Succinylcholine Metabolite Succinic Acid Alters Steady State Activation in Muscle Sodium Channels

Wild-type α subunits of human skeletal muscle sodium channels (hSkM1 13) were heterologously expressed in human embryonic kidney (HEK293) cells (American Tissue Culture Collection CRL 1573). The plasmid pRc (Invitrogen, San Diego, CA) was used for mammalian transfection. Plasmids containing the α subunits were transfected into HEK293 cells using the calcium phosphate precipitation method. 14Permanent expression was achieved by selection for resistance to the aminoglycoside antibiotic geneticin G418 (Life Technology, Eggenstein, Germany). 15Transfected cells were a gift from Professor Lehmann-Horn, Ulm, Germany. The clone has been used in several investigations. 15,16Successful transfection and expression was confirmed electrophysiologically.

Each experiment consisted of recordings in bath solution (control and washout) and in test solution containing succinic acid (Sigma Chemicals, Deisenhofen, Germany) in different concentrations (0.05–10 mm) derived from a 10-mm stock in bath solution. Patch electrodes contained 130 mm CsCl², 2 mm MgCl², 5 mm EGTA, and 10 mm Hepes; bath solution contained 140 mm NaCl, 1 mm MgCl², 4 mm KCl, 2 mm CaCl², 5 mm Hepes, and 5 mm dextrose. Succinic acid–containing solutions were adjusted to pH 7.4 by addition of CsOH, and osmolarity was at 290 mOsm. Temperature during data acquisition was maintained at 20°C (Temperature Control System, List Medical, Darmstadt, Germany).

HEK cells also express endogenous Na⁺channels that conduct with amplitudes ranging from 50 to 350 pA (mean, 112 ± 12 pA). 15To minimize a possible contribution of endogenous Na⁺channels, but also to avoid series resistance errors, we only analyzed currents with peaks (I^peak) ranging between 1.5 and 6 nA.

Standard whole-cell voltage-clamp current recordings 17were obtained at 20°C. Each patched cell was exposed to one concentration of succinic acid only; for each concentration, experiments derived from at least six patches were evaluated. For data acquisition and further analysis, we used the EPC9 digitally controlled amplifier in combination with Pulse and Pulse Fit software (HEKA Electronics, Lambrecht, Germany). The EPC9 provides automatic subtraction of capacitive and leakage currents by means of a prepulse protocol. Cell capacities ranged from 9 to 15 pF, and residual series resistance (after 50% compensation) ranged from 1.2 to 2.5 MΩ. Experiments during which an increase in series resistance occurred were rejected. The time constant of voltage settling in the membrane (residual series resistance × cell capacitance) was generally < 35 μs.

For steady state voltage dependence of activation, peak currents (I) elicited by families of depolarizing pulses from a prepulse potential of −120 mV (15 ms duration, holding potential −100 mV) to the indicated test potentials (−50 mV, −45 mV, and so on, in 5-mV increments to +45 mV) were plotted against the test potentials; conductance (g) was obtained using g = I/(V − V^rev), where V is the test potential and V^revis the reversal potential. The reversal potential (or zero-current potential) V^revwas obtained by extrapolation assuming that the decline in peak current amplitude after depolarization to more positive test potentials (with respect to the test potential at which maximum current was observed) is related to the difference between test potential and reversal potential according to Ohm’s law. Thus, V^revis the zero current point on the abscissae derived from the linear peak current–voltage relationship after depolarizations to test potentials ranging from +15 to +45 mV in 5-mV increments according to I^Vtest= (dI^{+15:+45 mV}/dV^{+15:+45 mV}) × V^test+ I^{0 mV}. For V^test= V^rev, I^Vtest= 0, and V^rev=−I^{0 mV}/(dI^{+15:+45 mV}/dV^{+15:+45 mV}).

Boltzmann fits [g/g^max= (1 + exp(−zF(V − V^0.5)/RT))⁻¹] to the conductance–voltage plots yielded the voltage at half-maximum conductance (V^0.5) and the slope factor z^a. Concentration–response curves for drug effects on voltage dependence of activation were obtained by plotting the mean difference in V^0.5in the indicated test concentration and V^0.5in the corresponding control experiment [δV^0.5= V^0.5(test ) − V^0.5(control )] against the applied concentration of succinic acid. Small time-dependent shifts in V^0.5occurring in a series of control experiments in bath solution (n = 12) were subtracted from the shifts observed in the presence of the drug. Hill fits [δV^0.5=δV^{0.5 max}/(1 + (EC⁵⁰/[C])^nH)] to the data yielded the concentration for half-maximum effect on voltage-dependent activation and the Hill coefficient n^Hdescribing stoichiometry of drug binding to the channel; δV^{0.5 max}is the maximum δV^0.5in the highest concentration of succinic acid applied (10 mm).

The voltage dependence of Na⁺-channel inactivation was obtained from a double-pulse protocol, in which the cell membrane was first conditioned by a 20-ms prepulse starting at −150 mV to prepulse potentials ranging from −150 mV to −5 mV, immediately followed by a 4-ms test pulse to 0 mV. The peak current (I^Na) in response to this step was normalized to I^Naelicited by the test pulse at the most negative prepulse potential (−150 mV). Boltzmann fits to the resulting current–voltage plots yielded the voltage at half-maximum inactivation (V^0.5) and the slope factor z: I/I^max=[1 + exp(−zF(V − V^0.5)/RT)]⁻¹.

Time constants of inactivation, τ^h, at 0 mV were obtained from single exponential fits to the decaying currents after 40-ms voltage jumps from −100 mV to 0 mV: I(t) = a⁰+ a¹exp(−t/τ^h1).

Data are presented as mean ± SD. Because small time-dependent shifts in steady state activation and inactivation occur independent of drug administration, the main goal was to demonstrate the significance of succinic acid–induced effects with respect to a series of control experiments (n = 12), exposed to bath solution only in the same time interval after seal formation (3–5 min). Changes in parameters of interest (V^0.5of activation and inactivation, slope factor, maximum current and conductance, time constant of inactivation) derived from single experiments (n = 6) in each drug concentration were tested for statistical significance. The data partly showed positive skew indicative of non-normal distribution. The small sample size in each group mandated the use of the nonparametric Mann—Whitney U test;P < 0.05 was considered significant. To avoid problems related to multiple hypotheses testing, we applied the method of multiple comparisons with a priori  ordered hypotheses. 18The test is based on the assumption that, if the null hypothesis is rejected, there is a positive monotonic relation between concentration and effect. As a consequence, the hypotheses to be tested could be ordered in advance, starting with the highest concentration of succinic acid. If the test results differed significantly from the control data, the effects of the next lower concentration were evaluated. The evaluation was stopped as soon as the first insignificant result was obtained. The advantage of this procedure, compared with other approaches to multiple testing, is that the level of type I error is kept at α= 0.05 for each statistical test. Furthermore, if the procedure starts with a significant result, the global null hypothesis is also rejected. Because of the time-dependent shifts in steady state activation, the starting values before drug application were not always reached during washout. Thus, washout was considered successful when the test results during washout did not differ significantly from the results obtained from the control experiment in bath solution in the same time interval after seal formation (5–8 min).

Control values for V^0.5of channel activation were −14.8 ± 5.3 mV (n = 57), and the slope factor z^awas 3.7 ± 0.4. Time constant of inactivation (τ^h) in the controls before succinic acid (10 mm) administration (n = 9) was 0.46 ± 0.09 ms, V^0.5of channel inactivation was −59.2 ± 3.6 mV, and the slope factor zⁱwas 2.5 ± 0.2. During application of bath solution only (n = 12), the time-dependent shift in V^0.5was −2.1 ± 1.6 mV for steady state activation and −0.6 ± 0.9 mV for steady state inactivation.

Peak currents elicited by families of depolarizing voltage steps from a prepulse potential of −120 mV were enhanced in the presence of succinic acid compared with the corresponding starting values (figs. 1A and 1B). Enhancement of maximum current ranged from 13% to 37% of control currents (table 1). Statistical significance of this effect with respect to the time-dependent control experiments could be shown for 10 mm succinic acid.

Conductance–voltage plots obtained from the peak currents during each depolarizing voltage step represent the fraction of available sodium channels (with respect to the maximum of open channels in the indicated test solution) that open at each test potential. The test potential at half-maximum activation (V^0.5) reflects the position of the curve along the voltage axis. Succinic acid shifted the midpoints of steady state activation in the direction of more negative test potentials, indicating that a greater proportion of channels open during smaller depolarizations in the presence of the drug (fig. 1C). The maximum amount of this shift was −6.3 ± 1.7 mV. Taking into account the small negative shift in the controls, a drug-related shift of −4.2 mV was observed in 10 mm succinic acid. Washout was positive for concentrations > 0.5 mm. The concentration–effect curves are depicted in figure 1D. A Hill fit to data yielded an EC⁵⁰value of 0.39 mm succinic acid (the small shift of −2.1 mV observed in the control experiments was considered to be point 0); the Hill coefficient n^Hwas 1.02. The slope factor z^aremained unchanged in the presence of 10 mm succinic acid (3.8 ± 0.5; ^P= 0.6) with respect to the starting value (3.7 ± 0.4).

Extrapolated values for V^revwere in good agreement between control (77.3 ± 9.4 mV), test (72.9 ± 12 mV) and washout (76.5 ± 8.5 mV) experiments (n = 57). No significant results were obtained for the enhancement of maximum conductance in the presence of succinic acid (table 2).

Succinic acid (10 mm) did not affect voltage-dependent inactivation or time course of channel inactivation: V^0.5for steady state inactivation in 10 mm succinic acid was shifted by −2.3 ± 2.0 mV into hyperpolarizing direction (to −61.5 ± 2.8 mV); this effect was not significant with respect to the −0.6 ± 0.9 mV shift in control solution (P = 0.08). The slope factor zⁱin the presence of drug was 2.3 ± 0.2 versus  2.5 ± 0.2 in the controls (P = 0.7). Time constant of inactivation (τ^h) in 10 mm succinic acid was 0.36 ± 0.05 ms versus  0.45 ± 0.09 ms in the controls (P = 0.2).

Our results show that succinylcholine metabolite succinic acid facilitates voltage-dependent activation of human muscle sodium channels in a concentration-dependent manner without affecting the rate and voltage dependence of inactivation. The control data we obtained from heterologously expressed α subunits of human skeletal muscle sodium channel confirm gating kinetics previously described for those channels. 15,16 

The observed effect of succinic acid on voltage-dependent activation is qualitatively comparable to the effect of certain anions on the voltage dependence of gating in frog skeletal muscle: For frog muscle fibers, sodium salts added to the external medium have been shown to potentiate the strength of a twitch in the order Cl⁻< Br⁻< NO³⁻< I⁻< SCN⁻. 19,20The primary cause of twitch potentiation by anions such as nitrate or thiocyanate was attributed to a shift of the voltage dependence of contractile activation toward more negative potentials. 21,22Dani et al.  23showed that anions in the external solution shifted the voltage dependence of gating in frog skeletal muscle with little effect on the steepness or maximum sodium permeability. The sequence of effectiveness was acetate < Br⁻< NO³⁻< SO⁴²⁻< SCN⁻< ClO⁴⁻. However, those anions shifted activation less than inactivation with an order of magnitude ranging from −3.7 ± 0.4 mV for Br⁻to −12.7 ± 0.7 mV for ClO⁴⁻when all the chloride in the bath was replaced by the given anion. Removal of negatively charged sialic acid residues from cells expressing rSkM1 sodium channels showed opposite effects, shifting the voltage dependence of gating into depolarizing direction. 24Thus, we assume that the acidic moiety of succinic acid is not important, and the observed effects should rather be attributed to the succinate anion. In addition, decreasing the external pH to 4.5 would shift Na⁺channel activation into depolarizing direction, whereas increasing the pH to 10 itself would cause a −8 mV hyperpolarizing shift. 25Nevertheless, the effects of succinic acid/succinate are far more specific: For the hyperpolarizing shift of steady state activation, we found an EC⁵⁰value of 0.4 mm, whereas the anion effects described previously were observed at > 100× higher concentrations. In the concentration range examined, succinic acid/succinate did not affect voltage-dependent inactivation. Thus, we speculate that the substance interferes more specifically with the voltage sensor of activation. The amino acid sequences of the known voltage-operated cation channels (Na⁺, K⁺, Ca²⁺) show striking similarities. One of the transmembrane α helices that is highly conserved in all known voltage-gated cation channels contains regularly spaced, positively charged amino acid residues. This helix has been implicated as the voltage sensor in these channels. 26It is thus tempting to speculate that succinic acid might exert similar effects on those channels.

For this study we used α subunits of the human skeletal muscle sodium channel heterologously expressed in a mammalian cell line (HEK293). Despite the lack of β subunits, the suitability of this preparation for studying channel gating kinetics has been verified experimentally: The α subunit is the primary pore-forming subunit of the channel and functions as a ion channel when expressed alone. 27Expression of hSkM1 α subunits in Xenopus  oocytes resulted in Na⁺currents with abnormally slow inactivation. In contrast, expression of the same α subunit in a mammalian cell line (tsA201, a transformed HEK293 cell line) resulted in channels exhibiting normal (with respect to experiments in native tissue) rapid activation and inactivation and retaining sensitivity to tetrodotoxin and μ-conotoxin. 28Other factors that might modify channel function in vivo , such as post-translational modification or phosphorylation of channel proteins, 27are beyond the scope of this model.

The results obtained for succinic acid at the molecular level are difficult to frame into clinical terms because they probably do not translate directly into the same effect at the tissue or organismal levels. First, small increases in sodium channel conductance have been shown to result in substantial changes in the action potential firing threshold of the squid giant axon, rendering the membrane hyperexcitable. 29Thus, a relatively small effect at the molecular level might be amplified in vivo . Second, gating processes are highly temperature-sensitive, and effects of a substance at 20°C might be accelerated at body temperature. To get an idea of the clinical relevance in vivo , we can compare the effects observed in the presence of succinic acid to the effect of decreasing of the Ca²⁺concentration in bath solution on voltage-dependent activation: Lowered [Ca²⁺]^ocauses nerve and muscle hyperexcitability, a condition seen in the clinic in patients with hypoparathyroidism. 25The amount of hyperpolarizing shift observed in high concentrations of succinic acid is comparable to the effect that lowering external [Ca²⁺] from 2 mm to 0.5 mm in frog Ringer’s solution exerts on the voltage dependence of activation in frog skeletal muscle. 25 

On application of a routine dose of succinylcholine, succinylcholine blood levels of 62 μg/ml (0.17 mm) have been reported. 30Thus, equimolar concentrations of succinic acid may be reached, 31given that cholinesterase activity is not impaired. Furthermore, in addition to being the product of succinylcholine hydrolysis, succinic acid is an endogenous dicarboxylic acid (citric acid cycle intermediate) ubiquitous in all body tissues. Tissue and serum levels of succinic acid show a pronounced increase after ischemia 32and hypoxia, up to serum levels of 0.1–0.2 mm. 33Hypoxic tissue is considered to be the source of succinate due to a reversal of the oxidative pathway from succinate to oxaloacetate. Serum levels are supposed to increase slowly, secondary to a breakdown of the permeability barrier to succinate between the cell and the circulation. 33Thus, we have to assume that succinic acid concentrations at the sarcolemma might even exceed serum levels during systemic hypoxia.

Despite the lack of direct evidence at the present time, it is tempting to speculate that elevated tissue and serum levels of succinic acid might have mediated the aberrant response to succinylcholine seen in patients with massive trauma 6,7and exsanguinating hemorrhage, 8as well as succinylcholine-provoked hyperkalemia in rabbits subject to hemorrhage. 34 

In conclusion, our finding that activation of voltage-operated muscle sodium channels is facilitated in the presence of succinic acid now yields a new hypothetical explanation for succinylcholine-induced muscle hyperexcitability despite complete neuromuscular blockade and hypersensitivity to succinylcholine-induced depolarization in cases of severe hypoxia and hypotension. Further studies, however, are required to elucidate the potential role of succinic acid causing membrane hyperexcitability in vivo , i.e. , a quantitative assessment of succinic acid in tissues and serum during shock and hypoxia preceding succinylcholine-provoked hyperkalemia.

The authors thank Professor Lehmann-Horn (Ulm) for providing transfected cells, Birgitt Nentwig (Department of Anesthesiology) for taking care of the cell culture, and Jobst Kilian (Department of Neurology) for technical support.