The endocannabinoid system impacts seizures in a mouse model of Dravet syndrome

Abstract

Dravet syndrome is a catastrophic childhood epilepsy with multiple seizure types that are refractory to treatment. The endocannabinoid system regulates neuronal excitability so a deficit in endocannabinoid signaling could lead to hyperexcitability and seizures. Thus, we sought to determine whether a deficiency in the endocannabinoid system might contribute to seizure phenotypes in a mouse model of Dravet syndrome and whether enhancing endocannabinoid tone is anticonvulsant. Scn1a^+/- mice model the clinical features of Dravet syndrome: hyperthermia-induced seizures, spontaneous seizures and reduced survival. We examined whether Scn1a^+/- mice exhibit deficits in the endocannabinoid system by measuring brain cannabinoid receptor expression and endocannabinoid concentrations. Next, we determined whether pharmacologically enhanced endocannabinoid tone was anticonvulsant in Scn1a^+/- mice. We used GAT229, a positive allosteric modulator of the cannabinoid (CB₁) receptor, and ABX-1431, a compound that inhibits the degradation of the endocannabinoid 2-arachidonoylglycerol (2-AG). The Scn1a^+/- phenotype is strain-dependent with mice on a 129S6/SvEvTac (129) genetic background having no overt phenotype and those on an F1 (129S6/SvEvTac x C57BL/6J) background exhibiting a severe epilepsy phenotype. We observed lower brain cannabinoid CB₁ receptor expression in the seizure-susceptible F1 compared to seizure-resistant 129 strain, suggesting an endocannabinoid deficiency might contribute to seizure susceptibility. GAT229 and ABX-1431 were anticonvulsant against hyperthermia-induced seizures. However, subchronic ABX1431 treatment increased spontaneous seizure frequency despite reducing seizure severity. Cnr1 is a putative genetic modifier of epilepsy in the Scn1a^+/- mouse model of Dravet syndrome. Compounds that increase endocannabinoid tone could be developed as novel treatments for Dravet syndrome.

Keywords: Cannabinoid receptor; Dravet syndrome; Endocannabinoid system; Epilepsy.

Figure 1.. Strain-specific cannabinoid receptor expression.

Relative Cnr1 transcript levels in the (A) cortex and (B) hippocampus of wildtype (WT, open bars) and Scn1a^+/− (grey bars) mice on 129 and F1 (129 x B6) background strains. Cnr1 expression was measured in primary RNA pools using RT-ddPCR and expressed as a ratio of Tbp. Data represents mean ± SEM, with n = 7–11 per group. Hippocampal Cnr1 expression was significantly lower in F1 compared to 129 mice (*p < 0.05, two-way ANOVA). (C) Densitometry analysis of CB₁ receptor expression as a ratio to β-tubulin for individual male (blue) and female (purple), wildtype (WT, open bars) and Scn1a^+/− (shaded bars) mice on 129 and F1 (129 x B6) background strains. A significant strain-dependent effect was observed (p < 0.05, three-way ANOVA). Data represents mean ± SEM, with n = 3–4 per group. (D) Males and females in panel C were combined because no-sex dependent effect was observed. Hippocampal CB₁ receptor expression was significantly lower in male and female F1 compared to 129 mice (*p < 0.05, two-way ANOVA). Data represents mean ± SEM, with n = 7–8 per group. (E) Western blot analysis of CB₁ receptor levels in hippocampal membrane preparations from WT and Scn1a^+/− (KO) mice on 129 and F1 (129 x B6) background strains. Representative blots from two biological replicates are shown for each background strain and genotype. Image was cropped to improve conciseness. Full-length Western blot image can be found in Supplemental Figure 1.

Figure 2.. Hippocampal concentrations of major endocannabinoids.

(A) Chemical structures of endocannabinoids. Hippocampal concentrations of (B) 2-arachidonoylglycerol, (C) 1-arachidonoylglycerol and (D) anandamide in male (blue) and female (purple), wildtype (WT, open bars) and Scn1a^+/− (shaded bars) mice on 129 and F1 (129 x B6) background strains. A significant strain by sex interaction was observed for both 2-AG and 1-AG (left panels; p < 0.05; three-way ANOVA). Genotypes (WT and Scn1a^+/−) were combined for respective groups (right panels). Significant interaction effects were present from both 2-AG and 1-AG (p < 0.05; two-way ANOVA). Statistical comparisons of 1-AG concentrations were made using 1/concentration values. No differences across groups were observed for anandamide concentrations. Dashed line represents the limit of detection for anandamide (LOD 0.125 pmol/mg brain), with data points on the line having anandamide concentrations below the limit of detection. Data represents mean ± SEM, with n = 4–5 per group.

Figure 3.. GAT229 and ABX-1431 are anticonvulsant against hyperthermia-induced seizures in F1.Scn1a^+/− mice.

Concentration-time curve for GAT229 in mouse (A) plasma and (B) brain following a 5 mg/kg i.p. injection. Data are expressed as means ± SEM, with n = 4–5 per time point. (C) Threshold temperature of individual mice for GTCS induced by hyperthermia following acute i.p. treatment with vehicle (grey bar) or varying doses of GAT229 (green bars). GAT229 (30 mg/kg and 100 mg/kg) significantly elevated the temperature thresholds (41.7 ± 0.2°C and 41.8 ± 0.1°C, respectively) for thermally-induced seizures compared to vehicle (41.1 ± 0.1°C). The average temperatures of seizure induction are depicted by the bars and error bars represent SEM, with n = 18–22 per group (*p < 0.05, **p < 0.01; logrank Mantel-Cox). Concentration-time curve for ABX-1431 in mouse (D) plasma and (E) brain following a 5 mg/kg i.p. injection. Data are expressed as means ± SEM, with n = 4–5 per time point. (F) Threshold temperature of individual mice for GTCS induced by hyperthermia following acute i.p. treatment with vehicle (grey bar) or varying doses of ABX-1431 (blue bars). ABX-1431 (10 mg/kg) significantly elevated the temperature threshold (41.3 ± 0.2°C) for thermally-induced seizures compared to vehicle (40.6 ± 0.1°C). The average temperatures of seizure induction are depicted by the bars and error bars represent SEM, with n = 17 per group (***p < 0.001; logrank Mantel-Cox). Hippocampal concentrations of (G) 2-AG and (H) AEA in F1.Scn1a^+/− mice from thermally induced seizure experiments. There were no difference in endocannabinoid concentrations between male and female vehicle-treated mice so sexes were combined for all groups. ABX-1431 treatment resulted in a dose-dependent increase in 2-AG concentrations (*p < 0.05, **p < 0.01, Brown-Forsythe ANOVA followed by Dunnett’s T3 post hoc). No differences were observed for anandamide concentrations with acute ABX-1431 treatment (Kruskal-Wallis test). Data represents mean ± SEM, with n = 5–10 per group. (I) Threshold temperature of individual mice for GTCS induced by hyperthermia following acute treatment with ABX-1431 (blue bar) and GAT229 (green bar) administered individually or in combination (blue-grey bar). ABX-1431 (10 mg/kg) and GAT229 (30 mg/kg) significantly elevated the temperature threshold for thermally-induced seizures. The average temperatures of seizure induction are depicted by the bars and error bars represent SEM, with n = 16 per group (*p < 0.05; logrank Mantel-Cox).

Figure 4.. ABX-1431 is proconvulsant against spontaneous seizures of F1.Scn1a^+/− mice.

(A) Generalized tonic-clonic seizure (GTCS) frequency of individual untreated and ABX-1431-treated mice. Treatment was administered orally via supplementation in chow at P18 following the induction of a single hyperthermia-induced seizure. Unprovoked, spontaneous GTCS were quantified over a 60-h recording period. ABX-1431 treatment significantly increased the frequency of GTCS, with n = 17 per group (**p < 0.01, Mann-Whitney test). (B) Proportion of spontaneous GTCS with (grey bars) or without (white bars) full tonic hindlimb extension. ABX-1431 treatment significantly reduced the severity of GTCS (**p < 0.005, Fisher’s exact test). Total number of seizures for each treatment is presented in parentheses above the bar. (C) Survival curves comparing untreated and ABX-1431-treated mice. Treatment began at P18 and survival was monitored to P30. ABX-1431 had no effect on survival of F1.Scn1a^+/− mice (Mantel-Cox logrank). (D) Hippocampal concentrations of 2-AG in F1.Scn1a^+/− mice untreated or treated with ABX-1431. Concentrations of 2-AG are significantly increased following subchronic ABX-1431 treatment (*p < 0.05, Welch’s t-test). Data represents mean ± SEM, with n = 4 per group. (E) Densitometry analysis of CB₁ receptor expression as a ratio to β-tubulin for individual F1.Scn1a^+/− mice untreated or treated with ABX-1431. CB₁ receptor expression was significantly reduced following subchronic ABX-1431 treatment (**p < 0.01, Student’s t-test). Data represents mean ± SEM, with n = 4 per group. Full-length Western blot image can be found in Supplemental Figure 3.