Tonic Blocking Action of Meperidine on Na⁺and K⁺Channels in Amphibian Peripheral Nerves

The patch–clamp method 19was applied to sciatic nerves 17of the clawed toad Xenopus laevis . Animals were killed by decapitation. These procedures were reported to the local veterinarian authority and were in accordance with German guidelines. Nerves were dissected and desheathed mechanically and incubated with 3 mg/ml collagenase (Worthington type CLS II; Biochrom, Berlin, Germany) in Ringer’s solution for 135 min and subsequently with 1 mg/ml protease (type XXIV; Sigma Chemical Co., St. Louis, MO) in Ca²⁺-free Ringer’s solution for 35 min. The temperature was kept constant at 24°C. Next, the preparation was washed in Ca²⁺-free Ringer’s solution, cut into 3-mm-long segments, and put into a culture dish that had been coated with laboratory grease (Glisseal Blue; Borer Chemie, Solothurn, Switzerland). After 30 min of rest, axons were used for up to 10 h.

Experiments were performed with different external solutions containing (in mM) 110 NaCl, 2.5 KCl, 2 CaCl², 5 BES (N,N-bis[2-hydroxyethyl]-2-aminoethanesulfonic acid; Ringer’s solution), and either 100 nM tetrodotoxin (Ri-TTX; Sigma, Deisenhofen, Germany) or 10 mM tetraethylammoniumchloride (Ri-TEA), all adjusted to p  H 7.4 with tris[hydroxymethyl]aminomethane (TRIS) base. In some experiments, an external solution with an increased concentration of K⁺has been used (105 mM K^o), containing 105 KCl, 13 NaCl, 2 CaCl², 5 BES, and 100 nM TTX.

The internal solution contained (in mM) 105 KCl, 13 NaCl, 5 BES, 3 EGTA adjusted to p  H 7.2 with TRIS base (KClⁱ) or CsCl instead of KCl for the investigation of Na⁺current (CsClⁱ). Meperidine was obtained as a commercial preparation (Dolantin; Hoechst, Bad Soden, Germany) that did not contain other ingredients except water. Lidocaine was obtained from Sigma (Deisenhofen, Germany). Control solution and different test solutions were applied to excised outside-out patches by a multiple barrel perfusion system.

All experiments were performed using outside-out patches formed from the membrane of partially demyelinated axons. Patch pipettes were pulled from borosilicate glass tubes (GC 150; Clark Electromedical Instruments, Pangbourne, United Kingdom), coated with Sylgard 184 (Dow Corning, Seneffe, Belgium), and fire-polished before the experiment. The pipettes had a resistance of 4–20 MΩ. Voltage clamp was performed by an EPC 7 patch–clamp amplifier (List, Darmstadt, Germany). Voltage-dependent currents were filtered at 3 kHz with a four-pole low-pass Bessel filter, digitized at 10 kHz with a Labmaster TM-40 analog-digital/digital-analog board (Scientific Solutions, Solon, OH) and stored on the hard disk of an IBM-compatible computer. Traces of voltage-independent currents were stored on videotape using a modified PCM-501ES pulse code modulation unit (Sony, Tokyo, Japan) and processed offline later. Recordings were made at 14 ± 1°C. Membrane potentials (E) are given for the inner side with respect to the outer side of the membrane.

Na⁺currents were investigated in outside-out patches formed from the membrane of the nodal region of the axon. Ri-TEA was used as the external and CsClⁱas the pipette solution. In voltage–clamp mode, the membrane holding potential was set to −90 mV. To remove fast inactivation of the Na⁺channels completely, a 50-ms prepulse to −130 mV was applied before a 50-ms depolarizing test pulse to −40 mV elicited Na⁺currents. To reduce noise originating from random channel openings in subsequent traces, 20 traces of Na⁺currents were averaged before the peak Na⁺current was measured to quantify fractional block in given drug concentrations. These data were used to construct concentration–inhibition curves, which were then fitted with to obtain half-maximal inhibiting concentration (IC⁵⁰) values.

For the investigation of voltage-dependent K⁺currents (K^DR), Ri-TTX was used as the bath and KClⁱas the pipette solution. Holding potential was −90 mV; a 50-ms test pulse to 0 mV elicited K⁺currents. Twenty traces were averaged, and the steady-state K⁺outward current was measured during the last 20 ms of the averaged trace. Concentration–effect relations were evaluated similarly for Na⁺channels.

The calcium-dependent K⁺channel (K^Ca) was investigated in outside-out patches at a holding potential of +40 mV. At this potential, other K^DRinactivated after 1 min and thereafter made no contribution to the current. The bath solution was Ri-TTX, and the pipette solution was KClⁱ, which contained 100 μM CaCl²to activate K^Cacurrents. For analysis, current traces of at least 30 s in control solution and in different concentrations of drug were filtered at 1 kHz and digitized at 5 kHz offline from videotape. The mean current of a trace was then calculated as the average of all sample points in the trace. Dividing the mean current in drug by the mean current in control solution and subtracting the result from 1 yields fractional block.

To study flicker K⁺background channels (K^fli), outside-out patches were formed on thin axons (≈5 μm outer diameter). The holding potential was −90 mV, and no test pulse was applied. The bath solution was 105 mM K^o, and the pipette solution was KClⁱ. Fractional block was measured as described earlier.

Current records were analyzed using pClamp 5.5.1 software (Axon Instruments, Burlingame, CA). Data points and error bars are given either as mean ± SEM or as fitted values ± SE (standard error of fit). Statistical analysis, fits, and preparation of the figures were accomplished with Fig.P 6.0 software (Biosoft, Cambridge, United Kingdom).

To quantify block, concentration–effect curves were constructed by measuring the fractional block (fb) in dependence of certain drug concentrations (c). A nonlinear least-squares fit of the data points to

was then performed to obtain the IC⁵⁰values and the maximal achievable block (b^max).

To determine whether meperidine and lidocaine act at the same binding site on the Na⁺channel, we performed competition experiments with these two agents. 20In these experiments, Na⁺channels were preblocked with certain concentrations of lidocaine (c^lido). Concentration–effect relations with the partially blocked channels reveal apparent IC⁵⁰values for meperidine (IC^{50,meperidine,app}) depending on the fraction of channels blocked previously by lidocaine and on the number of binding sites. Competition forthe same binding site results in a shift of the IC^{50,meperidine,app}value according to

where IC^50,lidoand IC^{50,meperidine}are the IC⁵⁰values for lidocaine and meperidine, respectively, as evaluated from concentration–inhibition relations. If the two substances act at different binding sites, no shift should occur; however, possible allosteric interactions have not been considered in this model.

With outside-out patches, formed at the nodal membrane of partially demyelinated axons, voltage-dependent Na⁺currents were investigated after blocking K⁺currents with external TEA and internal Cs⁺. The Na⁺peak inward current in response to a test pulse was in the range of 20–150 pA, corresponding to approximately 20–150 channels simultaneously open at the peak. 17Random openings of the small number of Na⁺channels present in the outside-out patches resulted in fluctuations in amplitude of successive traces, which had to be smoothed by averaging procedures before measurement of peak current. This hindered investigation of use-dependent (phasic) block, which is measured as an increase of Na⁺current inhibition during successive membrane depolarizations at high frequency. We therefore were able to measure tonic block at only low depolarization frequencies.

Meperidine reduced the peak amplitude in a concentration-dependent manner (fig. 1A) in the same concentration range as the local anesthetic agent lidocaine (fig. 1B). The block was always fully reversible for both substances. To compare the potencies of meperidine and lidocaine, fractional block of the peak amplitude was plotted against the concentration of blocker (fig. 1C). Concentrations of meperidine ranged from 1 nM to 3 mM to investigate its effect in a broad range. Fitting to the data revealed IC⁵⁰values of 164 ± 14 μM (n = 10) for meperidine and 172 ± 9 μM (n = 5) for lidocaine. Block was complete at high concentrations of meperidine and lidocaine (maximal block [f^bmax] fixed to 1 for fitting).

Local anesthetic block is most likely mediated by binding to specific receptors on voltage-gated Na⁺channels. 21Meperidine has a similar Na⁺channel–blocking effect in the same concentration range as local anesthetic agents, and therefore it can be postulated that meperidine acts at the same binding site. To address this question, we performed competition experiments. Concentration–effect curves for lidocaine and for meperidine after lidocaine preincubation were constructed as described in Materials and Methods and in figure 1. IC⁵⁰values found for meperidine without and with lidocaine incubation are plotted against concentration of lidocaine (fig. 1D). The apparent potency of meperidine is clearly independent of the concentration of lidocaine, suggesting that there is no competition for a common binding site on the Na⁺channel. These results imply that meperidine acts on the peripheral nerve Na⁺channel via  a different molecular mechanism than local anesthetic agents do.

There is great diversity in K⁺channels in the axonal membrane, 22and the effect of meperidine on some of these channels was investigated in the present study. K^DRof the peripheral nerve have been studied in outside-out patches formed at the nodal or paranodal region of the peripheral nerve. A test pulse to 0 mV from a holding potential of −90 mV elicited K^DR. Similar to Na⁺currents, K^DRwere reversibly blocked in a concentration-dependent manner (fig. 2), but block was not complete even at high concentrations of meperidine. The IC⁵⁰was evaluated as 194 ± 24 μM (n = 8), and an f^bmaxvalue of 0.83 ± 0.03 was reached. We did not perform further investigation of subtypes of K^DR.

To investigate K^Ca, 100 μM CaCl²was added to the pipette solution, and the membrane potential was maintained continuously at +40 mV to activate the channels. After inactivation of the K^DR(1–2 min), single K^Cas were observed. Meperidine also had a blocking effect on these channels at high concentrations as demonstrated in figure 3A. The fast flickery block of this channel produced by meperidine indicates rapid binding (blocked) and unbinding (unblocked) of the drug in the microsecond range. Because the flickering is too fast to be resolved by the system (because of filtering), it manifests as a reduction in amplitude of the channel trace (fig. 3A).

Figure 3Bshows the effect of meperidine on single voltage-independent (background) K⁺channel found mainly in thin myelinated nerve fibers. 23Because of its flickery appearance in control solution, the channel was termed flicker K⁺channel. This channel shares features with the cloned K⁺channel rTASK, 24which belongs to a group of recently discovered two-pore domain background K⁺channels. A noteworthy feature of the flicker K⁺channel is its high sensitivity to lipophilic amide local anesthetic agents, such as ropivacaine and bupivacaine, which might play a role in differential nerve block. 12Block of this channel by meperidine is concentration dependent and reversible and appears as a reduction in single channel amplitude as seen in amplitude histograms. Flickering block by meperidine overlaps the intrinsic flickering of the channel (fig. 3B).

The concentration–effect relations of meperidine for the Na⁺channel and all K⁺channels investigated are plotted in figure 4. IC⁵⁰values as derived from nonlinear least-squares fit were 194 ± 24 μM (n = 8) for K^DR, with b^max= 0.83 ± 0.03. The flicker K⁺and the K^Cawere blocked completely at high concentrations of meperidine (b^maxfixed to 1), with IC⁵⁰values of 139 ± 15 μM (n = 6) and 161 ± 18 μM (n = 5), respectively.

Meperidine block of Na⁺channels and K^DRhas been tested for antagonism by naloxone. In these experiments, partial block of Na⁺channels and K^DRwas induced by 100 and 1,000 μM meperidine, respectively. Addition of naloxone did not influence meperidine block of both channel types (fig. 5).

Among clinically used opioids, meperidine exerts the strongest local anesthetic–like effect 25and has been used successfully for local anesthesia. 1–7In spinal anesthesia, meperidine produces a segmental motor and sensory block comparable to lidocaine. 1,26In in vitro  experiments on rabbit vagal nerves, meperidine inhibited compound action potentials of A and C fibers with IC⁵⁰values of ≈100 μg/ml (350 μM). 8Single dorsal root axons required ≈700 μM meperidine to achieve conduction block in ≈60% of the fibers. 9The IC⁵⁰value of 164 μM for axonal Na⁺channel block in this study is in good agreement with the earlier observations of compound action potentials block if one takes into account that, because of the conduction safety factor, ≈80% of the Na⁺channels have to be blocked before loss of conduction occurs in a peripheral nerve fiber.

In voltage–clamp experiments, potencies of meperidine and other opioids for peripheral nerve impulse blockade have been evaluated. 27Highly potent opioids, however, including fentanyl and sufentanil, failed to block nerve conduction in several experimental studies, even at high concentrations. 9,28,29Electrophysiologic effects of opioids on the central and peripheral nervous system have been reviewed by Duggan and North. 30 

There is a diversity of opioid receptors in the central nervous system. 31In the sensory peripheral nervous system, the existence of opioid binding sites has been proved in cell somata, i.e. , rat dorsal root ganglion neurons, and in the central terminals of these afferent fibers. 32Further, the amount of μ-, δ-, and κ-receptor encoding mRNA expressed in dorsal root ganglion cells has been found to be intense. 33However, there seems to be no electrophysiologic importance of opioid receptors in the soma of dorsal root ganglion neurons. 34Opioid receptors have been shown in sensory peripheral nerve axons, 33and axons are known to convey opioid receptors by axonal transport into inflamed tissues. 35 

In frog sciatic nerve, Hunter and Frank 16demonstrated the contribution of opioid receptors to Na⁺current block. In their experiments, naloxone partially antagonized a morphine-induced conduction block. In contrast, Hu and Rubly 14proposed that the blocking effect of morphine on Na⁺and K⁺currents in frog sciatic nerve is not mediated by opioid receptors. In their experiments, high concentrations of opioids were needed to achieve nerve block. Intravenous application of opioids, which results only in low concentrations in nervous tissues, does not have an effect on peripheral nerve C fibers. 29 

In our experiments, a Na⁺channel–blocking effect mediated via  opioid receptors is unlikely for two reasons. First, the potency of meperidine in blocking Na⁺channels is low. The IC⁵⁰value of 164 μM as found in this study was 80 times higher than concentrations in plasma needed to achieve postoperative analgesia mediated via  opioid receptors, which is ≈500 ng/ml (2 μM). 36An effect on peripheral Na⁺channels in this low concentration range has not been observed for meperidine in our experiments. Second, meperidine block of Na⁺channels could not be antagonized by naloxone, a drug that displaces opioids from the binding site at the opioid receptor. The lack of antagonizing ability by naloxone is interpreted as an nonopioid receptor–mediated effect.

Opioid action on ion channels is mediated via  intracellular second messenger systems. In neurons of the locus coeruleus and plexus submucosus, it has been shown that μ- and δ-opioid receptors can activate inward rectifier K⁺channels by intracellular second messenger systems, thus hyperpolarizing the cell. 37In substantia gelatinosa neurons of guinea pigs, μ- and κ-receptors produce an increase in K⁺conductance. 38Calcium channels are regulated by G^oproteins coupled to opioid receptors, which leads to a reduction of excitability and transmitter release. 39 

In our studies, however, only direct interactions of opioids with ion channels without emphasis on intracellular second messenger mechanisms have been investigated. An activating effect on K⁺channels has not been observed in a broad concentration range in our experiments.

Meperidine has a blocking effect on Na⁺channels in a similar concentration range as lidocaine, which explains the local anesthetic potency of this opioid. Competition experiments between the two agents, however, demonstrated different mechanisms in blocking Na⁺channels.

Meperidine shows no selectivity in blocking different peripheral nerve ion channels (all channels are blocked in a similar concentration range). Similarly, high concentrations of morphine were able to block Na⁺and K⁺currents equipotently in the squid giant axon 40and myelinated nerve. 14The maximal achievable block of K^DRof 83% is probably a result of the ion channel diversity contributing to this current, 17,41and certain channel subtypes may not be sensitive to meperidine. We did not further discriminate between distinct types of K^DRin this study, however.

In contrast, local anesthetic agents inhibit different channel types at different concentrations. They generally have higher potencies in blocking voltage-dependent Na⁺channels than K^DRin peripheral nerves. 18,42This selectivity for Na⁺channel block also has been demonstrated by the axonal patch–clamp method for other substances that have a local anesthetic effect, such as ketamine 43and droperidol (Bräu ME, unpublished data, 1997). For droperidol, competition with lidocaine has been shown in these experiments, revealing a common blocking mechanism for these two agents. Furthermore, lipophilic amide-linked local anesthetic agents, such as bupivacaine and ropivacaine, have much higher potencies in blocking flicker K⁺channels 12than voltage-sensitive Na⁺channels.

The exact physiologic role of axonal flicker K⁺channels remains unresolved; however, it is speculated that these channels set the resting membrane potential in thin myelinated nerve fibers. 23Block of flicker K⁺channels may enhance block of Na⁺channels in small axons by depolarizing the resting membrane potential and thus driving more Na⁺channels into their inactivated state. 12 

The successful use of meperidine for local anesthesia is probably attributable to its low affinity to opioid receptors. Because of this, commercial preparations of meperidine are highly concentrated—50 mg/ml (176 mM)—compared with 0.05 mg/ml (0.15 mM) for fentanyl. Therefore, local application of the preparation of meperidine yields sufficiently high concentrations of drug at the nerve for Na⁺channel blockade, whereas potent opioids such as fentanyl do not. Concentrations in plasma achieved during local anesthesia with meperidine, however, are too low to have an analgesic effect. 44 

The opioid meperidine has a local anesthetic–like activity because of its Na⁺channel–blocking potency. Its affinity to peripheral nerve Na⁺channels is comparably low, as for the local anesthetic agent lidocaine. Its action, however, is mediated by a different binding site on the Na⁺channel but not via  opioid receptors. Current through different K⁺channels of the peripheral nerve is also inhibited by meperidine; inhibiting concentrations are all in a similar range, suggesting a nonselective blocking effect. Nonspecific binding to membrane proteins or nonspecific membrane effects therefore may be the mechanism of action of meperidine on nerve excitability.

The authors thank B. V. Safronov for reading and critically discussing the manuscript.