Stereoselectivity of Bupivacaine in Local Anesthetic–sensitive Ion Channels of Peripheral Nerve 

Sciatic nerves of freshly killed Xenopus laevis  were dissected. The decapitation procedures were reported to the local veterinarian authority and are in accordance with German guidelines. Nerve fibers were enzymatically prepared as described by Jonas et al. , 17leading to dissociation and retraction of their myelin sheath. Na⁺and flicker K⁺currents were recorded using the outside-out configuration of the patch-clamp method. 18 

Micropipettes were pulled from borosilicate glass tubes (GC150; Clark Electromedical Instruments, Pangbourne, UK), coated with sylgard 184 (Dow Corning, Seneffe, Belgium), and fire-polished to give a resistance of 5–10 MΩ for Na⁺channel and 15–30 MΩ for flicker K⁺channel recording.

External solution contained 110 mM NaCl, 2.5 mM KCl, 2 mM CaCl², 5 mM N,N-bis[2-hydroxyethyl]-2-aminoethanesulfonic-acid (BES) and 10 mM tetraethylammonium chloride (TEA)(Ri-TEA) for Na⁺channel recording and 105 mM KCl, 13 mM NaCl, 2 mM CaCl², 5 mM BES, and 0.1 μM tetrodotoxin (high-K^o) for flicker K⁺channel recording; both solutions were adjusted to pH 7.4 with tris[hydroxymethyl]-aminomethane (TRIS).

Pipette solution for Na⁺channel recording (CsClⁱ) contained 110 mM CsCl, 13 mM NaCl, 3 mM ethylene glycol-bis[β-aminoethyl ether] N,N,N′,N′-tetraacetic acid (EGTA), 5 mM BES, adjusted to pH 7.2 with TRIS. For flicker K⁺channel recording CsCl was replaced by 105 mM KCl (KClⁱ). All chemicals were obtained from Sigma, Deisenhofen, Germany. Temperature was kept constant at 14 ± 1°C.

Control and test solutions were rapidly applied to outside-out patches with a multiple-barrel perfusion system. Bupivacaine enantiomers were gifts from ASTRA Pain Control, Södertälje, Sweden, and were dissolved in distilled water to give 10-mM stocks.

Currents were recorded using an EPC 7 patch-clamp amplifier (List, Darmstadt, Germany), filtered with a four-pole low-pass Bessel filter at 1 kHz, digitized with a Labmaster TM-40 AD/DA board (Scientific Solutions, Solon, OH), and stored on a personal computer. PClamp 5.5.1 software (Axon Instruments, Burlingame, CA) was used for data acquisition and evaluation.

For Na⁺channel recordings, outside-out patches were formed from nodal membrane areas. The membrane potential was held at −90 mV in the voltage-clamp mode of the amplifier. Inward Na⁺currents were evoked at 10-s intervals by applying a 50-ms prepulse to −130 mV to remove fast inactivation, followed by a 50-ms test pulse to −40 mV. Noise of recordings was reduced by averaging 20 successive traces before Na⁺current measurement. Fractional inhibition (f^I) of averaged peak Na⁺current was calculated measuring the peak current amplitude in different drug concentrations (c) with respect to control to give concentration–inhibition curves.

Flicker K⁺channels were continuously recorded in outside-out patches formed from thin fibers (about 5 μm). Membrane potential was held at −90 mV and no test pulse was applied. Fractional inhibition was calculated measuring the mean current through multiple-channel patches from a 30-s recording under different drug concentrations with respect to control.

Concentration–inhibition curves were constructed by plotting the fractional current inhibition against blocker concentration. Half-maximal blocking concentrations (IC⁵⁰) were derived from nonlinear least-squares fits of the Hill function,

to the data points of the curves. The Hill coefficient (n) was set to 1, assuming a one-to-one reaction of the local anesthetic molecule with the channel, unless mentioned otherwise.

Data for the concentration-dependent block of the flicker K⁺channel by R(+)-bupivacaine were additionally fitted with the sum of two Hill functions,

where F¹and F²(F¹+ F²= 1) are the fractions of current with the half-maximal blocking concentrations IC^50,1and IC^50,2, respectively.

Competition experiments were performed to rule out the existence of two different binding sites for R(+)-bupivacaine and S(−)-bupivacaine at the flicker K⁺channel. When two drugs (A and B) compete for the same receptor, with both drugs reacting in a one-to-one fashion, the presence of drug B reduces the receptor occupation by drug A so that its apparent dissociation constant is enlarged. In the presence of increasing concentrations of the competitive agonist B, the concentration-inhibition curve for A is shifted along the concentration axis to higher concentrations, without a change in shape and height of the curve. The apparent IC⁵⁰of drug A can be calculated as follows:

Assuming two different binding sites to which drugs A and B bind independently in a one-to-one reaction and assuming the channel is blocked if either one, the other, or both sites are occupied, a simplified model predicts no change in the dissociation constant of drug A in the presence of increasing concentrations of drug B.

Kinetic data for S(−)-bupivacaine were derived from closed- and open-time histograms. To generate closed- and open-time histograms, outside-out patches with only one active flicker K⁺channel were analyzed. Currents were filtered at 200 Hz and recorded with a sample rate of 1 ms to eliminate the flickering in the open state of the channel. Then, a semiautomatic procedure was used to detect the idealized channel events. Closings less than 30 ms and openings less than 15 ms were ignored. Open- and closed-time histograms were constructed with a bin width of 5 ms and fitted with single exponentials using unweighted, nonlinear least-squares fitting to give open-(τ^o) and closed-time constants (τ^c). From the time constants, the rate constants can be easily calculated with

Kinetic data for R(+)-bupivacaine were estimated from current recordings of multiple-channel patches during wash-in and wash-out of the local anesthetics. These recordings were filtered at 60 Hz and fitted with a single exponential using unweighted, nonlinear least-squares fitting to give wash-in (τ^on) and wash-out time constants (τ^off). The rate constants for dissociation k⁻¹can be derived directly from wash-out time constants with

The association rate constant is calculated as a function of τ^onand τ^offas

Data points are given as mean ± SEM, IC⁵⁰values are given as fitted values ± standard error of the fit. Fig.P 5.0 software (Biosoft, Cambridge, UK) was used for creating figures and performing fitting procedures.

Average Na⁺currents in outside-out patches, evoked as described, had a peak amplitude of 20–150 pA. Assuming a single-channel current of 1 pA per channel at −40 mV, 17this corresponds to 20–150 channels simultaneously open at peak. Because of this small number of channels present in the outside-out patches, amplitude fluctuations originating from random channel openings in subsequent traces were considerable and had to be smoothed by averaging procedures before peak current measurement. These procedures, however, impeded investigation of phasic block, which is measured as the increase of Na⁺current inhibition during successive membrane depolarizations at high frequency. Therefore, only tonic block measured at low depolarization frequencies was investigated.

The enantiomers R(+)-bupivacaine and S(−)-bupivacaine reversibly blocked the peak amplitude of axonal Na⁺currents (fig. 1) in a concentration-dependent manner. Concentration–inhibition experiments gave IC⁵⁰values for tonic block of 29 ± 3 μM (n = 10) and 44 ± 3 μM (n = 14) for R(+)-bupivacaine and S(−)-bupivacaine, respectively, revealing a weak stereoselectivity with a stereopotency ratio (+/−) of 1.5.

Flicker K⁺channels were easily identified by their flickery appearance in high-K^osolution at a holding potential of −90 mV. At that potential, this voltage-insensitive channel is active, and disturbances from voltage-activated channels are not to be expected. It was shown previously that high-K^o, in comparison to physiologic Ringer's solution, had no influence on the potency of externally applied bupivacaine. 15 

Externally applied bupivacaine enantiomers reversibly reduced the open probability of the channel in a concentration-dependent manner. R(+)-bupivacaine induced long closings in the range of seconds, and S(−)-bupivacaine induced short closings in the range of milliseconds (fig. 2).

Concentration–inhibition experiments gave IC⁵⁰values of 0.15 ± 0.02 μM (n = 8) and 11 ± 1 μM (n = 7) for R(+)-bupivacaine and S(−)-bupivacaine, respectively, revealing a stereopotency ratio (+/−) of 73 (fig. 3).

The data points from concentration–inhibition experiments were all fitted with a constrained Hill coefficient of 1 under the assumption of a single binding site (equation 1). Fitting of the data with a variable Hill coefficient revealed values close to unity for block of the Na⁺channel by R(+)-bupivacaine and S(−)-bupivacaine (0.97 ± 0.13 and 0.90 ± 0.10, respectively) and for block of the flicker K⁺channel by S(−)-bupivacaine (0.97 ± 0.05, fits not shown). Best fits for the concentration-dependent block of the flicker K⁺channel by R(+)-bupivacaine, however, have been obtained under the assumption of two independent binding processes (equation 2), revealing a fraction F¹of 0.73 exhibiting-high affinity binding (IC^50,1= 0.06 ± 0.01 μM) and a fraction F²of 0.27 exhibiting binding with a lower affinity (IC^50,2= 1.9 ± 0.9 μM).

To rule out the existence of two different binding sites for (+)-enantiomers and (−)-enantiomers, simultaneous block by R(+)-bupivacaine and S(−)-bupivacaine was analyzed for competition. For reasons of simplicity, we conducted and analyzed these experiments under the assumption of one binding site for R(+)-bupivacaine, with an apparent IC⁵⁰of 0.15 μM. Concentration–inhibition experiments for S(−)-bupivacaine in the presence of 0.3, 0.6, and 1 μM R(+)-bupivacaine revealed apparent IC⁵⁰values for S(−)-bupivacaine of 33 ± 3, 66 ± 7, and 76 ± 18 μM, respectively. Predicted values obtained from equation 3, assuming a competitive interaction between R(+)-bupivacaine and S(−)-bupivacaine are 30, 50, and 77 μM, respectively (fig. 4). The experimental data clearly support competition of the enantiomers for a single binding site.

To characterize the blocking kinetics that underlie stereoselective action, we investigated the association and dissociation rate constants k¹and k⁻¹of block for R(+)-bupivacaine and S(−)-bupivacaine. As during control conditions, the flicker K⁺channel had an open probability of more than 0.95; the kinetics of the induced block by S(−)-bupivacaine could be conveniently characterized. Closed- and open-time histograms were derived from single-channel patches (figs. 5A and B, respectively). The mean closed and open times τ^cand τ^o, respectively, were evaluated by fitting the closed- and open-time histograms with single exponentials (see Materials and Methods). From the time constants, the rate constants can be easily calculated (equations 4 and 5). The long closed times induced by R(+)-bupivacaine would necessitate analysis of single-channel recordings, with a duration of up to 60 min, a duration that exceeds the usual lifetime of an outside-out patch in this preparation. To obtain kinetic data for R(+)-bupivacaine, the time constants τ^onand τ^offwere estimated from the current recordings of a multiple-channel patch during wash-in and wash-out of the drugs (fig. 6). This method is not applicable for S(−)-bupivacaine because its offset kinetics are too fast to be resolved by the system of solution exchange used in this study. The wash-in and wash-out records were fitted beginning at the moment when solution exchange was completed. Values for τ^onand τ^offwere obtained by monoexponential fits of the current traces. The rate constants for dissociation k⁻¹can be derived directly from wash-out time constants (equation 6); the association rate constant is calculated as a function of τ^onand τ^off(equation 7). As shown in figure 6, duration of solution exchange was in a time range close to the process of wash-in of R(+)-bupivacaine. Thus, onset kinetics of R(+)-bupivacaine might be faster than it appears from the current decay, and we might have underestimated association rate constants for R(+)-bupivacaine. The kinetic constants are given in table 1. The equilibrium constants (K^D) calculated from corresponding rate constants are comparable with the IC⁵⁰values obtained in concentration–inhibition experiments and therefore justify the different methods that were applied to obtain kinetic data for R(+)-bupivacaine and S(−)-bupivacaine.

As already suggested by figure 2, the closings induced by S(−)-bupivacaine are much shorter (k⁻¹= 8.3 s⁻¹) than those induced by R(+)-bupivacaine (k⁻¹= 0.13 s⁻¹). In other words, unbinding of S(−)-bupivacaine from the binding site is much faster compared to R(+)-bupivacaine. The corresponding ratio k^{−1,S(−)-bupivacaine}/ k^{−1,R(+)-bupivacaine}is 64, which is close to the stereopotency ratio obtained in concentration–inhibition experiments. Association rate constants for S(−)-bupivacaine and R(+)-bupivacaine differ by 2.3-fold.

The Na⁺channel investigated in this work displays a weak stereoselectivity for tonic block by bupivacaine enantiomers with a stereopotency ratio (+/−) of 1.5. This is in accordance with results from peripheral nerve conduction block measured as tonic block of compound action potentials revealing a stereopotency ratio (+/−) of 61.6. Na⁺currents in cells of a rat pituitary cell line (GH³cells) have a similar sensitivity and stereoselectivity to bupivacaine enantiomers, with a ratio of 1.6 for R(+)-bupivacaine over S(−)-bupivacaine at a holding potential of −70 mV. 19 

Two different modes account for Na⁺current inhibition by local anesthetics: Tonic inhibition is measured as the decrease in Na⁺current at low-depolarization frequencies and is thought to be accomplished by binding of local anesthetics to channels in the resting state and inactive states at rest. Use-dependent, or phasic, inhibition is measured as the further decrease in Na⁺current during brief membrane depolarizations repeated at high frequency. This phenomenon is thought to result from increased binding of local anesthetics to open and inactivated states of the channel. Resting channels have a low affinity for local anesthetics, whereas open or inactivated channels have a higher affinity. The changes between low- and high-affinity channels have been explained with voltage-dependent conformational changes that modulate the affinity of the local anesthetic binding site (“modulated receptor hypothesis”20,21) or with voltage-dependent gates that govern local anesthetic access to a binding site with constant affinity (“guarded receptor hypothesis”22).

In this work, interpretations concerning Na⁺channel block are limited insofar as only tonic block could be investigated. However, stereoselectivity of bupivacaine to decrease compound action potentials in isolated frog peripheral nerve is only slightly enhanced for phasic over tonic block with the (+)-isomer being more potent than the (−)-isomer. 6 

Interestingly, block of the inactivated state of the Na⁺channel in guinea pig ventricular myocytes displayed a stereoselectivity with a ratio of 1.7 for R(+)-bupivacaine over S(−)-bupivacaine that is comparable to neuronal Na⁺channels. Tonic block was not stereoselective. 11It seems to be a common feature of Na⁺channels not to show more than a twofold stereoselectivity for the enantiomers of bupivacaine.

The most striking result of this work is the high stereoselectivity of the flicker K⁺channel block, with a stereopotency ratio of 73 for R(+)-bupivacaine over S(−)-bupivacaine. The high stereoselectivity underlies the extremely high and specific affinity of bupivacaine to this channel.

The existence of two different binding sites for (+)-enantiomers and (−)-enantiomers could be clearly ruled out by competition experiments. The theoretical IC⁵⁰value for the racemic mixture of bupivacaine, assuming equal presence of enantiomers and competition for one binding site, is calculated to be 0.3 μM and is in good agreement with previously reported experimental values for the racemic mixture (0.21 μM 15).

Comparing the potencies of R(+)-bupivacaine and S(−)-bupivacaine is complicated because, for concentration-dependent block of the flicker K⁺channel by R(+)-bupi-vacaine, best fits have been obtained under the assumption of two independent binding processes, whereas concentration-dependent block by S(−)-bupivacaine was adequately described under the assumption of one binding site. Interestingly, a similar phenomenon was seen in a study that investigated the molecular basis for stereoselective bupivacaine block of the human cardiac K⁺channel hKv1.5, 23which is preferentially blocked by R(+)-bupi-vacaine, with a stereoselectivity ratio of 7 over S(−)-bupivacaine. 24Concentration-dependent block of mutations T505I, T505S, and T477S by either enantiomer revealed biphasic concentration–inhibition curves that were not apparent in wild-type channels. This observation was not dependent on stereoselective properties because these were preserved for mutations T477S and T505S, but not for T505I. The favored hypothesis to explain this phenomenon was the existence of multiple open states with different bupivacaine affinities. Mutations T505I, T505S, and T477S would enhance a drug-induced shift in gating to reveal an open state with lower affinity at higher bupivacaine concentrations.

However, based on kinetic and pharmacologic properties, the flicker K⁺channel does not seem to be related to hKv1.5. Thus, different mechanisms are likely to underlie stereoselective properties and biphasic concentration–effect curves of these two K⁺channels. Further conclusions regarding the molecular structure of the local anesthetic binding site at the flicker K⁺channel cannot be drawn.

Unlike in a study that investigates heterologously expressed channels, in our preparation, we cannot exclude the existence of two different channel populations with different affinities for R(+)-bupivacaine. Unfortunately, we were not able to obtain adequate closed- and open-time histograms for R(+)-bupivacaine to answer the question of the existence of either two channel populations or different channel conformations with different affinities for R(+)-bupivacaine. However, both binding processes apparent for R(+)-bupivacaine show higher affinities (183-fold and 5.8-fold) compared to binding of S(−)-bupivacaine.

In addition to the overall equilibrium constants, rate constants of association and dissociation give additional information about the process of drug–receptor interaction, although great care has to be taken in comparing affinities and rate constants of drugs for a receptor and deriving general features concerning the structure of the receptor and the forces that constitute the interaction. 25 

According to our data, the potencies of the enantiomers are predominantly dependent on the dwell times of the molecules. This is a phenomenon suggested in the rate theory of drug action. 26It could account for the fact that the binding site is easily and equally accessible for both molecules, whereas the more potent molecule has a more stable interaction with the binding site and therefore stays bound longer at the site. In other words, the stereoselective difference between the bupivacaine enantiomers to block the flicker K⁺channel derives from the difference in the rate constants of dissociation, implying that there is equal access to the binding site for both enantiomeric forms.

For the open-channel block of the human cardiac K⁺channel Kv1.5, it was shown, in contrast, that the difference in potencies of R(+)- over S(−)-bupivacaine derived from a difference in the association rate constants, suggesting that the less potent molecule is less likely to access the binding site in an optimal configuration. 24One possible explanation was that the S(−)-enantiomer needs to adopt a less favored conformation with a higher free energy profile to bind at the site compared to the R(+)-enantiomer.

Likewise, the calculated dissociation rate constants for the slightly stereoselective block of the inactivated state of the cardiac Na⁺channel were shown to be similar, whereas the association rate constants seemed to determine the stereoselective difference between the twofold more potent R(+)-bupivacaine over S(−)-bupivacaine. 11 

In our experiments, we found a 2.3-fold difference in association rate constants for S(−)-bupivacaine over R(+)-bupivacaine. In contrast to the discussed data for cardiac K⁺channels (Kv1.5) and cardiac Na⁺channels, in our investigation, the association rate constant for the less potent S(−)-bupivacaine was faster compared to the association rate constant for R(+)-bupivacaine. This again indicates that the association process does not account for the stereoselective difference in potency between R(+)-bupivacaine and S(−)-bupivacaine and may even attenuate it.

In conclusion, our kinetic data show that, not only quantitatively, but also qualitatively, block of the flicker K⁺channel by bupivacaine enantiomers represents a unique example for stereoselectivity with a tight fit between R(+)-bupivacaine and the binding site.

However, the pharmacologic relevance of the flicker K⁺channel block remains elusive. The channel has only been described in Xenopus laevis  so far, and a putative human correlate might be less sensitive to bupivacaine. Furthermore, the role of the flicker K⁺channel in neuronal excitability is unclear. Because the channel is mainly found in thin, myelinated fibers, it was suggested that block of this channel might lead via  membrane depolarization to an enhanced conduction block of thin fibers, giving a possible explanation for differential nerve block. 15Recently, the effect of bupivacaine was analyzed on voltage-clamped intact Xenopus laevis  nodes of Ranvier, 27in which no effect of bupivacaine on the voltage-independent leak current or the resting potential was found. However, the axons that were studied ranged in diameter from 10 μm to 20 μm and thus were too large to contain flicker K⁺channels.

Given a conduction block of thin fibers by bupivacaine, differential nerve block should exhibit a high stereoselectivity, favoring R(+)-bupivacaine. However, clinical studies do not support this theory. The 0.75% S(−)-bupivacaine applied for extradural anesthesia gave a significantly longer sensory block compared to racemic bupivacaine and no difference in onset time, maximum spread of sensory block, or intensity of motor block between the groups. 5Another investigation found a shorter onset of pain relief after the main dose of extradurally applied 0.5% racemic bupivacaine compared to 0.5% ropivacaine (the S(−)-enantiomer of 1-propyl-2′,6′-pipecoloxylidide, a structural homologue of bupivacaine, 1-butyl-2′,6′-pipecoloxylidide) but no significant differences in duration, onset of pain relief with the subsequent dose, quality of analgesia, or spread of sensory and motor block. 28The majority of studies reported that ropivacaine produced a less intense motor block than racemic bupivacaine. 29,30 

Recently, a new family of K⁺channels has been described. Their molecular specialties are two-pore domains and four-transmembrane segments. Members of this family have been cloned from mammalian and human tissue and are abundant, particularly in the brain, heart, and lungs. 31–34Similar to the flicker K⁺channel, these channels are K⁺-selective background channels and most likely play a pivotal role in the control of the resting membrane potential and thus in the modulation of electrical activity of excitable cells. Among these channels, electrophysiologic and pharmacologic properties of the TWIK (tandom pore domain weak inwardly rectifying K⁺channel)-related acid-sensitive K⁺channel (rTASK)34best resemble those of the flicker K⁺channel. 14,15However, the IC⁵⁰for racemic bupivacaine to block rTASK is, at 68 μM, remarkably higher than the IC⁵⁰for racemic bupivacaine to block the flicker K⁺channel. 34 

Further investigation will show whether the flicker K⁺channel belongs to this specific family of K⁺channels or whether other channels that share structural features with the flicker K⁺channels are targets for highly stereoselective side effects in local anesthesia.

The authors thank Dr. Rune Sandberg of Astra Pain Control, Södertälje, Sweden, for the generous gift of bupivacaine enantiomers, and Dr. Gary R. Strichartz, Boston, for helpful discussions.