Voltage modulates halothane-triggered Ca²⁺ release in malignant hyperthermia-susceptible muscle

Halogenated gaseous drugs are extensively used for inducing reversible loss of consciousness in general anesthesia. These drugs are thought to perform their function through weak interactions with specific protein targets such as GABA-A and NMDA receptors in brain regions involved in the control of sleep and arousal (Franks, 2008).

A possible devastating side effect of volatile anesthetics is the induction of malignant hyperthermia (MH) in genetically predisposed individuals. MH is a hypermetabolic crisis resulting from an increase in the cytoplasmic Ca²⁺ concentration in skeletal muscle leading to muscle rigidity, hyperthermia, acidosis, tachycardia, hyperkalemia, and rhabdomyolysis (Rosenberg et al., 2015). It originates from a drug-induced leakiness of the SR, mostly caused by mutations in the Ca²⁺ release channel (ryanodine receptor RyR1; Hwang et al., 2012; Guerrero-Hernández et al., 2014). In recent years, several of these human mutations have been introduced into the mouse genome (Chelu et al., 2006; Yang et al., 2006), providing excellent murine models to study the functional implications on the organismic and cellular level.

A functional hallmark of MH susceptibility is a significantly higher sensitivity to a variety of contraction-inducing drugs (Hopkins, 2011). This is the basis of standardized diagnostic tests (in vitro contracture tests [IVCTs]) that use force recording on muscle biopsies and use caffeine and halothane as triggering drugs (Larach, 1989; Rosenberg et al., 2002; Hopkins et al., 2015). The animal model used in the present study is based on human RyR1 mutation Y522S (Y524S in the mouse genome) and shows strong and lethal MH-like episodes caused by either application of volatile anesthetics or elevated ambient temperature (Chelu et al., 2006).

RyR1 is located in the membrane of the terminal cisternae of the SR and is in mechanical contact with the dihydropyridine receptor (DHPR; L-type Ca²⁺ channel, CaV1.1) of the adjacent transverse tubular membrane (Franzini-Armstrong and Protasi, 1997; Van Petegem, 2015). The DHPR serves as a sensor of the cell membrane potential. Action potentials initiated at the postsynaptic face of the neuromuscular junction spread into the transverse tubular system and cause rapid release of Ca²⁺ from the SR, which activates contraction (Lanner et al., 2010; Hernández-Ochoa et al., 2016). The Ca²⁺ release results from the voltage-induced conformational signal transmission from the DHPR to the RyR1 and does not require any entry of Ca²⁺ from the outside through the L-type Ca²⁺ channel (Armstrong et al., 1972; Spiecker et al., 1979; Melzer et al., 1995; Dayal et al., 2017), contrary to the situation in heart muscle (Eisner et al., 2017).

Because RyR1, the central protein in excitation-contraction coupling (ECC), is voltage-activated via the DHPR, MH mutations in RyR1 can be expected to affect depolarization-induced Ca²⁺ release. That this is indeed the case has been demonstrated on various occasions. The first hint was the demonstration by Gallant and co-workers that external potassium-induced contraction exhibited a lower threshold in muscle of MH-susceptible (MHS) pigs compared with controls (Gallant and Donaldson, 1989; Gallant and Lentz, 1992). In agreement with this observation, our group demonstrated that the voltage threshold for Ca²⁺ release in MHS muscle cells was lower than in normal myocytes (Dietze et al., 2000; Andronache et al., 2009). Given the higher voltage sensitivity of the ECC system in muscle cells with mutant RyR1, it can be speculated, but has not yet been demonstrated, that halogenated anesthetics may further lower the threshold of voltage-activated Ca²⁺ release.

The basis of our study, therefore, is the hypothesis that MH-inducing gaseous drugs, when acting directly on the RyR1, will alter characteristics of voltage-activated Ca²⁺ release. These could show up as alterations of the time course and amplitude of voltage-triggered Ca²⁺ release and as a change in its voltage dependence.

We pursued this question by recording Ca²⁺ signals in single isolated muscle fibers of WT and MH-susceptible mice and applying the volatile anesthetic drug halothane. The membrane potential was altered by eliciting action potentials and using a voltage clamp method.

Male heterozygous WT/Y524S mice (Chelu et al., 2006) and age-matched homozygous WT littermates were used for the experiments. For simplicity, the genotypes will be referred to as YS (mutant) and WT, respectively. The animals were bred in the specific pathogen–free facility of the Animal Research Center of Ulm University.

All experimental procedures performed on mice were in accordance with German animal protection laws, approved by the local animal welfare committee, and conducted under the project license O.47 of the Institutional Animal Care and Use Committee of Ulm University (Tierforschungszentrum, Universität Ulm) with approval by the regional administrative authority (Regierungspräsidium Tübingen). Adult mice were killed by CO₂ application and rapid cervical dislocation. Interosseous muscles were dissected from the hind limbs in Ringer’s solution. Single muscle fibers were enzymatically dissociated in Ringer’s solution containing 2 mg/ml collagenase (Liu et al., 1997; Ursu et al., 2005).

Ringer’s solution for extracellular stimulation experiments was (mM) 145 NaCl, 5 KCl, 1 MgCl₂, 2.5 CaCl₂, 10 HEPES, 0.1 N-benzyl-p-toluene sulfonamide (BTS), and 10 glucose, pH 7.4. External (bathing) solution for voltage clamp experiments was (mM) 130 TEA-OH, 130 HCH₃SO₃, 1 MgCl₂, 2.5 CaCl₂, 5 4-aminopyridine, 10 HEPES, 0.001 tetrodotoxin, 10 glucose, and 0.05 BTS, pH 7.4.

Cells were loaded with 5 µM fura-2-AM for 45 min. Remaining external dye was washed out with the experimental solution, and the cells were incubated for 30 min to allow for full intracellular cleavage of the acetoxymethyl ester. The myosin II blocker BTS was added to suppress muscle contraction.

Measurements of action potential–triggered fura-2 Ca²⁺ signals were performed as described (Braubach et al., 2014) using an inverted fluorescence microscope (Axiovert 100; Zeiss) equipped with a photomultiplier tube and extracellular stimulation system. Rectangular current pulses (0.5-ms duration) were applied through two stainless steel electrodes immersed in the bath solution close to the cell. Screening for responsive cells was performed using sequential pulses of different voltage (6, −7, 8, and −9 V). Then, pulses of equal size but opposite sign, spaced 500 ms apart, were applied, and their amplitudes were gradually increased. Measurements were confined to cells that produced equal all-or-none reactions to both the positive and the negative stimuli. The final pulse voltage for further measurements was set to 1 V above the excitation threshold. Fluorescence was excited consecutively with UV light of 360 nm (near isosbestic wavelength) and 380 nm (bandwidth 10 and 15 nm, respectively). Fibers were stimulated during the 380-nm irradiation interval.

Fluorescence ratio signals R = F380/F360 (for 380 and 360 nm excitation, respectively) were calculated and converted to free Ca²⁺ concentration ([Ca²⁺]), including temporal deconvolution (Carroll et al., 1995), using dissociation constant and off-rate constant of the dye of 278 nM and 50/s, respectively. Ratios for Ca²⁺-free (R_min) and Ca²⁺-saturated (R_max) conditions were 4.0 and 0.7, respectively, based on calibrations. Most experiments were performed at room temperature (between 20°C and 25°C).

In voltage clamp experiments, we used a two-electrode system (SEC-05L-TEC; NPI-Electronic). The experiments were performed in an external solution (see Experimental solutions) providing conditions to eliminate all major ionic currents to ensure space clamping. Both intracellular electrodes were sharp micropipettes (ca. 7 MΩ) filled with 3 M KCl and were inserted close to each other near the center of the fiber. Rectangular voltage steps were applied from a holding potential of −80 mV.

To prepare halothane-air mixtures of defined concentrations, we used a vaporizer (Halothan-Vapor 19.3; Dräger) and measured the resulting halothane percentage with a calibrated infrared sensor (IRIS; Dräger). The anesthetic-containing gas mixture was equilibrated with Ringer’s solution in a glass beaker that was part of a circular solution flow system. The halothane-containing Ringer’s solution was passed at 2 ml/min through the experimental chamber on the microscope stage. The inlet to the chamber passed a custom-made electrical heating device that we used in some experiments to raise the temperature of the applied solution up to 35°C. Temperature was recorded with a miniature sensor positioned close to the cells under investigation.

We report concentration in percentage units (recorded with the calibrated IRIS system). In separate experiments, we took samples from the recording chamber and analyzed them using gas chromatography. The results confirmed that concentrations measured in the dish were proportional to readings from the IRIS systems with a conversion factor of 0.51 mM/%.

To estimate properties of Ca²⁺ removal and release during action potential–induced activation, we used an activation protocol consisting of a single pulse followed by several repetitive stimuli and a model to determine Ca²⁺ binding to major sites in the fiber as described (Braubach et al., 2014). Binding to Ca²⁺-specific (T) sites and Ca²⁺–Mg²⁺ (P) sites of troponin C was calculated as published (Robertson et al., 1981; Baylor and Hollingworth, 2003). The fixed rate constant and concentration values used for the calculations were as follows: for T-sites: k_on,T,Ca = 115/µM/s, k_off,T,Ca = 230/s; for P-sites: k_on,P,Ca = 300/µM/s, k_off,P,Ca = 0.6/s, k_on,P,Mg = 0.1/µM/s, and k_off,P,Mg = 2/s. [T]_tot and [P]_tot, the total concentrations of T- and P-sites, were 240 µM each. [Fura]_tot was 100 µM. Fast Ca²⁺ binding to ATP was described by a component proportional to free Ca²⁺ (scaling factor F = 3.6; Baylor and Hollingworth, 2003).

Any additional slow binding and slow Ca²⁺ sequestration (to the SR, mitochondria, etc.) was modeled using the combination of a saturable (S) and a nonsaturable (NS) first-order system whose constants ([S]_tot, k_on,S, k_off,S and k_NS) were adjusted by least-squares-fitting the relaxation time courses as described in previous work (Melzer et al., 1986, 1987; Ursu et al., 2005; Braubach et al., 2014). Any model that provides a good simultaneous fit (for varying amounts of released Ca²⁺) to the relaxation phases when release has stopped can be expected to describe overall removal of Ca²⁺ well and can be used to calculate Ca²⁺ release during the depolarizing stimuli. The Ca²⁺ removal model analysis was performed using software written in Delphi (Borland) and Visual Basics for Applications implemented in Excel (Microsoft; Ursu et al., 2005). Further analysis and statistical calculations were performed using the programming language R (R Core Team, 2014). Data in the text are presented as mean values ± SEM. Means of experimental group data were compared by two-sided unpaired t test, not assuming equal variances. Differences were considered significant when p-values were <0.05.

In a first series of experiments, we superfused isolated muscle fibers of both WT and YS mice with Ringer’s solution equilibrated with 3% halothane. Fig. 1 shows the averaged response of 14 WT and 20 YS fibers. Even though the WT cells showed a visible increase in free myoplasmic [Ca²⁺] on halothane application, the YS cells responded much more strongly. Mean free [Ca²⁺] values were significantly different at all measurement points.

Our experiments were generally performed at room temperature because the survival rate was very low when working with heated solutions, particularly in combination with the anesthetic drug. However, we explored the additional effect of temperature by heating up the inflowing solution to 35°C 10 min after the onset of halothane application, as demonstrated in Fig. 1 A. In each case, the temperature rise induced a further clear increase in the free myoplasmic [Ca²⁺].

Because washout using the circular perfusion system was relatively slow, we applied fresh, completely halothane-free solution from a separate channel of the perfusion system when the reversibility of halothane effects was tested. As can be seen in Fig. 1 A, reversibility was almost perfect in both WT and YS fibers challenged with halothane.

Action potentials are the physiological electrical signals causing Ca²⁺ release and contraction in muscle. Therefore, we were interested in determining whether action potential–induced Ca²⁺ release was changed in the situations described in the preceding paragraph. All-or-none responsiveness was always tested at the start of the experiment, to ensure that the Ca²⁺ signals resulted from action potentials. To determine Ca²⁺ release, we applied a pulse protocol consisting of repetitive stimuli. A single pulse was followed by a silent period of 500 ms and four 50-Hz tetani, lasting 120 ms each, separated by 150-ms breaks (Fig. 1, B and C). The measurement contains five long relaxation intervals at different levels of released Ca²⁺ and therefore intracellular binding site saturation. It can be used to characterize Ca²⁺ removal. This pulse protocol was applied before and during the application of halothane (indicated by the dashed vertical lines in Fig. 1). We subjected the resulting fluorescence signals to an analysis generally known as removal model fitting (Melzer et al., 1986). In brief, the method we applied uses a Ca²⁺ binding and transport model and kinetic constants, partly available from the literature (see Materials and methods) and partly optimized by least squares fitting, to finally estimate the time course and extent of Ca²⁺ release (Braubach et al., 2014). The fit is performed in the relaxation phases of the repetitive fluorescence transients assuming that Ca²⁺ release has then ceased and the decay is exclusively caused by binding and transport. If this is not the case, part of the residual release activity would be attributed to slower removal and would consequently lower the estimated release, as we demonstrated by simulations (Timmer et al., 1998). Starting the model fitting early during relaxation of the fluorescence ratio signals produced only small deviations between the measured fluorescence signal and the fitted curve in WT fibers, similar to previous results in another murine preparation (Braubach et al., 2014). In YS fibers and in the presence of halothane, the fit clearly deviated from the measurement. We consider this an indication of a component of extra release during relaxation. We therefore used the mean best-fit parameters (Fig. 2 legend) obtained in the 14 WT fibers with no halothane to calculate Ca²⁺ release flux from signals obtained with halothane and in mutant fibers. As an additional precaution, the fit in WT fibers was started later, when the fluorescence ratio signals had relaxed to approximately half their amplitude. The changes in [Ca²⁺] and Ca²⁺ release flux kinetics are summarized in Fig. 2. Shown are the averages (from different cells) of signals caused by the leading single pulse of the multipulse protocol (Fig. 1, B and C).

Fig. 2 (A and B) shows the free [Ca²⁺] transients, measured before (continuous lines) and during the halothane application (dashed lines). The WT transients (Fig. 2 A) exhibit the characteristics reported previously by others for type II fibers (Calderón et al., 2009). In particular, a fast and slow phase of relaxation is noticeable. Interestingly 3% halothane further delayed mainly the slow part, without significantly changing the mean peak of the signals. In YS fibers (Fig. 2 B), the mean amplitude of the transients was smaller and the relaxation phase slower than in WT fibers. Both amplitude and relaxation speed were further decreased by halothane.

Fig. 2 (C and D) shows the estimated Ca²⁺ release flux signals corresponding to the traces in Fig. 2 (A and B). To compare their time course, the signals were normalized to the peaks.

The result indicates a small residual component of release with delayed turnoff as the cause of the slow phase in free [Ca²⁺]. Furthermore, it is the prolongation of this phase that seems to be the main cause of the alterations in relaxation speed of the [Ca²⁺] signals. The relaxation phases of both [Ca²⁺] transients and Ca²⁺ release flux signals were fitted using a dual exponential function (Table 1). Consistently, the time constant τ₂ (column 6) of the slow phase of relaxation was significantly increased (without a significant change of its fractional amplitude A₂; see column 4) when comparing measurements in halothane with those before halothane application. Columns 8 and 10 of Table 1 show means of the absolute values of free [Ca²⁺] and Ca²⁺ flux in the baseline and at the peak, respectively.

The procedures described so far allowed investigation of the isolated fibers under close to physiological conditions (noninvasive Ca²⁺ recording, action potential activation of Ca²⁺ release). To pursue the question of whether and how the membrane potential modulates the effect of the anesthetic drug, we performed further experiments using the two-electrode voltage clamp technique, which allowed us to both stabilize the membrane at defined holding potentials during the drug application and apply differently depolarizing voltage steps to test for the ability of releasing Ca²⁺. Unlike in our previous study of voltage-controlled Ca²⁺ release in YS muscle fibers (Andronache et al., 2009), we did not dialyze the cell but used sharp microelectrodes to minimally disturb the cytoplasm.

We subjected voltage-clamped fibers to a rigorous protocol that included the application of halothane and steady changes in the holding membrane potential, as demonstrated in Fig. 3. After impaling the microelectrodes and recording the basal Ca²⁺ level, the solution flow (2 ml/min) was started. The flow itself usually caused only very minor changes in free [Ca²⁺]. 15 min after starting the perfusion, a steady depolarization to −70 mV was applied for 5 min, followed by a 10-mV hyperpolarization to −90 mV for another 5 min before the membrane voltage was set back to −80 mV. After another time interval of 10 min, the vaporizer system was turned on and adjusted such that the reading on the IRIS detector showed 0.2%. Ten minutes after starting the application of halothane, the same excursions of the membrane holding potential described before were applied. In the YS example (red) of Fig. 3, free [Ca²⁺] rose about fourfold shortly after applying halothane. After a partial decline, it started slowly rising again. When comparing the response to the depolarizing and hyperpolarizing voltage shifts, a dramatic difference to the no-halothane situation can be noticed. The small 10-mV depolarization caused a near doubling of the [Ca²⁺] level induced by halothane alone, and the hyperpolarization reduced the concentration markedly below this level. The black symbols in Fig. 3 show a typical example of applying the same protocol to a WT muscle fiber. Here, almost no change can be seen on superfusion with the halothane-containing solution, and the voltage responses are very small.

To quantify the voltage-induced Ca²⁺ signals, we fitted sloping lines to the three measurements before and after each voltage excursion. The line-subtracted free [Ca²⁺] changes were then analyzed to compare the voltage effects. Two series of experiments were performed, one with 0.2% halothane as shown in Fig. 3, and one with 0.5% halothane. The results of the individual measurements resembled the example shown. In Fig. 4, we pooled the results from the two series of experiments. The figure compares the holding potential–induced changes for halothane versus no halothane and demonstrates a highly significant difference in YS fibers, whereas WT fibers showed a very small difference.

The results presented in Figs. 3 and 4 clearly indicate that small changes in the resting potential can dramatically alter the effectiveness of halothane to stimulate Ca²⁺ release in YS muscle fibers. This result would be expected if there were a significant alteration in the voltage dependence of Ca²⁺ release caused by halothane in the way that the threshold for activation is shifted to more negative potentials. We pursued this possibility by combining the protocol of Fig. 3 with applying short voltage steps, depolarizing the membrane to different potentials before and after the application of halothane. To avoid excessive additional stress to the fibers, we applied only two pulses separated by a time interval of 1 s. The first one depolarized to −50 mV, which, in WT fibers, is close to but generally below the threshold of detection. The second one depolarized to −30 mV, which consistently leads to partial activation in WT fibers.

If the suspected voltage shift occurred, a rise in amplitude of the first response, relative to the second one, should be expected. Fig. 5, displaying mean free [Ca²⁺] signals for WT and YS at 0, 0.2, and 0.5% halothane, shows that this is in fact the case. Because the mutation is known to shift the voltage dependence of activation as reported by us and others (Dietze et al., 2000; Ávila and Dirksen, 2001; Andronache et al., 2009), the smaller pulse was always above the threshold in YS in contrast to WT. When halothane was applied, the −50-mV depolarization caused a response in WT where there was none before. Moreover, in the YS fibers a strong potentiation of the −50-mV response relative to the −30-mV response can be noticed. A similar potentiation occurred in the WT fibers at 0.5% halothane. This indicates that halothane indeed caused a shift of the activation curve to more negative potentials in both cases. In contrast to WT fibers, the signals in the YS fibers declined in amplitude parallel to the rise in baseline [Ca²⁺].

In Fig. 6, the Ca²⁺ signals of Fig. 5 are converted to estimates of Ca²⁺ release flux. For this, we used the mean model parameters obtained with the removal fit analysis from action potential–induced signals in WT fibers. In addition to the relative potentiation of the −50-mV response, these measurements indicate a remarkable slowing of the kinetics of release deactivation, compatible with the results from action potential–triggered signals (Fig. 2). Because the relaxation phases could not be fitted by two exponentials in these cases, we quantified the kinetic change by determining the half time of relaxation (Table 2). The mean half time in halothane-free solution was almost four times larger in YS compared with WT (24.4 vs. 6.5 ms) but reached similar values in 0.5% halothane (53.6 vs. 50.8 ms).

Fig. 7 uses the peak values of the mean flux signals shown in Fig. 6 for an approximate reconstruction of the alterations in voltage-dependent activation. For simplicity, the analysis assumes Boltzmann-type activation curves of equal steepness, with a constant steepness parameter k. We used k = 6.6 mV, which is close to the values determined for WT and YS in our previous study (Andronache et al., 2009). With this assumption, an approximate quantification of the alterations in fractional activation by halothane could be performed (Fig. 7). The procedure uses the fact that the ratio of the values at the two fixed voltages (given a canonical Boltzmann-type voltage dependence) shows a unique sigmoidal dependence on the voltage of half-maximal activation V_0.5 (Fig. 7 A). The filled circles in Fig. 7 A indicate the peak ratios of the measurements at the different conditions. Their abscissa values show the estimates of the V_0.5 values. This analysis predicts a shift in V_0.5 of −26.6 mV in WT (black) and −30.0 mV in YS (red) for 0.2% halothane. For the measurements with 0.5% halothane, the identical analysis leads to shifts in V_0.5 of −34.1 mV for WT and 30.3 mV for YS. Fig. 7 B displays the fractional voltage dependencies derived from this analysis with the normalized flux amplitudes at the two test voltages included as filled circles.

Skeletal muscle fiber depolarization is the trigger for rapid Ca²⁺ release during normal contractile performance. On the other hand, volatile anesthetics are potent chemical triggers of Ca²⁺ release in malignant hyperthermia-susceptible muscle. Here we investigated the cross relation between these two triggers in one of the available murine models of MH using halothane, an anesthetic trigger agent whose application is part of the standardized in vitro contracture test for MH susceptibility. We found that halothane lowers the threshold for voltage activation and that depolarization enhances the effectiveness of halothane.

Early fura-2 Ca²⁺ measurements in normal and MHS muscle cells of pigs and humans (Iaizzo et al., 1988, 1989b) showed altered caffeine and halothane responses but no significant difference in the resting free [Ca²⁺] level, in contrast to measurements using Ca²⁺-selective microelectrodes (Lopez et al., 1986). In the murine model used here, the initial mean resting [Ca²⁺] in the experiments of Fig. 1 was significantly higher (P = 0.001) in the group of YS fibers compared with WT cells (42.2 ± 4.24 vs. 21.8 ± 1.43 µM). YS fibers could also be well distinguished from their WT counterparts by their much stronger response to halothane application. Nevertheless, WT fibers, too, showed an increase (even though much smaller) in resting [Ca²⁺] when halothane was applied, consistent with the notion that the sensitivity to volatile anesthetics is inherent to the Ca²⁺ release system and not introduced by the MH mutation.

A drawback in our study is the fact that, for stability reasons, we had to perform our measurements at a temperature lower than physiological temperature. The single set of experiments in which we increased temperature on top of a preequilibration with 3% halothane (Fig. 1) demonstrates that halothane responses are enhanced at higher temperatures. Therefore, the results reported here will differ quantitatively but probably not qualitatively at higher temperatures.

Our method of converting the Ca²⁺ transients to Ca²⁺ release flux using the Ca²⁺ removal model approach shows that in YS fibers the overall amplitude of the flux is decreased considerably during halothane application and the slow phase of release is prolonged (Table 1). Given the elevated resting Ca²⁺ concentration, the decrease in amplitude likely results from a partial steady-state depletion of the SR leading to activation of store-operated Ca²⁺ entry. As has been reported by Manno et al. (2013), flexor digitorum brevis fibers of the YS mouse exhibit partial depletion already under nontriggering conditions.

Both action potential–triggered Ca²⁺ signals and voltage step-activated Ca²⁺ transients showed a consistently slower relaxation back to baseline. A change in relaxation kinetics, as observed here, can have several possible causes: (a) Ca²⁺ release could persist longer than normally even after repolarization, (b) the sequestration mechanisms (buffers with slow kinetics, Ca²⁺ transport mechanisms) could be less effective, and finally, (c) an increase in the cytoplasmic concentration of a fast equilibrating Ca²⁺ binding protein would decrease the amplitude and, consequently, slow down relaxation kinetics. The experiments could not unambiguously identify the reason for the slower relaxation. However, halothane has been demonstrated to favor Ca²⁺ permeation through isolated and reconstituted RyR1 channels (Nelson, 1992; Jiang et al., 2008). On the other hand, at low concentration, as used here, it has been shown to enhance the SR Ca²⁺ pump rather than inhibiting it (Karon et al., 1999). Also, it is difficult to see how halothane should mobilize a rapidly acting Ca²⁺ buffer. Therefore, we consider it most likely that the slower Ca²⁺ relaxation results from a slower turnoff of RyR1 permeation by repolarization, and the analysis was performed accordingly. It is interesting to note that the change in relaxation kinetics resembles the effect of low concentrations of caffeine in frog muscle fibers, which was attributed to a Ca²⁺-induced component of Ca²⁺ release (Klein et al., 1990). Consistent with this explanation is the reported enhancement by volatile anesthetics of Ca²⁺-induced Ca²⁺ release in skinned fibers (Matsui et al., 1991). Therefore, the prolongation can probably be attributed to the kinetic alteration of RyR1 gating induced by halothane. Although somewhat less pronounced, WT fibers, too, showed this effect, again in agreement with the view that sensitivity to halothane is inherently present and not introduced but strongly enhanced by the MH mutation.

In our voltage clamp experiments, we performed both long-lasting small changes in the holding potential and short-pulse depolarizations of larger amplitude. Together, these measurements showed that the application of halothane, even at relatively low concentrations, caused a very significant alteration in the activation voltage dependence of ECC. In WT muscle and YS muscle under drug-free conditions, the threshold for voltage activation is kept distant from the resting membrane potential, ensuring that fluctuations in the resting membrane potential will not immediately lead to Ca²⁺ stress. Relatively mild steady depolarizations have been shown to cause partial SR depletion even in WT fibers (Robin and Allard, 2013). In contrast, application of small concentrations of halothane shifts the threshold to more negative voltages than the normal resting values, with the consequence that even the normal resting potential will now lead to a substantial [Ca²⁺] elevation that cannot be compensated for by the intrinsic Ca²⁺ sequestration mechanisms.

One can also look at the situation from another point of view by considering the membrane potential as a strong modulator of the effectiveness of the drug in releasing Ca²⁺ from the SR. This is most evident from the results of Fig. 4. Unlike under drug-free conditions, the small alterations in the holding potential got translated to large changes in free myoplasmic [Ca²⁺] when halothane was present at low, clinically relevant concentrations. The result of Fig. 1 A suggests that these effects would be even larger at higher temperatures. Membrane depolarization likely occurs in heavily active muscle, for instance through the outward transfer of intracellular potassium ions (Allen et al., 2008), and heat development is a natural consequence of muscle metabolism. Therefore, it is easy to envisage that membrane depolarization together with warming and reactive oxidative chemical species can form an important part of the destructive feed-forward cycle (Durham et al., 2008) that has been suggested to lead into the fulminant MH crisis. Notably, Fig. 4 shows that hyperpolarization will have the opposite effect. Therefore, any intervention that hyperpolarizes or even only stabilizes the muscle membrane potential can be expected to have a beneficial effect in an MH crisis. A simple SR leak induced by the anesthetic drug in parallel to the ECC pathway of Ca²⁺ release would not show this effect. Despite the apparent partial loss of kinetic control, at least a large part of the leak remains under the command of the transverse tubular voltage sensor.

Similar conclusions can be drawn from previous experiments investigating potassium-conditioned caffeine contractures (Gallant et al., 1995). Caffeine is the second chemical tool (in addition to halothane) that is used in the IVCT diagnosis for MH susceptibility. A pre-depolarization to approximately −40 mV by 30 mM K⁺ lowered the threshold for the caffeine response in muscle preparations of both normal and MHS pigs (Gallant et al., 1995). Therefore, the membrane potential may also be a source of variability in the diagnostic IVCT. This question has also been addressed in porcine and human biopsies by determining the effect of spontaneous depolarization on the contractile response (using halothane, caffeine, or both). The results vary, showing enhancement (Gallant et al., 1986), no effect (Iaizzo and Lehmann-Horn, 1989), and inhibition (Adnet et al., 1992) of the drug response. The reasons for this variability are not clear. They may originate from different conditions that caused the membrane to depolarize in the cited studies and a variable direct effect on Ca²⁺ release.

In addition to their effect on the RyR1 protein, anesthetics might have other channel targets in muscle (Franks, 2008) that could alter the membrane potential and boost the response. However, in MH-susceptible pigs, halothane contractures were not accompanied by fiber depolarization (Iaizzo et al., 1989a). Furthermore, elegant in vivo experiments by Eltit et al. (2013) on MH-susceptible mice, using Ca²⁺-sensitive microelectrodes, measured [Ca²⁺] and membrane potential in parallel and demonstrated that the resting potential remained unchanged during anesthetic-induced [Ca²⁺] elevation.

We discuss our results in terms of a primary modulation of the Ca²⁺ efflux from the SR. Basic principles require that any steady-state alteration of cytoplasmic free [Ca²⁺], as found here and in muscle fibers of another MH-susceptible mouse (Eltit et al., 2013), must have a contribution from Ca²⁺ entry from the extracellular space (Ríos, 2010). Thus, changing the holding potential might affect free cytoplasmic [Ca²⁺] also by directly driving a Ca²⁺ flux through voltage-independent pathways of the cell membrane. However, the alterations in free [Ca²⁺] that we observed when changing the holding potential (Figs. 3 and 4) are opposite to what one would expect if they arose simply from changing the driving force for Ca²⁺ entry through channels in the sarcolemma or the T-tubules. The changes are obviously linked to a voltage gating effect, very likely the DHPR-controlled EC coupling pathway.

That MH-susceptible muscle exhibits a higher sensitivity to membrane depolarization has first been reported by Gallant and Donaldson (1989) and Gallant and Lentz (1992) using potassium depolarization on mutant porcine muscle preparations. This effect has been confirmed by measuring Ca²⁺ signals in cultured and mature muscle cells by us and others (Dietze et al., 2000; Ávila and Dirksen, 2001; Andronache et al., 2009). We suggested a reaction scheme that appreciates the allosteric conformational coupling between the DHPR and the RyR1 to explain the lower voltage threshold in MH-susceptible muscle. The model assumes a lowering of the energy of the open state of the RyR1 as the primary effect of the mutation. Because the RyR1 is mechanically coupled to the DHPR, the altered closed-open equilibrium of the RyR1 will “pull” at the voltage-dependent equilibrium between resting and active states of the DHPR and therefore favor the active state with the consequence of a displacement of the voltage activation curve to more negative potentials. The effect of an anesthetic drug on the EC coupling voltage dependence could be explained in a similar way if the drug binding further stabilizes the open conformation of the RyR1. In fact, this view is in line with current models of anesthetic action on neuronal targets. There it is thought that the drugs bind to specific conformations of the target channels, thus favoring functional states (Franks, 2008). If those conformations exhibit binding sites for physiological ligands, the ligand concentration dependence of channel activation will be altered in a similar way as the voltage dependence of Ca²⁺ release is altered in muscle.

In this scenario, the alteration in voltage dependence is a simple consequence of the coupled gating reactions of DHPR and RyR1. In addition, or alternatively, the coupling process itself might be changed as in the case of the chaotropic anion perchlorate. This substance likewise shifts the voltage dependence of activation (Gomolla et al., 1983; Lüttgau et al., 1983; González and Ríos, 1993; Csernoch et al., 1999) by affecting the ryanodine receptor (Ma et al., 1993; Anderson et al., 1997). However, it also increases the sensitivity of Ca²⁺ release to the voltage sensor input signal, reflected in an alteration of the “transfer function,” i.e., the relationship between voltage sensor charge movement and Ca²⁺ release (González and Ríos, 1993; Ma et al., 1993; Ríos et al., 1993, 2015; Csernoch et al., 1999). Caffeine, too, left-shifts voltage-dependent activation (Lüttgau and Oetliker, 1968; Shirokova and Ríos, 1996a; Csernoch et al., 1999). A direct comparison of both drugs in mammalian muscle revealed differences in their transfer functions (Csernoch et al., 1999), indicating that they alter EC coupling in different ways. The effects of both caffeine and perchlorate have also been compared in WT versus MHS preparations by studying ryanodine receptor open probability and potassium-induced contractures (Anderson et al., 1997). The alterations in voltage dependence resembled our result of Fig. 7 with halothane. The available data suggest that, in the presence of these drugs, small shifts in holding potential would lead to similar responses as shown in our study for halothane (Fig. 3).

Another modulatory contributor to the halothane mechanism that should be considered is the [Ca²⁺] elevation itself. Enhancing effects of myoplasmic Ca²⁺ on the DHPR voltage sensing process have been demonstrated (Csernoch et al., 1991; González and Ríos, 1993; Shirokova and Ríos, 1996b). Furthermore, changes in the EC coupling transfer function and the DHPR-mediated L-type Ca²⁺ current have been attributed to the elevation in [Ca²⁺] (Feldmeyer et al., 1993; González and Ríos, 1993; Balog and Gallant, 1999). Investigations studying charge movements in relation to Ca²⁺ release and manipulating intracellular free [Ca²⁺] would be necessary to further analyze the mechanistic details by which membrane potential and anesthetic drugs interfere with each other in muscle.

In conclusion, we present clear evidence for the cross-influence of a volatile anesthetic drug and the membrane potential in a mature MH-susceptible muscle preparation. The results indicate a strong effect of the drug on the main characteristics of voltage-activated Ca²⁺ release and a sensitive modulation of the effectiveness of a given drug concentration by small variations of the membrane potential. These results can be explained by allosteric effects in the macromolecular complex constituting the calcium release units of the transverse tubular–SR junction. The detailed mechanisms and the consequences for diagnostics and therapeutics of these findings deserve further investigation.

We are grateful to Dr. Susan L. Hamilton (Baylor College of Medicine, Houston, TX) for the gift of breeding pairs of Y524S mice and Dr. José R. López (Mount Sinai Medical Center, Miami Beach, FL) for information on the halothane application device. We thank Erhard Schoch (Institute of General Physiology, Ulm University) for valuable technical advice, Karin Fuchs (Institute of Applied Physiology, Ulm University) for expert technical help, and Michelle Bence for correcting the manuscript.

The authors declare no competing financial interests.

P. Elischer and M. Textor are recipients of a doctoral scholarship of the Medical Faculty of Ulm University. Deutsche Forschungsgemeinschaft research grant ME-713/18-2 to W. Melzer is gratefully acknowledged.

Author contributions: A. Zullo, acquisition of data, critical revision; M. Textor, acquisition of data, critical revision; P. Elischer, study conception, setup design; S. Mall, study conception, setup design; A. Alt, chemical analysis; W. Klingler, study conception, providing material; W. Melzer, study conception and design, analysis and interpretation of data, drafting of manuscript.

Eduardo Ríos served as editor.