Inhibitory effects of cannabidiol on voltage-dependent sodium currents

Abstract

Cannabis sativa contains many related compounds known as phytocannabinoids. The main psychoactive and nonpsychoactive compounds are Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD), respectively. Much of the evidence for clinical efficacy of CBD-mediated antiepileptic effects has been from case reports or smaller surveys. The mechanisms for CBD's anticonvulsant effects are unclear and likely involve noncannabinoid receptor pathways. CBD is reported to modulate several ion channels, including sodium channels (Nav). Evaluating the therapeutic mechanisms and safety of CBD demands a richer understanding of its interactions with central nervous system targets. Here, we used voltage-clamp electrophysiology of HEK-293 cells and iPSC neurons to characterize the effects of CBD on Nav channels. Our results show that CBD inhibits hNav1.1-1.7 currents, with an IC₅₀ of 1.9-3.8 μm, suggesting that this inhibition could occur at therapeutically relevant concentrations. A steep Hill slope of ∼3 suggested multiple interactions of CBD with Nav channels. CBD exhibited resting-state blockade, became more potent at depolarized potentials, and also slowed recovery from inactivation, supporting the idea that CBD binding preferentially stabilizes inactivated Nav channel states. We also found that CBD inhibits other voltage-dependent currents from diverse channels, including bacterial homomeric Nav channel (NaChBac) and voltage-gated potassium channel subunit Kv2.1. Lastly, the CBD block of Nav was temperature-dependent, with potency increasing at lower temperatures. We conclude that CBD's mode of action likely involves 1) compound partitioning in lipid membranes, which alters membrane fluidity affecting gating, and 2) undetermined direct interactions with sodium and potassium channels, whose combined effects are loss of channel excitability.

Keywords: Kv2.1; cannabidiol; cannabinoid; central nervous system (CNS); electrophysiology; epilepsy; neuron; phytocannabinoid; sodium channel; voltage-gated sodium channel.

Figure 1.

CBD and THC inhibit Nav currents. A, the IC₅₀ curves of CBD block on hNav1.1–1.7 and mNav1.6. The chemical structure of CBD is shown in the top left corner. B, THC IC₅₀ of hNav1.2 compared with CBD. The pulse protocols used and the chemical structure of THC are shown on the bottom right. Channels were exposed to each compound for 20 min. C, representative current traces at IC₅₀ in each sodium channel. Traces are taken from the concentrations that are closest to IC₅₀. D, table of IC₅₀ and Hill slope fitted parameters (n = 3–15 cells exposed at each concentration; the S.E. values quoted are errors of the fit).

Figure 2.

CBD effects on activation, steady-state fast inactivation (SSFI), resurgent, and persistent currents. A, conductance difference in hNav1. 1 in vehicle and 3.3 μm CBD concentration (vehicle: V_½ = −42.5 ± 0.8, slope = 3.1 ± 0.7, n = 11; CBD: V_½ = −39.6 ± 1.8, slope = 4.6 ± 1.5, n = 5). B, average current density of hNav1.1 in vehicle and CBD (vehicle: current density = −75.3 ± 8.9 pA/pF, n = 11; CBD: current density = −6.8 ± 3.0 pA/pF, n = 5). C, voltage dependence of activation as normalized conductance plotted against membrane potential (vehicle: V_½ = −41.9 ± 0.4 mV, slope = 3.6 ± 0.4, n = 7; CBD: V_½ =−40.3 ± 1.0 mV, slope = 7.2 ± 0.9, n = 5). D, voltage dependence of SSFI as normalized current plotted against membrane potential (vehicle: V_½ = −69.7 ± 0.2 mV, slope = 5.0 ± 0.2, n = 12; CBD: V_½ = −77.8 ± 0.3 mV, slope = 5.8 ± 0.3, n = 5). E, resurgent current block in hNav1.6 (vehicle: resurgent density = −33.9 ± 4.8 pA/pF, n = 11; CBD: resurgent density = −7.3 ± 1.2 pA/pF, n = 23; tetrodotoxin (TTX): resurgent density = −3.3 ± 0.7 pA/pF, n = 31). F, IC₅₀ of CBD block of peak and persistent currents in hNav1.6 mutant (N1768D) (peak: IC₅₀ = 10.0 ± 0.7 μm, slope = 2.0 ± 0.4, n = 3–11; persistent: IC₅₀ = 6.4 ± 1.0 μm, slope = 1.3 ± 0.2, n = 3–9).

Figure 3.

State-dependent inhibition of Nav current by CBD and effects on recovery from inactivation. A, pulse protocol showing 180 pulses run at 1 Hz at each holding potential and representative current traces. B, CBD potency at four holding potentials at pulse 180 (3 min) (IC₅₀: −100 mV = 12.7 ± 1.0 μm, −90 mV = 10.3 ± 0.5 μm, −80 mV = 6.7 ± 0.4 μm, −70 mV = 2.9 ± 0.6 μm; n = 2–6). C, apparent K_d at different voltages was well fit with a four-state model invoking different potencies for resting and inactivated-state block. D and E, recovery from inactivation in 3.7 μm CBD at: 300 ms (vehicle (Veh): τ_Fast = 0.00173 s, τ_Slow = 0.0688 s, n = 35; CBD: τ_Fast = 0.00654 s; τ_Slow = 0.516 s; n = 3) and 10 s (vehicle: τ_Fast = 0.0715 s, τ_Slow = 0.696 s, n = 33; CBD: τ_Fast = 0.272 s; τ_Slow = 8.72 s; n = 3). F, the slow components of recovery from inactivation in vehicle and CBD at 300 ms and 3 s are shown on the left y axis, and the fraction of slow to fast component of recovery from inactivation is shown on the right y axis.

Figure 4.

Lower temperatures increase kinetics and potency of CBD inhibitory effects. A–C, the kinetics of CBD block at 20, 28, and 33 °C at 50. 0, 25.0, 12.5, and 6.3 μm (20 °C: 6.3 μm = 51.0 ± 0.6 s, 12.5 μm = 42.2 ± 0.3 s, 25.0 μm = 21.6 ± 0.4 s, 50.0 μm = 23.3 ± 0.3 s; 28 °C: 6.3 μm = 162.6 ± 0.9 s, 12.5 μm = 87.7 ± 0.2 s, 25.0 μm = 42.8 ± 0.1 s, 50.0 μm = 30.4 ± 0.4 s; 33 °C: 6.3 μm = 299.0 ± 9.5 s, 12.5 μm = 137.0 ± 0.5 s, 25.0 μm = 79.2 ± 0.3 s, 50.0 μm = 84.1 ± 0.3 s, n = 10–14). The variability at the lower concentration of 6.3 μm at 33 °C is larger because of the slowing of CBD effect. D, time constants associated with the plots at A–C. E, IC₅₀ at the noted temperatures (IC₅₀ (μm): 20 °C = 2.1 ± 0.1, 28 °C = 3.4 ± 0.1, 33 °C = 4.7 ± 0.2, n = 2–11), the slope factor is fixed at 3.4. F, relationship of CBD potency as a function of temperature.

Figure 5.

CBD block of F1763A mutant, NaChBac, and Kv2.1. A, TTC block of hNav1.1 (hNav1.1, IC₅₀ = 1.6 ± 0.1 μm, slope = 0.7 ± 0.03; Phe¹⁷⁶³, IC₅₀ ∼ 44.1 μm) with and without the F1763A mutation. B, F1763A causes a slight decrease in CBD potency in hNav1.1 (hNav1.1: IC₅₀ = 2.5 ± 0.2 μm, slope = 2.0 ± 0.2, n = 2–6; F1763A: IC₅₀ = 4.8 ± 0.2 μm, slope = 4.1 ± 0.6, n = 3–811 cells exposed at each concentration). C, shows a comparison of the CBD inhibition of hNav1.6, NaChBac, and Kv2.1 (hNav1.6: IC₅₀ = 3.0 ± 0.1 μm, slope = 3.1 ± 0.2, n = 2–6; NaChBac: IC₅₀ = 1.5 ± 0.2 μm, slope = 2.8 ± 0.9, n = 1–4; Kv2.1: IC₅₀ = 3.7 ± 0.8 μm, slope = 1.1 ± 0.2, n = 1–5 11 cells exposed at each concentration). D, shows the state-dependent block of NaChBac tested at −55 and −100 mV (−100 mV: IC₅₀ = 1.5 ± 0.2 μm, slope = 2.8 ± 0.9, n = 1–4; −55 mV: IC₅₀ = 0.24 ± 0.05 μm, slope = 2.8 ± 0.9, n = 1–311 cells exposed at each concentration). E, current traces associated with the channels shown in C.

Figure 6.

CBD inhibits human iPSC neuronal Nav and Kv currents. A, concentration-response relationship for CBD inhibition of hNav1.2 channels obtained using manual patch clamp (IC₅₀ = 1.3 ± 0.1 μm, slope = 2.1 ± 0.4, n = 3–6). B, plot of normalized SSFI after a 100 ms prepulse before (V_½ = −50.0 ± 0.4 mV, slope = 6.4 ± 0.4, n = 3) and after (V_½ = −66.1 ± 0.7 mV, slope = 8.1 ± 0.6, n = 3) perfusion. C, open-state fast-inactivation time constants shown on log scale on the y axis at −20 mV for vehicles (−20 mV: 1.3 ± 0.1 ms, n = 3) and CBD (−20 mV: 1.3 ± 0.4 ms, n = 3) in iPSC neurons. D and E, representative current-voltage relationship recorded from iPSC neurons with a KF based internal solution before 1 μm perfusion of CBD (D) and after perfusion (E). Insets in the panel are zoomed-in view of currents at the test pulse used to assess availability.

Figure 7.

CBD reduces excitability in an action-potential model and schematic representation of CBD's mode of action. A, simulation of the effects of CBD on action potential morphology over a series of increasing current injection intensities. B, zoomed-in simulation of action potentials from the first interval shown in A. C, proposed mode of action of CBD involving interactions with both the membrane, resting and inactivated sodium channels.