Cannabidiol attenuates seizures and social deficits in a mouse model of Dravet syndrome

Abstract

Worldwide medicinal use of cannabis is rapidly escalating, despite limited evidence of its efficacy from preclinical and clinical studies. Here we show that cannabidiol (CBD) effectively reduced seizures and autistic-like social deficits in a well-validated mouse genetic model of Dravet syndrome (DS), a severe childhood epilepsy disorder caused by loss-of-function mutations in the brain voltage-gated sodium channel Na_V1.1. The duration and severity of thermally induced seizures and the frequency of spontaneous seizures were substantially decreased. Treatment with lower doses of CBD also improved autistic-like social interaction deficits in DS mice. Phenotypic rescue was associated with restoration of the excitability of inhibitory interneurons in the hippocampal dentate gyrus, an important area for seizure propagation. Reduced excitability of dentate granule neurons in response to strong depolarizing stimuli was also observed. The beneficial effects of CBD on inhibitory neurotransmission were mimicked and occluded by an antagonist of GPR55, suggesting that therapeutic effects of CBD are mediated through this lipid-activated G protein-coupled receptor. Our results provide critical preclinical evidence supporting treatment of epilepsy and autistic-like behaviors linked to DS with CBD. We also introduce antagonism of GPR55 as a potential therapeutic approach by illustrating its beneficial effects in DS mice. Our study provides essential preclinical evidence needed to build a sound scientific basis for increased medicinal use of CBD.

Keywords: Dravet syndrome; Scn1a; autism; cannabidiol; epilepsy.

Fig. 1.

CBD reduced seizures in DS mice. (A) Dose dependence of CBD reduction in duration of thermally induced seizures (SI Materials and Methods): vehicle, 31.5 ± 1.6 s, n = 116 seizures; 100 mg/kg CBD, 12.2 ± 1.9 s, n = 36 seizures; 200 mg/kg CBD, 14.8 ± 2.9 s, n = 21 seizures). F(5,115) = 11.24, P < 0.001, one-way ANOVA; post hoc comparisons, P < 0.001. (B) Dose dependence of CBD reduction in seizure severity (SI Materials and Methods): vehicle, 4.66 ± 0.07; 100 mg/kg CBD, 3.36 ± 0.20; 200 mg/kg CBD, 3.52 ± 0.36. F(5,115) = 7.65, P < 0.001, one-way ANOVA; post hoc comparisons, P < 0.001. (C) Mean number of seizures/mouse/hour (bars) after twice-daily injection of vehicle (blue) or 100 mg/kg CBD (red) from postnatal day 21–27: vehicle, 0.31 ± 0.10 seizures/mouse/hour, n = 6; CBD, 0.09 ± 0.04, n = 7. F(1,13) = 7.99, P = 0.03, two-way ANOVA. ***P < 0.001.

Fig. S1.

CBD does not impair locomotor activity in DS mice. (A) Summary plot showing the total distance traveled in the open field during a 10-min test session. CBD (100 mg/kg) reduced distance traveled in WT [t(12) = 2.90, P = 0.013; paired-t test] and DS mice [t(14) = 3.53, P = 0.003]. (B) Summary plot showing the mean velocity as the subject traversed through the center of the chamber. CBD had no effect on the mean velocity of mice of either genotype as they traveled through the center of the chamber. (C) Summary plot showing the mean latency to fall on the accelerating rotarod as an average of each of the 3 d. There was significant interaction between treatment and genotype [F(3,75) = 6.00, P = 0.017], revealing that CBD impaired performance in DS mice only, but DS mice treated with CBD still performed better than WT mice treated with either vehicle or CBD (all P < 0.05, post hoc comparisons). *P < 0.05; **P < 0.01.

Fig. 2.

CBD improved social deficits in DS mice. (A) Effects of CBD in the Three-Chamber Test of Social Interaction. Preference ratio (PR) was measured as in SI Materials and Methods. WT with vehicle: PR = 0.59 ± 0.02, n = 47; t(46) = 5.09, P < 0.001, one-sample t test; DS with vehicle: PR = 0.49 ± 0.03, n = 53; t(52) = 0.43, P = 0.67. CBD improved PR in DS mice but not in WT: genotype by dose interaction, F(3,169) = 2.75, P = 0.045, two-way ANOVA. CBD showed a trend toward increased PR at 10 mg/kg: 0.55 ± 0.04, n = 11, P = 0.09, post hoc comparisons. CBD significantly increased the PR at 20 mg/kg: 0.60 ± 0.03, n = 11, P = 0.006. (B) Effect of CBD on locomotor activity in the Open Field Test (SI Materials and Methods). DS mice are hyperactive compared with WT [main effect of genotype, F(1,91) = 116.25, P < 0.001, two-way ANOVA]. CBD had no effect on distance traveled in either genotype [main effect of dose, F(3,91) = 1.19, P = 0.32]. (C and D) Effect of CBD on frequency of social interactions (C) and escape behaviors (D) in the Reciprocal Interaction Test (SI Materials and Methods) compared using independent-samples t-tests. (C) Social interactions. WT with vehicle, 42.9 ± 2.8, n = 14; DS with vehicle, 42.9 ± 3.2, n = 13; WT with 10 mg/kg CBD, 38.7 ± 1.7, n = 9; DS with 10 mg/kg CBD, 53.9 ± 4.2. n = 8, t(19) = 2.12, P = 0.048. (D) Defensive escapes: WT with vehicle, 6.21 ± 0.83, n = 14; DS with vehicle, 13.1 ± 2.0, n = 13; t(25) = 3.26, P = 0.003; WT with 10 mg/kg CBD, 5.67 ± 1.68, n = 9, P > 0.28; DS with 10 mg/kg CBD, 5.88 ± 1.11, n = 8, t(19) = 2.66, P = 0.015. ^#P = 0.09; *P < 0 05; **P < 0.01.

Fig. 3.

CBD increased inhibitory neurotransmission to DGCs. (A) Voltage-clamp recordings (V_h = −60 mV) of sIPSCs in DGCs (E_Cl = 0 mV) in WT and DS mice (SI Materials and Methods). (B) sIPSC frequency: WT, 4.5 ± 0.75 Hz, n = 7; DS: 2.5 ± 0.5 Hz, n = 10; t(15) = 2.34, P = 0.034, independent-samples t test. sIPSC amplitude: WT: 31.5 ± 1.8 pA; DS: 30.7 ± 3.2 pA; P = 0.83. (C) Voltage-clamp recordings (V_h = −60 mV) of DGC sIPSCs (E_Cl = 0 mV) in CNQX (20 µM) and APV (50 µM), GABAzine (GBZ), or TTX (500 nM). (D) Voltage-clamp recordings (V_h = −60 mV) of DGC sEPSCs (E_Cl = −60 mV) alone, in GBZ, or in TTX. (E) Percentage change in frequency of sIPSCs and EPSCs. sIPSCs: CBD, 34.6 ± 13.6%; t(6) = 2.55, P = 0.043, n = 7, one-sample t test; CBD + GBZ, 0% change, n = 3; CBD + TTX, −1.8 ± 4.7%, n = 7. sEPSCs: CBD, −30.7 ± 5.1%; t(4) = 6.0, P = 0.004, n = 5, one-sample t test; CBD + GBZ, 0.92 ± 23.5%; t(4) = 0.05, P = 0.97, n = 5; CBD + TTX, 11.0 ± 13.6%, n = 6. Frequencies, amplitudes, and decay of miniature IPSCs and EPSCs (mIPSCs and mEPSCs) were unchanged by CBD. mIPSCs: frequency, −1.8 ± 4.7%, n = 7; amplitude, 5.7 ± 7.0%; decay, 10.0 ± 12.6%. mEPSCs: frequency, 11.0 ± 13.6%, n = 6; amplitude, 0.4 ± 4.1%; decay, −3.5 ± 7.1%; all *P > 0.05, one-sample t tests.

Fig. 4.

Differential effects of CBD on action potential firing in excitatory and inhibitory neurons. (A) Recordings of APs in DGCs (SI Materials and Methods) in the presence of subrheobase current (M = 30.83 ± 5.76 pA, n = 12) in vehicle, CBD (16 µM), or CBD + GBZ. Percent change in frequency: CBD, −58.2 ± 13.1%; t(4) = 4.44, P = 0.01, n = 5; CBD + GBZ, 11.6 ± 23.5%; t(6) = 0.50, P = 0.64, n = 7, one-sample t test). (B) Percentage maximum AP frequency in the presence of vehicle (blue) or 16 µM CBD (red) with increasing current injection in PV-positive interneurons identified with td-Tomato fluorescence (SI Materials and Methods). (Inset) Current-clamp recordings at 75 pA injection amplitude. Rheobase: DS + vehicle: 121.7 ± 19.4 pA, n = 6; DS + CBD: 95 ± 19.5 pA, n = 6; t(5) = 2.61, P = 0.048, paired-t test. CBD increased AP frequency at increasing current injection amplitudes [main effect of condition, F(1,200) = 7.87, P = 0.04, repeated-measures ANOVA]. (C) Percentage maximum AP frequency in vehicle (blue) or 16 µM CBD (red) at increasing current injection steps in DGCs. Rheobase: vehicle: 20.8 ± 4.0 pA; CBD: 16.7 ± 3.3 pA, n = 6, P > 0.4, CBD reduced. AP firing frequency [main effect of condition, F(1,120) = 12.65, P = 0.04, two-way repeated-measures ANOVA]. *P < 0.05.

Fig. 5.

CBD enhancement of sIPSC frequency is independent of CB1 receptors. (A) Voltage-clamp recordings (V_h = −60 mV) of DGC sIPSCs (E_Cl = 0 mV; SI Materials and Methods) in the presence of CNQX (20 µM) and APV (50 µM), the CB1 receptor antagonist AM281 (1 µM), or AM281 and CBD (16 µM). (B) Percentage change in sIPSC frequency. AM281 enhanced sIPSC frequency from baseline [14.0 ± 4.4%, n = 6; t(5) = 2.94, P = 0.032, one-sample t test]. CBD further enhanced sIPSC frequency compared with AM281 alone [30.5 ± 6.52%, n = 6, t(5) = 4.67, P = 0.005]. *P < 0.05; **P < 0.01.

Fig. 6.

Role of GPR55 in the CBD-induced increase in inhibitory transmission. (A) Current-clamp recordings of DGCs (SI Materials and Methods) in the presence of subrheobase-stimulating current (M = 28.0 ± 10.0 pA, n = 10). (B) Voltage-clamp recordings (V_h = −60 mV) of DGC sIPSCs (E_Cl = 0 mV; SI Materials and Methods) in CID16020064 (CID, 10 µM) or CBD (16 µM) in the presence of CID, the AMPA receptor antagonist CNQX (20 µM), and the NMDA receptor antagonist APV (50 µM). Asterisks indicate sIPSC. (C) Percentage change in AP frequency: CID 10 µM, −44.18 ± 11.1%, t(9) = 4.00, P = 0.003, n = 10; CID (10 µM) + CBD, −9.8 ± 15.2%, n = 9, t(8) = 0.64, P = 0.64, one-sample t test. Percentage change in IPSC frequency: 10 µM CID, 16.4 ± 9.3%, n = 9, t(8) = 1.77, P = 0.1. CID 10 µM blocked the CBD-induced enhancement of sIPSC frequency [−6.74 ± 7.42%, n = 9, t(8) = 0.91, P = 0.39, one-sample t test]. CID (1 µM), 27.9 ± 9.3%, n = 9, t(8) = 3.00, P = 0.02. CID (1 µM) also blocked the CBD-induced enhancement of sIPSC frequency [2.54 ± 12.4%, n = 11, t(10) = 0.21, P = 0.84]. (D) Percentage maximum AP frequency in 10 µM CID (blue) or 16 µM CBD (red) at increasing current injection steps in PV-positive interneurons identified with td-Tomato fluorescence. Rheobase: vehicle, 64.0 ± 9.27 pA, n = 5; CBD: 62.0 ± 5.83, n = 5; t(4) = 0.54, P = 0.62, paired t test. CID blocked the effect of CBD on AP frequency [main effect of condition, F(1,160) = 2.31, P = 0.20, repeated-measures ANOVA]. ^#P = 0.07; *P < 0.05; **P < 0.01.