Interaction of Ropivacaine with Cloned Cardiac Kv4.3/KChIP2.2 Complexes

Chinese hamster ovary cells transiently transfected with cloned human potassium channels were grown as nonconfluent monolayers in MEM Eagle’s Alpha medium (Life Technologies, Paisley, Scotland) containing 10% fetal calf serum (Biochrom, Berlin, Germany), penicillin (100 U/ml), and streptomycin (100 mg/ml) (Life Technologies). The cells were cultured in 50-ml flasks (NUNC, Roskilde, Danmark) at 37°C in a humidified atmosphere (5% CO²). Cells were subcultured in monodishes (35 mm diameter; NUNC) 1 day before transfection and at least 48 h before electrophysiologic experiments were performed. Transfection was performed using 3.5 μl of the lipofectamine reagent (Gibco-BRL, Rockville, MD) according to the manufacturer’s protocol and 0.1 μg Kv4.3 plasmid DNA (GenBank No. NM 172198) and 1.0 μg KChIP2.2 plasmid DNA (GenBank No. NM 173191). To identify transfected cells, enhanced green fluorescent protein was used. Only those cells were investigated that emitted green fluorescence when stimulated by ultraviolet light. Transfection efficiency was between 60 and 80%.

Whole cell currents were measured with the patch clamp method29using an EPC-9 amplifier (HEKA Elektronik, Lambrecht, Germany) and Pulse software version 8.11 (HEKA Elektronik). Patch electrodes with a pipette resistance between 1.8 and 4.0 MΩ were pulled from borosilicate glass capillary tubes (World Precision Instruments, Saratoga, FL) and filled with the following solution: 160 mm KCl, 0.5 mm MgCl, 10 mm HEPES, 2 mm Na-ATP (all from Sigma, Deissenhofen, Germany); pH 7.2 (adjusted with KOH). The external solution applied to the cells was of the following composition: 135 mm NaCl, 5 mm KCl, 2 mm CaCl², 2 mm MgCl², 5 mm HEPES, 10 mm sucrose, 0.1 mg/ml phenol red (all from Sigma); pH 7.4 (adjusted with NaOH). The capacity of the cells was 29.8 ± 16.1 pF (n = 20). Series resistance was 2.5–5 MΩ and actively compensated for by at least 85%. A p/4 leak subtraction protocol was used in the study. S  (−)- and R  (+)-ropivacaine (AstraZeneca, Södertalje, Sweden) were dissolved in the extracellular solution. They were superfused on the cells by a hydrostatically driven perfusion system.

The holding potential of the cell membrane was −80 mV. Current activation was analyzed by depolarization of the membrane potential for 1,000 ms from −40 to +60 mV in 10-mV steps. Steady state inactivation was determined by depolarizing the membrane to +40 mV for 25 ms from prepulses of a duration of 4,000 ms increasing from −75 and −15 mV in steps of 5 mV. For measuring the recovery from inactivation, the initial depolarization of the membrane potential to +40 mV for 1,500 ms was followed by repolarizing the membrane potential to −80 mV for different lengths of time increasing from 10 to 5,120 ms by doubling the time spent at −80 mV after each individual depolarization. The repolarization was followed by a depolarizing step of 25 ms to +40 mV. Deactivating currents were evoked with a tail current protocol. After a 200-ms-long hyperpolarizing prepulse to −100 mV, the membrane was depolarized to +20 mV for 4 ms and then repolarized to a potential of −100 mV for 150 ms. Experiments were performed at room temperature. The recorded signal was filtered at 2 kHz and stored on a personal computer for later analysis with a sampling rate of up to 20 kHz.

The current–time relation observed with the different protocols for potassium channel activation was used to determine the maximal outward current (I^max) as well as to quantify the charge (Q) crossing the membrane during the duration of the depolarization induced by the rectangle protocol. Q is equivalent to the time integrals of current traces and was determined using Pulse software version 8.11. Inhibition was quantified as Inhibition = 1 − Q^drug/((Q^control+ Q^{wash out})/2) as well as by Inhibition = 1 − I^maxdrug/((I^maxcontrol+ I^maxwashout)/2). The concentration–response curves were fitted with the Hill equation e/e^max= c^γ/(IC^50γ+ c^γ) using nonlinear regression. Here, e = effect, e^max= maximal effect, c = anesthetic concentration, γ= Hill coefficient, and IC⁵⁰= concentration of half-maximal effect. Regression analysis was performed using Kaleidagraph (Synergy Software, Reading, PA). The apparent rate constants for binding (k) and unbinding (l) were obtained from the equation 1/τ^B= k ×[D]+ l, where τ^Brepresents the time constant of the drug-induced macroscopic current decline during depolarization to +40 mV, and [D] represents the drug concentration. Determination of k and l was performed by linear regression analysis of 1/τ^Bversus  different drug concentrations. The whole cell conductance was calculated using the following formula: G^max= I^max/(V^m− E^K), where I^maxis the maximum current at the end of each test potential, V^mis the membrane potential, and E^Kis the Nernst potential for potassium (−87.54 mV). The means of the whole cell conductance were mathematically described by a Boltzmann function, y = M³/(1 + exp((M¹− M⁰)/M²)), where M⁰is the membrane potential, M¹is the activation midpoint, M²is the slope factor, and M³is the maximal whole cell conductance. Time-dependent inactivation was mathematically described using biexponential functions. Recovery from inactivation was determined by relating the maximal size of the outward current during the test pulse to the size of maximal outward current during the prepulse. The time dependent increase of this ratio was mathematically described with y = 1− A¹exp(−t/τ¹) − A²exp(−t/τ²), where t is the time (ms), τ¹and τ²are the system time constants, and A¹and A²are the amplitudes of each component of the exponential. Data points are given as mean ± SD unless stated otherwise. Statistical significance (P < 0.05) was tested using analysis of variance and Tukey-Kramer multiple comparisons test (Graph Pad Prism, San Diego, CA) or two-sided paired and unpaired Student t  tests as appropriate (Excel; Microsoft, Redmond, WA). The number n is the number of experiments.

Kv4.3/KChIP2.2 currents activated with a midpoint of 1 ± 2 mV (n = 6). The time constants of current decline decreased with voltage and were described with a biexponential function yielding time constants of inactivation at +40 mV of 39 ± 14 and 90 ± 43 ms (n = 30). The midpoint of voltage-dependent inactivation was −43 ± 5 mV (n = 6). Recovery from inactivation occurred with a time constant of 50 ± 9 ms (n = 7). The sensitivity of Kv4.3/KChIP2.2 currents to the pharmacologic effect of S  (−)-ropivacaine was investigated first at a test potential of +40 mV. At this potential, the ion channels were fully activated. Original current traces (fig. 1A) demonstrate that S  (−)-ropivacaine inhibited Kv4.3/KChIP2.2 in a reversible manner. Concentration-dependent inhibition of Kv4.3/KChIP2.2 channels was quantified as inhibition of the peak current response (I^max) and as the reduction of the charge transfer through Kv4.3/KChIP2.2 channels during the entire duration of the test pulse (Q). The inhibitory drug effects were reversible at all concentrations. The concentration–response data were mathematically described with Hill-functions (fig. 1B). The IC⁵⁰values for inhibition of I^maxand Q by S  (−)-ropivacaine were 159 ± 33 μm (n = 34) and 117 ± 21 μm (n = 30), respectively (table 1).

The local anesthetic induced a reversible crossing of the currents at drug concentrations between 8 μm and 250 μm with the currents under control conditions (fig. 1A). The time point at which the crossover occurred increased with concentration from 135 ± 14 ms at 8 μm (n = 6) to 328 ± 65 ms at 250 μm (n = 3). The resulting stimulation of Kv4.3/KChIP2.2 channels occurring after crossing of the currents amounted to, on average, 158 ± 81% (n = 24). Inhibition of the charge between the peak current and the crossing of currents (Q^cross) occurred with a significantly lower IC⁵⁰value (62 ± 8 μm, n = 33) than inhibition of Q (fig. 1Band table 1).

S  (−)-ropivacaine accelerated macroscopic current decline (fig. 1C). This effect was concentration dependent and reversible. Concentration-dependent decrease of the fast inactivation time constant (τ¹) by S  (−)-ropivacaine was mathematically described by a Hill function (fig. 1D) yielding an IC⁵⁰value of 77 ± 11 μm (n = 30) (table 1). Concentration-dependent changes of τ²were not analyzed because the slower time constant (τ²) was difficult to be resolved under high drug concentrations (250 μm) because of the ratio of the size of the time constant to the duration of the test pulse as well as because of the small size of the residual current. The rate constants for binding (k) and unbinding (l) were determined by linear regression analysis of 1/τ¹versus  drug concentration (see Materials and Methods) and were k = 0.39 × 10⁶m⁻¹s⁻¹and l = 26.2 s⁻¹(n = 21, r = 0.99).

After establishing the concentration–response relation of the effects of S  (−)-ropivacaine on Kv4.3/KChIP2.2 currents, the effects of the local anesthetic on the gating behavior of Kv4.3/KChIP2.2 channels were analyzed in more detail. For this purpose, the effects of 83 μm S  (−)-ropivacaine on voltage-dependent activation, steady state inactivation, and recovery from inactivation were investigated (figs. 2 and 3). The concentration was chosen because it is between the IC⁵⁰values for inhibition of Q and I^maxon one site and τ¹and Q^crosson the other (table 1). Figure 2shows the original current traces elicited by the activation protocol (fig. 2A), the inactivation protocol (fig. 2B), and the recovery protocol (fig. 2C). S  (−)-ropivacaine shifted the midpoint of current activation in the depolarizing direction by 7 ± 3 mV (P < 0.05, n = 6; table 2) and caused an increase of the slope of the activation curve by 2 ± 1 (P < 0.05, n = 6; fig. 3Aand table 2). The local anesthetic did not influence voltage dependence of inactivation (fig. 3Band table 2), but it slowed recovery from inactivation (fig. 3Cand table 2). Under control and washout conditions, the recovery from inactivation followed a monoexponential process (time constant 54 ± 11 ms, mean of control and wash out, n = 7), whereas a biexponential function was necessary to adequately describe the recovery from inactivation under the influence of S  (−)-ropivacaine. The corresponding time constants τ^rec1and τ^rec2were 40 ± 6 and 1,222 ± 882 ms, respectively (n = 7). Both time constants were significantly different from the time constant under control and washout conditions (table 2). S  (−)-ropivacaine altered the time course of deactivating currents (tested at −100 mV) by increasing the slower time constant from 0.96 ± 0.25 to 1.7 ± 0.7 ms (n = 8 paired experiments, P < 0.05). The slowing of current deactivation induced a crossing of deactivating currents.

The accelerated current decline induced by S  (−)-ropivacaine suggested that the reduction of peak current would be a nonequilibrium measure of block. Voltage dependence of Kv4.3/KChIP2.2 inhibition was therefore analyzed as the reduction of Q. Inhibition of Q increased with voltage from 27 ± 16% at −10 mV to 39 ± 18% at +40 mV (fig. 3D; P < 0.05, n = 6). Also, the time to crossing increased with voltage from 138 ± 31 ms at −10 mV to 214 ± 71 ms at +40 mV (n = 5). The time constants of S  (−)-ropivacaine induced fast macroscopic current decline (τ¹) decreased with voltage from 27 ± 10 ms at −10 mV to 4 ± 1 ms at +40 mV (fig. 3E; P < 0.05, n = 6). The voltage-dependent distribution of the amplitudes of inactivation time constants was also changed by the local anesthetic. Under control and washout conditions, the contribution of τ¹to channel inactivation declined with voltage from 94 ± 3% at 0 mV to 47 ± 17% at +40 mV (P < 0.05, n = 6), whereas the contribution of τ¹to channel inactivation remained constant over voltage under the influence of S  (−)-ropivacaine (fig. 3F; P > 0.05, n = 6).

To determine whether inhibition of Kv4.3/KChIP2.2 channels by ropivacaine is stereospecific, the inhibitions by S  (−)- and R  (+)-ropivacaine were compared at a concentration close to the steepest part of the concentration–response curve of S  (−)-ropivacaine (83 μm). The original current recordings (fig. 4A) demonstrate that both isomers differed in their effect on the current decay. I^maxwas inhibited by both isomers to the same extent (45 ± 8% for S  (−)-ropivacaine, n = 7, vs.  43 ± 8% for R  (+)-ropivacaine, n = 7 paired experiments; P > 0.05), whereas inhibition of Q by S  (−)-ropivacaine was significantly larger than inhibition by R  (+)-ropivacaine (fig. 4B). This difference resulted from a larger residual current under the influence of the R  (+)-isomer due to isomer-specific changes of the inactivation time constants. Both isomers decreased the fast time constant of current decay (τ¹), whereas only R  (+)-ropivacaine significantly increased τ²(figs. 4C and D). Analysis of the time course of the development of block expressed as the proportion of the control current at any given time after the start of the depolarizing pulse ((I^control− I^drug)/I^control) revealed that block of Kv4.3/KChIP2.2 channels by S  (−)-ropivacaine developed with a time constant significantly slower than the block by R  (+)-ropivacaine (τ^onset= 6.1 ± 1.5 ms for S  (−)-ropivacaine versus τ^onset= 5.2 ± 1.2 ms for R  (+)-ropivacaine, n = 7 paired experiments; P < 0.05).

Finally, we investigated whether coexpression of KChIP2.2 alters the pharmacologic effect of S  (−)-ropivacaine and R  (+)-ropivacaine on Kv4.3 channels. The original current traces demonstrate (fig. 5A) that neither isomer induced the crossing phenomenon in Kv4.3 channels as observed with Kv4.3/KChIP2.2 channels. Inhibition of Q by either isomer differed between Kv4.3 and Kv4.3/KChIP2.2 channels (fig. 5B; n = 6 for Kv4.3 and n = 7 for Kv4.3/KChIPP2.2; P < 0.05). Inhibition of Kv4.3 channels by S  (−)- and R  (+)-ropivacaine was also stereospecific (fig. 5B; n = 6 paired experiments; P < 0.05), and the time course of the development of block of Kv4.3 channels differed between S  (−)-ropivacaine and R  (+)-ropivacaine as well (τ^onset= 2.3 ± 0.4 ms for S  (−)-ropivacaine vs. τ^onset= 2.6 ± 0.5 ms for R  (+)-ropivacaine, n = 7 paired experiments; P < 0.05). Block of Kv4.3 channels by either isomer thus developed with a time constant significantly smaller (P < 0.05) than the respective time constant of the block of Kv4.3/KChIP2.2 channels (fig. 5C).

In this study, the effects of the amino-amide local anesthetic S  (−)-ropivacaine on complexes formed by human Kv4.3/KChIP2.2 K channel subunits transiently expressed in a mammalian cell line were investigated. The cardiac ion channel complex gave rise to an A-type current with biophysical properties in accord with those reported previously.22,24 S  (−)-ropivacaine reversibly inhibited Kv4.3/KChIP2.2 channel conductance, it shifted the midpoint of current activation into the depolarizing direction, and the local anesthetic slowed recovery from inactivation. Each individual effect as well as the sum of effects resulted in a reduction of I^tocurrent. Conductance block rather than gating changes contributed most to the suppression of Kv4.3/KChIP2.2 channels.

S  (−)-ropivacaine accelerated macroscopic current decline of the channels in a concentration-dependent and reversible manner. This is a characteristic feature of open K channel block.30,31It has previously been described as characteristic for the inhibition of Kv1.5 channels by amide local anesthetics such as S  (−)-ropivacaine.1Because of their structural homology within the pore and S6 regions, molecular requirements reported for Kv1.5 channel block may help to interpret the results of Kv4.3/KChIP2.2 channel block as well. Amide local anesthetics exert their effect on Kv1.5 by interacting with threonine 477 within the tetraethylammonium binding domain in the pore region and also by interacting with threonine 505, lysine 508, and valine 512 in the S6 segment of Kv1.5.2Threonine 477 in Kv1.5 subunits corresponds to threonine 358 in Kv4.3 subunits, and valine 512 corresponds to valine 401. S  (−)-ropivacaine may thus interact with this region of Kv4.3 with remarkably similar rate constants of binding and unbinding as for interaction with Kv1.5.2Hydrophobic substitution of threonine 505 in Kv1.5 (such as T505V) drastically reduces stereoselective block.2Because Kv4.3 channels carry a valine at the corresponding position, the very subtle stereoselective effect of ropivacaine on Kv4.3/KChIP2.2 as well as on Kv4.3 channels is compatible with these mutagenesis studies.2The experimental results of our study are consistent with the idea that ropivacaine blocks open Kv4.3/KChIP2.2 channels by interacting with amino acids in the pore region and the S6 segment of the Kv4.3 subunit.

Valine 399 and valine 401 in Kv4.3 channels have furthermore been suggested to play a critical role in the modulatory effects of KChIP2b on Kv4.3 α subunits.24By binding to the N-terminal domain,22KChIP2b shifts the midpoint of voltage-dependent Kv4.3 channel activation into the hyperpolarizing direction,24it slows initial inactivation of Kv4.3 channels,24and KChIP2b speeds up recovery from inactivation.24,S  (−)-ropivacaine seemed to interfere with the effects of KChIP2.2 on Kv4.3 channels. The local anesthetic caused a depolarizing shift of voltage-dependent current activation, it accelerated macroscopic current decline, and it slowed recovery from channel inactivation. The crossing of the current traces under drug conditions with the current under drug-free conditions would also be consistent with a reversal of the effects of KChIP on Kv4 channel gating.22,23This interpretation is furthermore supported by the drug-induced change of the voltage-dependent distribution of the amplitudes of inactivation time constants. In the presence of the local anesthetic, the distribution of the amplitudes of inactivation time constants is similar to the distribution of Kv4.3 channel amplitudes when expressed without the β subunit.23Interference of the effects of S  (−)-ropivacaine and KChIP2.2 may finally explain why the block of Kv4.3/KChIP2.2 developed significantly slower than block of Kv4.3 channels. Whether the reversal of KChIP effects is the result of a direct or indirect interaction of the drug molecule with the β subunit must be further investigated. Nonetheless, the results of this study underscore and extent the observation that β subunits modify the response of cardiac Kv channels to local anesthetics.32 

Based on these considerations, the interference of effects can be summarized in the following gating scheme (fig. 6). Kv4 channels inactivate from a preopen closed state (C) by a mechanism involving conformational changes at the internal vestibule of the pore23,33as well as from the open state (O) by a mechanism involving the N-termini.23,25The interaction of KChIPs with the N-termini impairs inactivation from the open state (O) by preventing the channel’s N-termini to interact with the internal mouth of the pore.23As a consequence, inactivation from the preopen closed state predominates in complexes formed by Kv4 and KChIP subunits.23By interacting with the channel pore, S  (−)-ropivacaine introduces a blocked open state (B) of Kv4.3/KChIP2.2 channels reminiscent of N-type inactivation30,31and thus reverses KChIP2.2 effects on Kv4.3 channels. The crossing of currents under control and drug conditions during the depolarization of the membrane to +40 mV may indicate that blocked channels must reopen before the drug leaves the channel pore. This has been modeled to underlie tedisamil block of I^toin myocytes isolated from human heart34and has also been suggested for disopyramide block of transient outward current in ventricular myocytes from rat heart.35Because opening of blocked channels is less likely to occur at more negative membrane potentials, one possible explanation for the slowed recovery from inactivation is trapping of the drug after blocking the open channels.34,35 

The inhibitory effects of S  (−)-ropivacaine on Kv4.3/KChIP2.2 channels may contribute to prolongation of the QTc interval of the electrocardiogram by altering the initial phase of the action potential waveform. Prolongation of initial action potential repolarization due to inhibition of I^tohas furthermore been suggested to impair ventricular excitation contraction coupling.27Suppression of I^toby S  (−)-ropivacaine may thus result in complex changes of myocardial function. The magnitude of I^todiffers between endocardial, midmyocardial, and epicardial layers of the human ventricle. These differences result at least in part from different levels of myocardial KChIP expression.17,18The difference in S  (−)-ropivacaine sensitivity between Kv4.3 channels and channel complexes formed by Kv4.3 and KChIP2.2 may imply that ventricular layers with a lower level of KChIP expression may be more vulnerable to the inhibitory action of S  (−)-ropivacaine.

The results of our study indicate that not only bupivacaine15,16but also S  (−)-ropivacaine is an inhibitor of I^toat concentrations that induce cardiotoxic side effects in vivo .36,37Knudsen et al.  8demonstrated the absence of severe cardiac side effects of S  (−)-ropivacaine at free arterial plasma concentrations of 0.56 μg/ml. Feldman et al.  36demonstrated cardiovascular toxicity of S  (−)-ropivacaine after bolus application in dogs at plasma concentrations of 70 μg/ml. These reports indicate that at concentrations below 2 μm, cardiotoxic side effects are unlikely to occur and that cardiotoxic effects including severe ventricular arrhythmia are present at concentrations of S  (−)-ropivacaine of approximately 250 μm. In accord with these in vivo  concentrations, S  (−)-ropivacaine only inhibits Kv4.3/KChIP2.2 channels at the toxicologically relevant concentration. However, I^tois not the only K current affected during local anesthetic intoxication.1–5To clarify the precise role of I^toin cardiotoxic action of amino-amide local anesthetics, further study is warranted.

In summary, the amino-amide local anesthetic S  (−)-ropivacaine inhibited complexes formed by Kv4.3/KChIP2.2 in a concentration-dependent, reversible, and stereospecific manner. The results of our study are consistent with the idea that S  (−)-ropivacaine blocks Kv4.3/KChIP2.2 channels from the open state. S  (−)-ropivacaine seemed to partially reverse the effects of KChIP2.2 on Kv4.3 channels. The β subunit altered the pharmacologic response of the α subunit. Inhibition of Kv4.3/KChIP2.2 channel complexes by the local anesthetic may contribute to the generation of severe ventricular arrhythmia as well as to the impairment of excitation–contraction coupling during events of intoxication.

The authors thank Olaf Pongs, Ph.D. (Director, Institute of Neural Signal Transduction, Center for Molecular Neurobiology, University of Hamburg, Hamburg, Germany), for continuous support. The authors also thank Andrea Zaisser (Medical Technical Assistant, Institute of Neural Signal Transduction, Center for Molecular Neurobiology, University of Hamburg) and Dirk Isbrandt, M.D. (Institute of Neural Signal Transduction, Center for Molecular Neurobiology, University of Hamburg), for providing particular support for this study.