac potassium channels is a common and widely accepted way of testing the cardiac safety of pharmacological agents.7 8 22 Impairment of repolarizing ventricular currents and suppression of ventricular currents that counterbalance sudden depolarizations may render the myocardium susceptible to severe arrhythmia.11 Possible modulating effects of genetic variations in ion channel subunits on local anaesthetic sensitivity have not been reported previously. The data presented here therefore provide information on the local anaesthetic sensitivity of an important cardiac potassium channel, the cardiac safety of clinically used local anaesthetics and the
Friederich et al.
The concentration of local anaesthetic needed to halfmaximally inhibit HERG channels was used to compare the sensitivity of HERG, HERG/MiRP1 and HERG/ MiRP1T8A to local anaesthetics. Co-expression of HERG with MiRP1 did not alter the sensitivity of HERG channels. The interaction of the auxiliary subunit MiRP1 with HERG channels consequently does not perturb the binding of amide local anaesthetics to HERG channels. The same lack of difference in local anaesthetic sensitivity between HERG and HERG/MiRP1 also holds true for the comparison of HERG/MiRP1 with HERG/MiRP1T8A. The T8A single nucleotide polymorphism (SNP) occurs in 1.6% of healthy Caucasians14 and has been suggested to constitute an important risk factor for drug-induced ventricular arrhythmia.14 However, our results do not provide any evidence that this SNP may render HERG channels more susceptible to the pharmacological action of local anaesthetics. Our experimental results do not suggest a role for genetic testing of this common MiRP1 gene polymorphism in the prevention of local-anaestheticinduced cardiac arrhythmia. In summary, the amide local anaesthetics bupivacaine, levobupivacaine and ropivacaine inhibited cloned repolarizing ventricular potassium channels from human heart at toxicologically relevant concentrations. All three anaesthetics reduced outward currents in response to a simulated action potential and impaired the capability of these cardiac ion channels to respond to a sudden depolarization. There were no indications of qualitative differences in channel inhibition between bupivacaine, levobupivacaine and ropivacaine. The sensitivity of HERG channels was not altered by coexpression with MiRP1, and the common SNP producing MiRP1T8A did not in¯uence local anaesthetic sensitivity of HERG/MiRP1 channels. As drug-induced inhibition of repolarizing potassium channels from cardiac ventricles is a well-known cause of severe cardiac arrhythmia and sudden death, our results support the notion that none of the local anaesthetics investigated may be considered as safe with regard to cardiotoxic side-effects. MiRP1T8A seems not to confer an increased risk of these severe cardiac side-effects to carriers of this common polymorphism.
Acknowledgements We thank technical assistants Annerose Schmidt-Darlison and Andrea Zaisser at the Institute of Neural Signal Transduction for skilful cell culture and cell transfection. We are grateful to O. Pongs PhD, Director of the Institute of Neural Signal Transduction, for his generous support and for critically reading the manuscript. Levobupivacaine and ropivacaine were kindly provided by AstraZeneca, SoÈdertalje, Sweden. This study was supported by grants from the Fresenius Stiftung, Bad Homburg, Germany (PF), the Deutsche Forschungsgemeinschaft, Bonn, Germany (FR-1625±1/ 1), and the Foundation Leducq, Paris, France (DI).
100
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
in¯uence of a proposed arrhythmia susceptibility polymorphism14 on local anaesthetic sensitivity of a cardiac potassium channel complex that has repeatedly been associated with drug-induced ventricular arrhythmia.12 14 Bupivacaine, levobupivacaine and ropivacaine reduced the charge crossing the membrane through activated HERG channels in a concentration-dependent manner. The drugs exhibited two effects on Imax of HERG channels activated by the simulated action potential: they suppressed the maximal amplitude of the current and they delayed the occurrence of Imax. Both effects will decrease the repolarizing outward current during phases 3 and 4 of the cardiac action potential. Bupivacaine, levobupivacaine and ropivacaine also impaired the capability of HERG channels to conduct outward currents in response to a sudden depolarization. This effect was also concentration dependent and reversible. The difference in inhibition between ±120 mV and ±40 mV and between ±120 mV and +40 mV would be compatible with a voltage-dependent block. However, inhibition of HERG by bupivacaine did not differ between ±40 mV and +40 mV. It may thus be postulated that the increased driving force acting on the drug molecule at higher membrane depolarizations may be offset by conformational changes of the ion channel protein such as that occurring during channel inactivation. The direction of HERG current ¯ux also differs between currents measured at ±120 mV and ±40 mV. It may therefore be hypothesized that the potassium ions passing through the ion channel pore may interfere with the action of the local anaesthetic on the HERG channel protein. In any case the magnitude of the inhibitory effect is largest at potentials where HERG channels pass the largest currents during the ventricular action potential. Both the prolongation of phases 3 and 4 of the ventricular action potential and a reduced ability to counteract sudden depolarizations are likely to facilitate the occurrence of clinically observed ventricular arrhythmia during local anaesthetic intoxication. Further study is needed to characterize the precise molecular mechanism of the drug±channel interaction. The results of our study demonstrate that all three drugs are potent inhibitors of human ventricular potassium channels at concentrations that induce polymorphic ventricular arrhythmia in vivo.2 Our results obtained with a stimulation protocol that approximates ventricular action potentials support the notion11 that HERG channels constitute cardiac targets relevant for the induction of ventricular arrhythmia by local anaesthetics. As inhibition of repolarizing potassium channel complexes in human myocardium will prolong the duration of ventricular action potentials, the effects of local anaesthetics on these potassium channels will also in¯uence local anaesthetic action on cardiac sodium channels. Prolongation of the action potential will reduce sodium channel availability by increasing the fraction of inactivated sodium channels. This increased inactivation will add to the direct inhibitory effect of local anaesthetics on human heart sodium channels.23
Local anaesthetic sensitivities of cloned HERG channels
References
101
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
1 Kasten GW. Amide local anesthetic alterations of effective refractory period temporal dispersion: relationship to ventricular arrhythmias. Anesthesiology 1986; 65: 61±6 2 Chang DH, Ladd LA, Copeland S, Iglesias MA, Plummer JL, Mather LE. Direct cardiac effects of intracoronary bupivacaine, levobupivacaine and ropivacaine in the sheep. Br J Pharmacol 2001; 132: 649±58 3 Roden DM, Lazzara R, Rosen M, Schwartz PJ, Towbin J, Vincent GM. Multiple mechanisms in the long-QT syndrome. Current knowledge, gaps, and future directions. The SADS Foundation Task Force on LQTS. Circulation 1996; 94: 1996±2012 4 Sanguinetti MC, Jiang C, Curran ME, Keating MT. A mechanistic link between an inherited and an acquired cardiac arrhythmia: HERG encodes the IKr potassium channel. Cell 1995; 81: 299± 307 5 Splawski I, Shen J, Timothy KW, Vincent GM, Lehmann MH, Keating MT. Genomic structure of three long QT syndrome genes: KVLQT1, HERG, and KCNE1. Genomics 1998; 51: 86±97 6 Trudeau MC, Warmke JW, Ganetzky B, Robertson GA. HERG, a human inward recti®er in the voltage-gated potassium channel family. Science 1995; 269: 92±5 7 Keating MT, Sanguinetti MC. Molecular and cellular mechanisms of cardiac arrhythmias. Cell 2001; 104: 569±80 8 Mitcheson JS, Chen J, Lin M, Culberson C, Sanguinetti MC. A structural basis for drug-induced long QT syndrome. Proc Natl Acad Sci USA 2000; 97: 12329±33 9 Lipka LJ, Jiang M, Tseng GN. Differential effects of bupivacaine on cardiac K channels: role of channel inactivation and subunit composition in drug-channel interaction. J Cardiovasc Electrophysiol 1998; 9: 727±42 10 Gonzalez T, Longobardo M, Caballero R, Delpon E, Tamargo J, Valenzuela C. Effects of bupivacaine and a novel local anesthetic, IQB-9302, on human cardiac K+ channels. J Pharmacol Exp Ther 2001; 296: 573±83 11 Gonzalez T, Arias C, Caballero R, et al. Effects of levobupivacaine, ropivacaine and bupivacaine on HERG channels: stereoselective bupivacaine block. Br J Pharmacol 2002; 137: 1269±79
12 Abbott GW, Sesti F, Splawski I, et al. MiRP1 forms IKr potassium channels with HERG and is associated with cardiac arrhythmia. Cell 1999; 97: 175±87 13 Mazhari R, Greenstein JL, Winslow RL, Marban E, Nuss HB. Molecular interactions between two long-QT syndrome gene products, HERG and KCNE2, rationalized by in vitro and in silico analysis. Circ Res 2001; 89: 33±8 14 Sesti F, Abbott GW, Wei J, et al. A common polymorphism associated with antibiotic-induced cardiac arrhythmia. Proc Natl Acad Sci USA 2000; 97: 10613±18 15 Roden DM, Balser JR, George AL Jr, Anderson ME. Cardiac ion channels. Annu Rev Physiol 2002; 64: 431±75 16 Gonzalez T, Navarro-Polanco R, Arias C, et al. Assembly with the Kvbeta1.3 subunit modulates drug block of hKv1.5 channels. Mol Pharmacol 2002; 62: 1456±63 17 Isbrandt D, Friederich P, Solth A, et al. A novel mutation in the KCNE2 (MiRP1) gene that alters HERG channel kinetics. J Mol Med 2002; 80: 524±32 18 Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. P¯ugers Arch 1981; 391: 85±100 19 Barry PH, Lynch JW. Liquid junction potentials and small cell effects in patch-clamp analysis. J Membr Biol 1991; 121: 101±17 20 Hancox JC, Levi AJ, Witchel HJ. Time course and voltage dependence of expressed HERG current compared with native `rapid' delayed recti®er K current during the cardiac ventricular action potential. P¯ugers Arch 1998; 436: 843±53 21 Smith PL, Baukrowitz T, Yellen G. The inward recti®cation mechanism of the HERG cardiac potassium channel. Nature 1996; 379: 833±6 22 Weerapura M, Nattel S, Chartier D, Caballero R, Hebert TE. A comparison of currents carried by HERG, with and without coexpression of MiRP1, and the native rapid delayed recti®er current. Is MiRP1 the missing link? J Physiol 2002; 540: 15±27 23 Nau C, Wang SY, Strichartz GR, Wang GK. Block of human heart hH1 sodium channels by the enantiomers of bupivacaine. Anesthesiology 2000; 93: 1022±33
DOI: 10.1093/bja/aeh026
LABORATORY INVESTIGATIONS Local anaesthetic sensitivities of cloned HERG channels from human heart: comparison with HERG/MiRP1 and HERG/MiRP1T8A P. Friederich1*, A. Solth1 2, S. Schillemeit2 and D. Isbrandt2 1
Department of Anaesthesiology, University Hospital Hamburg Eppendorf, Martinistr. 52, D-20251 Hamburg, Germany. 2Institute of Neural Signal Transduction, Center for Molecular Neurobiology, University of Hamburg, Germany
Methods. Whole-cell patch-clamp recordings and site-directed mutagenesis were combined to compare local anaesthetic sensitivities of cloned and mutated human potassium channel subunits. The ion channels were activated by a protocol that approximated ventricular action potentials. Results. The amide local anaesthetics bupivacaine, levobupivacaine and ropivacaine inhibited HERG channels at toxicologically relevant concentrations, with IC50 values of 20 (SEM 2) mM (n=29), 10 (1) mM (n=40) and 20 (2) mM (n=49), respectively. Hill coef®cients were close to unity. There were no indications of qualitative differences in channel inhibition between the three anaesthetics. The putative subunit MiRP1 did not alter local anaesthetic sensitivity of HERG channels. The common single nucleotide polymorphism producing MiRP1T8A did not increase local anaesthetic sensitivity of HERG/MiRP1 channels. Conclusions. Amide local anaesthetics target HERG and HERG/MiRP1 channels with identical potency. The effects on these ion currents may signi®cantly contribute to local-anaestheticinduced cardiac arrhythmia. MiRP1T8A does not seem to confer an increased risk of severe cardiac side-effects to carriers of this common polymorphism. Br J Anaesth 2004; 92: 93±101 Keywords: anaesthetics local; anaesthetics, toxicity; complications, arrhythmia Accepted for publication: July 21, 2003
Accidental intravascular injection of long-acting amide local anaesthetics such as bupivacaine may cause prolongation of ventricular action potentials, monitored as prolongation of the QT interval, and severe ventricular arrhythmia known as torsades de pointes.1 2 Prolongation of the QT interval and the induction of torsades de pointes ventricular arrhythmia are well known clinical manifestations of the long QT syndrome.3 The inherited form of this disease may be associated with mutations in several cardiac ion channel
genes such as the HERG (human ether-aÁ-go-go) gene.4 5 HERG channels underlie the repolarizing cardiac potassium current IKr.4 6 They are critical for the maintenance of normal rhythm in the human heart.4 Mutations in the HERG gene may lead to dysfunctional HERG channels and a reduced IKr. This correlates with the prolongation of ventricular action potentials, as well as an increase in susceptibility to ventricular arrhythmia.7 Inhibition of HERG channels has also been associated with drug-induced
Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2004
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
*Corresponding author. E-mail: [email protected] Background. Myocardial potassium channels are complexes formed by different subunits. The subunit composition may in¯uence the cardiotoxic action of local anaesthetics. The effects of amide local anaesthetics on HERG channels co-expressed with the putative subunit MiRP1 have not been established. It is also unclear if the common polymorphism MiRP1T8A that predisposes individuals to drug-induced cardiac arrhythmia increases local-anaesthetic sensitivity of HERG/MiRP1 channels. This may suggest the presence of genetic risk factors for local-anaesthetic-induced cardiac arrhythmia.
Friederich et al.
long QT syndrome and ventricular ®brillation.8 Long-acting amide local anaesthetics may therefore exhibit cardiotoxic effects in part by inhibiting HERG channels.9±11 In human myocardium HERG channels may associate with minKrelated peptide 1 (MiRP1) to form IKr.12 13 This auxillary subunit encoded in KCNE2 has been reported to alter the pharmacological sensitivity of HERG channels.12 Several mutations in hKCNE2, as well as the common T8A polymorphisms of MiRP1, have been shown to predispose individuals to drug-induced cardiac arrhythmia.12 14 KCNE2 may thus also constitute a susceptibility gene for localanaesthetic-induced ventricular arrhythmia. Myocardial potassium channels are complexes formed by different subunits.15 The subunit composition may in¯uence the effects of local anaesthetics on these membrane proteins.12 16 The effects of amide local anaesthetics on complexes formed by HERG and MiRP1 remain to be compared. It is therefore unclear if co-expression of HERG and MiRP1 alters the local anaesthetic sensitivity of HERG channels. If this were the case, MiRP1T8A might confer an increased sensitivity to HERG/MiRP1 channels, which may be indicative of an additional risk for carriers to develop local-anaesthetic-induced cardiac arrhythmia. Furthermore, it may suggest a role for genetic testing in the prevention of local-anaesthetic-induced cardiac arrhythmia. This study was designed to investigate if the local anaesthetic sensitivity of HERG channels is altered by the presence of MiRP1 or MiRP1T8A.
CTACTTTATCCAATTTCGCACAGACGCTGGAAGACGTCTTCCGAAG-3¢) and the pcDNA6-BGH primer (Invitrogen, Groningen, The Netherlands). Successful mutagenesis and the MiRP1T8A cDNA sequence were veri®ed by sequencing. The respective cDNAs were cloned into pcDNA6 expression vectors for transfection of CHO cells.
Transfection of CHO cells
Electrophysiology of potassium channels Whole-cell currents were measured using the patch clamp method18 with an EPC-9 ampli®er and Pulse software version 8.11 (HEKA Elektronik, Lambrecht, Germany). Patch electrodes with a pipette resistance of 2.0±4.0 MW were pulled from borosilicate glass capillary tubes (World Precision Instruments, Saratoga, USA) and ®lled with a solution comprising 160 mM KCl, 0.5 mM MgCl2, 10 mM HEPES, 2 mM Na-ATP (all from Sigma, Deissenhofen, Germany), adjusted to pH 7.2 with KOH. The external solution applied to the cells was 135 mM NaCl, 5 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 5 mM HEPES, 10 mM sucrose, phenol red 0.1 mg ml±1 (all from Sigma, Deissenhofen, Germany) adjusted to pH 7.4 with NaOH. The capacitance of the cells was 18.2 (SD 5.7) pF (n=20). Series resistance was 2.5±7.5 MW and was actively compensated for by at least 80%. A leak subtraction protocol was not used in this study. The data were therefore only recorded from cells in which tight mega-ohm or giga-ohm seals had been formed. As leak currents tend to increase during the course of an experiment, we only analysed data that did not show more than a 6% difference in any tested parameter between control and wash-out conditions. The liquid junction potential was calculated according to the Henderson equation.19 It accounted for 3.86 mV and was not corrected for. Bupivacaine (Sigma, Deissenhofen, Germany), levobupivacaine (AstraZeneca, SoÈdertalje, Sweden) and ropivacaine (AstraZeneca, SoÈdertalje, Sweden) were dissolved in the extracellular solution. The drugs were superfused onto the cells by a hydrostatically driven perfusion system. We studied the activity of cardiac potassium channels using a pulse protocol that approximates a cardiac ventricular action
Methods Cell culture Chinese hamster ovary (CHO) cells stably or transiently transfected with cloned human potassium channels were grown as non-con¯uent monolayers in MEM Eagle's Alpha medium (Life Technologies, Paisley, Scotland) containing 10% fetal calf serum (Biochrom, Berlin, Germany), penicillin 100 IU ml±1 and streptomycin 100 mg ml±1 (Life Technologies, Paisley, Scotland). The cells were cultured in 50 ml ¯asks (NUNC, Roskilde, Denmark) at 37°C in a humidi®ed atmosphere (carbon dioxide 5%). Cells were subcultured in 35 mm diameter monodishes (NUNC, Roskilde, Denmark) for at least 24 h before electrophysiological experiments were performed.
Cloning of ion channel subunits and cDNAcontaining expression vectors and in vitro mutagenesis Fragments encoding the complete open reading frames of the KCNH2 or KCNE2 genes were ampli®ed from human cardiac cDNA by the reverse transcriptase polymerase chain reaction (PCR) employing primers speci®c for the particular gene.17 The amino acid exchange T8A was introduced into KCNE2 by PCR using the primer KCNE2-T8A (5¢-ATGT94
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
CHO cells were seeded in 35 mm diameter monodishes 1 day before transfection. Transfection was performed using 3.5 ml of lipofectamine reagent (Gibco-BRL, Rockville, USA) according to the manufacturer's protocol and 1.0±2.0 mg of the respective ion channel plasmid DNA. Coexpression of HERG channels with MiRP1 and MiRP1T8A within the cell membrane was veri®ed using internal ribosomal entry site eGFP constructs of MiRP1 and MiRP1T8A, respectively. Only those cells transfected with HERG/MiRP1 or HERG/MiRP1T8A that emitted green ¯uorescence when stimulated by ultraviolet light were investigated.
Local anaesthetic sensitivities of cloned HERG channels
potential.20 The potassium channels were activated from a holding potential of ±80 mV by a step depolarization of the cell membrane to +60 mV and subsequent repolarization within 1000 ms back to the holding potential of ±80 mV (ramp protocol). The ramp protocol was followed by a 30 ms depolarization of the membrane to a test potential of +40 mV.17 Repetitive ramps were applied to determine that steady-state inhibition was reached. The voltage dependence of local anaesthetic action on HERG channels was investigated after a 2 s depolarizing test pulse to +60 mV by stepping the membrane potential for 0.5 s to potentials of ±120 mV, ±40 mV and +40 mV. Experiments were performed at room temperature. The recorded signal was ®ltered at 2 kHz and stored with a sampling rate of 5 kHz on a personal computer for later analysis.
Data analysis The current±time relationship observed with the ramp protocol for cardiac potassium channel activation was used to quantify the charge crossing the membrane during the ramp protocol (Qramp). Similarly, the current±time relationship observed during the step protocol for channel activation was used to quantify the charge crossing the membrane during the step protocol (Qstep). Inhibition of currents evoked by the ramp protocol was measured as inhibition of the maximal outward current and as reduction of Qramp. Inhibition of the current during the step protocol was analysed as a reduction of Qstep. The charge crossing the membrane is equivalent to the time integrals of current traces and was determined using Pulse software version 8.11 (HEKA Elektronik, Lambrecht, Germany). Inhibition was quanti®ed by the ratio of time integrals of current traces in the presence of the drug to the mean of the time integrals before application and after wash out of the drug (inhibition=1± Qdrug/[(Qcontrol+Qwash out)/2]) and by the ratio of the size of the current in the presence of the drug to the mean of the size of the current before application and after wash out of the drug (inhibition=1±Imax drug/[(Imax control+Imax wash out)/2]). The concentration±response curves were ®tted to the Hill equation (e/emax=cg/[IC50g+cg], where e=effect, emax=maximal effect, c=anaesthetic concentration, g=Hill coef®cient and IC50=concentration of half-maximal effect) using nonlinear regression. Regression analysis was performed using Kaleidagraph (Syngery Software, Reading, Pennsylvania, USA). Data points are given as mean (SD) unless stated otherwise. Statistical analysis was performed with an unpaired Student's t-test; n is the number of experiments.
Local anaesthetic sensitivity of HERG The current traces in Figure 3A demonstrate the effects of bupivacaine, levobupivacaine and ropivacaine on HERG currents after stimulation by the ramp protocol. The drugs reversibly inhibited HERG currents in a concentrationdependent manner. Inhibition was immediate and resulted from the wash in of the drugs. Wash out occurred within the same time scale as the wash in (5±10 s). The inhibitory effects were quanti®ed as a reduction of Qramp. Concentration-dependent reduction of Qramp was described
Results Properties of HERG, HERG/MiRP1 and HERG/ MiRP1T8A evoked by ramp protocols HERG outward currents were evoked from a holding potential of ±80 mV by a ramp protocol descending over 95
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
1000 ms from a membrane potential of +60 mV back to the holding potential of ±80 mV (Fig. 1A). This protocol was chosen because it approximates cardiac ventricular action potentials.20 Activation of HERG channels by the ramp protocol resulted in a bell-shaped current response, with an Imax of 0.5 (0.2) nA (n=20) at a membrane potential of ±40 mV. The size of Imax corresponded to a current density of 32 (12) pA/pF (Fig. 1B). Qramp was 356 (107) pC (n=20) (Fig. 1C). The time of the ramp protocol necessary to evoke peak HERG current was 711 (41) ms (n=15) (Fig. 1D). The current responses of HERG/MiRP1 and HERG/ MiRP1T8A were also bell-shaped, with Imax values of 0.7 (0.5) nA (n=10) and 0.8 (0.4) (n=10), respectively. The size of Imax corresponded to current densities of 35 (11) pA/pF in the case of HERG/MiRP1 and 33 (12) in the case of HERG/ MiRP1T8A (Fig. 1B). Qramp was 376 (203) pC for HERG/ MiRP1 and 397 (183) for HERG/ MiRP1T8A (Fig. 1C). The time to peak current response did not differ between HERG, HERG/MiRP1 and HERG/MiRP1T8A channels (Fig. 1D). The ramp was followed by an extra 30 ms test pulse to +40 mV (Fig. 2A). This additional depolarizing step assays an important property of HERG channels.21 Deactivation of HERG channels is much slower than their recovery from inactivation. This additional depolarizing step evoked rapidly inactivating outward currents passing through HERG, HERG/MiRP1 and HERG/ MiRP1T8A. The densities of the currents were 120 (46) pA/pF (n=18), 114 (67) pA/pF (n=9) and 114 (36) pA/pF (n=10), respectively (Fig. 2B). The current densities did not differ signi®cantly. The time course of inactivation of HERG, HERG/MiRP1 and HERG/ MiRP1T8A was adequately ®tted by a single exponential function with time constants (t) of 3.0 (0.4) ms (n=14), 3.1 (0.4) ms (n=9) and 3.0 (0.6) ms (n=9), respectively (Fig. 2C). The ratio of the size of the current in response to the step depolarization to the preceding ramp protocol (Imax step/Imax ramp) was 3.4 (0.8) (n=5), 3.3 (1.3) (n=6) and 3.6 (0.7) (n=13) for HERG, HERG/MiRP1 and HERG/ MiRP1T8A channels, respectively. The ratios did not differ between these channels (Fig. 2D). The ramp protocol combined with the step depolarization highlights two important physiological functions of cardiac HERG channels.17 We used this protocol to analyse local anaesthetic action on HERG, HERG/MiRP1 and HERG/MiRP1T8A channels.
Friederich et al.
Under control conditions Imax was reached after 723 (SD 36) ms of the test pulse (mean of control and wash out, n=19). The occurrence of Imax was signi®cantly delayed by bupivacaine 30 mM, to 766 (33) ms (n=7, P<0.05), levobupivacaine 10 mM, to 796 (49) ms (n=6, P<0.05) and ropivacaine 25 mM, to 803 (10) ms (n=6, P<0.05). As a consequence, inhibition of HERG channels was time dependent. Relating the size of Imax under control conditions to the size of the current at the same time of the ramp protocol during the drug effect yielded an inhibition of Imax by bupivacaine, levobupivacaine and ropivacaine of 60 (4)% (n=7), 48 (6)% (n=6) and 53 (4)% (n=6), respectively. Relating the size of Imax during drug action to the size of the current at the same time of the ramp protocol under control conditions yielded a signi®cantly smaller inhibition (P<0.05) by all three local anaesthetics: 55 (3)% by bupivacaine (n=7), 39 (3)% by levobupivacaine (n=6) and
mathematically by ®tting Hill functions to the concentration±response data (Fig. 3B, Table 1). The mean IC50 values of bupivacaine, levobupivacaine and ropivacaine were 20 (SEM 2) mM (n=29), 10 (1) mM (n=40) and 20 (2) mM (n=49), respectively (Fig. 3B, Table 1). The Hill coef®cients were close to unity (Table 1). Inhibition of HERG channels was also quanti®ed as inhibition of Imax. The IC50 values for inhibition of Imax were essentially identical to those for reduction of Qramp: 21 (SEM 1) mM (n=29), 11 (1) mM (n=40) and 22 (1) mM (n=56) for bupivacaine, levobupivacaine and ropivacaine, respectively. The mean Hill coef®cients were again close to unity: 0.95 (SEM 0.03), 0.95 (0.1) and 1.0 (0.02) for bupivacaine, levobupivacaine and ropivacaine, respectively. All three anaesthetics in¯uenced the shape of the current trace in response to the 1000 ms ramp protocol (Fig. 3A).
96
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
Fig 1 (A) HERG currents in response to a ramp protocol. Original current traces and the stimulation protocol are shown. Activation of these ion channels by the ramp protocol resulted in a bell-shaped current response. (B) The current densities did not differ between HERG, HERG/MiRP1 and HERG/MiRP1T8A channels expressed in Chinese hamster ovary cells. (C) The currents passing through the channels were quanti®ed as charge ¯ow (area under the current±time curve). Charge ¯ow did not differ between HERG, HERG/MiRP1 and HERG/MiRP1T8A channels. (D) MiRP1 and MiRP1T8A did not in¯uence the time to peak current response of the HERG channels. Data are mean and SD.
Local anaesthetic sensitivities of cloned HERG channels
43 (7)% by ropivacaine (n=6). Accordingly, the shape of the remaining HERG current trace after drug application was altered. Inhibition of HERG was voltage dependent, as assayed by a deactivating pulse protocol. Bupivacaine 30 mM inhibited the maximal currents by 40 (6)% at ±120 mV (n=7), by 51 (3)% at ±40 mV (n=7) and by 49 (5)% at +40 mV (n=7). Inhibition was signi®cantly different between membrane potentials of ±120 mV and ±40 (P<0.05) and between membrane potentials of ±120 mV and +40 mV (P<0.05) but was not signi®cantly different between membrane potentials of ±40 mV and +40 mV. The inhibitory effects of the local anaesthetics on HERG currents evoked by the step depolarization following the
ramp was also reversible and concentration dependent. The effects were quanti®ed as a reduction of Qstep. Fitting Hill functions to the concentration±response data (Fig. 3C) yielded IC50 values for the reduction of Qstep 2±3 times higher than for the inhibition of Qramp. Hill coef®cients were again close to unity (Table 1).
Local anaesthetic sensitivity of HERG/MiRP1 and HERG/MiRP1T8A The in¯uence of MiRP1 on local anaesthetic sensitivity of HERG channels and the in¯uence of the T8A polymorphism of KCNE2 on local anaesthetic sensitivity of IKr were 97
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
Fig 2 (A) HERG/MiRP1 currents in response to a ramp protocol followed by a rectangle depolarization. Original current traces and the stimulation protocol are shown. Activation of these ion channels by the additional depolarizing step evoked inactivating outward currents. (B) The current densities of the inactivating outward currents did not differ between HERG, HERG/MiRP1 and HERG/MiRP1T8A channels expressed in Chinese hamster ovary cells. (C) The time course of inactivation of HERG, HERG/MiRP1 and HERG/ MiRP1T8A channels was adequately ®tted by a single exponential function. The time constants of channel inactivation did not differ signi®cantly. (D) Current amplitudes evoked by the step depolarization were threefold larger than those evoked by the ramp protocol. MiRP1 and MiRP1T8A did not in¯uence the ratio of peak current responses of the HERG channels. Data are mean and SD.
Friederich et al.
MiRP1T8A (Qstep) accounted for 33 (13)% (n=6), 26 (9)% (n=13) and 29 (5)% (n=9) by bupivacaine, levobupivacaine and ropivacaine, respectively (Fig. 4C). The degree of inhibition of HERG/MiRP1T8A was not signi®cantly different from that seen for HERG/MiRP1 channels.
investigated at concentrations close to the IC50 values for inhibition of HERG channels (Qramp) of all three local anaesthetics. If MiRP1 or MiRP1T8A alter the local anaesthetic sensitivities of HERG channels, this effect can best be detected at the steepest part of the concentration±response curve for inhibition of HERG channels. All three anaesthetics reversibly inhibited both HERG/MiRP1 and HERG/ MiRP1T8A channels (Fig. 4A). Inhibition of HERG/MiRP1 (Qramp, Fig. 4B) accounted for 55 (3)% (n=10) by bupivacaine 30 mM, 48 (6)% (n=7) by levobupivacaine 10 mM and 49 (8)% (n=12) by ropivacaine 25 mM. Inhibition of Qstep (Fig. 4C) accounted for 37 (6)% (n=5), 27 (8)% (n=4) and 29 (9)% (n=8) by bupivacaine, levobupivacaine and ropivacaine, respectively. The extent of inhibition of HERG/MiRP1 was not signi®cantly different from that of HERG channels (P>0.05). Inhibition of HERG/MiRP1T8A (Qramp) accounted for 55 (6)% (n=7), 43 (5)% (n=14), and 50 (6)% (n=9) by bupivacaine, levobupivacaine and ropivacaine (Fig. 4B). The inhibitory effect on HERG/
Discussion
Fig 3 (A) HERG current traces after activation by ramp protocols (shown beneath the current traces). Currents are shown under control conditions and with bupivacaine 30 mM (left), levobupivacaine 10 mM (middle), ropivacaine 25 mM (right), and after wash out. Local anaesthetics reduced the size of the time integrals of current traces (Qramp), inhibited the maximal outward current (Imax) and increased the time needed to reach the peak current. (B) Concentration±response curves for the inhibition of Qramp by bupivacaine, levobupivacaine and ropivacaine. The rank order of potency for the inhibition of Qramp was levobupivacaine > bupivacaine = ropivacaine, with pIC50 values of 4.7 M, 5 M and 4.7 M, respectively. (C) The concentration±response curve for the reduction of Qstep by bupivacaine, levobupivacaine and ropivacaine. Local anaesthetics reduced the size of the time integrals of the current traces (Qstep). pIC50 values were 4.4 M, 4.5 M and 4.3 M, respectively (see Table 1).
98
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
The present work explored the pharmacological sensitivity of cloned repolarizing potassium channel complexes from human cardiac ventricle to the action of three clinically used amide local anaesthetics. The cardiac ion channels were activated by a protocol that approximates ventricular action potentials.20 This protocol has previously been used to study drug action on HERG and HERG/MiRP114 and facilitates interpretation of the clinical signi®cance of our results. The responses of the channels to the ramp protocol were in agreement with predictions derived from their respective roles during ventricular action potentials.7 The ramp
Local anaesthetic sensitivities of cloned HERG channels
protocol was followed by a sudden depolarization, highlighting the capability of these ion channels to counteract Table 1 Hill parameters for bupivacaine, levobupivacine and ropivacaine derived from concentration±response curves for the inhibition of HERG channels. The channels were activated with a ramp protocol (Qramp) followed by a step depolarization (Qstep). IC50=concentration at half-maximal effect. Data are mean (SEM).
Inhibition Qramp Number of experiments Maximum effect (%) Hill coef®cient IC50 (mM) Inhibition Qstep Number of experiments Maximum effect (%) Hill coef®cient IC50 (mM)
Bupivacaine
Levobupivacaine
Ropivacaine
29 94 (2) 0.95 (0.08) 20 (2)
40 91 (2) 0.97 (0.09) 10 (1)
49 93 (2) 0.99 (0.07) 20 (2)
34 90 (5) 1.14 (0.18) 43 (8)
28 94 (7) 0.75 (0.14) 32 (10)
26 93(5) 1.01 (0.12) 50 (8)
Fig 4 (A) HERG/MiRP1T8A current traces after activation by ramp protocols (shown beneath the current traces). Currents are shown under control conditions and with bupivacaine 30 mM (left), levobupivacaine 10 mM (middle), ropivacaine 25 mM (right) and after wash out of the drugs. Local anaesthetics reduced the size of the time integrals of current traces (Qramp), inhibited the maximal outward current (Imax) and increased the time needed to reach peak current. (B) Comparison of inhibition (Qramp) of HERG, HERG/MiRP1 and HERG/MiRP1T8A by bupivacaine 30 mM, levobupivacaine 10 mM and ropivacaine 25 mM. Inhibition of these channel complexes by local anaesthetics was not signi®cantly different between HERG, HERG/MiRP1 and HERG/MiRP1T8A. (C) Comparison of inhibition (Qstep) of HERG, HERG/MiRP1 and HERG/MiRP1T8A by bupivacaine 30 mM, levobupivacaine 10 mM and ropivacaine 25 mM. Inhibition of these channel complexes by all three anaesthetics was not signi®cantly different between HERG, HERG/MiRP1 and HERG/MiRP1T8A.
99
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
sudden depolarizations such as those observed during ventricular extrasystoles.17 The design of our study thus allowed investigation of local anaesthetic action on HERG, HERG/MiRP1, and HERG/MiRP1T8A by a protocol that accounts for the complex biophysical response of these ion channels during ventricular action potentials.20 Establishing drug sensitivity of cardiac potassium channels is a common and widely accepted way of testing the cardiac safety of pharmacological agents.7 8 22 Impairment of repolarizing ventricular currents and suppression of ventricular currents that counterbalance sudden depolarizations may render the myocardium susceptible to severe arrhythmia.11 Possible modulating effects of genetic variations in ion channel subunits on local anaesthetic sensitivity have not been reported previously. The data presented here therefore provide information on the local anaesthetic sensitivity of an important cardiac potassium channel, the cardiac safety of clinically used local anaesthetics and the
Friederich et al.
The concentration of local anaesthetic needed to halfmaximally inhibit HERG channels was used to compare the sensitivity of HERG, HERG/MiRP1 and HERG/ MiRP1T8A to local anaesthetics. Co-expression of HERG with MiRP1 did not alter the sensitivity of HERG channels. The interaction of the auxiliary subunit MiRP1 with HERG channels consequently does not perturb the binding of amide local anaesthetics to HERG channels. The same lack of difference in local anaesthetic sensitivity between HERG and HERG/MiRP1 also holds true for the comparison of HERG/MiRP1 with HERG/MiRP1T8A. The T8A single nucleotide polymorphism (SNP) occurs in 1.6% of healthy Caucasians14 and has been suggested to constitute an important risk factor for drug-induced ventricular arrhythmia.14 However, our results do not provide any evidence that this SNP may render HERG channels more susceptible to the pharmacological action of local anaesthetics. Our experimental results do not suggest a role for genetic testing of this common MiRP1 gene polymorphism in the prevention of local-anaestheticinduced cardiac arrhythmia. In summary, the amide local anaesthetics bupivacaine, levobupivacaine and ropivacaine inhibited cloned repolarizing ventricular potassium channels from human heart at toxicologically relevant concentrations. All three anaesthetics reduced outward currents in response to a simulated action potential and impaired the capability of these cardiac ion channels to respond to a sudden depolarization. There were no indications of qualitative differences in channel inhibition between bupivacaine, levobupivacaine and ropivacaine. The sensitivity of HERG channels was not altered by coexpression with MiRP1, and the common SNP producing MiRP1T8A did not in¯uence local anaesthetic sensitivity of HERG/MiRP1 channels. As drug-induced inhibition of repolarizing potassium channels from cardiac ventricles is a well-known cause of severe cardiac arrhythmia and sudden death, our results support the notion that none of the local anaesthetics investigated may be considered as safe with regard to cardiotoxic side-effects. MiRP1T8A seems not to confer an increased risk of these severe cardiac side-effects to carriers of this common polymorphism.
Acknowledgements We thank technical assistants Annerose Schmidt-Darlison and Andrea Zaisser at the Institute of Neural Signal Transduction for skilful cell culture and cell transfection. We are grateful to O. Pongs PhD, Director of the Institute of Neural Signal Transduction, for his generous support and for critically reading the manuscript. Levobupivacaine and ropivacaine were kindly provided by AstraZeneca, SoÈdertalje, Sweden. This study was supported by grants from the Fresenius Stiftung, Bad Homburg, Germany (PF), the Deutsche Forschungsgemeinschaft, Bonn, Germany (FR-1625±1/ 1), and the Foundation Leducq, Paris, France (DI).
100
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
in¯uence of a proposed arrhythmia susceptibility polymorphism14 on local anaesthetic sensitivity of a cardiac potassium channel complex that has repeatedly been associated with drug-induced ventricular arrhythmia.12 14 Bupivacaine, levobupivacaine and ropivacaine reduced the charge crossing the membrane through activated HERG channels in a concentration-dependent manner. The drugs exhibited two effects on Imax of HERG channels activated by the simulated action potential: they suppressed the maximal amplitude of the current and they delayed the occurrence of Imax. Both effects will decrease the repolarizing outward current during phases 3 and 4 of the cardiac action potential. Bupivacaine, levobupivacaine and ropivacaine also impaired the capability of HERG channels to conduct outward currents in response to a sudden depolarization. This effect was also concentration dependent and reversible. The difference in inhibition between ±120 mV and ±40 mV and between ±120 mV and +40 mV would be compatible with a voltage-dependent block. However, inhibition of HERG by bupivacaine did not differ between ±40 mV and +40 mV. It may thus be postulated that the increased driving force acting on the drug molecule at higher membrane depolarizations may be offset by conformational changes of the ion channel protein such as that occurring during channel inactivation. The direction of HERG current ¯ux also differs between currents measured at ±120 mV and ±40 mV. It may therefore be hypothesized that the potassium ions passing through the ion channel pore may interfere with the action of the local anaesthetic on the HERG channel protein. In any case the magnitude of the inhibitory effect is largest at potentials where HERG channels pass the largest currents during the ventricular action potential. Both the prolongation of phases 3 and 4 of the ventricular action potential and a reduced ability to counteract sudden depolarizations are likely to facilitate the occurrence of clinically observed ventricular arrhythmia during local anaesthetic intoxication. Further study is needed to characterize the precise molecular mechanism of the drug±channel interaction. The results of our study demonstrate that all three drugs are potent inhibitors of human ventricular potassium channels at concentrations that induce polymorphic ventricular arrhythmia in vivo.2 Our results obtained with a stimulation protocol that approximates ventricular action potentials support the notion11 that HERG channels constitute cardiac targets relevant for the induction of ventricular arrhythmia by local anaesthetics. As inhibition of repolarizing potassium channel complexes in human myocardium will prolong the duration of ventricular action potentials, the effects of local anaesthetics on these potassium channels will also in¯uence local anaesthetic action on cardiac sodium channels. Prolongation of the action potential will reduce sodium channel availability by increasing the fraction of inactivated sodium channels. This increased inactivation will add to the direct inhibitory effect of local anaesthetics on human heart sodium channels.23
Local anaesthetic sensitivities of cloned HERG channels
References
101
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on March 9, 2015
1 Kasten GW. Amide local anesthetic alterations of effective refractory period temporal dispersion: relationship to ventricular arrhythmias. Anesthesiology 1986; 65: 61±6 2 Chang DH, Ladd LA, Copeland S, Iglesias MA, Plummer JL, Mather LE. Direct cardiac effects of intracoronary bupivacaine, levobupivacaine and ropivacaine in the sheep. Br J Pharmacol 2001; 132: 649±58 3 Roden DM, Lazzara R, Rosen M, Schwartz PJ, Towbin J, Vincent GM. Multiple mechanisms in the long-QT syndrome. Current knowledge, gaps, and future directions. The SADS Foundation Task Force on LQTS. Circulation 1996; 94: 1996±2012 4 Sanguinetti MC, Jiang C, Curran ME, Keating MT. A mechanistic link between an inherited and an acquired cardiac arrhythmia: HERG encodes the IKr potassium channel. Cell 1995; 81: 299± 307 5 Splawski I, Shen J, Timothy KW, Vincent GM, Lehmann MH, Keating MT. Genomic structure of three long QT syndrome genes: KVLQT1, HERG, and KCNE1. Genomics 1998; 51: 86±97 6 Trudeau MC, Warmke JW, Ganetzky B, Robertson GA. HERG, a human inward recti®er in the voltage-gated potassium channel family. Science 1995; 269: 92±5 7 Keating MT, Sanguinetti MC. Molecular and cellular mechanisms of cardiac arrhythmias. Cell 2001; 104: 569±80 8 Mitcheson JS, Chen J, Lin M, Culberson C, Sanguinetti MC. A structural basis for drug-induced long QT syndrome. Proc Natl Acad Sci USA 2000; 97: 12329±33 9 Lipka LJ, Jiang M, Tseng GN. Differential effects of bupivacaine on cardiac K channels: role of channel inactivation and subunit composition in drug-channel interaction. J Cardiovasc Electrophysiol 1998; 9: 727±42 10 Gonzalez T, Longobardo M, Caballero R, Delpon E, Tamargo J, Valenzuela C. Effects of bupivacaine and a novel local anesthetic, IQB-9302, on human cardiac K+ channels. J Pharmacol Exp Ther 2001; 296: 573±83 11 Gonzalez T, Arias C, Caballero R, et al. Effects of levobupivacaine, ropivacaine and bupivacaine on HERG channels: stereoselective bupivacaine block. Br J Pharmacol 2002; 137: 1269±79
12 Abbott GW, Sesti F, Splawski I, et al. MiRP1 forms IKr potassium channels with HERG and is associated with cardiac arrhythmia. Cell 1999; 97: 175±87 13 Mazhari R, Greenstein JL, Winslow RL, Marban E, Nuss HB. Molecular interactions between two long-QT syndrome gene products, HERG and KCNE2, rationalized by in vitro and in silico analysis. Circ Res 2001; 89: 33±8 14 Sesti F, Abbott GW, Wei J, et al. A common polymorphism associated with antibiotic-induced cardiac arrhythmia. Proc Natl Acad Sci USA 2000; 97: 10613±18 15 Roden DM, Balser JR, George AL Jr, Anderson ME. Cardiac ion channels. Annu Rev Physiol 2002; 64: 431±75 16 Gonzalez T, Navarro-Polanco R, Arias C, et al. Assembly with the Kvbeta1.3 subunit modulates drug block of hKv1.5 channels. Mol Pharmacol 2002; 62: 1456±63 17 Isbrandt D, Friederich P, Solth A, et al. A novel mutation in the KCNE2 (MiRP1) gene that alters HERG channel kinetics. J Mol Med 2002; 80: 524±32 18 Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. P¯ugers Arch 1981; 391: 85±100 19 Barry PH, Lynch JW. Liquid junction potentials and small cell effects in patch-clamp analysis. J Membr Biol 1991; 121: 101±17 20 Hancox JC, Levi AJ, Witchel HJ. Time course and voltage dependence of expressed HERG current compared with native `rapid' delayed recti®er K current during the cardiac ventricular action potential. P¯ugers Arch 1998; 436: 843±53 21 Smith PL, Baukrowitz T, Yellen G. The inward recti®cation mechanism of the HERG cardiac potassium channel. Nature 1996; 379: 833±6 22 Weerapura M, Nattel S, Chartier D, Caballero R, Hebert TE. A comparison of currents carried by HERG, with and without coexpression of MiRP1, and the native rapid delayed recti®er current. Is MiRP1 the missing link? J Physiol 2002; 540: 15±27 23 Nau C, Wang SY, Strichartz GR, Wang GK. Block of human heart hH1 sodium channels by the enantiomers of bupivacaine. Anesthesiology 2000; 93: 1022±33