Lipophilic and Stereospecific Interactions of Amino-amide Local Anesthetics with Human Kv1.1 Channels

Human Kv1.1 channels (accession No. L02750) were cloned by Hind  III and Kpn  I into the eucaryotic expression vector pcDNA3 (Invitrogen, Karlsruhe, Germany) and stably expressed in human embryonic kidney 293 cells. Human embryonic kidney 293 cells were cultured in 50-ml flasks (NUNC, Roskilde, Denmark) at 37°C in Dulbecco medium (GIBCO, Invitrogen, Carlsbad, CA) with 10% fetal calf serum (Biochrom, Berlin, Germany), 100 U/l penicillin, and 100 mg/ml streptomycin (GIBCO) in a humidified atmosphere (5% CO²). The cells were grown as nonconfluent monolayers and subcultured in monodishes (35-mm diameter; NUNC) at densities between 2 and 3 × 10⁴cells per dish before the electrophysiologic experiments.

Whole cell currents were measured with the voltage clamp method of the patch clamp technique25using an EPC-9 amplifier (HEKA Elektronik, Lambrecht, Germany) and Pulse software version 8.53 (HEKA Elektronik). The currents were filtered at 1 kHz, and series resistance was actively compensated by at least 80%. A p/4 leak subtraction protocol was used in the study, with the exception of the experiments for frequency dependence. The patch electrodes were fabricated from borosilicate glass capillary tubes with filament (World Precision Instruments, Saratoga, FL) using a Sutter P-97 puller (Sutter Instrument Company, Novato, CA). The pipettes had a resistance of 2–4 MΩ and were filled with a solution containing the following electrolyte concentrations: 160 mm KCl, 0.5 mm MgCl², 10 mm HEPES, and 2 mm Na²ATP, with a pH of 7.25 adjusted with KOH. The extracellular solution consisted of 135 mm NaCl, 5 mm KCl, 2 mm CaCl², 2 mm MgCl², 5 mm HEPES, 10 mm sucrose, and 0.01 mg/ml phenol red, with a pH of 7.4 adjusted with NaOH. All experiments were performed at room temperature. The local anesthetics were dissolved in the watery extracellular solution, yielding stock solutions of 1 mm for dextrobupivacaine, bupivacaine, levobupivacaine, and dextroropivacaine. For levoropivacaine, a stock solution of 3 mm was prepared, and for mepivacaine, we used a stock solution of 10 mm (all chemicals were purchased from Sigma, Deissenhofen, Germany, except for levobupivacaine, dextrobupivacaine, and dextroropivacaine, which were gifts from AstraZeneca, Södertalje, Sweden). The drugs were applied on the cells by a perfusion system driven by hydrostatic pressure with Teflon tubing (N Research Europe, Guemlingen, Switzerland). The inflow of the perfusion system did not have significant influence of the currents.

The holding potential during all experiments was −80 mV. The current–voltage relation was established by depolarizing the cell membrane in steps of 10 mV for 100 ms between the membrane potentials −70 and +60 mV. To characterize the concentration-dependent effects of the local anesthetics, outward currents of Kv1.1 channels were measured at +40 mV at the end of the test potential. The inhibition of the outward currents by local anesthetic agents was determined as the inhibition of the maximum current and calculated by the following formula: inhibition = 1 − (I^drug/((I^control+ I^washout)/2)). The ratio of the maximum current (I) under influence of the local anesthetic agents and the mean of maximum current under control and washout conditions was subtracted from 1. The data of the concentration–response curves were mathematically described by Hill functions: f = 1/{1 + (IC⁵⁰/c)^h}, where IC⁵⁰is the concentration of the half-maximal inhibition, c is the drug concentration, and h is the Hill coefficient, using Kaleidagraph software (Synergy-Software, Reading, PA). SEs of calculated Hill parameters were used as defined by Kaleidagraph. The whole cell conductance was calculated using the following formula: G^max= I^max/(V^m− E^K), where I^maxis the maximum current of each test potential, V^mis the membrane potential, and E^Kis the Nernst potential for potassium. The conductance–voltage relation was mathematically described by a Boltzmann equation: I = I^max/{1 + exp((V^mid− V^m)/k)}, where V^midis the voltage of half-maximal activation, V^mis the membrane potential, and k is the slope factor. The fitting procedure was performed with Kaleidagraph software. Frequency dependence of Kv1.1 channels was analyzed with repetitive 30-ms pulses to +40 mV with three different interpulse durations (100, 300, and 1,000 ms). The remaining current under the influence of the local anesthetic agents in the steady state phase at the end of the test potential was normalized to control values. The normalized current values were plotted against a time axis according to the different interpulse durations. Time constants of current inactivation were determined by fitting current decay with a monoexponential function using Pulsefit software version 8.53.

Statistical significance was tested using analysis of variance and the Tukey–Kramer multiple comparisons test (Graph Pad; Prism, San Diego, CA) or two-sided paired and unpaired Student t  test as appropriate (Excel; Microsoft, Redmond, WA). Data points are given as mean ± SD unless stated otherwise; n values indicate the number of experiments.

Original current recordings of Kv1.1 channels exhibited typical delayed rectifier currents devoid of fast inactivation (figs. 1A–F). The currents activated at membrane potentials more positive than −40 mV, with a midpoint of current activation (V^mid) of −26 ± 1 mV (n = 5) and a slope factor of 7.1 ± 0.5 (n = 5).

The inhibitory effects of dextrobupivacaine, racemic bupivacaine, levobupivacaine, dextroropivacaine, levoropivacaine, and racemic mepivacaine were analyzed at a membrane potential of +40 mV and quantified as the reduction of the steady state outward current at the end of each test potential. All local anesthetics inhibited currents through Kv1.1 channels (figs. 1A–F). Inhibition of Kv1.1 channels was concentration dependent, stereoselective, and reversible. The concentration–response data for inhibition of Kv1.1 channels were mathematically described by Hill equations (fig. 2). Dextrobupivacaine inhibited Kv1.1 channels with the highest potency. The IC⁵⁰value was 41 ± 3 μm, and the corresponding Hill coefficient was 1.0 ± 0.1 (mean ± SEM; n = 20). The sensitivity of Kv1.1 channels for the different local anesthetic agents decreased in the following order: dextrobupivacaine > bupivacaine > levobupivacaine > dextroropivacaine > levoropivacaine > mepivacaine (fig. 2; for overview, see table 1). Stereoselectivity of the effects was analyzed by comparing the current inhibition of Kv1.1 channels by dextrobupivacaine and levobupivacaine, as well as by dextroropivacaine and levoropivacaine at a membrane potential of +40 mV. Both dextroisomers inhibited Kv1.1 channels more potently than the respective levoisomers did. For dextrobupivacaine (30 μm) and levobupivacaine (30 μm), Kv1.1 channel inhibition was 74 ± 9% (n = 5) versus  59 ± 5% (n = 6) (P < 0.05), and for dextroropivacaine (100 μm) and levoropivacaine (100 μm), it was 38 ± 3% (n = 6) versus  17 ± 5% (n = 5) (P < 0.05). All local anesthetic agents shifted the voltage dependence of Kv1.1 channel activation into the hyperpolarizing direction (figs. 3A–F). The shift of the activationmidpoint (V^mid) into the hyperpolarizing direction by all local anesthetic agents significantly increased with concentration (figs. 4A–F). The effects were dependent on the lipophilic properties of the drugs.

For the analysis of the voltage dependence of inhibition, the size of the outward current at the end of each test potential (I^max) under the influence of the different local anesthetic agents was related to the mean of control and washout current. The extent of inhibition was plotted against the membrane potential. All local anesthetic agents inhibited Kv1.1 channels in a voltage-dependent manner (figs. 5A–F; analysis of variance; P < 0.05). Inhibition of Kv1.1 channels was not frequency dependent, because the stimulation frequency of the depolarizing pulse did not significantly influence local anesthetic–induced inhibition (figs. 6A–F).

Apart from mepivacaine, all local anesthetics induced inactivation-like behavior of Kv1.1 currents characterized by an acceleration of macroscopic current decline (figs. 1A–F and 7A). Comparing inhibition of the peak current at the beginning of the depolarizing test pulse and inhibition of the outward current at the end of each test pulse (fig. 7B) revealed that apart from mepivacaine, inhibition of the peak currents by the local anesthetics was significantly smaller than inhibition of the steady state currents (fig. 7B). The time constants of inactivation were determined at concentrations close to the IC⁵⁰value of the respective drug by fitting a monoexponential function to the declining current at a membrane potential of +60 mV. Both dextroisomers induced a faster current decline than the respective levoisomers did. Dextrobupivacaine (30 μm) and levobupivacaine (30 μm) inactivated Kv1.1 channels with time constants of 3.3 ± 1.1 ms (n = 6) and 9.3 ± 0.4 ms (n = 5), respectively (P < 0.05). Racemic bupivacaine (30 μm) yielded a time constant of 7.1 ± 2.2 ms (n = 6) (fig. 7C). The time constants for dextroropivacaine (100 μm) and levoropivacaine (100 μm) were 4.5 ± 1.1 ms (n = 5) and 37.9 ± 5.2 ms (n = 5), respectively (P < 0.05).

Lipophilicity of the different local anesthetic agents26,27correlated significantly with the number of CH²groups in the side chain (y = 1.63 + 0.01x ; r = 0.90, P < 0.05). The IC⁵⁰values for Kv1.1 channel inhibition and the shift of V^midwere correlated with the length of the N-substituent of the homolog series of the local anesthetics (figs. 8A and B). Both the inhibitory potency and the potency to shift of the midpoint of current activation significantly correlated with the number of CH²groups (r = 0.96, P < 0.05; and r = 0.94, P < 0.05, respectively).

This study established the effects of different amino-amide local anesthetics on human Kv1.1 channels. The biophysical properties of Kv1.1 wild-type channels used in our study were in accord with biophysical properties of these channels reported in previous studies.23,28Kv1.1 channels exhibited fast activating currents devoid of fast inactivation. The effects of racemic mixtures of bupivacaine and mepivacaine as well as the R  (+) and the S  (−) enantiomers of bupivacaine and ropivacaine on Kv1.1 channels were analyzed. All local anesthetic agents studied in this investigation inhibited Kv1.1 currents in a concentration-dependent, reversible, stereoselective, voltage-dependent, and frequency-independent manner. Voltage-dependent inhibition and the observed inactivation-like behavior are characteristic features of open channel block and are compatible with a direct interaction of the blocker with the pore region of Kv1.1 channels.6,11,12,22,23,29 

Binding sites for local anesthetics on voltage-dependent potassium channels have been postulated to reside in the internal pore of the ion channel as well as in the region of the S6 helix and the S5–S6 linker of the channel protein.6,11,22,23The structural homology of the pore region in Kv1.1, Kv1.5, and Kv4.3 allows interpreting the results of Kv1.1 channel block with reference to Kv1.515as well as to Kv4.3 channels.11Amino-amide local anesthetic binding to Kv1.5 channels involves lipophilic interaction at the position leucin 508 and valine 512 in the S6 segment of the channel as well as interaction with threonine 477 within the internal tetraethylammonium binding domain.15Because the corresponding positions in Kv1.1 channels carry the same amino acids (L402, V406, and T371), our results are consistent with the idea that amino-amide local anesthetics inhibit Kv1.1 channels by interacting with these lipophilic amino acids in the S6 region. The correlation of lipophilicity of the local anesthetics as determined by correlating the number of methyl groups within the side chains with the potency to inhibit Kv1.1 channels adds further evidence to the idea that the interaction of amino-amide local anesthetics with these voltage-dependent potassium channels is lipophilic.23Despite potential subtype specificity in local anesthetic action on Kv channels,23lipophilic interaction of local anesthetics with this group of ion channels, thus, seems to constitute a common pharmacologic mechanism. In parallel to clinical knowledge, inhibition of Kv1.1 channels follows the same correlation of lipophilicity and analgetic potency.30To obtain equipotent analgesia in patients, higher concentrations of mepivacaine and ropivacaine have to be administered compared with bupivacaine.30 

All local anesthetics induced a concentration-dependent shift of the activation midpoint into the hyperpolarizing direction. This shift correlated with lipophilicity as well and may be explained by interference of local anesthetics with the coupling of voltage sensing and channel opening of Kv1.1 channels.31Interaction of local anesthetics with Kv1.1 channels is, therefore, unlikely to be restricted to single amino acids in the pore region of the channels. Rather, it seems to result from a more complex interaction with the channel protein. The interaction may likely involve regions of the S6 segment that are necessary for coupling the activation gate of these potassium channels to their voltage sensor residing in the S4 region.23,31Recently, a concerted contribution of the S4–S5 linker and the S6 segment to the modulation of voltage-dependent potassium channels by n-alkanols has been described.31A similar interaction of local anesthetics with the coupling of the activation gate to the voltage sensor of Kv1.1 channels rather than a direct interaction with the voltage sensor may, therefore, offer a possible explanation for the shift of the activation midpoint. Taken together, the effects on current inhibition and activation gating allow hypothesizing that local anesthetics, by binding to lipophilic amino acids L402 and V406 in the pore region, alter channel conductance and modulate gating by interference with the activation gate.

Stereoselective drug binding implies that the drug receptor is able to discriminate between the stereoisomers of a drug molecule.32Franqueza et al .15reported that bupivacaine block of Kv1.5 channels is stereoselective, with the R  (+) enantiomer being sevenfold more potent than the S  (−) enantiomer. In Kv1.5 channels, the stereoselective effects of bupivacaine enantiomers were reduced by substituting the amino acid threonine at position 505 by valine.15Threonine at the corresponding position 399 of Kv1.1 channels may suggest stereospecific interaction with Kv1.1 channels as well. Although this was the case, inhibition of Kv1.1 channels by amino-amide local anesthetics displayed a four- to five-times-weaker stereoselectivity than Kv1.5 channels. This quantitative difference in stereoselectivity between Kv1.1 and Kv1.5 channels supports the hypothesis that differences regarding the pharmacologic effects of local anesthetic action on potassium channels exist.8,16,23These difference likely result from subtype and even species differences in channel structure.23The difference in stereoselectivity between Kv1.1 and Kv1.5 channels may, furthermore, imply that despite structural similarities, the degree of lipophilic and polar interactions differ between different voltage-dependent potassium channel subtypes. Because the pore region is well conserved among different voltage-dependent potassium channels, such as Kv1.1 and Kv1.5, other regions of the protein or the lipid–protein interface may be involved in drug channel interaction. In this context, it is interesting to note that the three-dimensional orientation of R  (+) as compared with S  (−) ropivacaine reduces the difference in potency between S  (−) ropivacaine and S  (−) bupivacaine introduced by an additional CH²group in the latter by a factor of two. Furthermore, Kv1.1 is the only potassium channel that has been systematically studied with the same pattern of stereospecificity as sodium channels.33 

In view of the physiologic significance of Kv1.1 channels, interaction with amino-amide local anesthetics may cause increased neuronal excitability.9Because Kv1.1 channels are predominately expressed in myelinated nerve fibers and large dorsal root ganglia neurons,19binding of local anesthetics to Kv1.1 channels may impair and delay the repolarization of the resting membrane potential of neurons.9This would slow the recovery of inactivated sodium channels and therefore decrease the availability of these channels, which would consequently increase the refractory period for action potential firing.9Inhibition of Kv1.1 channels may therefore potentiate conduction block induced by sodium channel inhibition. A similar mechanism has been postulated for n-butyl-p-aminobenzoate, a new local anesthetic that blocks the delayed rectifier current of dorsal root ganglia neurons.9On the other hand, central nervous system toxicity may also result from interaction of local anesthetics with potassium channels.8,16Because of limitations of a heterologous expression system that was intentionally used to establish concentration–response curves in relation to clinical concentrations of the different local anesthetic agents, the obtained results of this study do not allow any valid conclusion to be drawn about clinical implications. However, it is striking that Kv1.1 channels would be inhibited at neurotoxic plasma concentrations34,35reached after accidental intravascular injection of a dose of the local anesthetic used for regional anesthesia. Furthermore, the total plasma concentration of the different local anesthetics measured during convulsing36,37correlates with the IC⁵⁰values of Kv1.1 channel inhibition (r = 0.93, P < 0.05). Also, the lower potency of S  (−) ropivacaine and S  (−) bupivacaine as compared with racemic bupivacaine to inhibit Kv1.1 channels agrees with evidence for higher safety of S  (−) ropivacaine and S  (−) bupivacaine over racemic bupivacaine resulting from animal toxicity studies.37,38Finally, similar qualitative and quantitative differences in stereoselectivity observed in sodium channels33as well as in Kv1.1 channels may support the hypothesis that inhibition of Kv1.1 channels is important. Different molecular fingerprints of local anesthetics on individual potassium channels as evidenced by different degrees of sensitivity as well as stereoselectivity of effects11–13,15,16may, thus, contribute to a molecular explanation of the complex clinical effects of these widely used therapeutic agents.

In summary, our results demonstrate that block of Kv1.1 channels by amino-amide local anesthetics is concentration dependent, voltage dependent, stereoselective, frequency independent, and reversible. Inhibition likely results from an open state–dependent blocking mechanism. Our results support the idea of a lipophilic interaction of amino-amide local anesthetics with Kv1.1 channels because the lipophilicity of the local anesthetic agents determines the potency to induce gating changes as well as to inhibit Kv1.1 channels. Nonetheless, because stereospecific effects reduce the increased potency resulting from the introduction of a single methyl group by a factor of two, also polar interactions may influence the potency to inhibit Kv1.1 channels. Inhibitory concentrations of amino-amide local anesthetics significantly correlate with systemic concentrations measured during neurotoxic side effects of these pharmacologic agents. Resolving whether human Kv1.1 channels play a significant role in local anesthetic drug effects, therefore, warrants further study. Regardless of future results, our study suggests that human Kv1.1 channels constitute an important biophysical model for elucidating molecular mechanisms underlying clinical effects or side effects of amino-amide local anesthetics.

The authors thank Andrea Zaisser (Technician, Institute of Neuronal Signal Transduction, Center for Molecular Neurobiology, University Medical Center Hamburg, Hamburg, Germany) for cell culture and Cornelia Siebrands, Ph.D. (Postdoctoral Fellow, Department of Biochemistry and Molecular Biology I, University Medical Center Hamburg), for critically reading the manuscript. The authors are grateful for the support of Olaf Pongs, Ph.D. (Director, Institute of Neuronal Signal Transduction, Center for Molecular Neurobiology, University Medical Center Hamburg).