Voltage-dependent sodium channels play an essential role in the propagation of action potentials in the heart as well as in the entire central and peripheral nervous system. Although in the past sodium channels had been regarded as less structurally diverse than other ion channels , different sodium channels isoforms have since been identified. Cardiac sodium channels are structurally and functionally distinct from those of neuronal or skeletal muscle origin . Addressing the question whether different sodium channel isoforms differ in their responses to clinical therapeutic agents may constitute a first step towards designing more specific drugs and therapies. Concerning the relative insensitivity of the cardiovascular system to general anaesthetic action compared with the central nervous system (CNS) (the heart still functions sufficiently at concentrations, where the patient is unconscious) the question arises whether cardiac sodium channels are less sensitive to general anaesthetic action than sodium channels in the CNS and which of their molecular properties might be responsible.
Molecular functions and pharmacological interactions of sodium channels can be studied with two separate techniques that allow the recordings of electrophysiological properties of single sodium channel molecules: the patch-clamp technique and the planar lipid bilayer technique. The patch-clamp technique has been successfully employed to study different human sodium channels in native cells and in expression systems (for reviews see, e.g [3–5]). However, when anaesthetic effects on sodium channels from different tissues are compared, it is necessary to allow for the possibility that differences in anaesthetic sensitivity may not only arise from changes in the amino acid sequence of the ion channel [6–8] but also from alterations after post-translational modification . Thus, differences in anaesthetic potency may not become apparent when sodium channels are expressed heterologously in foreign cells using molecular biological methods. In addition, due to methodological limitations, material removed during surgery can be used for a few hours only, allowing but a few single channel experiments from each patient.
Therefore, we used native human cardiac sodium channels reconstituted into planar lipid bilayers in the presence of batrachotoxin. Besides requiring only small amounts of fresh or frozen human tissue, this method proved to be a reliable and reproducible method for the direct examination of single human sodium channels . Because pharmacological agents such as anaesthetics might act by causing changes in the physiochemical properties of the lipid membrane, experimental control of the lipid environment of the channel becomes important. This option is available in the planar bilayer approach. It is currently the only technique that has been successfully used to study intensively the interactions of different drugs with human brain sodium channels from a single preparation [10,11].
In this initial study, sodium channels from human ventricular muscle were fused with planar lipid bilayers and exposed to pentobarbital. Pentobarbital, a thiopental analogue, was chosen as the use of an intravenous (i.v.) compound allowed the maintenance of a constant anaesthetic concentration during the long-duration experiments required in these studies. The effects of pentobarbital on single channel conductances, single channel fractional open times and the steady-state activation behaviour of single sodium channels were investigated. They were compared with data on pentobarbital block of human brain sodium channels, of cardiac sodium channels from other species, and of human cardiac channels from recombinant systems or expressed human cell lines under the same experimental conditions.
Materials and methods
With the approval of the local Committees on Human Rights in Research, human ventricular muscle samples were obtained from four patients with congestive heart failure undergoing heart transplantation. The source of tissue was the patient's heart, which was removed prior to the transplantation and was considered surgical waste. Samples were extracted from macroscopically unaffected parts of the left ventricular muscle directly after explantation. Samples were immediately frozen at −80°C. Membrane preparation was as described for canine heart  and stored at −80°C.
Most materials and experimental methods are described elsewhere [12,13]; a brief description is given below. All experiments were conducted at room temperature (22–24°C) in either symmetrical 500 mmol L−1 NaCl or symmetrical 100 mmol L−1 NaCl buffered at pH 7.4 with 10 mmol L−1 HEPES (United States Biochemical); no corrections were made for temperature differences between experiments. Planar bilayers were formed from neutral phospholipid solutions containing (4:1) 1-palmitoyl-2-oleoyl-phosphatidylethanolamine and 1-palmitoyl-2-oleoyl-phosphatidylcholine (Avanti Polar Lipids, Birmingham, AL, USA) in decane (5% wt/v, 99.9% pure; Wiley Organics, Columbus, OH, USA). Batrachotoxin (BTX) was a gift from Dr J. Daly. Pentobarbital (acid form) was purchased from Sigma (Sigma Chemical Co., St. Louis, MO, USA).
Teflon chambers were prepared and used as previously described ; the chambers were divided by a Teflon partition into a cis compartment to which the preparation was added, and a trans compartment. The partition had a hole, approximately 300 µm, in its centre; planar bilayers were formed over this aperture. Sodium channels were incorporated into the bilayers in the presence of 0.5 µmol L−1 batrachotoxin; the electrophysiological sign convention was used in the presentation of results.
Channel currents were recorded under voltage-clamp conditions and filtered at 50 Hz. Time-averaged conductances were measured by a computer. After incorporation of a sodium channel into the bilayer, control currents were measured for at least 60 min and the sidedness (extra- and intracellular side) of the channel experimentally determined.
In some experiments, steady-state activation properties were determined as described . In the range of channel activation, channel fractional open time (fo) could be described by a two-level Boltzmann distribution with one open and one closed state.
The fractional open time (fo) is characterized by the steady-state mid-point potential (Va, the potential at which the channel is open 50% of the maximal fraction fmax), and a valence of the effective gating charge (za), related to the slope of the sigmoidal curve (Figure 1b) (V = membrane potential, F = Faraday constant, R = gas constant, T = absolute temperature).
After the control recording period, pentobarbital was added from an ethanol stock solution to the aqueous compartment facing either the extracellular or the intracellular side of the channel. Pentobarbital block (Block[PTB]) was calculated according to:
where g[PTB] is the time-averaged conductance in the presence of pentobarbital and g[control] is the time-averaged conductance in the absence of pentobarbital. The protocol of the control measurements was repeated with increasing concentrations of pentobarbital. The same channels were studied in the presence of pentobarbital for up to 6 h, and the experiments usually ended when the bilayer ruptured.
Standard deviations (SD) were used in the presentation of averages; paired-sample (Student's) t- tests (P < 0.05) were used to examine significance.
Even with the sodium channel activator batrachotoxin present, incorporation of single channels into the lipid bilayer are rare events under our experimental conditions. Out of 100 experiments (each yielding a total recording time from several membranes of 245–512 min) 49 experiments showed incorporation of sodium channels. In some experiments incorporation of additional channels (16) or unstable background conditions (10) led to an exclusion from the study, in 13 experiments the lipid membranes broke during the repeated administration of drugs. Therefore, only the 10 remaining experiments in which a single channel was successfully exposed to pentobarbital were included in this study. However, 10 additional experiments were added to the control data.
These channels could be clearly identified as sodium channels as they were sodium selective and blockable by tetrodotoxin, a sodium channel specific inhibitor. In addition, they had a comparatively low tetrodotoxin sensitivity, characteristic of the cardiac origin of the sodium channels .
Figure 2a shows the recordings of one single cardiac sodium channel activated by batrachotoxin. The channel remains open most of the time because batrachotoxin impairs channel fast inactivation . Unlike the sodium channel of the human brain, which is open about 95% of the time and has only singular and short closures , the cardiac channel showed frequent closures, lasting not only hundreds of milliseconds, but up to seconds. This results in a fractional open time of 0.85 (± 0.04 SD, n = 13). The fractional open time was independent of membrane potentials at depolarized potentials (−45 mV to +45 mV).
After the addition of pentobarbital (Figure 2b–d), sodium channels underwent more frequent transitions between a fully open and a fully closed state (‘flickering’). This flicker became too rapid for full resolution. Therefore, this effect was quantified by averaging the current over time (Figure 3a); from these data the fractional open times were calculated. Pentobarbital induced a block of the channels, which was concentration-dependent and independent of membrane potential between −45 mV and +45 mV. The data were fitted to a rectangular hyperbola and an IC50 value of 690 µmol L−1 was estimated, with maximal suppression of 100% (Figure 3b, filled symbols); the Hill coefficient was 1.2.
After addition of pentobarbital (Figure 2b–d), long channel closures (> 0.2 s) were still present and were not changed in frequency, duration or noise level. Under control conditions the frequency was 14.1 min−1 ± 5.4 min−1 and the duration 571 ms ± 58 ms (n = 13) which was similar to the frequency of 13.3 min−1 ± 4.4 min−1 and duration of 498 ms ± 71 ms observed after the addition of 670 µmol L−1 pentobarbital (n = 4).
In 100 mmol L−1 NaCl electrolyte, the channel fractional open time decreased to 0.81 (± 0.02 SD) from 0.85 in 500 mmol L−1 NaCl. Pentobarbital under these conditions again induced a concentration-dependent block of fractional open time with an IC50 value of 406 µmol L−1 and a maximal suppression of 100% (Figure 3b, open symbols); the Hill coefficient was 1.0.
Under the influence of pentobarbital the channel current-voltage relationship stayed symmetrical, the single channel conductance was unchanged (24.7 pS, symmetrical 500 mmol L−1 NaCl solutions; see Figure 3a).
Reduction of NaCl concentration from symmetrical 500 mmol L−1 to symmetrical 100 mmol L−1 resulted in a reduction in single channel conductance. Single channel conductance decreased from 25.0 pS to 19.0 pS (± 0.9, n = 7). Pentobarbital had no significant effect on either single channel conductance.
At membrane potentials more hyperpolarized than −45 mV the fractional open time (but not the single channel conductance) became dependent on membrane potential, decreasing successively with hyperpolarized potentials until the channel was totally closed at a potential more negative than −120 mV (Figure 4a). This effect represents steady-state activation, which – due to batrachotoxin modification – is shifted to more hyperpolarized potentials .
In this range of channel activation, channel fractional open time (fo) could be described by a two-level Boltzmann distribution (Figure 4a) characterized by the steady-state mid-point potential Va = −102.6 mV ± 4.2 mV and za = 5.8 ± 1.0 (n = 9; 500 mmol L−1 NaCl).
With increasing concentrations of pentobarbital the mid-points of activation were shifted to more negative potentials (Figure 4b, filled symbols, and Table 1). At all examined pentobarbital concentrations these shifts were significant.
Pentobarbital in 100 mmol L−1 NaCl electrolyte caused similar shifts in the mid-points of steady-state activation (Figure 4b, open symbols, and Table 1). At all concentrations these shifts were again significant.
The human cardiac sodium channels examined here have the qualitative biophysical characteristics expected for batrachotoxin-modified sodium channels in planar lipid bilayers, but they differ in several physiological and pharmacological properties from human brain sodium channels. Cardiac sodium channels gate at more negative membrane potentials, they spend more time in the closed state and they are considerably less sensitive to tetrodotoxin. These differences must result from the molecular structure of the sodium channels themselves because they were compared within identical membranes under identical experimental conditions, an advantage of the planar lipid bilayer technique. In addition, this technique allows the observation of the steady-state properties of human sodium channels for hours without functional degradation , an advantage that is particularly important in the examination of drug actions. Finally, the bilayer system has been successfully applied in dissecting different molecular components of anaesthetic action without the need of resorting to controversial mathematical modelling .
Despite these differences, human sodium channels from brain and ventricular muscle show a number of striking similarities in their response towards the anaesthetic pentobarbital. The pentobarbital block on cardiac fractional open time had an IC50 of 690 µmol L−1, which was not significantly different from human brain sodium channels under bilayer conditions (660 µmol L−1) with the two concentration-response curves nearly overlapping (Figure 1a). In this regard the human cardiac sodium channel does not deviate from 10 other sodium channels which show variations of IC50 values of a factor of four or less , in contrast to the subtype variability of other molecular targets of anaesthetics such as GABAA receptors  and acetylcholine receptors .
The IC50 value of channel block doubled when the symmetrical 100 mmol L−1 NaCl electrolyte was changed to 500 mmol L−1 symmetrical NaCl. While such electrolyte dependence on anaesthetic sensitivity has not been reported for sodium channels before, it has been observed for voltage-gated (Kv type) potassium channels . At clinical concentrations of pentobarbital (88 µmol L−1) , the blocking effect on sodium channels in bilayers is 18% in 100 mmol L−1 NaCl and 9% in 500 mmol L−1 NaCl. Without knowledge of the neuronal networks involved in anaesthesia it is not yet possible to judge whether this block is sufficient to be of clinical importance. However, effects on sodium channel inactivation as little as 2% may be clinical relevant and give cause to myotonia .
Pentobarbital caused a hyperpolarizing shift in the mid-point of the steady-state activation curves of cardiac sodium channels. This shift was significant for all three examined pentobarbital concentrations under both NaCl concentrations. A shift of similar magnitude was observed for human brain sodium channels which was not significantly different (Figure 1b).
There are studies suggesting the involvement of sodium channels in cardiac pathology in certain conditions. Anomalous repolarization in congestive heart failure might in part be due to an increased steady-state inward sodium current . Muller-Ehmsen and co-workers found  that there is ‘an enhanced sensitivity of the failing human myocardium towards sodium-channel modulation, …. this may be due to an altered Na(+)-homeostasis in human heart failure.’ These pathologies might be caused either by changes in factors modulating sodium channel function or by changes in sodium channel structure itself. Because there are no modulating factors in the bilayer system, any changes in sodium channel function observed in the bilayer system would be due to changes in sodium channel structure. For obvious reasons we cannot obtain healthy human cardiac tissue. However, the basic properties of human cardiac sodium channel in the bilayer in this study are very similar to those of cardiac sodium channels from other species [25,26]. Although we find also that the anaesthetic sensitivity of these human cardiac sodium channels are not different from that of human CNS sodium channels, we cannot completely rule out that human cardiac sodium channels from healthy tissue have a different anaesthetic sensitivity.
In conclusion, our results suggest that despite pharmacological and electrophysiological differences between human brain and human cardiac sodium channels in bilayers their responses to the anaesthetic pentobarbital are similar both when channel conductances and channel steady-state activation are considered. Therefore, this cannot explain the observed differences in anaesthetic sensitivity of the CNS and the cardiovascular system. It is still possible that the inactivation properties of sodium channels, their residing in different membrane environments or being part of different neuronal networks could explain such differences . The finding of channel block by anaesthetics depending on the electrolyte composition is novel for sodium channels but not for potassium channels  and should be investigated further.
We thank Z. Dorner and A. Brambeer for technical assistance and Dr B. Rehberg for many helpful discussions.
1 Hille B. Ionic Channels of Excitable Membranes
, 2nd edn. Sunderland, UK: Sinauer Assoc Inc., 1992.
2 Catterall WA. Common modes of drug action on Na+ channels: local anesthetics, antiarrythmics and anticonvulsants. Trends Pharmacol Sci
1987; 8: 57–65.
3 Balser JR. Structure and function of cardiac sodium channels
. Cardiavasc Res
1999; 42: 327–338.
4 Goldin AL. Diversity of mammalian voltage-gated sodium channels
. Ann N Y Acad Sci
1999; 868: 38–50.
5 Lehmann-Horn F, Jurkat-Rott K. Voltage-gated ion channels
and hereditary disease. Physiol Rev
1999; 79: 1317–1372.
6 Guo XT, Uehara A, Ravindran A, Bryant SH, Hall S, Moczydlowski E. Kinetic basis for insensitivity to tetrodotoxin and saxitoxin in sodium channels
of canine heart
and denervated rat skeletal muscle. Biochemistry
1987; 26: 7546–7556.
7 Mandel G. Tissue-specific expression of the voltage-sensitive sodium channel. J Membr Biol
1992; 125: 193–205.
8 Wang DW, George Jr AL, Bennett PB. Comparison of heterologously expressed human cardiac and skeletal muscle sodium channels
. Biophys J
1996; 70: 238–245.
9 Recio-Pinto E, Thornhill WB, Duch DS, Levinson SR, Urban BW. Neuraminidase treatment modifies the function of electroplax sodium channels
in planar lipid bilayers. Neuron
1990; 5: 675–684.
10 Frenkel C, Duch DS, Urban BW. Molecular actions of pentobarbital
isomers on sodium channels
from human brain cortex. Anesthesiology
1990; 72: 640–649.
11 Frenkel C, Duch DS, Urban BW. Effects of i.v. anaesthetics on human brain sodium channels
. Br J Anaesth
1993; 71: 15–24.
12 Recio-Pinto E, Duch DS, Levinson SR, Urban BW. Purified and unpurified sodium channels
from eel electroplax in planar lipid bilayers. J Gen Physiol
1987; 90: 375–395.
13 Duch DS, Recio-Pinto E, Frenkel C, Urban BW. Human brain sodium channels
in bilayers. Brain Res
1988; 464: 171–177.
14 Wartenberg HC, Duch DS, Urban BW. Distinct molecular sites of anaesthetic
action: Pentobarbitone block of human sodium channels
is alleviated by removal of fast inactivation. Br J Anaesth
1999; 82: 76–80.
15 Wasserstrom JA, Liberty K, Kelly J, Santucci P, Myers M. Modification of cardiac Na+ channels by batrachotoxin: effects on gating, kinetics, and local anesthetic binding. Biophys J
1993; 65: 386–395.
16 Frenkel C, Wartenberg HC, Duch DS, Urban BW. Steady-state properties of sodium channels
from healthy and tumorous human brain. Brain Res Mol Brain Res
1998; 59: 22–34.
17 Rehberg B, Wittmann M, Urban BW. Sodium channels
(from rat, mouse and man) in neuroblastoma cells and different expression systems have similar sensitivities to pentobarbital. Neuroscience Lett
1999; 264: 81–84.
18 Davies PA, Hanna MC, Hales TG, Kirkness EF. Insensitivity to anaesthetic
agents conferred by a class of GABA (A) receptor subunit. Nature
1997; 385: 820–823.
19 Eterovic VA, Li L, Ferchmin PA et al.
The ion channel of muscle and electric organ acetylcholine receptors: differing affinities for noncompetitive inhibitors. Cell Mol Neurobiol
1993; 13: 111–121.
20 Friederich P, Urban BW. The inhibition of human neuronal K+ currents by general
anesthetic agents is altered by extracellular K. Brain Res Mol Brain Res
1998; 60: 301–304.
21 Urban BW, Duch DS, Frenkel C, Rehberg B, Wartenberg HC. [Do general
anesthetics act on specific receptors?]. Anasthesiol Intensivmed Notfallmed Schmerzther
1995; 30: 375–382.
22 Cannon SC, Corey DP. Loss of Na+ channel inactivation by anemone toxin (ATX II) mimics the myotonic state in hyperkalaemic periodic paralysis. J Physiol (Lond)
1993; 466: 501–520.
23 Undrovinas AI, Maltsev VA, Sabbah HN. Repolarization abnormalities in cardiomyocytes of dogs with chronic heart
failure: role of sustained inward current. Cell Mol Life Sci
1999; 55: 494–505.
24 Muller-Ehmsen J, Frank K, Brixius K, Schwinger RH. Increase in force of contraction by activation of the Na+/Ca (2+)-exchanger in human myocardium. Br J Clin Pharmacol
1997; 43: 399–405.
25 Renaud JF, Kazazoglou T, Lombet A et al.
The Na+ channel in mammalian cardiac cells. Two kinds of tetrodotoxin receptors in rat heart
membranes. J Biol Chem
1983; 258: 8799–8805.
26 Zamponi GW, Doyle DD, French RJ. State-dependent block underlies the tissue specificity of lidocaine action on batrachotoxin-activated cardiac sodium channels
. Biophys J
1993; 65: 91–100.
27 Weigt HU, Rehmert GC, Bosnjak ZJ, Kwok WM. Conformational state-dependent effects of halothane on cardiac Na+ current. Anesthesiology
1997; 87: 1494–1506.
28 Rehberg B, Duch DS, Urban BW. The voltage-dependent action of pentobarbital
on batrachotoxin-modified human brain sodium channels
. Biochim Biophys Acta
1994; 1194: 215–222.