Pentobarbital inhibition of human recombinant α_1A P/Q-type voltage-gated calcium channels involves slow, open channel block

Abstract

BACKGROUND AND PURPOSE

Pre-synaptic neurotransmitter release is largely dependent on Ca²⁺ entry through P/Q-type (Ca_V2.1) voltage-gated Ca²⁺ channels (PQCCs) at most mammalian, central, fast synapses. Barbiturates are clinical depressants and inhibit pre-synaptic Ca²⁺ entry. PQCC barbiturate pharmacology is generally unclear, specifically in man. The pharmacology of the barbiturate pentobarbital (PB) in human recombinant α_1A PQCCs has been characterized.

EXPERIMENTAL APPROACH

PB effects on macroscopic Ca²⁺(I_Ca) and Ba²⁺(I_Ba) currents were studied using whole-cell patch clamp recording in HEK-293 cells heterologously expressing (α_1A)_human(β_2aα₂δ-1)_rabbit PQCCs.

KEY RESULTS

PB reversibly depressed peak current (I_peak) and enhanced apparent inactivation (fractional current at 800 ms, r₈₀₀) in a concentration-dependent fashion irrespective of charge carrier (50% inhibitory concentration: I_peak, 656 µM; r₈₀₀, 104 µM). Rate of mono-exponential I_Ba decay was linearly dependent on PB concentration. PB reduced channel availability by deepening non-steady-state inactivation curves without altering voltage dependence, slowed recovery from activity-induced unavailable states and produced use-dependent block. PB (100 µM) induced use-dependent block during physiological, high frequency pulse trains and overall depressed PQCC activity by two-fold.

CONCLUSION AND IMPLICATIONS

The results support a PB pharmacological mechanism involving a modulated receptor with preferential slow, bimolecular, open channel block (K_d = 15 µM). Clinical PB concentrations (<200 µM) inhibit PQCC during high frequency activation that reduces computed neurotransmitter release by 16-fold and is comparable to the magnitude of Ca²⁺-dependent facilitation, G-protein modulation and intrinsic inactivation that play critical roles in PQCC modulation underlying synaptic plasticity. The results are consistent with the hypothesis that PB inhibition of PQCCs contributes to central nervous system depression underlying anticonvulsant therapy and general anaesthesia.

Introduction

Voltage-activated Ca²⁺ currents (VACC) have diverse physiological and pharmacological properties and have been classified into five broad categories (L-, P/Q-, N-, R- and T-type) based on features of activation, inactivation and susceptibility to peptide blockers. The associated channel proteins have been identified and labelled using the Ca_Vx nomenclature (Catterall et al., 2003). These channels are expressed as a membrane complex consisting of a primary α₁ pore-forming subunit accompanied by β, α₂δ and γ accessory subunits. Expression of the α₁ subunit alone is sufficient to form a functional channel, but the presence of β and α₂δ subunits are required for normal channel expression and gating properties. Ca²⁺ entry through P/Q-type (Ca_V2.1) voltage-gated Ca²⁺ channels (PQCC) is a dominant trigger for neurotransmission at most fast conventional synapses in the mammalian central nervous system (CNS), and are integral components of the pre-synaptic neurotransmitter release machinery. In addition, pre-synaptic neurotransmitter release is proportional to up to the 4th power of Ca²⁺ entry. Therefore, PQCCs represent a high gain system for the regulation of synaptic transmission in the CNS in which small fractional changes in Ca²⁺ entry through these channels may profoundly affect neuronal signalling (for reviews, see Catterall, 2000; Catterall and Few, 2008).

Barbiturates depress the CNS providing for clinical utility as general anaesthetics and anticonvulsants. CNS depression arises from their well-known enhancement of inhibitory post-synaptic Cl^- currents carried by GABA_A receptors (Evans, 1979; Mathers and Barker, 1980; Olsen, 1988; Macdonald et al. 1989), and also by inhibition of excitatory pre-synaptic neurotransmitter release (Brooks and Eckert, 1947; Weakly, 1969; Richards, 1972; Baudoux et al., 2003) involving reduced Ca²⁺ entry into the pre-synaptic nerve terminal (Morgan and Bryant, 1977; Leslie et al. 1980; Baudoux et al., 2003). Therefore, inhibition of pre-synaptic PQCCs may contribute to CNS depression produced by barbiturates. The barbiturate pentobarbital (PB) inhibits excitatory neurotransmitter release in brain slices by a mechanism involving alterations in PQCC function (Kitayama et al., 2002). Previous investigations have generally reported that PB enhances apparent inactivation of N-, L-type or mixed populations of neuronal VACCs (Nishi and Oyama, 1983; Ikemoto et al. 1986; Gross and Macdonald, 1988; Gundersen et al., 1988;ffrench-Mullen et al., 1993) leading to the proposal that PB and other barbiturates modulate voltage-dependent inactivation (VDI) (Nishi and Oyama, 1983; Ikemoto et al. 1986; Gross and Macdonald, 1988; Gundersen et al., 1988). PB also inhibits rat, reticular thalamocortical, T-type currents and those of recombinant Ca_V3.3 (Joksovic et al., 2005), Ca_V3.1 and Ca_V3.2 channels (Todorovic et al., 2000). There is only a single investigation, to our knowledge, that selectively examined PB modulation of PQCCs, as assayed by changes in infrequent peak I_Bas, which reported little depression at clinically relevant concentrations (Hall et al., 1994). Overall, the pharmacological mechanism governing PB modulation of PQCCs remains unknown, as well as its possible role in CNS depression. Finally, it is important to examine PB pharmacology in human PQCCs since PB anaesthetic potency appears to be species dependent (Spector, 1956; Richards, 1972; Franks and Lieb, 1993).

To address these questions, we investigated the inhibitory effects of PB on human recombinant α_1A PQCCs heterologously expressed in HEK-293 cells, which manifest inactivation and facilitation that closely resembles the behaviour of pre-synaptic PQCCs recorded from the Calyx of Held synapse in the rat brainstem (Lee et al., 2000). This approach provides a homogeneous channel population, avoids the uncontrollable complexities of the synaptic milieu and provides for the study of a human α_1A subunit in a heterologous channel. We studied the (α_1A)_human(β_2aα₂δ-1)_rabbit channel complex which probably reflects the function of the human channel since the primary α_1A isoform is derived from man (Soong et al., 2002) and the α subunit includes the conduction pore, voltage sensors and most of the binding sites for second messengers, drugs and toxins (Catterall, 2000). We included the β_2a subtype to facilitate investigation of PB effects on inactivation since it produces slow inactivation allowing resolution of changes in current time course, simplifies the inactivation mechanism by precluding preferential closed-state inactivation (Patil et al., 1998; Lee et al. 2000) and also results in a channel complex prevalent in many regions of the brain (Stea et al., 1994; Tanaka et al. 1995). PQCCs manifest calcium-dependent inactivation (CDI) that is triggered by elevated global cytosolic Ca² and which can only be observed under experimental conditions that employ near-physiological cytosolic Ca²⁺ concentrations (Lee et al., 2000; Liang et al. 2003). Therefore, to investigate the effects of PB on CDI, we used mild Ca²⁺ buffering in whole-cell, patch-clamp, pipette solutions, which produces near-physiological cytosolic Ca²⁺ concentrations (Lee et al., 2000; Liang et al. 2003).

Methods

The nomenclature used conforms to the British Journal of Pharmacology's ‘Guide to Receptors and Channels’ (Alexander et al., 2008).

Cell culture and transient transfection

Transformed HEK-293 cells, purchased from American Type Culture Collection (Bethesda, MD, USA), were plated on 12 × 12 mm glass coverslips in 60 × 15 mm Falcon dishes (Becton Dickinson, Lincoln Park, NJ, USA) and cultured in minimum essential medium (MEM) supplemented with 10% fetal bovine serum and penicillin (100 IU·mL⁻¹), streptomycin (0.1 mg·mL⁻¹) and glutamine (2 mM) all from Invitrogen (Carlsbad, CA, USA). After incubation in 37°C, 5% CO2 for 48 h, cells were transiently transfected using a lipofection technique described previously (Gingrich et al., 2009) with cDNAs encoding wild-type human α_1A pore-forming subunit (Soong et al., 2002), rabbit β_2a (Perez-Reyes et al., 1992) and rabbit skeletal muscle α₂δ-1 (Tomlinson et al., 1993) accessory subunits and the respiratory syncytial virus T antigen. The β_2a subunit also encoded a green fluorescent protein (GFP) sequence following an internal ribosomal entry site, so as to identify successfully transfected HEK-293 cells by fluorescence (Chaudhuri et al., 2007). Briefly, aliquots of lipofection reagent (Lipofectamine, Life Technologies, Carlsbad, CA, USA) and appropriate plasmids (1:1:1:0.5 by weight, α_1A/β_2a/α₂δ-1/T antigen) were mixed in a modified, serum-free medium (Optimem, Invitrogen), and incubated at room temperature for 10 min. Cells were washed with PBS (Invitrogen), and supplemented MEM was replaced with Optimem, followed by addition of liposome-plasmid containing solution. After a 3–4 h incubation period (37°C, 5% CO2), cells were washed with PBS and returned to supplemented MEM for further incubation. Cells were ready for electrophysiological recording 48 h after transfection.

Electrophysiology

Whole-cell barium (I_Ba) and calcium (I_Ca) currents were recorded at room temperature 48–96 h post-transfection from individual cells presumably co-expressing β_2a and GFP as identified by fluorescence microscopy. Both Ba²⁺ and Ca²⁺ were used as charge carriers in order to differentiate between effects on VDI and CDI. Cells were voltage clamped using the whole-cell configuration of the patch-clamp technique at room temperature. Data were collected using an Axopatch 200A amplifier (Axon Instruments, Union City, CA, USA) connected to an IBM-compatible computer running the Matlab (Mathworks, Natick, MA, USA) environment and software of our own design.

Extracellular solutions contained 130 mM N-methyl-D-glucamine, 130 mM aspartic acid, 1 mM MgCl₂-H₂O, 10 mM glucose, 10 mM 4-aminopyridine, 10 mM HEPES, 10 mM BaCl₂ or CaCl₂, pH 7.3–7.4 with aspartic acid. Internal solutions contained 135 mM caesium methansulphonate, 5 mM CsCl, 1 mM MgCl₂, 4 mM Mg-adenosine triphosphate, 5 mM HEPES and 0.5 mM EGTA, pH 7.3–7.4 with CsOH. Buffering with 0.5 mM EGTA provided near-physiological internal calcium buffering (free [Ca²⁺] ≅ 100 nM). Currents were sampled at 10 kHz and filtered at 2 kHz. Currents were measured (−90 mV holding potential, HP) in control and in the presence of PB (0–500 µM). Leak and capacitive transients were subtracted using a P/8 protocol. Cells were continuously superfused with solutions delivered by a dual-barreled, local solution changer, which involved a closed system constructed entirely of stainless steel, glass, polytetrafluoroethylene tubing and glass syringes with a syringe pump. This system provides complete cell bathing solution changes within 2 s.

By using the internal and external solutions described, pipette resistances were 2–5 MΩ, which were series compensated by greater than 60%. The barbiturate PB (Sigma-Aldrich, St. Louis, MO, USA) was diluted from a 5 mM stock solution on the day of experimentation and was loaded into the local solution changer. Current–voltage (I–V) relationships in control solution were obtained at the beginning and end of each experimental period. Results obtained from cells with outward currents or atypical I–V relationships at either time point were discarded. Currents were obtained in control, drug, wash sequence unless otherwise stated and data were included only when the measures of current amplitude and inactivation in the wash response were within 10% of initial control. These stringent inclusion criteria were necessary to ensure a stable preparation required for detailed time course analysis. As a result, data were collected from less than 5% of cells that were successfully voltage clamped, which typically showed recording stability for <30 min. Cells appeared to poorly tolerate the near-physiological cytosolic Ca²⁺ concentrations obtained from mild buffering conditions, which probably contributed to the difficulty in obtaining long periods of preparation stability. Experiments involving voltage protocols promoting inactivation were repeated no more frequently than every 60 s to allow for recovery of activity-induced unavailable channels between stimuli in both control and drug (see Figure 6). Therefore, protocols involving a wide range of voltages were collected in two subsets involving depolarizing voltages that were either odd and/or even multiples of 10 mV.

Figure 6.

I_Ba recovery from voltage-dependent inactivation and pentobarbital (PB) block. (A) I_Ba recovery families triggered by indicated voltage protocol (B, right) used to assess recovery (at 1 s and 5 s, then at 5 s intervals thereafter) from an inactivating conditioning pulse (+10 mv) in control (CTL, left) and 100 µM pentobarbital (PB, right) from individual cells expressing P/Q-type voltage-gated Ca²⁺ channels. (B) Peak test pulse I_Ba normalized to that of the pre-pulse current plotted versus time after conditioning pulse end (recovery time) with bi-exponential curve fits (curved solid lines) in CTL and 100 µM PB (means ± SEM, n = 5). Left, inset, histogram showing the fast and slow time constants (τ_fast and τ_slow) and amplitudes (A_fast and A_slow) for bi-exponential function fits (estimated parameter value ± SE) of recovery responses (see Methods). Asterisks indicate statistically significant differences between control and PB (P <0.05).

Data analysis

Curve fitting and statistical analysis were performed with ORIGIN software (OriginLab, Northampton, MA, USA). The data are shown as mean ± SEM, unless otherwise stated. Paired and unpaired t-tests were applied as appropriate and significance was taken at P <0.05. The maximum channel conductance (G_peak) was calculated using [G = I/(V−V_rev)] where V is the voltage of the activating depolarization, I is the current maximum during the activating depolarization and V_rev is the current reversal potential as estimated by extrapolating the quasi-linear portion of the current–voltage (I–V) relation to zero current at positive voltages. Voltage dependence of activation was quantified by first relating peak G normalized by maximum G to the activating voltage, which was followed by a quantitative description of this relation using a two-state Boltzmann equation of the form [Normalized G = (1 − P)/{1 + exp([V−V_1/2]/dx)} + P] where V is voltage, V_1/2 is the midpoint voltage, P is the plateau value, and dx is the slope factor. Concentration–inhibition relations were fit with a logistic equation of the form [Y = 1/{1 + IC₅₀/[PB]^slope}] where Y is the response variable, [PB] is the PB concentration, IC₅₀ is the concentration at 50% inhibition from control and slope is the factor related to the Hill coefficient. Response time courses were fit with a bi-exponential function [A₁• exp(–t/τ₁) +A₂• exp(–t/τ₂) + B] where A_n is the nth component amplitude, B is a constant representing a plateau, t is time and τ_n is the nth component time constant.

Results

PB depressed peak current and enhanced apparent inactivation

We were interested in examining overall PB effects on human α_1A PQCCs as well as specifically investigating the role of inactivation in the pharmacological mechanism. PQCCs manifest both VDI and CDI (Catterall and Few, 2008). We began by investigating the PB effects on the current peak and decay triggered by depolarizations long enough (800 ms) to induce both VDI and CDI. Figure 1A (left) shows current responses from an individual cell expressing human PQCCs with Ca²⁺ as the charge carrier (I_Ca). Control I_Ca activated rapidly, reaching a peak within 50 ms, followed by monotonous decay reflecting both CDI and VDI, which together reduced current by nearly threefold, consistent with previous results (Lee et al., 2000). Fractional current remaining at the end of the 800 ms depolarization (r₈₀₀) was taken as a measure of inactivation. 50 µM PB depressed peak I_Ca and accelerated current decay reflecting enhanced apparent inactivation as quantified by r₈₀₀ depression. A higher PB concentration (500 µM) potentiated both effects (Figure 1A, lower left) indicating concentration dependence. The onset and offset of PB effects occurred within ∼50 s given a voltage stimulus period of 60 s, initiation of cell bathing solution exchanges within ∼5 s of the preceding voltage stimulus and solution exchange kinetics of ∼2 s (Figure 1B, inset). The findings in individual cells were confirmed in concentration–response relationships in grouped data over a range of concentrations (Figure 1B). Inactivation enhancement (r₈₀₀) was marked over the clinically relevant concentrations (<200 µM). Depression of peak I_Ca was less sensitive to PB relative to inactivation as reflected by a 10-fold greater IC₅₀ and ∼50% shallower slope of the fitted logistic equation. Notably, peak current depression (50 µM) was nearly twofold greater than previously reported in rodent cerebellum (Hall et al., 1994). CDI is triggered by Ca²⁺ binding to calmodulin that is pre-associated with the PQCC complex (Lee et al., 1999; DeMaria et al. 2001; Erickson et al., 2001). Therefore, to investigate these effects in the absence of CDI, Ca²⁺ was replaced by Ba²⁺ as the charge carrier since Ba²⁺ poorly activates calmodulin (Chao et al., 1984). In Figure 1A (right), control I_Bas manifests minimal decay indicating near elimination of CDI and demonstrate that PB produced similar effects on I_Ba as for I_Ca, at both concentrations in individual cells. Concentration–response relationships in grouped data for I_Ba were similar to those of I_Ca. These findings indicate that Ca²⁺ binding to calmodulin and subsequent CDI play little role in PB action, and point to selective PB enhancement of VDI. PB effects on peak current and inactivation were reversible and data were included only when washout peak current and r₈₀₀ returned to near-baseline control values prior to PB exposure (see Methods). Figure 1A (lower trace) shows representative control/wash pairs for both charge carriers where washout responses nearly superimpose with those of baseline controls. PB is structurally related to the non-anaesthetic barbituric acid, which was employed as a negative control. Barbituric acid (500 µM) failed to significantly alter peak I_Ba and r₈₀₀ (normalized to control: peak, 1.02 ± 0.034; r₈₀₀, 1.01 ± 0.04, compared to unity P < 0.37 and 0.39, respectively, n = 3, data not shown).

Figure 1.

Concentration-dependence of the peak current depression and enhancement of apparent inactivation induced by pentobarbital. (A) Currents triggered by test pulses (protocol above) with Ca²⁺(I_Ca) and Ba²⁺(I_Ba) as charge carriers in control (CTL) and pentobarbital (PB) at indicated concentrations from individual cells expressing (α_1A)_human(β_2aα₂δ-1)_rabbit channels. Depolarization produced downward current deflection marking channel activation, which was followed by current decay reflecting inactivation. Solid horizontal line marks baseline and tail currents were clipped for display purposes. Both 50 µM (above) and 500 µM (below) PB depressed peak current (I_peak), accelerated current decay and apparent inactivation (shown by fractional current at 800 ms, r₈₀₀ indicated by horizontal arrow) for both charge carriers. The responses to PB were replotted after normalization to control I_peak (norm PB, dotted line) to illustrate potentiated current decay. PB 500 µM panels include a response following ∼110 s washout period (Washout, dashed lines) that are representative and which nearly superimpose on baseline control responses, indicating preparation stability and reversibility of PB effects. (B) I_peak and r₈₀₀ (fraction of control) concentration–response relationships for both charge carriers (means ± SEM, n = 3–7). Smooth curves are logistic function fits (estimated value ± SE) of grouped I_peak (IC₅₀ = 656 ± 56 µM, slope = 0.86 ± 0.06) and r₈₀₀ (IC₅₀ = 104 µM ±, slope = 1.37 ± 0.08) data (see Methods), see text. Inset shows representative time course of I_peak and r₈₀₀* (current magnitude at the end of the 800 ms pulse) from an individual cell during multiple exposures to 500 µM PB (open horizontal bars).

PB-induced inhibition of peak current was independent of activating voltage

PB depression of peak current may be explained by block of resting channels, as well as by alterations in activation or permeation. To address the latter two possibilities, we investigated PB effects on current–voltage (I–V) relationships. Figure 2A (upper traces) shows response families from individual cells triggered by 25 ms depolarizations over a range of voltages with Ca²⁺ (left) and Ba²⁺ (right) as charge carriers. I_Ca activation time course was slower than that for I_Ba consistent with previous results (Chaudhuri et al., 2007). We investigated the effects of 500 µM PB, which is near the IC₅₀ for peak current depression, so that its effects could be easily resolved. However, this concentration produced discernible current decay during 25 ms depolarizations (data not shown) that was not present in control with both charge carriers (Figure 2A, upper trace). Therefore, pulse durations were shortened individually for each charge carrier such that current decay during the pulse was no longer appreciable (Figure 2A, lower trace), thereby minimizing potential accumulation of unavailable states during the collection of an I–V response family. PB appeared to depress peak I_Ca and I_Ba from individual cells in a uniform fashion over the voltage range tested, which was also evident in I–V relationships in grouped data (Figure 2B) as reduced amplitude unattended by changes in form. To further explore possible effects on activation, we constructed normalized conductance–voltage (G–V) relationships (see Methods). In control, the Ba²⁺ relationship was shifted leftward by ∼2 mV relative to that for Ca²⁺ (Figure 2C) similar to previous reports (Lee et al., 1999). G–V relationships in control and 500 µM PB were similar with both charge carriers, lending further support to minimal PB effects on activation. The results suggest that peak current depression arises from the reduction of channels available for activation at resting membrane potentials, which may occur when resting channels undergo direct blockade or alternatively are promoted to VDI states.

Figure 2.

Effects of pentobarbital on current– and conductance–voltage relationships. (A) I_Ca (left) and I_Ba (right) families triggered by a range of depolarization voltages (protocols above) in control (CTL, above) and 500 µM pentobarbital (PB, below) from single cells expressing P/Q-type voltage-gated Ca²⁺ channels. Pulse depolarizations shortened in the presence of PB to minimize inactivation, see text. Straight lines mark baselines and tail currents were clipped for clarity. Stimulus period was 15 s. (B) Control current–voltage (I–V) relationship plots for normalized peak I_Ca (left) and I_Ba (right) (means ± SEM, n = 4–6). Responses to 500 µM PB normalized by CTL rendered I–V relationships that are uniformly attenuated for both charge carriers. (C) Plots of normalized conductances for Ca²⁺ (G_Ca, left) and Ba²⁺ (G_Ba, right) versus voltage representing activation curves which demonstrate similar relationships for CTL and PB. Curved lines are Boltzmann function fits of control responses (G_Ca: V_1/2 = −2.4 mV, slope = 4.7; G_Ba: V_1/2 = −0.6 mV, slope = 4.2), see Methods.

Calcium-dependent facilitation and deactivation were insensitive to PB

PQCCs exhibit Ca²⁺-dependent facilitation (CDF) in which Ca²⁺ entry during a conditioning pre-pulse potentiates subsequent activation, which involves Ca²⁺ binding to calmodulin independent of G-protein regulation (Lee et al., 2000; 2003; DeMaria et al., 2001). To investigate possible PB effects on CDF, we studied I_Cas triggered by a test pulse with and without a preceding conditioning pulse using a two-pulse protocol (conditioning pulse, +20 mV, 25 ms; test pulse, +10 mV, 50 ms). Depolarizations were delivered from a HP of −90 mV and conditioning and test pulses were separated by a 10 ms period at HP. A +20 mV conditioning pulse was chosen to induce CDF because it triggers minimal relief of G-protein inhibition (Chaudhuri et al., 2007). In this paradigm, the conditioning pulse accelerated the activation time course of the test pulse I_Ca, which represents CDF. To quantify changes in the activation time course, we determined the time from 10 to 90% of peak I_Ca(t_10–90). Without the conditioning pulse, the test pulse triggered slowly activating I_Cas reaching a maximum at pulse end, which is consistent with previous reports (Chaudhuri et al., 2007) and responses obtained in experiments contributing to I–V relationships (Figure 2A, left upper trace). Conditioning pulse application significantly accelerated the test pulse I_Ca activation time course reducing t_10–90 from 13.4 ± 7.9 ms to 3.53 ± 0.51 ms (P < 0.05, n = 4, data not shown) consistent with previous reports (Lee et al., 2000; Chaudhuri et al. 2007). In the same cells, 100 µM PB depressed the test pulse I_Ca independent of the conditioning pre-pulse. The conditioning pre-pulse reduced t_10–90 from 12.6 ± 6.5 ms to 3.47 ± 0.48 ms (data not shown), which were not significantly different compared to control (P < 0.65 and P < 0.85, respectively, n = 4). To test whether PB affects voltage-mediated closure of activation gates, we investigated deactivation. Tail current deactivation (voltage paradigm of Figure 1A) time course was well fit by a mono-exponential function (not shown) in which associated time constants were unaffected by 100 µM PB (control, 0.29 ± 0.063 ms; PB, 0.3 ± 0.048 ms; P < 0.67, n = 5). The results suggest that PB produces little effect on CDF and deactivation.

PB enhanced apparent VDI

Our results so far show that PB effects are independent of charge carrier, which indicates that mechanisms activated by Ca²⁺ entry are not involved and suggests that PB causes block of resting channels or promotes VDI at resting membrane potentials. We next focused on VDI by employing Ba²⁺ as the charge carrier, which minimizes CDI. We examined PB effects on current decay over a range of activating voltages. Figure 3 (right) shows I_Bas triggered by selected depolarizations from an individual cell in control. Current decay reflecting VDI is absent at a triggering voltage of −20 mV, whereas 0 mV produced marked decay that was enhanced little by increasing the depolarization (+20 mV), demonstrating a plateau effect. Inactivation curves obtained by plotting r₈₀₀ depression versus voltage confirmed these observations in grouped data (Figure 3, left). The control inactivation curve reached a plateau near the peak of the G–V relationship (+20 mV) consistent with traditional open channel inactivation (Armstrong and Bezanilla, 1977; Bean, 1981). PB (100 µM) produced little depression of r₈₀₀ at −20 mV in individual responses from the same cell (Figure 3, right). However, r₈₀₀ depression at 0 mV and +20 mV was greater and, to a similar degree, relative to control. These observations indicate PB enhancement is voltage dependent reaching a plateau at depolarizing potentials exceeding 0 mV. The associated inactivation curve in grouped data (Figure 3, left) accords with these observations. The complement of the PB inactivation curve normalized by control yields the fraction of inactivated channels that are blocked by PB (Frxn PB Blocked) versus voltage. This function increases with voltage and reaches a plateau near the maximum of the G–V relationship, consistent with PB acting on channel states visited subsequent to activation, which include open, inactivated or nearby states. These observations are consistent with direct PB modulation of VDI or open channel block.

Figure 3.

Pentobarbital modulation of apparent voltage-dependent inactivation. Right, selected I_Bas from a cell expressing P/Q-type voltage-gated Ca²⁺ channels triggered by the voltage protocol indicated (above) in control (CTL, middle) and 100 µM pentobarbital (100 µM, below) normalized to peak current to illustrate differences in current time course. Noise appears greater in PB responses due to the greater magnification caused by current reduction by resting block. The current remaining at the end of the 800 ms depolarization (r₈₀₀) is indicated by the arrow, for the 100 µM PB response triggered by 0 mV. Left, voltage dependence of r₈₀₀ in control (r₈₀₀-CTL) and 100 µM PB (r₈₀₀-PB), fraction of inactivated channels blocked by PB (Frxn PB Blocked, complement of r₈₀₀ of 100 µM PB normalized to that of control), and G_Ba as replotted from Figure 2C in grouped data (means ± SEM, n = 3–8), see text.

PB depressed channel availability by deepening non-steady-state inactivation

To further explore the role of VDI in the PB action, we examined the dependence of channel availability on HP using a three pulse inactivation protocol (Figure 4B, inset). Figure 4A (left) shows selected I_Bas from an individual cell in control. Conditioning pulse voltages greater than −30 mV triggered currents and reduced channel availability as reported by the amplitude of the test pulse I_Ba. Grouped data support this finding (Figure 4B) where only voltages greater than −30 mV reduced channel availability. Increasing depolarization produced progressive decreases in channel availability, reaching a non-zero plateau near +20 mV, therefore representing a non-steady-state inactivation curve. The voltage range which reduced channel availability (−20 to +20 mV) coincided with that for channel activation and accords with a gating mechanism in which open, and not resting, states undergo VDI. This accords with previous reports in PQCCs with the β_2a accessory subunit (Patil et al., 1998) and indicates classical open-state inactivation (Armstrong and Bezanilla, 1977; Bean, 1981) without appreciable preferential closed-state inactivation.

Figure 4.

Voltage dependence of channel availability. (A) Selected I_Ba responses triggered by conditioning pulse voltages indicated in control (CTL, left) and pentobarbital (PB, right) from a cell expressing PQCCs, using the voltage protocol described in (B, inset). Test pulse current (I_TEST) displayed on expanded time scale for clarity as indicated by curved arrows. (B) the amplitude of I_TEST normalized to that of the pre-pulse current and plotted versus conditioning pulse voltage for control and PB in grouped data (means ± SEM, n = 3–9). These relationships show the voltage dependence of channel availability and represent non-steady-state inactivation curves. Solid smooth lines are Boltzmann fits of control and PB relationships (CTL: V_1/2 = −7.1 ± 2.1 mV, dx = 8.0 ± 2, P = 0.49 ± 0.03; PB: V_1/2 = −10.1 ± 1.9 mV, dx = 8.3 ± 1.4, P = 0.14 ± 0.03; estimated parameter value ± SE). Boltzmann fit of CTL normalized to that of PB (dotted line) illustrates the similar shape of these relationships. Normalized conductance in Ba²⁺ (G_Ba) and associated Boltzman function fit (dashed curve) were replotted from Figure 2C. PB effects on voltage-dependent inactivation occurred over a voltage range where channels activate, suggesting that channels states near opening are targeted by PB.

PB 100 µM potentiated the fraction of unavailable channels over the same voltage range as in control for individual cell responses and in grouped data (Figure 4A,B). Both control and PB inactivation curves were well fit (see Methods) by two state Boltzmann relationships with similar midpoint and slopes (control, V_1/2 = −7.1 ± 2.1 mV; dx = 8.0 ± 2; P = 0.49 ± 0.03; PB, V_1/2 = −10.1 ± 1.7 mV, dx = 8.28 ± 1.4, P = 0.14 ± 0.04; estimated parameter value ± SE). Therefore, PB deepened the inactivation curve without a voltage shift or change in slope, which is visually demonstrated by the similarity of the PB inactivation curve to that of the scaled control response (dotted line). The results point to two conclusions. First, transitions from resting closed to inactivated states are unlikely to underlie peak current depression. Second, PB acts on states associated with activation, open or inactivated, by a mechanism that involves either PB block of open channels or promotion of VDI states.

PB accelerated current decay consistent with slow, open channel block

To distinguish between mechanisms of VDI enhancement and open channel block, we analysed the effects of PB on the time course of I_Ba over a range of PB concentrations. The kinetic nature of I_Bas was revealed using semilogarithmic plotting, which transforms exponential components into linear phases. Figure 5 (left) shows that current responses in PB are composed primarily of a single exponential component with a time constant that is progressively reduced with increasing PB concentration. Plotting the reciprocal time constants, equal to associated rates, versus PB concentration (Figure 5, right) yielded a linear relationship. These findings can be parsimoniously accounted for by a mechanism involving a bimolecular binding with a single binding site that, when occupied, blocks the open state. In this analysis, the slope describes the blocking rate constant (k_b) and the ordinate intercept marks the unblocking rate (k_ub). The fitted parameters (k_b = 12.2[s × PB (mM)]⁻¹, k_ub = 0.2 s⁻¹) indicate that this is a slow binding reaction (computed K_d = 15 µM). The results indicate a simple open channel blocking mechanism involving bimolecular binding. However, a more complex mechanism involving allosteric modulation of VDI cannot be excluded.

Figure 5.

Concentration dependence of pentobarbital-induced acceleration of I_Ba decay. (A) Semi-logarithmic plots of I_Bas triggered by 800 ms test pulses (protocol, right panel above inset) at indicated PB concentration from individual cells expressing PQCCs. The decay time course was well fit by a mono-exponential function (straight lines) for each response and is labelled with the corresponding time constant (τ). Mono-exponential functions appear linear on these axes. (B) Plot of 1/τ versus PB concentration (means ± SEM, n = 3–6) shows a linear relationship (slope = 12.2[s × PB (mM)]⁻¹, intercept = 0.2 s⁻¹). Lower inset shows simple open channel blocking scheme where open channels (O) become blocked (O*) which is governed by a bimolecular binding reaction in which the blocking rate is the product of the blocking rate constant (k_b) and PB concentration ([PB]) and the unblock rate (k_ub). Upper inset shows voltage protocol, stimulus period of 60 s.

PB increased unavailable channel states and slowed recovery

Recovery from activity-induced, unavailable states was examined using a three pulse protocol (Figure 6B, right). Control responses from an individual cell (Figure 6A, left) showed that a conditioning (2 s) pulse rendered about 60% of channels unavailable, as revealed by the first test pulse I_Ba amplitude relative to that of the pre-pulse presumably due to VDI. Later, test pulses triggered I_Bas that charted a recovery time course that appeared to be biphasic in nature and where recovery was complete in approximately 60 s. PB 100 µM drove a greater fraction of channels into unavailable states (>80%) and appeared to slow the recovery time course. These observations in individual cells were confirmed in grouped data (Figure 6B). The control recovery time course was well fit (see Methods) by a bi-exponential function with fast and slow components where the fast component (τ_f≈ 3 s) was dominant (Figure 6B, inset). These findings are consistent with previous results (Lee et al., 2000). PB increased the fraction of unavailable channels by promoting both fast and slow components but slowed overall recovery primarily by increasing the magnitude (A_f) and time constant (τ_f) of the fast component (Figure 6B, inset).

PB produced use-dependent block

The dual PB effects of increasing activity-induced channel unavailability and slowing recovery from these states may give rise to use-dependent block. Use-dependent block occurs with repetitive stimulation in which the time between stimuli is insufficiently long to allow recovery of drug- and activity-induced unavailable states leading to their accumulation and incremental depression of subsequent currents (Courtney, 1975). We tested for simple use-dependent block by PB using a pulse train protocol (Figure 7, inset). In control, the pulse train (−10 mV) produced a small decrease in the current consistent with minor accumulation of VDI states. PB (100 µM) produced a marked use-dependent block in which more than 70% of channels were unavailable at the end of the pulse train. More depolarized potentials (+20 mV) caused enhancement of inactivation leading to significant accumulation of VDI states and marked reduction of available states in control. PB also produced use-dependent block over this pulse train as well, but fractionally less than with −10 mV depolarizations since channel availability was markedly reduced in control.

Figure 7.

Use-dependent block by pentobarbital (PB) in low-frequency pulse trains. Plots of peak I_Ba normalized to the peak of the first pulse current triggered by repetitive low-frequency (2 Hz) step depolarizations (400 ms) to −10 mV and +20 mV according to the indicated protocol (means ± SEM, n = 3–4). The decrease of the I_Ba over the train of pulses represents an accumulation of inactivated channels in control. The enhanced reduction of this current in the presence of PB represents use-dependent block, see text.

PB induced use-dependent block of high frequency, physiological pulse trains

PB use-dependent block (Figure 7) prompted us to consider whether use-dependent block is also manifest during physiological, high frequency stimuli. Therefore, we examined the effects of a clinically relevant PB concentration (100 µM) on I_Cas elicited by 200 Hz trains of 2 ms depolarizations (Figure 8A, inset), which roughly approximates neuronal action potential waveforms (Penington et al., 1992; Toth and Miller, 1995) and frequencies observed in the mammalian CNS (Ylinen et al., 1995). Tail current amplitudes (I_Tail) were taken as an indication of channel activation since peak I_Ca was difficult to determine when triggered by brief 2 ms pulses (see Figure 8A, exploded inset). In control, I_Tails from an individual cell initially increased by approximately 25% within the first 100 ms of the pulse train (Figure 8A), reflecting CDF. This degree of facilitation is similar to that observed pre-synaptically (Borst et al., 1995) and in other reports with recombinant PQCCs (Lee et al., 2000; Chaudhuri et al. 2007). Subsequently, I_Tail amplitudes progressively declined throughout the remainder of the pulse train representing accumulation of inactivated states. PB depressed the first I_Tail consistent with resting block. PB appeared to attenuate early I_Tail potentiation by CDF and accelerate the decrease of I_Tail following the peak presumably due to use-dependent blockade. These observations were confirmed in grouped data (Figure 8B, left). The peak I_Tail response relationship normalized to that of control (Figure 8B, right) indicates that PB inhibition was more than twofold over the pulse train. Neurotransmitter release varies with pre-synaptic Ca²⁺ influx by a power function of up to 4 (Dodge and Rahamimoff, 1967; Katz and Miledi, 1970; Augustine et al., 1987; Zucker and Regehr, 2002). Therefore, to explore the potential effects of PB PQCC inhibition on synaptic transmission, we computed neurotransmitter release (calculated release) as a 4^th-order function of channel activation (normalized I_Tail) (Figure 8B). The relationship shows that resting block inhibited calculated release with the first pulse by nearly 40%. Use-dependent block presumably suppressed calculated release further by >60% over the next 150 ms, which reached >90% by the end of the pulse train. Relative contributions of resting and use-dependent block to depression of I_Tail and calculated release over the course of the pulse train show that both contribute significantly where use-dependent block is predominant (Figure 8B, inset). The results suggest that PB, at clinical concentrations, may produce significant synaptic depression particularly during high frequency PQCC activation.

Figure 8.

Inhibition of peak I_Ca in physiological high-frequency pulse trains. (A) I_Ca families (left) triggered by pulse trains (inset, below) from an individual cell expressing P/Q-type voltage-gated Ca²⁺ channels in control (CTL) and 100 µM pentobarbital (PB) as indicated. The boundaries or envelopes of the current families are formed by tail currents that are reflective of peak I_Ca and channel activation during the triggering pulse. Upper inset, ellipse encircles portion of pulse trains that are replotted on expanded timescale to illustrate differences in the I_Ca time course in CTL and PB. Peak tail current (I_Tail) was taken as an indication of peak channel activation. (B) Left, I_Tail amplitude normalized to that of the first I_Tail in control and plotted versus pulse train duration (means ± SEM, n = 3) in CTL and PB. Every 10th (every 50 ms) response was evaluated in the first second and every 40th response (every 200 ms) for the remainder of the pulse train (3 s). The control response increased initially due to CDF, reached a peak, and then decayed. PB depressed channel activation over the entire pulse train. (B) Right, plot of normalized PB I_Tail and calculated synaptic release (calculated release) as a fraction of control. The calculated release was computed as the 4th power of the I_Tail response. Inset, plot of the contributions of resting and use-dependent block to the pulse train, both as a fraction of control, during the PB-mediated depression of I_Tail and calculated release.

Discussion and conclusions

Mechanism of PB block

PB inhibited I_peak in a concentration-dependent manner (Figure 1) which may be accounted for by block of resting channels, altered activation, rapid open channel block or transitions from resting to inactivated states. PB depressed I_peak without altering the voltage dependence of activation (Figure 2) arguing against altered activation. The kinetics of open channel block are more than hundred-fold slower than those of activation, discounting a significant contribution to peak current depression (Figure 5). PB deepened non-steady-state inactivation curves without changes in midpoint (V_1/2) and slope (dx) of Boltzmann function fits (Figure 4) which argues against transitions from resting or nearby closed states to inactivated states prior to opening, thereby supporting resting channel block. Furthermore, channel availability reported by non-steady-state inactivation curves decreased over a voltage range, which also triggers activation, consistent with open or nearby channel states undergoing block or enhanced inactivation. Taken together, these findings indicate strongly that PB blocks resting channels and that this underlies I_peak inhibition. Insight into the underlying binding reaction can be derived from the I_peak inhibition relationship assuming that it primarily reflects binding. The fitted logistic function to this relationship (Figure 1) yielded an IC₅₀ of ∼600 µM and a slope factor of ∼0.8 pointing to bimolecular binding with a K_d of 600 µM.

PB accelerated current decay reported by r₈₀₀ in a concentration-dependent fashion independent of charge carrier (Figure 1), which discounts a role of CDI. Current decay was enhanced by depolarization that reached a maximum near the plateau of the G–V relationship (Figure 3) and maximal peak open probability. PB decreased channel availability reported by non-steady-state inactivation curves over a voltage range that triggered activation (Figure 4). Overall, the results can be explained by either block of open or nearby channel states or, alternatively, enhanced VDI. To discriminate between these possibilities, we examined the nature of the PB-induced acceleration of mono-exponential current decay and found that associated rates were linearly dependent on PB concentration (Figure 5). PB also increased the fraction of activity-induced unavailable states and slowed their recovery primarily by increasing the magnitude and time constant of the fast component of a bi-exponential recovery time course (Figure 6). These findings may arise from a novel drug-bound blocked state with recovery kinetics slower but similar to that of the fast component in control, thereby rendering them kinetically un-resolvable. Alternatively, it may be explained by drug-induced slowing of recovery from fast inactivated states. Overall, these results lead us to a conclusion of open channel block, but a more complex mechanism involving allosteric modulation of VDI cannot be excluded.

The proposed mechanism of open channel block may involve an intrapore receptor that, when occupied, occludes the permeation pathway and renders the channel non-conducting. The concentration dependence of current decay rate (Figure 5) suggests that the underlying binding is bimolecular (K_d = 15 µM), and kinetically slow. The apparent macroscopic kinetics of the presumed open channel block are slow with time constants in the order of fractional seconds. Simple open channel block that is manifest in macroscopic currents is determined by the underlying microscopic binding reaction and the open state probability. The open state probability of PQCCs is low (Sather et al., 1993; Chaudhuri et al. 2007) and reported to be <0.05 with Ba²⁺ as the charge carrier (+20 mV) (Chaudhuri et al., 2007). Therefore, microscopic binding rates may be more than an order of magnitude greater than those inferred from macroscopic currents. Overall, the findings support a modulated receptor mechanism, where the binding affinity (K_d = 15 µM) of open channel block is 40-fold greater than for resting block.

Comparison with previous studies of PB and neuronal VACCs

Nishi and Oyama (1983) studied VACCs in Helix F-1 neurones and found that barbiturates (thiopentone, PB and phenobarbitone) depressed peak current without shifting the current–voltage relationship, accelerated apparent inactivation and leftward shifted a two-pulse inactivation curve. The authors concluded that barbiturates, including PB, produced inhibition by enhancing VDI. Similar results have been reported by Gundersen et al. (1988) in ω-conotoxin-sensitive, dihydropyridine-resistant VACCs, thereby probably N-type (Ca_V2.2), obtained from the expression of human temporal lobe mRNA in Xenopus; by Gross and Macdonald (1988) for N-type currents in mouse dorsal root ganglia; and by Ikemoto et al. (1986) in Aplysia neurones. A leftward shift of the inactivation curve is a key finding in support of barbiturate enhancement of VDI. In contrast, we found that PB failed to shift the inactivation curve of PQCCs in accord with an open channel blocking mechanism. This discrepancy may be due to preparation differences or reflect distinct barbiturate pharmacology between PQCCs and Ca_V2.2 channels, as reported for other VACCs (Gross and Macdonald, 1988).

Ffrench-Mullen et al. (1993) investigated the effects of barbiturates (phenobarbital, 5-(2-cyclohexylideneethyl)-5-ethyl barbituric acid, PB and its optical entantiomers: R(−)-PB, S(+)-PB) on a mixed population of VACCs from hippocampal CA1 cells. The IC₅₀s for the optical entantiomers, R(−)-PB and S(+)-PB, were 3.5 µM and >1000 µM, respectively, indicating stereospecific inhibition. R(−)-PB blocked peak current in a voltage-dependent fashion sensing 40% of the membrane electric field from the extracellular surface, was pH dependent and accelerated a slow component of bi-exponential inactivation. Voltage-dependent inhibition of peak current contrasts with the results of this and earlier studies (Nishi and Oyama, 1983;Gundersen et al., 1988). We were interested in investigating the effects of optical enantiomers of PB in our preparation but were unable to acquire the molecules.

There is only one investigation, to our knowledge, that has selectively examined PB modulation of PQCCs (Hall et al., 1994). The authors reported little reduction (6.4%) of peak I_Ba by 50 µM PB in ex vivo rat cerebellar Purkinje cells and the investigation did not address inactivation. Acknowledging that PB anaesthetic potency is greater in human than in rat (Spector, 1956; Franks and Lieb, 1994), we explored PB action on PQCCs with a human primary α_1A subunit to begin to investigate the possible underpinnings of this species difference. We found in this study a nearly twofold greater reduction (11%) in peak I_Ba. The discrepancies in peak current blocking potency of PB between these studies may be due to preparation differences, but may also reflect genuine species-dependent pharmacological phenotypes since α_1A peptide homology differs by ∼10% between rat and human (Zhuchenko et al., 1997). Whether this variation in protein structure of the α₁ subunit of PQCCs underlies the differences in PB potency awaits future structure-function analysis.

Pharmacology of other PQCC inhibitors

PQCCs are inhibited by the peptide toxins Agelenopsis aperta spider ω-agatoxin-IVA (Mintz et al., 1992), and marine cone snail Conus magusω-conotoxin-MVIIC (Hillyard et al., 1992), which are irreversible blockers (Olivera et al., 1994; Garcia-Palomero et al. 2000). SNX482, isolated from the venom of an African tarantula, Hysteocrates gigas, produces reversible inhibition of PQCCs (Arroyo et al., 2003). Small organic blocking molecules include antazoline (Milhaud et al., 2002), and the benzimidazolyl-substituted tetraline derivative, mibefradil (Fang and Osterrieder, 1991; Bezprozvanny and Tsien, 1995; Aczel et al., 1998). Mibefradil has been the most extensively investigated and its inhibition of PQCC involves accelerated apparent inactivation with (Fang and Osterrieder, 1991; Bezprozvanny and Tsien, 1995) or without (Aczel et al., 1998) leftward shifting of the inactivation curve leading to proposed mechanisms of either enhanced VDI or open channel blockade, respectively. Aczel et al. (1998) also reported use-dependent block with this drug, slowed recovery from activity-induced unavailable channels and slow drug-induced current decay occurring over fractions of seconds (Figure 3A, Aczel et al., 1998)). These results are remarkably similar to those of PB inhibition of PQCCs in this study. Mibefredil is highly lipophilic with titratable hydrogen ion sites at a benzimidazolyl group (pK_a = 4.8) and the tertiary amine (pK_a = 5.5). PB is also highly lipophilic and, like other barbiturates, the ring nitrogens can shed a hydrogen ion (pK_a = 8.0) through resonance stabilization by a flanking carbonyl group leading to an anionic molecule. Given this acid-base chemistry, both molecules are largely uncharged at physiological pH and remain lipophilic. The similarities between these two organic molecules at inhibiting PQCCs make it interesting to speculate that both may share a pharmacological mechanism involving preferential binding of the neutral drug form to a lipophilic binding pocket of open channels culminating in blockade. This speculation awaits identification of binding sites mediating PQCC inhibition for each of these drugs.

PB effects on pre-synaptic neurotransmitter release

We are aware of two reports addressing PB modulation of pre-synaptic neurotransmitter release. Kitayama et al. (2002) investigated the effects of PB and other anaesthetic agents on K⁺-evoked glutamate release from rat cerebrocortical slices and found that glutamate release was completely blocked by selective inhibitors of PQCCs, and potently inhibited by PB (50% inhibitory concentration: 8.3 µM). The authors concluded that inhibition of glutamate release by clinical concentrations of PB is mainly due to direct PQCC inhibition, although a role of GABA_A receptor activation could not be excluded. Baudoux et al. (2003) studied excitatory post-synaptic potentials (EPSP) and Ca²⁺ transients in axonal varicosities and fine axon collaterals of hippocampal CA1 neurones. PB (150 µM) depressed Ca²⁺ transients recorded from synaptic boutons and reduced the amplitude and frequency of spontaneous EPSPs. The authors proposed that PB inhibition of pre-synaptic Ca²⁺ transients contributes to its depressant effect on excitatory synaptic transmission in the CNS. Overall, the above results are consistent with the notion that PB inhibits pre-synaptic Ca²⁺ influx leading to reduced release of neurotransmitter and synaptic depression.

Possible clinical relevance of PB inhibition of PQCCs

The range of free PB concentrations that produce anticonvulsant activity in man is 50–100 µM (Raines et al. 1979; Heyer and Macdonald, 1982), and for anaesthesia, the range is slightly higher at 50–150 µM (Richards, 1972; Heyer and Macdonald, 1982; Franks and Lieb, 1994). PB resting block depressed peak current by 16% at clinically relevant concentrations (100 µM), which is similar in absolute magnitude to the effects of CDF, G-protein modulation and intrinsic inactivation, all of which play critical roles in PQCC modulation of underlying synaptic plasticity (Catterall, 2000; Tedford and Zamponi, 2006). The magnitude of this effect corresponds to a nearly twofold reduction in computed neurotransmitter release, assuming a fourth-order relationship between peak Ca²⁺ entry and release (Dodge and Rahamimoff, 1967; Katz and Miledi, 1970; Augustine et al., 1987; Zucker and Regehr, 2002). Additional PQCC depression may occur with high frequency repetitive stimulation sufficient to induce use-dependent block. We observed use-dependent block during physiological high frequency (200 Hz) pulse trains which, when combined with resting block, produced overall PQCC inhibition of 21, 25, 32 and 49% at 150, 500, 1000 and 3000 ms time points during the pulse train, respectively (Figure 8). These produced corresponding depressions of computed synaptic release of 61, 68, 79 and 93%, respectively. Therefore, volleys of high frequency action potential trains may cause further synaptic depression through use-dependent block of PQCCs. The relevance of use-dependent block to anaesthetic action is supported by reports of firing frequencies in the mammalian CNS of up to 200 Hz (Ylinen et al., 1995). Epileptogenic foci in the cortex and CA3 hippocampal region fire high-frequency bursts of action potentials with frequencies reaching 500 Hz (Ward, 1969; Calvin, 1972). These higher firing frequencies would presumably induce greater use-dependent block and synaptic depression in the associated local neural circuit, thereby extinguishing the CNS dysrhymthia before it is able to generalize. Use-dependent block plays a critical role in the inhibition of voltage-gated channels by some clinical anticonvulsants (Ragsdale and Avoli, 1998) and cardiac antidysrhythmics (Rosen and Wit, 1983), but is less common in the clinical utility of inhibitors of neuronal VACCs. The findings are consistent with the hypothesis that PQCC inhibition contributes to the CNS depression produced by PB during general anaesthesia and anticonvulsant therapy.

PQCCs are generally thought to play a dominant role in neurotransmitter release at most fast synapses in the mammalian CNS (Catterall and Few, 2008), which include both excitatory and inhibitory connections. However, at some hippocampal synapses, release of the inhibitory neurotransmitter γ-aminobutyric acid is governed predominantly by Ca_V2.2 channels (N-type currents) (Poncer et al., 1997; 2000). Overall, the contribution of PB-induced inhibition of PQCCs to clinical anaesthesia and anticonvulsant therapy remains an open question and awaits elucidation of critical neuronal circuits modulated to produce the clinical effect.

Glossary

Abbreviations

CDI

Ca²⁺-dependent inactivation

PB

pentobarbital

PQCC

P/Q-type voltage-gated Ca²⁺ channel

VACC

voltage-activated Ca²⁺ current

VDI

voltage-dependent inactivation

Statement of conflicts of interest

None.

Supplemental material

Supporting Information: Teaching Materials; Figs 1–8 as PowerPoint slide.