Differential Inhibition of Neuronal Sodium Channel Subtypes by the General Anesthetic Isoflurane

Abstract

Volatile anesthetics depress neurotransmitter release in a brain region- and neurotransmitter-selective manner by unclear mechanisms. Voltage-gated sodium channels (Na_vs), which are coupled to synaptic vesicle exocytosis, are inhibited by volatile anesthetics through reduction of peak current and modulation of gating. Subtype-selective effects of anesthetics on Na_v might contribute to observed neurotransmitter-selective anesthetic effects on release. We analyzed anesthetic effects on Na⁺ currents mediated by the principal neuronal Na_v subtypes Na_v1.1, Na_v1.2, and Na_v1.6 heterologously expressed in ND7/23 neuroblastoma cells using whole-cell patch-clamp electrophysiology. Isoflurane at clinically relevant concentrations induced a hyperpolarizing shift in the voltage dependence of steady-state inactivation and slowed recovery from fast inactivation in all three Na_v subtypes, with the voltage of half-maximal steady-state inactivation significantly more positive for Na_v1.1 (−49.7 ± 3.9 mV) than for Na_v1.2 (−57.5 ± 1.2 mV) or Na_v1.6 (−58.0 ± 3.8 mV). Isoflurane significantly inhibited peak Na⁺ current (I_Na) in a voltage-dependent manner: at a physiologically relevant holding potential of −70 mV, isoflurane inhibited peak I_Na of Na_v1.2 (16.5% ± 5.5%) and Na_v1.6 (18.0% ± 7.8%), but not of Na_v1.1 (1.2% ± 0.8%). Since Na_v subtypes are differentially expressed both between neuronal types and within neurons, greater inhibition of Na_v1.2 and Na_v1.6 compared with Na_v1.1 could contribute to neurotransmitter-selective effects of isoflurane on synaptic transmission.

Introduction

Volatile anesthetics have been in clinical use for over 170 years, but their mechanisms of action are poorly understood (Hemmings et al., 2005). They have well described effects on multiple protein targets involved in neurotransmission, including ligand-gated and voltage-gated ion channels (Hemmings et al., 2005; Franks, 2006), but the contributions of these molecular actions to their specific neurophysiological effects are less clear. Volatile anesthetics induce a therapeutic state of unconsciousness, amnesia, and immobility, but they can also produce serious side effects, including cardiovascular and respiratory depression and developmental neurotoxicity (Hemmings et al., 2005; Franks, 2006; Jevtovic-Todorovic, 2016). Understanding the effect-specific cellular and molecular targets for volatile anesthetics is therefore critical for potential mechanism-based development of novel anesthetics or approaches to mitigating toxicities.

Voltage-gated Na⁺ channels (Na_vs) are crucial for mediating cellular membrane excitability, including initiation and propagation of action potentials (APs) (Hodgkin and Huxley, 1952; Catterall et al., 2005; Hu et al., 2009; Clay, 2013). They are also targets for the effects of volatile anesthetics (Rehberg et al., 1996; Hemmings, 2009; Herold and Hemmings, 2012; Herold et al., 2014; Covarrubias et al., 2015). Multiple Na_v subtypes are inhibited by volatile anesthetics (Ouyang et al., 2003, 2009; Shiraishi and Harris, 2004; OuYang and Hemmings, 2007), which can lead to reduced presynaptic AP amplitude in rat calyceal neurons (Wu et al., 2004), isolated neurohypophysial nerve terminals (Ouyang et al., 2003), and inhibition of neurotransmitter release (Westphalen and Hemmings, 2003a, 2006; Wu et al., 2004). Modulation of Na_v function also contributes to general anesthetic potency in vivo. Activation of Na_v in the spinal cord by intrathecal infusion of veratridine reduces the immobilizing effect of isoflurane in rats; conversely, blocking Na_v by intrathecal infusion of the highly selective Na_v blocker tetrodotoxin (TTX) enhances the immobilizing effect of isoflurane (Zhang et al., 2008, 2010).

Volatile anesthetics inhibit neurotransmitter release in a brain region- and neurotransmitter-selective manner (Westphalen and Hemmings, 2006; Westphalen et al., 2010, 2011; Baumgart et al., 2015). For example, isoflurane inhibits release of glutamate more potently than release of GABA from cultured hippocampal neurons (Westphalen and Hemmings, 2006; Baumgart et al., 2015) and differentially inhibits neurotransmitter release between brain regions and the spinal cord (Westphalen et al., 2010, 2011). Na_v subtypes show subcellular, regional, and neurotransmitter-selective expression in the central nervous system (CNS) (Lai and Jan, 2006; Ogiwara et al., 2007; Lorincz and Nusser, 2008b; Johnson et al., 2017). Given their sensitivity to volatile anesthetics, the differential effects of anesthetics on neurotransmitter release might be explained by their selective inhibition of specific presynaptic Na_v subtypes. Of the nine identified Na_v subtypes (Na_v1.1–Na_v1.9), Na_v1.1, Na_v1.2, and Na_v1.6 are highly expressed in the CNS (Black and Waxman, 1996; Wood and Baker, 2001). Volatile anesthetics display differential subtype-selective inhibition of Na_v1.2 (one of the predominant brain subtypes), Na_v1.4 (the predominant skeletal muscle subtype), and Na_v1.5 (the predominant cardiac subtype) (OuYang and Hemmings, 2007). However, the role of subtype-selective effects of volatile anesthetics on neuronal Na_v subtypes in their presynaptic effects is unknown. We therefore examined the effects of isoflurane on the three major neuronal Na_v subtypes expressed in the adult mammalian brain using whole-cell patch-clamp electrophysiology in a homogenous neuronal cell line. Selective inhibition of Na_v subtypes would provide a neurophysiological basis for Na_v inhibition in the region- and neurotransmitter-selective effects of volatile anesthetics on neuronal excitability and synaptic transmission.

Materials and Methods

Materials.

Isoflurane was obtained from Abbott Laboratories (North Chicago, IL). TTX was from Sankyo Kasei (Tokyo, Japan). All other compounds for solution preparation were from Sigma-Aldrich (St. Louis, MO).

cDNA Constructs.

Na⁺ currents from three neuronal Na_v subtypes (Na_v1.1, Na_v1.2, and Na_v1.6) were recorded and analyzed by heterologous expression of the respective subtype in a neuronal background using the ND7/23 neuroblastoma cell line (Herold et al., 2009). Each subtype was rendered TTX resistant by mutation of a single amino acid located in the extracellular toxin binding domain that has been shown to have no effect on gating properties, channel kinetics, or anesthetic sensitivity (Herzog et al., 2003; Leffler et al., 2005; Cestèle et al., 2013; Purtell et al., 2015). We refer to the mutant channels as Na_v1.1, Na_v1.2, and Na_v1.6 below. Wild-type human Na_v1.1 (accession number NM_001165963), kindly provided by A. L. George, Jr. (Northwestern University, Evanston, IL), was mutated to F383S (Cestèle et al., 2013). Wild-type rat Na_v1.2 (accession number NM_012647), kindly provided by W. Catterall (University of Washington, Seattle, WA), was mutated to F385S (Leffler et al., 2005; Purtell et al., 2015). Wild-type mouse Na_v1.6 (accession number NM_001077499), kindly provided by S. G. Waxman (Yale University, New Haven, CT), was mutated to Y371S (Herzog et al., 2003).

Cell Culture and Transfection.

Rodent ND7/23 neuroblastoma cells (Sigma-Aldrich) were plated on 12-mm glass coverslips and incubated at 37°C in a humidified atmosphere of 5% CO₂/95% O₂ in Dulbecco’s modified Eagle’s medium supplemented with 10% (v/v) fetal bovine serum, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (Sigma-Aldrich). This cell line provides a neuronal background for expression of Na_v by providing critical factors such as auxiliary subunits that facilitate functional expression of the channel protein (Herold et al., 2009). Transfected TTX-resistant Na_v subtypes were studied in the presence of 250 nM TTX to block endogenous Na⁺ currents (Herold et al., 2009; Purtell et al., 2015). ND7/23 cells were cotransfected with 2.5 μg Na_v1.1, Na_v1.2, or Na_v1.6 cDNA together with 0.7 μg pEGFP-N1 cDNA (Clontech, Mountain View, CA) using Lipofectamine LTX (Invitrogen, Carlsbad, CA) to allow identification of enhanced green fluorescent protein–transfected cells by fluorescence microscopy. Electrophysiological studies were conducted 48 hours after transfection.

Measurement of Sodium Currents.

Whole-cell patch-clamp recordings of Na⁺ currents (I_Na) were obtained at room temperature (23–24°C) using an Axopatch 200B amplifier (Axon Instruments, Burlingame, CA) digitized with a Digidata 1321A interface and analyzed using pClamp 10.2 software (Axon Instruments). Whole-cell I_Na was sampled at 50 kHz and low-pass filtered at 10 kHz. Pipettes were filled with internal solution as follows: 120 mM CsF, 10 mM NaCl, 10 mM HEPES, 10 mM EGTA, 10 mM TEA-Cl, 1 mM CaCl₂, and 1 mM MgCl₂ adjusted to pH 7.3 (by CsOH) and to 334 mOsm/kg (with sucrose). Pipette resistance when filled was 1.5–3.0 MΩ. The external solution contained the following: 140 mM NaCl, 10 mM HEPES, 3 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 20 mM TEA-Cl, 5 mM d-glucose, and 0.00025 mM TTX adjusted to pH 7.4 (with NaOH) and to 330 mOsm/kg (with sucrose). Liquid-junction potential was not corrected. To minimize space-clamp and series resistance errors, only cells expressing 1–8 nA peak current were analyzed. Capacitance transients were cancelled and voltage error was minimized using 70%–80% series resistance correction. Series resistance was typically 2–5 MΩ; data were discarded if series resistance was >8 MΩ. Recordings began 5 minutes after attaining the whole-cell patch to allow equilibration of pipette solution and cytosol. Leak currents were subtracted using the P/4 method (Bezanilla and Armstrong, 1977) applied after test pulses. Stimulation protocols were applied in control external solution containing 250 nM TTX, and again after a 2-minute perfusion with external solution containing isoflurane.

Simulation of Isoflurane Effects on APs.

A computational model using NEURON software 7.4 (http://www.neuron.yale.edu/neuron/) was used to simulate the effects of isoflurane on Na_v1.1, Na_v1.2, and Na_v1.6 function by adjusting reported Na_v electrophysiological parameters (Herzog et al., 2001) to the parameters we obtained. Simulation of APs was modified from the model (Akemann et al., 2009) to be mediated by Na_v1.1, Na_v1.2, or Na_v1.6, with soma length and width set as 25 μm and R_a as 80 Ω/cm. The electrophysiological properties of original pas (passive) and hh (Hodgkin-Huxley) channels were set to default NEURON values, and resting membrane potential of the soma was set at −70 mV. Because experimental recordings were performed at 23 to 24°C, temperature-dependent effects of isoflurane on simulated Na_v kinetics were modeled as described (Collins and Rojas, 1982) based on the kinetic model of m³h, where Q₁₀ = 2.34 for m and Q₁₀ = 2.9 for h to simulate APs mediated by Na_v1.1, Na_v1.2, or Na_v1.6 at 37°C. The relationship between AP amplitude and probability of transmitter release was modeled using a previously established nerve terminal model (Graham and Redman, 1994). The simulation of excitatory postsynaptic currents (EPSCs) was modified from Graham et al. (2001) and based on the observed effects of isoflurane on AP amplitude and firing frequency mediated by Na_v1.1, Na_v1.2, or Na_v1.6 at 37°C. The control presynaptic stimulus was 100 milliseconds at 100 Hz. The stimulus under isoflurane was based on the effects of isoflurane on AP frequency mediated by each Na_v subtype.

Isoflurane Application.

A saturated stock solution of isoflurane (Abbott Laboratories) in external solution (12 to 13 mM) was diluted to the desired final concentrations in gas-tight glass syringes. For electrophysiological recording, isoflurane solutions were perfused using a pressure-driven microperfusion system (ALA BPS-8; ALA Scientific, Westbury, NY) with the perfusion pipette tip positioned 100–150 μm away from the recorded cell. Isoflurane concentrations sampled at the perfusion pipette tip were determined by gas chromatography using a Shimadzu GC-2010 Plus gas chromatograph (Shimadzu, Tokyo, Japan) after extraction into octane (Herold et al., 2009). Isoflurane (0.3 mM) was used as the predicted minimum alveolar concentration (MAC; equivalent to the EC₅₀ for immobilization) in rats after temperature adjustment to 23°C (Taheri et al., 1991; Ouyang et al., 2009).

Data and Statistical Analysis.

Data were analyzed using GraphPad Prism 6 (GraphPad Software Inc., San Diego, CA) and SPSS 22.0 (SPSS Science Software Inc., Chicago, IL). Conductance (G) values were derived from the I-V relationship using the equation G = I/(V − V_rev), where I is the peak I_Na at a given voltage (V) and V_rev is the measured Na⁺ reversal potential. Voltage of half-maximal activation (V_1/2act) was obtained by fitting data for each cell to a Boltzmann equation of the form G/G_max = 1/[1 + exp (V_1/2act − V/k)], where G/G_max is the normalized fractional conductance and k is the slope factor. Voltage of half-maximal fast inactivation (V_1/2inact) was measured by fitting normalized steady-state I_Na values to a Boltzmann equation of the form I_Na/I_Na max = 1/[1 + exp (V_1/2inact − V/k)]. Half-maximal inhibitory concentration (IC₅₀) values were obtained by least-squares fitting to the Hill equation: Y = 1/(1 + 10([logIC₅₀ − X] × h), where Y is the effect, X is the measured concentration of isoflurane, and h is the Hill slope. Time course data were fitted to the monoexponential function Y = exp (−τ × n) + A_P, where τ is the time constant, A_P is the plateau, and n is stimulus number based on complete recovery time. Data (expressed as means ± S.D.) were assessed for normality using the Shapiro–Wilk test and analyzed using the two-tailed paired t test or ANOVA with post hoc testing as indicated, with P < 0.05 as the threshold for statistical significance.

Results

Neuronal Na_v Subtypes have Distinct Electrophysiological Properties.

Voltage of half-maximal activation (V_1/2act) did not differ between Na_v1.1, Na_v1.2, and Na_v1.6 under control conditions (Table 1). Voltage of half-maximal inactivation (V_1/2inact) for Na_v1.1 was significantly more depolarized than for Na_v1.2 or Na_v1.6 (Table 1). From a hyperpolarized holding potential of −100 mV, recovery from fast inactivation (τ) was significantly slower for Na_v1.2 than for Na_v1.6. From a physiologic holding potential of −70 mV, τ for recovery was significantly slower for Na_v1.2 than for Na_v1.1 or Na_v1.6 (Table 2).

TABLE 1.

Effects of isoflurane on voltage-dependent activation and inactivation of neuronal Na_v subtypes

Data are expressed as means ± S.D. Numbers in parentheses are the percentage change compared with the control.

Subtype	V1/2act			V1/2inact
	CTL	ISO	Δ	CTL	ISO	Δ
	mV
Nav1.1	−14.4 ± 3.4	−15.4 ± 3.3	−1.0 ± 1.1 (7)	−49.7 ± 3.9*,**	−54.6 ± 4.9***	−4.9 ± 2.7 (10)
Nav1.2	−15.8 ± 1.4	−16.5 ± 0.8	−0.6 ± 0.7 (4)	−57.5 ± 1.2*	−61.3 ± 1.4***	−3.8 ± 1.7 (7)
Nav1.6	−15.3 ± 2.0	−16.3 ± 2.1	−0.9 ± 0.5 (7)	−58.0 ± 3.8**	−64.2 ± 5.3***	−6.2 ± 2.3 (11)

CTL, control; ISO, isoflurane at 0.48 ± 0.03 mM (∼1.6 MAC).

^*P < 0.05 (Na_v1.1 vs. Na_v1.2); **P < 0.05 (Na_v1.1 vs. Na_v1.6); ***P < 0.05 (CTL vs. ISO).

TABLE 2.

Effects of isoflurane on neuronal Na_v subtype recovery from fast inactivation

Data are expressed as means ± S.D. Numbers in parentheses are the percentage change compared with the control.

Subtype	τrecovery, Vh −100 mV			τrecovery, Vh −70 mV
	CTL	ISO	Δ	CTL	ISO	Δ
	ms
Nav1.1	1.6 ± 0.2	1.9 ± 0.3**	0.3 ± 0.2 (16)	6.1 ± 0.8***	7.4 ± 1.2**	1.3 ± 0.7 (21)
Nav1.2	1.9 ± 0.5*	2.2 ± 0.6**	0.3 ± 0.2 (17)	8.4 ± 2.4*,***	10.1 ± 2.9**	1.7 ± 0.8 (20)
Nav1.6	1.2 ± 0.2*	1.7 ± 0.2**	0.4 ± 0.3 (38)	4.7 ± 0.7*	6.6 ± 1.0**	1.8 ± 0.4 (39)

CTL, control; ISO, isoflurane at 0.58 ± 0.04 mM (∼1.9 MAC); τ_recovery, half-time for recovery form inactivation; V_h, holding potential.

^*P < 0.05 (Na_v1.2 vs. Na_v1.6); **P < 0.05 (CTL vs. ISO); ***P < 0.05 (Na_v1.1 vs. Na_v1.2).

Isoflurane Does Not Affect Sodium Channel Activation.

Channel activation was evoked by a series of 10-millisecond voltage steps from –70 to +60 mV preceded by a 100-millisecond prepulse to –100 mV (Purtell et al., 2015). Isoflurane (0.48 ± 0.03 mM; ∼1.6 MAC) did not significantly affect current-voltage (I-V) relationships or voltage dependence of half-maximal activation (V_1/2act) of Na_v1.1 (n = 6), Na_v1.2 (n = 6), or Na_v1.6 (n = 5) (all P > 0.05 by paired two-tailed t test) (Fig. 1). Peak I_Na was elicited at 0 mV for all three Na_v subtypes for both control and isoflurane conditions. Values of V_1/2act for the control and 1.6 MAC isoflurane conditions are shown in Table 1.

Fig. 1.

Effects of isoflurane on Na_v activation. Isoflurane did not significantly affect the voltage of half-maximal activation (V_1/2act) of Na_v1.1 (n = 6), Na_v1.2 (n = 6), or Na_v1.6 (n = 5). See Table 1 for individual values. Channels were activated by a series of voltage steps from −70 to +60 mV preceded by a 100-millisecond prepulse to −100 mV. (A–C) Representative traces for the control (CTL) or isoflurane (ISO) (0.48 mM, 1.6 MAC) for Na_v1.1, Na_v1.2, and Na_v1.6. (D–F) Current-voltage relationships (I-V) for Na_v1.1, Na_v1.2, and Na_v1.6. (G–I) Activation curves for Na_v1.1, Na_v1.2, and Na_v1.6, for the control or isoflurane (0.48 mM isoflurane). (J–L) Voltage of half-maximum activation (V_1/2act) for Na_v1.1, Na_v1.2, and Na_v1.6 for the control or isoflurane (0.48 mM). Data are expressed as means ± S.D. n.s., not significant by paired t test (P > 0.05).

Isoflurane Enhances Voltage-Dependent Inactivation.

The effect of isoflurane on steady-state fast inactivation was determined by a double pulse protocol with a 300-millisecond prepulse ranging from −110 to −10 mV in 10-mV steps, followed by depolarization to 0 mV to elicit peak I_Na. Normalized I_Na/I_Na max values reflected the fraction of channels inactivated during the prepulse. Isoflurane (0.48 ± 0.03 mM; ∼1.6 MAC) shifted the voltage dependence of steady-state inactivation toward more hyperpolarized potentials for Na_v1.1 (n = 6), Na_v1.2 (n = 6), and Na_v1.6 (n = 5) (all P < 0.05, by paired t test) (Fig. 2, G–I). The magnitude of the voltage shift was similar for all three Na_v subtypes tested (Table 1).

Fig. 2.

Effects of isoflurane on steady-state fast inactivation of neuronal Na_v subtypes. Isoflurane (0.48 mM; 1.6 MAC) shifted the voltage of half-maximal inactivation (V_1/2inact) in the hyperpolarizing direction for all three isoforms (n = 6 for Na_v1.1 and Na_v1.2; n = 5 for Na_v1.6; P < 0.05 by two-tailed, paired t test). Steady-state inactivation was determined by eliciting currents at 0 mV after a 300-millisecond prepulse to voltages of −110 to −10 mV in 10-mV steps (see inset). (A–C) Representative families of current traces in the absence (CTL) or presence of isoflurane (ISO) for Na_v1.1, Na_v1.2, and Na_v1.6. (D–F) Isoflurane shifted the voltage dependence of inactivation of Na_v1.1, Na_v1.2, and Na_v1.6 in the hyperpolarized direction. Data are expressed as means ± S.D. (G–I) Effects of isoflurane on the voltage of half-maximal inactivation (V_1/2inact). *P < 0.05 by two-tailed, paired t test.

Isoflurane Differentially Inhibited Peak I_Na.

Inhibition by isoflurane was greater for Na_v1.2 and Na_v1.6 than for Na_v1.1 at a physiologically relevant holding potential of −70 mV. We tested the degree of Na_v inhibition from holding potentials (V_h) of −120, −110, or −70 mV or V_1/2inact (voltage of half-maximal inactivation measured for each individual cell prior to the control recording). V_1/2inact was −51.2 ± 1.2, −58.8 ± 1.8, or −59.0 ± 1.3 mV for Na_v1.1, Na_v1.2, or Na_v1.6, respectively (Fig. 3). From a holding potential of −120 mV, isoflurane (0.49 ± 0.03 mM; ∼1.6 MAC) did not inhibit peak I_Na for any Na_v subtype. From a holding potential of −110 mV, isoflurane inhibited peak I_Na of Na_v1.2 by 8.9% ± 2.9% and of Na_v1.6_R by 8.8% ± 3.9%, whereas Na_v1.1 was not significantly inhibited (Fig. 4). From a physiologic holding potential of −70 mV, isoflurane inhibited peak I_Na of Na_v1.2 by 16.5% ± 5.5% and of Na_v1.6 by 18.0% ± 7.8%, with no effect on Na_v1.1. From a holding potential of V_1/2inact, isoflurane significantly inhibited peak I_Na of all three Na_v subtypes to a similar extent (by ∼30%). The concentration-dependent inhibition of peak I_Na by isoflurane was well fitted to a Hill equation, and potency for inhibition was voltage dependent (Fig. 5, A–C). From a holding potential of −70 mV, Na_v1.2 and Na_v1.6 were more sensitive to isoflurane (IC₅₀ = 1.0 ± 0.2 and 1.1 ± 0.1 mM, respectively) compared with Na_v1.1 (IC₅₀ = 2.0 ± 0.1 mM; P < 0.05 by ANOVA, Fig. 5). From a holding potential of V_1/2inact, the IC₅₀ values of isoflurane for all three Na_v subtypes were similar (0.9 ± 0.2 mM for Na_v1.1, 0.8 ± 0.2 mM for Na_v1.2, and 0.9 ± 0.2 mM for Na_v1.6; P > 0.05 by ANOVA).

Fig. 3.

Inhibition of peak Na⁺ current (I_Na) during wash-in and washout of isoflurane. Effect of 0.49 mM (1.6 MAC) isoflurane with an alternating pulse protocol to elicit peak I_Na by a prepulse to −110 mV, −70 mV, or V_1/2inact (voltage of half-maximal inactivation) (n = 6). V_1/2inact was determined for each individual cell prior to control recording. (A) Representative traces for Na_v1.1 for the control (black traces) or isoflurane (ISO; colored traces). No inhibition was observed for Na_v1.1 from a holding potential (V_h) of −110 or −70 mV. (B) Representative traces for Na_v1.2 for the control or isoflurane. (C) Representative traces for Na_v1.6 for the control or isoflurane. (D–F) Inhibition by isoflurane of Na_v subtypes. Data are expressed as means ± S.D.

Fig. 4.

Inhibition of peak Na⁺ current (I_Na) by isoflurane. From a holding potential (V_h) of −110 mV, isoflurane (0.49 ± 0.03 mM; ∼1.6 MAC) inhibited peak I_Na of Na_v1.2 and Na_v1.6, but not of Na_v1.1. From a holding potential of −70 mV, isoflurane inhibited peak I_Na of Na_v1.2 and Na_v1.6, but not of Na_v1.1. From a holding potential of V_1/2inact (voltage of half-maximal inactivation), isoflurane inhibited peak I_Na of Na_v1.1, Na_v1.2, and Na_v1.6 with similar efficacy (P > 0.05 by ANOVA). Data are expressed as means ± S.D. (n = 5 to 6). *P < 0.05 vs. Na_v1.1. n.s., not significant (P > 0.05).

Fig. 5.

Concentration-dependent effects of isoflurane on Na_v subtypes. IC₅₀ values for isoflurane inhibition of Na_v subtypes from holding potentials of −70 mV or V_1/2inact. (A–C) Data for concentration-dependent inhibition of peak I_Na by isoflurane were well fitted to a Hill equation with significant voltage-dependent inhibition (P < 0.05 by paired t test). The shaded area indicates the clinical concentration range of isoflurane (0.15–0.6 mM; 0.5–2.0 MAC). (D–F) Representative Na⁺ current traces for the control or isoflurane (ISO). (G) From a holding potential of −70 mV, Na_v1.2 (n = 20) and Na_v1.6 (n = 17) were more sensitive to isoflurane inhibition compared with Na_v1.1 (n = 19). From a holding potential of V_1/2inact, IC₅₀ values were similar for all three Na_v subtypes. *P < 0.05 by ANOVA with post hoc Bonferroni correction. Data are expressed as means ± S.D. (n = 3–5 for each point). n.s., not significant (P > 0.05).

Isoflurane Differential Affects Recovery from Fast Inactivation.

Neuronal firing frequency depends in part on the speed of Na_v gating through its resting, open, and inactivated states, which is voltage dependent. We measured the effects of isoflurane on the time course of recovery from the fast-inactivated state. Peak I_Na was recorded in response to two 5-millisecond pulses to 0 mV, where the duration between pulses varied from 1 to 200 milliseconds. Recovery time courses were well fitted by a monoexponential function in control and isoflurane conditions, indicating that channels predominantly entered a single fast-inactivated state. Isoflurane (0.58 ± 0.04 mM; ∼1.9 MAC) slowed the recovery time course (r_ecovery) of Na_v1.1, Na_v1.2, and Na_v1.6 (Table 2). Isoflurane significantly increased the time required for full channel recovery at a hyperpolarized holding potential of −100 mV (Fig. 6) in Na_v1.1, Na_v1.2, and Na_v1.6 (all n = 6, paired two-tailed t test). These effects were enhanced at a holding potential of −70 mV in Na_v1.1, Na_v1.2, and Na_v1.6 (all n = 6, paired t test).

Fig. 6.

Effect of isoflurane on recovery from fast inactivation. Isoflurane at 0.58 mM (1.9 MAC) significantly slowed recovery from fast inactivation. (A–C) Representative traces recorded in the absence (CTL, left) or presence of 0.58 mM isoflurane (ISO, right) for Na_v1.1 (A), Na_v1.2 (B), and Na_v1.6 (C). Currents were evoked by a paired-pulse protocol in which the time between the two 5-millisecond pulses (to 0 mV) was varied from 1 to 200 milliseconds. Holding potentials (V_h) were −100 mV (upper trace) or −70 mV (lower trace). (D–F) Normalized peak current (Pulse2/Pulse1) plotted against duration of the interpulse interval for a V_h of −100 mV (left) or −70 mV (right) for Na_v1.1 (D), Na_v1.2 (E), and Na_v1.6 (F). (G–I) Time constant (τ) for recovery from inactivation determined from monoexponential fits of data from individual cells in the absence (CTL) or presence of isoflurane from a holding potential of −100 mV (n = 6) or −70 mV (n = 6). Data are presented as means ± S.D. *P < 0.05 vs. CTL by two-tailed, paired t test.

Effects of Isoflurane on Simulated APs and Synaptic Transmission.

We used the NEURON algorithm to simulate the voltage dependence of Na_v1.1-, Na_v1.2-, and Na_v1.6-mediated Na⁺ currents in a model neuron based on the experimentally derived gating properties of each channel (Fig. 7, A–C). Simulated voltage-dependent activation and inactivation values of Na_v1.1, Na_v1.2, and Na_v1.6 were similar to those obtained experimentally by patch-clamp electrophysiological recordings (Fig. 7, D–F). Values of V_1/2act for Na_v1.1, Na_v1.2, and Na_v1.6 were all −15 mV, and values of V_1/2inact for Na_v1.1, Na_v1.2, and Na_v1.6 were −50, −58, and −58 mV, respectively.

Fig. 7.

Simulation of Na⁺ currents mediated by Na_v1.1, Na_v1.2, or Na_v1.6. (A–C) Representative I_Na for each neuronal Na_v subtype. Simulation of Hodgkin-Huxley model Na⁺ channel gating by NMODL. Currents at 24°C (left) were similar to I_Na recorded experimentally and the currents at 37°C (right) were corrected. The electrophysiological parameters of Na_v at 24°C were based on empirical data obtained from recordings. Temperature correction of Na_v was based on the kinetic model of m³h, where Q₁₀ = 2.34 for m and Q₁₀ = 2.9 for h (Collins and Rojas, 1982). (D–F) Simulated activation curves for Na_v1.1 (D), Na_v1.2 (E), and Na_v1.6 (F) in the absence (CTL; black traces) or presence (ISO; colored traces) of isoflurane. The simulated currents at both 24°C and 37°C are shown. (G–I) Simulated inactivation curves for Na_v1.1 (G), Na_v1.2 (H), and Na_v1.6 (I) in the absence (black traces) or presence (colored traces) of isoflurane. The simulated currents at both 24°C and 37°C are shown. CTL, control; ISO, isoflurane.

We also performed simulations to determine the temperature dependence of isoflurane effects on macroscopic Na_v gating properties. After temperature correction to 37°C, values of V_1/2act for Na_v1.1, Na_v1.2, and Na_v1.6 were all −23 mV, and values of V_1/2inact for Na_v1.1, Na_v1.2, and Na_v1.6 were −60, −68, and −69 mV, respectively (Fig. 7, G–I). At a temperature of 24°C, isoflurane (0.48 mM; ∼1.6 MAC) shifted the inactivation curves of Na_v1.1, Na_v1.2, and Na_v1.6 in the hyperpolarized direction to V_1/2inact values of −55, −61, and −64 mV, respectively, with no effect on V_1/2act (Fig. 7, G–I). At a temperature of 37°C, isoflurane (∼1.6 MAC) shifted the inactivation curves of Na_v1.1, Na_v1.2, and Na_v1.6 in the hyperpolarized direction to V_1/2inact values of −65, −72, and −75 mV, respectively, with no effect on V_1/2act (Fig. 7, G–I).

Inhibition by isoflurane of Na_v1.1, Na_v1.2, and Na_v1.6 from a holding potential of −70 mV was used for simulation of single APs at 37°C. For an AP evoked by a current stimulus of 0.15 nA (Fig. 8A), the simulated AP half-width was greater for Na_v1.2 (1.3 milliseconds) than for Na_v1.6 (0.7 milliseconds) or Na_v1.1 (0.8 milliseconds). Isoflurane inhibited AP initiation mediated by Na_v1.2 or Na_v1.6, whereas APs mediated by Na_v1.1 were minimally affected (Fig. 8B). When AP trains were evoked with a longer and larger-amplitude depolarizing stimulus (250-millisecond current injection at 0.5 nA, Fig. 8C), the simulated AP frequency was higher for Na_v1.6 (92 Hz) than for Na_v1.1 (84 Hz) or Na_v1.2 (88 Hz). Isoflurane reduced AP frequency to 80 Hz (−5.0%), 72 Hz (−18.2%), or 71 Hz (−22.8%) for Na_v1.1, Na_v1.2, and Na_v1.6, respectively (Fig. 8D). Simulation of EPSCs was based on the effects of isoflurane on AP amplitude and firing frequency as a prediction of effects on synaptic transmission. Isoflurane depressed single AP amplitude from 131, 132, and 132 mV to 128, 66, and 61 mV for Na_v1.1, Na_v1.2, and Na_v1.6, respectively (Fig. 8B). The depression by isoflurane of EPSCs was significant for Na_v1.2- and Na_v1.6-mediated transmission but minimal for Na_v1.1-mediated transmission (Fig. 8F).

Fig. 8.

Simulated effects of isoflurane on action potentials and synaptic transmission mediated by Na_v1.1, Na_v1.2, or Na_v1.6. Channel gating kinetics were temperature corrected to 37°C. (A) Effects of isoflurane (ISO) on single AP morphology evoked by 0.15 nA for 10 milliseconds. (B) Effects of isoflurane on half-width and amplitude of single APs mediated by Na_v1.1, Na_v1.2, or Na_v1.6. (C) Effects of isoflurane on AP trains evoked by a larger stimulus of 0.5 nA for 250 milliseconds. (D) Effects of isoflurane on AP frequency mediated by Na_v1.1, Na_v1.2, or Na_v1.6. (E) The relationship between AP amplitude and probability of transmitter release was modified from an established nerve terminal model (Graham and Redman, 1994). (F) Simulated effects of isoflurane on postsynaptic EPSCs mediated by Na_v1.1, Na_v1.2, or Na_v1.6.

Discussion

In a combined electrophysiological and simulation study, we determined the major neuronal Na_v subtype-specific effects of isoflurane on channel gating properties, AP firing, and synaptic transmission. Although voltage-dependent activation of Na_v was similar between the three major neuronal subtypes, their voltage dependence of inactivation differed, with that of Na_v1.1 being more positive compared with Na_v1.2 and Na_v1.6. This is consistent with previous reports that the V_1/2inact of Na_v1.1 is more positive than that of Na_v1.6 when heterologously expressed in human embryonic kidney cells (Patel et al., 2015; Fruscione et al., 2018). These differences in inactivation gating lead to greater isoflurane-induced fast inactivation and inhibition of peak I_Na for Na_v1.2 and Na_v1.6 compared with Na_v1.1 at a physiologic holding potential. The greater anesthetic sensitivity of Na_v1.2 and Na_v1.6 provides a plausible mechanism for brain region- and neurotransmitter-selective effects of isoflurane on synaptic transmission due to differential expression of Na_v subtypes between neuron subtypes and within neuronal compartments (Wood and Baker, 2001; Johnson et al., 2017).

All clinically used volatile anesthetics inhibit the major Na_v subtypes, including brain Na_v1.2, skeletal muscle Na_v1.4, and cardiac Na_v1.5 (OuYang and Hemmings, 2007). Inhibition by volatile anesthetics of Na_v1.5 contributes to their modulation of electrophysiological properties including APs in cardiomyocytes (Eskinder et al., 1993; Raatikainen et al., 1998). Of the nine identified Na_v subtypes (Na_v1.1–Na_v1.9), Na_v1.1, Na_v1.2, and Na_v1.6 are highly expressed in the CNS (Black and Waxman, 1996; Wood and Baker, 2001). Advantages of expressing neuronal Na_v subtypes in the ND7/23 neuroblastoma/dorsal root ganglion hybridoma cell line include robust expression in the presence of critical auxiliary subunits and other neuronal signaling pathways, minimal space-clamp issues compared with mature isolated neurons facilitating reliable voltage-control during electrophysiological recordings (Herold et al., 2009), and expression and function in a uniform neuronal background minimizing effects of neuronal subtype heterogeneity (John et al., 2004; Rogers et al., 2016).

Even small reductions in I_Na have large electrophysiological consequences due to nonlinear coupling (Fujiwara et al., 1988; Goldin, 2001; Engel and Jonas, 2005). Small differences in presynaptic AP shape can produce large changes in the timing and magnitude of presynaptic Ca²⁺ entry because the kinetics of Ca²⁺ channels are strongly voltage dependent (Clark et al., 1996; Clarke et al., 2016). Thus, AP-evoked Ca²⁺ entry can have a steep dependence on AP shape, which can translate into significant changes in Ca²⁺ entry and even more dramatic changes in transmitter release and postsynaptic currents (Bean, 2007). Volatile anesthetics inhibit Na_v and APs at clinically relevant concentrations (Westphalen and Hemmings, 2003b, 2006; Wu et al., 2004; Herold and Hemmings, 2012). Based on our findings, the simulated AP half-width is greater for Na_v1.2 and Na_v1.1 than for Na_v1.6, whereas the AP train frequency is higher for Na_v1.6 than for Na_v1.1 and Na_v1.2, consistent with Na_v1.6 as an important contributor to fast repetitive firing (Khaliq et al., 2003; Brackenbury et al., 2010). Our findings provide a mechanistic basis for previous observations that isoflurane inhibits neurotransmitter release and spontaneous firing patterns of hippocampal neurons in an activity-dependent manner (Fujiwara et al., 1988; Goldin, 2001; Wu et al., 2004; Purtell et al., 2015).

Isoflurane suppressed peak Na⁺ current, produced a hyperpolarizing shift in the voltage dependence of fast inactivation, and slowed recovery from fast inactivation of brain Na_v subtypes at a physiologically relevant membrane potential. These effects might involve more than one site of interaction of isoflurane with Na_v since volatile anesthetics appear to have multiple sites of interaction with voltage-gated ion channels including Na_v (Raju et al., 2013; Spurny et al., 2013; Sand et al., 2017). At a hyperpolarized holding potential, none of the Na_v subtypes were inhibited by isoflurane, consistent with a low affinity of isoflurane for the resting state of Na_v. From hyperpolarized or physiologic holding potentials, isoflurane did not inhibit Na_v1.1-mediated peak I_Na but did inhibit Na_v1.2 and Na_v1.6 at a physiologic membrane potential. Further studies are required to determine whether differences in anesthetic binding sites contribute to this subtype-selective sensitivity to isoflurane.

All neuronal Na_v subtypes were inhibited to a similar degree from holding potentials near their V_1/2inact, a potential at which 50% of channels are closed or inactivated, indicating that isoflurane has similar quantitative effects on Na_v1.1-, Na_v1.2-, and Na_v1.6-mediated currents when currents are measured at equivalent states of inactivation. Since the V_1/2inact is more positive for Na_v1.1 than for Na_v1.2 or Na_v1.6, isoflurane showed greater inhibition of Na_v1.2 and Na_v1.6 compared with Na_v1.1 at a holding potential of −70 mV, supporting preferential interaction with inactivated channels. At this holding potential, Na_v1.2 and Na_v1.6 have a higher proportion of channels in the inactivated state compared with Na_v1.1, leading to preferential inhibition of Na_v1.2 and Na_v1.6 by isoflurane. Thus, the differential sensitivity of neuronal Na_v subtypes to isoflurane can be explained by differences in voltage-dependent gating rather than by subtype-specific differences in drug sensitivity per se. These findings suggest conserved sites of interaction between isoflurane and the major neuronal Na_v subtypes, which are highly homologous (Black and Waxman, 1996; Wood and Baker, 2001). Identification of their pharmacologically relevant binding site(s) should clarify whether volatile anesthetics interact with structurally homologous sites in Na_v (Sand et al., 2017).

Neuronal excitability and synaptic transmission show physiologic and pharmacological differences between brain regions and between neurons of different neurotransmitter phenotype (Pinheiro and Mulle, 2008; Spruston, 2008). Some of this heterogeneity could derive from differences in relative expression of various neuronal Na_v subtypes having distinct voltage-dependent gating properties, with physiologic and pharmacological implications (Lai and Jan, 2006; Ogiwara et al., 2007; Lorincz and Nusser, 2008b; Johnson et al., 2017). Specific Na_v subtypes are selectively expressed on presynaptic and postsynaptic glutamatergic synapses in the hippocampus, consistent with subtype-specific roles in neurotransmitter release and synaptic plasticity (Johnson et al., 2017). Most excitatory neurons express a high density of Na_v1.6 and Na_v1.2 (Whitaker et al., 2000; Tian et al., 2014), whereas Na_v1.1 is preferentially expressed in inhibitory GABAergic interneurons (Ogiwara et al., 2007; Lorincz and Nusser, 2008a).

Neurotransmitter phenotype-specific expression and sensitivity to isoflurane inhibition of neuronal Na_v subtypes could underlie the greater isoflurane sensitivity of Na_v-dependent release of glutamate compared with GABA (Westphalen and Hemmings, 2006; Baumgart et al., 2015; Purtell et al., 2015). Further studies of the relationship between the functionally selective effects of isoflurane on Na_v subtypes and various pharmacological endpoints might provide leads for the development of more selective and potentially safer general anesthetic agents that selectively target specific Na_v subtypes. The function and gating kinetics of neuronal Na_v are also modulated by phosphorylation, calmodulin binding (Cantrell and Catterall, 2001; Pitt and Lee, 2016; Yan et al., 2017), and auxiliary subunits (Wimmer et al., 2015), which could further modulate Na_v sensitivity to general anesthetics.

This study has certain limitations. We used well characterized Na_v subtypes from different mammalian species (human, rat, and mouse), since their amino acid sequences are >98% conserved; species-specific differences in Na_v properties and anesthetic sensitivity are possible but unlikely and have not been reported (Whitaker et al., 2001; Lewis and Raman, 2011; Carrasco et al., 2017). The single point mutations to produce TTX resistance introduce a difference from wild-type channels with possible unanticipated effects on drug binding. However, these mutations are in the extracellular toxin binding domain and have not been shown to affect gating properties, channel kinetics, or anesthetic sensitivity compared with native hippocampal neuron channels (Herzog et al., 2003; Leffler et al., 2005; Purtell et al., 2015; Carrasco et al., 2017).

Although our simulation studies support our conclusions, the functional predictions have not been validated in intact neurons. Recordings of native I_Na from intact neurons are not an ideal model for comparing Na_v subtypes, since it is impossible to isolate the individual subtype and selectively expressing TTX-resistant subtypes in cultured neurons is complicated by the heterogeneity of neuron types. All electrophysiological recordings were performed at room temperature (23 to 24°C), as are most electrophysiological studies because of recording stability and reliability. Although we used a simulated model to correct the temperature to 37°C, the effects of isoflurane at 37°C were not measured directly, and whether isoflurane would more potently inhibit Na_v at 37°C is therefore unclear. Since the voltage dependence of Na_v inactivation was hyperpolarized at the physiologic temperature of 37°C, inhibition by isoflurane of Na⁺ currents would be enhanced compared with 24°C, with perhaps more marked effects on excitability and AP properties. Although all volatile anesthetics tested inhibit Na_v (OuYang and Hemmings, 2007), we only tested isoflurane and cannot extrapolate these effects to other clinically used volatile anesthetics, and it is possible that agent-specific effects exist.

In conclusion, Na⁺ currents mediated by the principal neuronal Na_v subtypes Na_v1.1, Na_v1.2, and Na_v1.6 exhibit different sensitivities to inhibition by isoflurane, leading to reduced inhibition of Na_v1.1 compared with Na_v1.2 and Na_v1.6. This differential sensitivity can be explained by fundamental differences in voltage-dependent gating properties between channel subtypes. Thus, differences in the expression of specific Na_v subtypes between brain regions, neurons, and synapses could underlie brain region- and neurotransmitter-selective effects of isoflurane, and perhaps other volatile anesthetics, on CNS function.

Abbreviations

AP

action potential

CNS

central nervous system

EPSC

excitatory postsynaptic current

I_Na

sodium current

MAC

minimum alveolar concentration

Na_v

voltage-gated sodium channel

TTX

tetrodotoxin

V_1/2act

voltage of half-maximal activation

V_1/2inact

voltage of half-maximal inactivation

Authorship Contributions

Participated in research design: Zhou, Herold, Hemmings.

Conducted experiments: Zhou, Johnson.

Contributed new reagents or analytic tools: Zhou, Herold.

Performed data analysis: Zhou, Herold, Hemmings.

Wrote or contributed to the writing of the manuscript: Zhou, Johnson, Herold, Hemmings.