Cannabidiol potentiates hyperpolarization-activated cyclic nucleotide-gated (HCN4) channels

Abstract

Cannabidiol (CBD), the main non-psychotropic phytocannabinoid produced by the Cannabis sativa plant, blocks a variety of cardiac ion channels. We aimed to identify whether CBD regulated the cardiac pacemaker channel or the hyperpolarization-activated cyclic nucleotide-gated channel (HCN4). HCN4 channels are important for the generation of the action potential in the sinoatrial node of the heart and increased heart rate in response to β-adrenergic stimulation. HCN4 channels were expressed in HEK 293T cells, and the effect of CBD application was examined using a whole-cell patch clamp. We found that CBD depolarized the V1/2 of activation in holo-HCN4 channels, with an EC50 of 1.6 µM, without changing the current density. CBD also sped activation kinetics by approximately threefold. CBD potentiation of HCN4 channels occurred via binding to the closed state of the channel. We found that CBD's mechanism of action was distinct from cAMP, as CBD also potentiated apo-HCN4 channels. The addition of an exogenous PIP2 analog did not alter the ability of CBD to potentiate HCN4 channels, suggesting that CBD also acts using a unique mechanism from the known HCN4 potentiator PIP2. Lastly, to gain insight into CBD's mechanism of action, computational modeling and targeted mutagenesis were used to predict that CBD binds to a lipid-binding pocket at the C-terminus of the voltage sensor. CBD represents the first FDA-approved drug to potentiate HCN4 channels, and our findings suggest a novel starting point for drug development targeting HCN4 channels.

Figure 1.

Effect of CBD on activation of holo-HCN4 channels. (A) Representative voltage protocol and activation current traces for mHCN4 construct before (black) and after (green) treatment with 5 µM CBD. The −100 mV trace is highlighted. (B) Average tail current values for treatment for mHCN4 before (black) and after application of 5 µM CBD (green). Values fit to a Boltzmann curve. Average V_1/2 for mHCN4 was −102.1 ± 1.1 (n = 9); average V_1/2 for 5 µM CBD was −87.5 ± 3.0 mV (n = 9). (C) Scatterplot of a shift in average ΔV_1/2 ± SEM for the vehicle (black) and increasing concentrations of CBD (green). One-way ANOVA and post-hoc Tukey’s test found a significant shift for 5, 10, and 20 µM CBD (***) versus the vehicle (P < 0.0001) and 1 µM CBD (P < 0.03). (D) ΔV_1/2 of activation after treatment with various concentrations of CBD. Data were fit with the Hill equation. EC₅₀ was 1.59 ± 0.21 µM with a slope of 2.50 ± 0.63. (E) Relative peak current from the end of activation epoch (I_{post-treatment}/I_{pre-treatment}) for the vehicle (black) and increasing concentrations of CBD (green). One-way ANOVA and post-hoc Tukey’s test found no significant differences in relative current levels after treatments (P > 0.2).

Figure 2.

Effect of CBD on activation kinetics of holo-HCN4 channels. (A) Superimposition of activation epoch (−120 mV) pre- (black) and post-treatment with 5 µM CBD (green), dotted line shows 0 current. Single exponential fit of activation epochs shown in dashed lines. (B) Average activation τ ± SEM between −140 and −100 mV. One-way ANOVA and post-hoc Tukey’s test found all concentrations of CBD significantly sped activation compared with vehicle (*P < 0.05; **P < 0.01; ***P < 0.001). Additionally at −120 mV, treatment with 5–20 µM CBD significantly sped activation compared to treatment with 1 µM CBD (P < 0.02). (C) Use-dependence 1/3 Hz protocol at −120 mV pre (black) and posttreatment with 5 µM CBD (green), sweep 1, 5, and 10 shown. The effect of CBD does not change over time. Dotted line shows 0 current. (D) Average τ_act ± SEM at −120 mV during 1/3 Hz protocol (n = 7), total of 20 sweeps. Unpaired Student’s t test found CBD significantly sped activation (P < 0.03), but that after treatment with 5 µM CBD sweep 1 τ_act is not significantly different than sweep 20 τ_act (P > 0.2). Thus, the effect of CBD does not change over the course of 60 s.

Figure S1.

Effect of cAMP on HCN4 channels. (A) Representative voltage protocol and activation current traces for mHCN4 without cAMP (i.e., apo) (grey, above) and mHCN4 with 30 µM cAMP (i.e., holo) (black, below). (B) Average tail current values for treatment with apo HCN4 (grey square) and holo HCN4 (30 µM cAMP) (black circle). Values fit to a Boltzmann curve. Average V_1/2 for apo HCN4 was −113.2 ± 2.2 mV (n = 7); average V_1/2 for holo HCN4 was −101.74 ± 0.75 mV (n = 39).

Figure 3.

CBD potentiation is distinct from cAMP and PIP₂. (A) Representative voltage protocol and traces of mHCN4 in the absence of cAMP before (black) and after (purple) treatment with 5 µM CBD. The −100 mV trace is highlighted. (B) Average tail current values for pre- (black square) and post-treatment with 5 µM CBD (purple square) in the absence of cAMP (apo-HCN4 channels), and treatment with 5 µM CBD in the presence of cAMP (green circle). Values were fit to the Boltzmann function. Average V_1/2 for pre-treatment (black square) was −113.2 ± 2.2 mV (n = 7); average V_1/2 for post-treatment with 5 µM CBD (purple square) was −97.6 ± 2.5 mV (n = 7); average V_1/2 for post-treatment with 5 µM CBD in presence of cAMP (green circle) was −87.5 ± 3.0 mV (n = 9). One-way ANOVA and post-hoc Tukey’s test found that the application of CBD significantly shifted V_1/2 (P < 0.003) by ∼15 mV, and the presence of cAMP and CBD significantly shifted V_1/2 (P < 0.04) a further 10 mV. (C) Representative traces of mHCN4 (black) and mHCN4 after application of diC8-PI(4,5)P₂ (orange). The −100 mV trace is highlighted. (D) Average tail current values for treatment with vehicle (black circle), 5 µM CBD (green circle), and 5 µM CBD with 10 µM diC8-PI(4,5)P₂ (orange diamond). Values fit to the Boltzmann function.

Figure 4.

Computed highest binding affinity binding site of CBD with the cAMP bound HCN4 structure. (A) Space-fill model of HCN4, CBD (green) highlighted by green circle. (B) Zoomed in view of lipid binding pocket bound by CBD. (C) CBD binding location in the VSD of one subunit. (D) CBD is predicted to be in close proximity to polar (Q401) and charged (R397) residues of S4. (E) Normalized superimposition of HCN4 Q401R activation epoch before (brown) and after application of 5 µM CBD (green) for representative trace.

Figure 5.

Schematic for the mechanism of action of CBD (green) on HCN4 channels. Top: Under normal physiological conditions, HCN4 channels would not open at weak driving forces (i.e., −80 mV). Bottom: When CBD is added, it stabilizes the S4 in the activated state, allowing HCN4 channels to open at weak driving forces. Figure created with https://BioRender.com.

Figure S2.

Comparison of single and double exponential fits. Left: Representative trace of holo mHCN4 after application of 5 µM CBD (green). Single exponential fits for −100, −120, and −140 mV traces shown as a dashed line. Right: Comparison of residuals from single (blue) and double (orange) exponential fits for a representative trace shown to left. Based on the lack of change in residuals with increase to a double exponential equation, single exponential fits were deemed sufficient to describe the activation curves.

Figure S3.

General regions for CBD (green) binding on an HCN4 cyroEM structure (PDB ID 7NP4). Calculated affinities for tetramer and monomer calculations listed on the right. Computed using AutoDock Vina, with exhaustiveness set to 1,000 for tetramer, 100 for monomer.

Figure S4.

CBD does not speed HCN4 Q401R activation. Representative traces were normalized to peak current and superimposed before (brown) and after (green) application of 5 µM CBD. (A–D) Out of a total of eight cells, four were chosen as representatives (A–D) across three separate transfections. D was chosen as the representative trace shown in Fig. 4 E. (E) The same cell as in D is shown after activation time from D was doubled to 8 s.

Figure S5.

Highest affinity binding site for cholesterol (blue) on HCN4 (PDB ID 7NP4) in the S4 lipid pocket. Calculated with AutoDock Vina.