Molecular Mechanisms of the Inhibitory Effects of Bupivacaine, Levobupivacaine, and Ropivacaine on Sarcolemmal Adenosine Triphosphate–sensitive Potassium Channels in the Cardiovascular System

In the cardiovascular system, K^ATPchannels are present in both cardiomyocytes and vascular smooth muscles and have a range of electrophysiologic properties because of their diverse molecular composition.9,18It was recently found that the cardiac K^ATPchannel isoform is composed of sulfonylurea receptor (SUR) 2A and inward-rectifier K⁺channel subunit (Kir) 6.2 (SUR2A/Kir6.2), whereas the vascular K^ATPchannel isoform may be a complex of SUR2B/Kir6.1.18,19 

The human Kir6.2, rat Kir6.1, rat SUR1, rat SUR2A, and rat SUR2B cDNAs were kindly provided by Susumu Seino, M.D., Ph.D. (Professor and Chairman, Department of Cellular and Molecular Medicine, Chiba University, Chiba, Japan). A truncated form of human Kir6.2 lacking the last 36 amino acids at the C terminus was obtained by polymerase chain reaction amplification as previous described.20Site-directed mutagenesis was carried out using a Site-Directed Mutagenesis system (Invitrogen Corp., Carlsbad, CA). Chimeric SURs (SUR1–2A, SUR1–2B, and SUR2–1) were made by exchanging the 42 C-terminal amino acids of SUR1 or SUR2A with those of SUR2A, SUR2B, and SUR1 with the overlap extension polymerase chain reaction method. All DNA products were sequenced by using a BigDye terminator cycle sequencing kit (Applied Biosystems, Foster City, CA) and an ABI PRISM 377 DNA sequencer (Applied Biosystems) to confirm the sequence.

COS-7 cells were plated at a density of 3 × 10⁵per 35-mm dish and cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum. A full-length Kir complementary DNA (cDNA) and a full-length SUR cDNA were subcloned into the mammalian expression vector pCMV6c. For electrophysiologic recordings, either wild-type or mutated pCMV6c Kir alone (1 μg), or pCMV6c Kir (1 μg) plus pCMV6c SUR (1 μg), were transfected into COS-7 cells with green fluorescent protein cDNA (pEGFP) as a reporter gene using the Lipofectamine and Opti-MEM 1 reagents (Life Technologies Inc., Rockville, MD) according to the manufacturer’s instructions. After transfection, cells were cultured for 48–72 h before being subjected to electrophysiologic recordings.

Membrane currents were recorded in the inside-out configurations using a patch clamp amplifier as described previously.20,21Transfected cells were identified by detecting the green fluorescence under a fluorescence microscope. The intracellular solution contained 140 mm KCl, 2 mm EGTA, 2 mm MgCl², and 10 mm HEPES (pH 7.3). The pipette solution contained 140 mm KCl, 1 mm CaCl², 1 mm MgCl², and 10 mm HEPES (pH 7.4). Recordings were made at 36 ± 0.5°C. Patch pipettes were pulled with an electrode puller (PP-830; Narishige, Tokyo, Japan). The resistance of pipettes filled with internal solution and immersed in the bath solution was 3–7 MΩ. The sampling frequency of the single-channel data was 5 KHz with a low-pass filter (1 KHz).

Channel currents were recorded with a patch clamp amplifier (CEZ 2200; Nihon Kohden, Tokyo, Japan) and stored in a personal computer (Aptiva; International Business Machine Corporation, Armonk, NY) with an analog-to-digital converter (DigiData 1200; Axon Instruments, Foster, CA). pClamp version 7 software (Axon Instruments) was used for data acquisition and analysis. The open probability (P^o) was determined from current amplitude histograms and was calculated as follows:

where t^jis the time spent at current levels corresponding to j = 0, 1, 2, N channels in the open state; T^dis the duration of the recording; and N is the number of channels active in the patch. Recordings of 2–3 min were analyzed to determine P^o. The channel activities were expressed as NP^o. Changes of NP^oin the presence of drugs were calculated as the relative NP^obetween the values obtained before and after drug treatment. When the concentration-dependent effects of bupivacaine, levobupivacaine, and ropivacaine were studied, these drugs were injected into the cell bath using a glass syringe to the final five concentrations in a cumulative manner (total volume injected was approximately 10–20 μl).

The drug concentration needed to induce half-maximal inhibition of the channels (IC⁵⁰) and the Hill coefficient were calculated as follows:

where y is the relative NP^o, [D] is the concentration of drug, Kⁱis IC⁵⁰, and nH is the Hill coefficient.

The following drugs were used: levobupivacaine (S  (−)-bupivacaine; Aldrich Chemical Co., Milwaukee, WI), ropivacaine (AstraZeneca, Osaka, Japan), bupivacaine, glibenclamide, diazoxide, and pinacidil (Sigma-Aldrich Japan, Tokyo, Japan). Bupivacaine, levobupivacaine, and ropivacaine were dissolved in distilled deionized water to yield stock solutions of 1 m from which further dilutions were made to obtain the desired final concentrations. Glibenclamide and pinacidil were dissolved in dimethyl sulfoxide (the final concentration of solvent was 0.01%); dimethyl sulfoxide at a concentration of 0.02% did not affect K^ATPchannel currents.

All data are presented as mean ± SD. Differences between data sets were evaluated either by repeated-measures one-way analysis of variance followed by Scheffé F test or by Student t  test. P < 0.05 was considered significant.

As in a previous study,20inwardly rectifying currents were obtained from excised inside-out patches in COS-7 cells transiently expressing two types of K^ATPchannel isoforms, the SUR2A/Kir6.2 and the SUR2B/Kir6.1 channels, with single-channel conductances of 70 ± 4 pS (n = 25) and 28 ± 3 pS (n = 24), respectively (fig. 1A). Figure 1Bshows representative examples of these currents. When the patch was excised into a nucleotide-free solution, the SUR2A/Kir6.2 channel showed marked current increases that were blocked by 1 mm ATP, whereas the SUR2B/Kir6.1 channel was not spontaneously activated. Both channels were stimulated by 100 μm pinacidil and inhibited by 3 μm glibenclamide (fig. 1B).

To assess the effects of bupivacaine, levobupivacaine, and ropivacaine on recombinant K^ATPchannels, we measured single-channel currents on inside-out patches in the presence of these drugs. Application of 1 mm bupivacaine, levobupivacaine, or ropivacaine to the intracellular membrane surface inhibited the SUR2A/Kir6.2 channel currents, with relative NP^odecreasing to 0.16 ± 0.08, 0.29 ± 0.08, and 0.31 ± 0.07, respectively (fig. 2). The SUR2B/Kir6.1 currents were also inhibited by bupivacaine, levobupivacaine, and ropivacaine at 1 mm, with relative NP^odecreasing to 0.35 ± 0.07, 0.48 ± 0.12, and 0.53 ± 0.14, respectively (fig. 3). The inhibitory effects of bupivacaine, levobupivacaine, and ropivacaine on K^ATPNP^owere reversible because the channel currents recovered after washout (figs. 2, 3). The effects were also concentration dependent (fig. 4).

To further probe the mechanism of these anesthetics, we next tested their effects on the SUR2B/Kir6.2 currents. The IC⁵⁰s and Hill coefficients of bupivacaine, levobupivacaine, and ropivacaine for the SUR2A/Kir6.2, SUR2B/Kir6.2, and SUR2B/Kir6.1 channels are summarized in table 1. Bupivacaine inhibited the SUR2A/Kir6.2, SUR2B/Kir6.2, and SUR2B/Kir6.2 channel activities with higher potency than levobupivacaine, indicating that the inhibitory effect of bupivacaine on K^ATPchannels might be stereoselective. In addition, all of the anesthetics inhibited the SUR2A/Kir6.2 channels with higher potency than the SUR2B/Kir6.2 and SUR2B/Kir6.1 channels, indicating that the inhibition is dependent on the SUR subtype. The blockades by bupivacaine, levobupivacaine, and ropivacaine did not significantly change the conductance of the recombinant K^ATPchannels in any case (data not shown).

Wild-type Kir6.2 alone does not show functional channel activity, whereas a C-terminal truncated pore-forming subunit of Kir6.2 (Kir6.2ΔC36) lacking the last 36 amino acids is capable of forming a functional channel in the absence of SUR.22This has proved to be a useful tool for discriminating the site at which various agents act on K^ATPchannels.

Adenosine triphosphate–sensitive inwardly rectifying currents were recorded from COS-7 cells transiently expressing the Kir6.2ΔC36 channel (fig. 5A), as was done previously.20 Figure 5Aalso shows that bupivacaine at 1 mm reversibly inhibited the Kir6.2ΔC36 currents, with relative NP^odecreasing to 0.34 ± 0.12. The concentration-dependent effects of bupivacaine, levobupivacaine, and ropivacaine on the Kir6.2ΔC36 currents are shown in figure 5B. These observations suggest that the high-affinity site as well as the stereoselective mechanism of bupivacaine is located on the Kir6.0 subunit. The IC⁵⁰and Hill coefficients of bupivacaine, levobupivacaine, and ropivacaine for Kir6.2ΔC36 currents (table 1) indicate that the inhibitory effect of these anesthetics on the Kir6.0 subunit was enhanced by coexpression with SUR2A but not with SUR2B.

We next used several site-specific mutations of Kir6.2ΔC36 channels to identify the regions of Kir6.2 that play critical roles in the bupivacaine-mediated inhibition of Kir6.2ΔC36 channel activity. Kir6.0 subunits possess two transmembrane domains, TM1 and TM2, linked by a pore loop, H5 (fig. 6A). The N and C termini are intracellular and contain the binding sites for intracellular ligands involved in ATP inhibition (arginine 50 [R50] and lysine 185 [K185]),23pH sensitivity (histidine 175 [H175]),24and channel gating (cysteine 166 [C166] and threonine 171 [T171])23,25(fig. 6B).

The effects of 1 mm bupivacaine and 1 mm levobupivacaine, which inhibit wild-type Kir6.2ΔC36 currents in a stereoselective manner, were tested on ATP-insensitive mutants (R50G, K185Q), a pH-insensitive mutant (H175A), and mutants that change the channel gating (C166S, T171A). Bupivacaine and levobupivacaine block Kir6.2ΔC36-R50G, -K185Q, and -H175A currents as effectively as Kir6.2ΔC36 currents (fig. 6C), indicating that the bupivacaine-inhibitory site is not identical to that for ATP (R50 and K185) and protons (H175). In contrast, both bupivacaine affinity and stereoselectivity were decreased by the Kir6.2ΔC36-C166S and -T171A mutations. This indicates that the cytosolic end of the second transmembrane domain of Kir6.2 may play an important role in both affinity and stereoselectivity for bupivacaine binding.

Sulfonylurea receptor 2A and SUR2B are splicing isoforms generated from a single gene that differ from each other only in the 42 amino acid segment at their C-terminal ends.26The C-terminal segment may therefore be responsible for the differential effects of bupivacaine on the K^ATPchannels containing these SUR2 isoforms. The C-terminal 42 amino acids (C42) of SUR2B possess only approximately 30% homology with those of SUR2A but approximately 70% with those of SUR1. Therefore, to clarify the role of C42, we examined the effect of the anesthetics on SUR1 and several chimeric molecules of SUR1 and SUR2 (fig. 7A).

Sulfonylurea receptor 1 and the chimeras SUR2-1, SUR1-2A, and SUR1-2B formed a fully functional K^ATPchannels with Kir6.2. The bupivacaine sensitivity of the Kir6.2ΔC36 channels was not enhanced by coexpression with SUR1 and SUR2-1 at almost similar concentrations to the SUR2B/Kir6.2 channel (fig. 7B). In contrast, SUR1-2A and SUR2A enhanced the bupivacaine sensitivity of Kir6.2ΔC36 channels, although SUR1-2B did not (fig. 7B). These results strongly indicate that SUR2A-C42 plays a specific functional role in the control of the inhibitory effect of bupivacaine on the K^ATPchannel.

In the current study, we expressed reconstituted SUR2A/Kir6.2 and SUR2B/Kir6.1 channels in COS-7 cells and demonstrated that bupivacaine, levobupivacaine, and ropivacaine inhibit these channels in a concentration-dependent manner. Bupivacaine was approximately threefold more potent than levobupivacaine and ropivacaine in blocking both channels. Furthermore, the potency of the local anesthetics to block SUR2A/Kir6.2 channels was approximately eightfold higher than the ability to block SUR2B/Kir6.1 channels. These results suggest that bupivacaine enantiomers and ropivacaine act in a stereoselective and tissue-specific way to inhibit sarcolemmal K^ATPchannels in the cardiovascular system.

Ropivacaine and levobupivacaine, both S  (−) pure enantiomer local anesthetics, exhibit a common chemical structure; they differ only in the length of the alkyl substituent of the tertiary nitrogen (position 1), which is a propyl group for ropivacaine and a butyl group for levobupivacaine. Several studies have shown that the length of this side chain is a structural determinant of the ability of these local anesthetics to inhibit cardiac Na⁺channels, transient outward K⁺channels, and the human Kv1.5 channels4,27,28: the longer the side chain, the higher the potency of the drug. In the current study, ropivacaine blockade of recombinant K^ATPchannel currents is of lower potency than that of levobupivacaine, suggesting that the ability of bupivacaine-type local anesthetics to block K^ATPchannels is also related to the length of their N -substituent.

The K^ATPchannel is a hetero-octamer composed of two subunits: the pore-forming subunit Kir6.0 (Kir6.1 or Kir6.2) and the regulatory subunit sulfonylurea receptor SUR (SUR1, SUR2A, or SUR2B).9Distinct from other inward-rectifier K⁺channels, Kir6.0 subunits do not express functional channels by themselves. However, removal of the last 26 or 36 C-terminal amino acids of Kir6.2 (Kir6.2ΔC26 or Kir6.2ΔC36) results in channels that show significant currents in the absence of SUR.22Our study shows that bupivacaine, levobupivacaine, and ropivacaine inhibit the Kir6.2ΔC36 channels as well as the reconstituted sarcolemmal K^ATPchannels. In addition, the Hill coefficient for the inhibitory effects of bupivacaine was close to unity for SUR2A/Kir6.2, SUR2B/Kir6.1, SUR2B/Kir6.2, and Kir6.2ΔC36 channels, indicating that only a single anesthetic molecule needs to bind for the channel to close. These results also suggest that the high-affinity site of bupivacaine is located on the Kir6.0 subunit. In addition, bupivacaine blockade of Kir6.2ΔC36 currents is more potent than blockade by levobupivacaine and ropivacaine. Therefore, the stereoselectivity of the blockade of K^ATPchannels by bupivacaine strongly indicates that it results from binding to a specific binding site on the Kir6.0 subunit.

It is well known that the intracellular domains of the Kir6.2 subunit possess binding sites for ATP (R50, K185) and protons (H175).23,24The bupivacaine inhibitory site, however, is not identical to the ATP- and proton-binding sites because mutations that markedly decrease the sensitivity to ATP (R50G, K185Q) and protons (H175A) were without significant effect on bupivacaine block.

Recent electrophysiologic studies have indicated that the second transmembrane domain of the Kir6.0 channel lines the intracellular mouth of the pore, because mutations within this region affect the channel gating kinetics.23,25The residues at cytosolic end of the second transmembrane domain of Kir6.2 are critical for the gating of Kir6.2ΔC36 because most mutations at this position, including C166S and T171A, substantially affected the single-channel kinetics; dramatic decrease occurred in the frequency of the long close channel state.23,25Mutations in this region, C166S and T171A, reduced the inhibitory effects of bupivacaine and levobupivacaine, which may result from changing the channel gating kinetics rather than altering the affinity of these anesthetics for its binding site. These mutations also completely abolished the stereoselective block observed in wild-type Kir6.2ΔC36 channels. An equivalent region (the cytosolic end of the sixth transmembrane segments) has been shown to participate in the gating of both Kv channels and cyclic nucleotide-gated channels.29–31Furthermore, mutations within this region of Kv channels have a similar effect on bupivacaine, markedly decreasing both the affinity and stereoselectivity of bupivacaine binding.14The bupivacaine binding site located at the intracellular end of the pore may therefore be a common feature of K⁺channels.

Our IC⁵⁰values for bupivacaine, levobupivacaine, and ropivacaine show that the SUR2A/Kir6.2 channels are more sensitive than the Kir6.2ΔC36 channels. These results indicate that although the primary site at which these anesthetics inhibit the K^ATPchannel resides on Kir6.2, the presence of SUR2A enhances the blockade. In contrast, the IC⁵⁰values for the SUR2B/Kir6.1 and SUR2B/Kir6.2 channels are similar to that for Kir6.2ΔC36 despite differences in the Kir6.0 subunit, suggesting that SUR2B does not regulate the sensitivity to anesthetics. SUR2A and SUR2B are generated from a single gene and differ only in their 42 amino acid residues at the C-terminal tail (C42; amino acids 1505–1546).26Therefore, these anesthetic-mediated inhibitions of the Kir6.0 subunit might be enhanced by the SUR2A tail but not affected by the SUR2B tail. The experiments using the SUR1 chimera whose C-terminal segment was replaced with that of either SUR2A or SUR2B strongly support this possibility. Therefore, the tissue specificity of the effects of these anesthetics on K^ATPchannels is likely due to the presence of different SURs.

High concentrations of bupivacaine act on the cardiovascular system as a depressant; the primary site of action is the cardiomyocytes, but arteriolar vasodilatation is also produced.1,2The cardiotoxic effects of bupivacaine are primarily explained by the blockade of voltage-sensitive Na⁺channels.3Bupivacaine binds much more potently to Na⁺channels during the action potential plateau when the Na⁺channels are inactivated.3Furthermore, hypoxia, acidosis, and hyperkalemia commonly occur during local anesthetic toxicity and greatly potentiate the cardiotoxicity of bupivacaine.6,7These conditions result in partial depolarization,32which increases the fraction of voltage-sensitive Na⁺channels in the inactivated state during diastole. Shortening of action potential duration and hyperpolarization of resting membrane potential have been suggested as potentially effective approaches in the treatment of bupivacaine-induced cardiotoxicity.4Sarcolemmal K^ATPchannels in cardiomyocytes, silent under physiologic conditions, are activated during metabolic stress, such as ischemia, acidosis, and hypoxia.8,9The opening of these channels results in membrane hyperpolarization and a shortening of the action potential duration,33resulting in a decrease in Ca²⁺influx through voltage-dependent Ca²⁺channels. Thus, bupivacaine-induced blockade of sarcolemmal K^ATPchannels, especially in the presence of hypoxemia and acidosis, leads to a prolongation of the action potential duration, which might be expected to enhance the severity of bupivacaine-induced cardiotoxicity. This is confirmed by the fact that the K^ATPchannel openers pinacidil and bimakalim attenuate the bupivacaine-induced atrioventricular conduction block.12In addition to cardiac effects, high concentrations of local anesthetics produce vasodilatation, which may enhance the bupivacaine-induced cardiovascular collapse.1,2In vascular smooth muscle cells, activation of K^ATPchannels, which leads to an increase in K⁺conductance and hyperpolarization of the cell membrane, results in vascular relaxation.26Therefore, the inhibitory effects of bupivacaine on sarcolemmal K^ATPchannels in vascular smooth muscle indicate that these channels are not involved in the vasodilatation effects of bupivacaine.

An experimental study performed in sheep demonstrated that intravascular injection of bupivacaine can result in transient plasma concentrations of the free form of bupivacaine of 4–12 μg/ml (12–36 μm),34assuming a blood-to-plasma concentration ratio of 0.73, which caused a serious depression of cardiac conduction that was difficult to reverse. The threshold concentrations at which bupivacaine (> 10 μm), levobupivacaine (> 30 μm), and ropivacaine (> 30 μm) inhibit reconstituted SUR2A/Kir6.2 channels are within this range. Some types of Kv channels in cardiomyocytes are also inhibited by these local anesthetics.4,5Therefore, the suppression of sarcolemmal SUR2A/Kir6.2 channel activities together with that of other K⁺channels, especially during metabolic stress, would be expected to be accompanied by prolongation of the action potential, which can lead to higher bupivacaine cardiotoxicity. On the other hand, the concentrations of these local anesthetics that inhibited vascular K^ATPchannel activities in the current study are very high (> 100 μm), suggesting that sarcolemmal SUR2A/Kir6.2 channels but not SUR2B/Kir6.1 channels may contribute to local anesthetic-induced cardiovascular toxicity.

Our study has several limitations. First, because one of the aims of the current study was to obtain an information about the structural determinants of stereoselective inhibition of K^ATPchannels using clinically available local anesthetics, we compared the inhibitory effects of racemic bupivacaine and S  (−)-bupivacaine on K^ATPchannel activities. Based on the report of Gonzalez et al.  17showing that the inhibitory effects of both R  (+)-bupivacaine and racemic bupivacaine on HERG K⁺channel activities were lower than those induced by S  (−)-bupivacaine, we expected it would be possible to estimate at least a part of stereoselective effects of bupivacaine on sarcolemmal K^ATPchannels in the cardiovascular system. However, the racemic mixture may exhibit properties that are not shared by either of the isomers, as a result of interactions between the isomers. Therefore, it should be more appropriate to compare the effects of R  (+)-bupivacaine and S  (−)-bupivacaine to know stereoselective inhibition of bupivacaine on K^ATPchannel activities. Second, as we discussed above, the threshold concentrations at which bupivacaine inhibited reconstituted SUR2A/Kir6.2 channels in the current study (>10 μm) are within plasma concentrations of bupivacaine (12–36 μm) obtained after intravascular injection of bupivacaine in sheep.34Because we used inside-out patch clamp configurations, the threshold concentrations of bupivacaine obtained in the current study should be compared with intracellular bupivacaine concentrations. However, we are unaware of any report evaluating intracellular bupivacaine concentrations especially in an accidental overdose of bupivacaine.

In conclusion, the Kir6.0 subunit mediates the inhibition of reconstituted sarcolemmal SUR2A/Kir6.2 and SUR2B/Kir6.1 channel activities by bupivacaine, levobupivacaine, and ropivacaine. These inhibitory effects of local anesthetics on K^ATPchannels in the cardiovascular system are stereoselective, with bupivacaine more potent than levobupivacaine and ropivacaine, and tissue specific, with local anesthetics blocking SUR2A/Kir6.2 channels more potently than SUR2B/Kir6.1 channels. Our results further suggest that determinant domains are located at the intracellular pore mouth of Kir6.0 subunit for stereoselectivity and in the 42 amino acids at the C-terminal tail of SUR2A subunit for tissue specificity.