Cannabidiol reduces atypical absence seizures and epileptic spasms in a Gabrb3+/D120N mouse model of Lennox–Gastaut syndrome Authors: Thomas Harman, Laura A. Sharman, Ka Lai Yip, Miguel Bedoya‐Pérez, Vaishali Janve, Jonathon C. Arnold Published: 2026 ABSTRACT AbstractObjectiveLennox–Gastaut syndrome (LGS) is a drug‐resistant developmental and epileptic encephalopathy (DEE). Preclinical drug development for LGS is constrained by a lack of syndrome‐relevant animal models. We aimed to evaluate a Gabrb3 +/D120N knock‐in (KI) mouse model of LGS by quantifying atypical absence seizures and epileptic spasms and assessing their sensitivity to antiseizure agents.MethodsVideo‐EEG recordings of adult (10‐week‐old) KI and wild‐type (WT) littermates were scored for atypical absence seizures, and the acute effects of ethosuximide (200 mg/kg), ulixacaltamide (60 mg/kg), and cannabidiol (CBD, 100 mg/kg) on seizure incidence and duration were evaluated using a within‐subjects, crossover design. Video recordings of postnatal day 16 (P16) KI and WT littermates were scored for epileptic spasms, and the effects of once‐daily dosing with vigabatrin (100 mg/kg) and CBD (100 mg/kg) from P13 to P15 were evaluated against vehicle.ResultsAdult KI but not WT mice exhibited spontaneous atypical absence seizures. CBD, ethosuximide, and ulixacaltamide reduced seizure incidence and duration. Epileptic spasms were more frequent in KI than in WT mice at P16. CBD and vigabatrin significantly reduced spasm frequency compared to the vehicle.Significance Gabrb3 +/D120N mice display robust atypical absence seizures and neonatal spasms that respond to antiseizure agents, supporting the predictive validity of this model as a preclinical platform for LGS drug discovery. CBD produced reductions in both atypical absence seizures and infantile spasms, suggesting that this model may be utilized as a translational tool for evaluating novel cannabinoid therapeutics for DEEs.Plain Language SummaryLennox–Gastaut syndrome (LGS) is a rare type of epilepsy that's hard to treat and poses a challenge for developing new drugs. Finding suitable animal models that accurately represent LGS is crucial. This article describes the development of a mouse model of LGS with a genetic mutation that increases seizures and epileptic spasms. We tested how different antiseizure drugs affect the mice. CBD, ethosuximide, and ulixacaltamide reduced seizure incidence and duration. CBD and vigabatrin also reduced spasm frequency in young mice. These promising results suggest the mouse model could be a valuable tool for drug discovery in LGS. Key points Gabrb3 +/D120N mice model spontaneous seizures associated with Lennox–Gastaut syndrome, exhibiting robust atypical absence seizures and neonatal epileptic spasms.Cannabidiol, ethosuximide, and the selective Cav3 inhibitor, ulixacaltamide, reduced atypical absence seizures.Cannabidiol and vigabatrin reduced neonatal epileptic spasms.Broad efficacy of cannabidiol across seizure types supports this model as a screening platform for novel cannabinoid therapies. Key points Gabrb3 +/D120N mice model spontaneous seizures associated with Lennox–Gastaut syndrome, exhibiting robust atypical absence seizures and neonatal epileptic spasms.Cannabidiol, ethosuximide, and the selective Cav3 inhibitor, ulixacaltamide, reduced atypical absence seizures.Cannabidiol and vigabatrin reduced neonatal epileptic spasms.Broad efficacy of cannabidiol across seizure types supports this model as a screening platform for novel cannabinoid therapies. Gabrb3 +/D120N mice model spontaneous seizures associated with Lennox–Gastaut syndrome, exhibiting robust atypical absence seizures and neonatal epileptic spasms. Gabrb3 +/D120N mice model spontaneous seizures associated with Lennox–Gastaut syndrome, exhibiting robust atypical absence seizures and neonatal epileptic spasms. Cannabidiol, ethosuximide, and the selective Cav3 inhibitor, ulixacaltamide, reduced atypical absence seizures. Cannabidiol, ethosuximide, and the selective Cav3 inhibitor, ulixacaltamide, reduced atypical absence seizures. Cannabidiol and vigabatrin reduced neonatal epileptic spasms. Cannabidiol and vigabatrin reduced neonatal epileptic spasms. Broad efficacy of cannabidiol across seizure types supports this model as a screening platform for novel cannabinoid therapies. Broad efficacy of cannabidiol across seizure types supports this model as a screening platform for novel cannabinoid therapies. INTRODUCTION Lennox–Gastaut syndrome (LGS) is a severe developmental and epileptic encephalopathy (DEE) that represents approximately 1–2% of all epilepsy cases and 19% of children with seizures presenting in infancy. 1 , 2 LGS is defined by multiple drug‐resistant seizure types, cognitive and behavioral impairments, and characteristic electroencephalogram (EEG) features, including slow spike‐and‐wave activity (SSW, ≤ 2.5 Hz). 1 Onset typically occurs within the first 8 years of life and persists into adulthood, with tonic and atypical absence seizures remaining frequent and pharmacoresistant. 1 , 3 Critically, LGS evolves from infantile epileptic spasms syndrome (IESS) in approximately 20–30% of cases, a developmental trajectory associated with particularly poor neurocognitive outcomes and increased mortality. 4 , 5 , 6 Despite the availability of several antiseizure medications (ASMs), often administered in combination, long‐term seizure freedom in LGS remains rare, highlighting the urgent need for novel therapies. 7 , 8 Progress in LGS drug discovery has been limited by a paucity of syndrome‐relevant preclinical models with robust construct, face, and predictive validity. 9 However, the recent identification of pathogenic variants in the γ‐aminobutyric acid type A (GABAA) receptor β3 subunit gene (GABRB3) as a cause of LGS provides a mechanistic foundation for genetic model development. 10 , 11 The heterozygous Gabrb3 +/D120N knock‐in mouse, which harbors a patient‐derived missense mutation, recapitulates core LGS features, including multiple seizure types, cognitive‐behavioral impairments, and EEG abnormalities. 12 Notably, neonatal Gabrb3 +/D120N mice also exhibit epileptic spasms, suggesting a conserved developmental trajectory from IESS to LGS; however, these behaviors have yet to be pharmacologically validated. Initial pharmacological characterization of this model focused on atypical absence seizures, as their high frequency enables robust quantification of drug efficacy. 12 , 13 These seizures are a clinical hallmark of LGS and contribute to cognitive impairments. 3 , 14 While the clinically utilized ASMs ethosuximide and clobazam, along with experimental cannabinoid‐related agents magnolol and WIN 55,212–2, attenuate atypical absence seizures in this model, the efficacy of the clinically approved LGS therapy, cannabidiol (CBD), remains unexplored. 12 , 13 , 15 Establishing the predictive validity of CBD in the Gabrb3 +/D120N model is a prerequisite for its implementation as a preclinical screening platform for next‐generation cannabinoid therapeutics. The present study aimed to validate the Gabrb3 +/D120N mouse as a translational platform for LGS‐targeted drug discovery and to characterize the effects of CBD in this model. We quantified atypical absence seizures and evaluated the efficacy of ethosuximide and the selective T‐type Ca2+ channel (Cav3) inhibitor, ulixacaltamide, agents known to suppress absence seizures in preclinical models. 15 , 16 We then quantified epileptic spasms and assessed the efficacy of vigabatrin, a first‐line ASM for IESS. 17 Finally, we evaluated CBD across both seizure types to determine its potential as a benchmark for novel cannabinoid‐related agents. We report the first pharmacological validation of epileptic spasms in the Gabrb3 +/D120N model and demonstrate the efficacy of CBD across multiple seizure types. These findings support the predictive validity of this model as a preclinical screening tool for DEE drug discovery and establish its utility in modeling an IESS‐to‐LGS developmental trajectory. This model provides a novel translational platform for evaluating novel therapeutics targeting the characteristic features and developmental progression of drug‐resistant epilepsy syndromes. METHODS Materials Ulixacaltamide, ethosuximide, and vigabatrin were purchased from Sigma‐Aldrich (St. Louis, USA). CBD was purchased from THC Pharm (Frankfurt, Germany). All drugs were prepared fresh on the morning of experiments in an ethanol‐Tween 80–0.9% saline vehicle (1:1:18 ratio) as previously described. 18 Animals All procedures were approved by the University of Sydney Animal Ethics Committee (approval numbers 2021/AE001952, 2021/AE001946) and conducted in accordance with the Australian Code for the Care and Use of Animals for Scientific Purposes. Gabrb3 +/D120N mice were obtained from the Jackson Laboratory (stock 035437). KI and WT littermates were generated by breeding Gabrb3 +/D120N males with female C57BL/6J mice (Jackson Laboratory stock 000664). Genotyping was performed via polymerase chain reaction as previously described. 12 Mice were housed in a pathogen‐free facility in ventilated cages of up to five adults per cage and maintained on a 12 h light–dark cycle with access to food and water ad libitum. Animals were habituated for at least 1 week prior to experimental procedures. Electroencephalogram recording and analysis Synchronized video‐EEG recordings were acquired from 10‐week‐old male and female mice using a Pinnacle Technology monitoring system (USA). EEG electrode implantation was performed as previously described. 12 , 13 Briefly, mice received meloxicam (5 mg/kg s.c.) 15 min before surgery and were anesthetized with isoflurane (4% induction, 1–2.5% maintenance) in a stereotactic frame. Body temperature was maintained at 37°C using a homeothermic monitoring system (ConductScience), and ophthalmic ointment (Lubrithal eye gel, Pet Chemist Pharmacy) was applied to avoid eye irritation. Following cranial incision, a prefabricated headmount (Pinnacle Technology) was affixed to the skull using cyanoacrylate adhesive, and stainless‐steel screws (0.10″ anterior, 0.12″ posterior) were inserted via bore holes to contact the brain surface. Electromyography (EMG) wires were inserted into the nuchal muscles. The incision was closed with non‐absorbable monofilament sutures, and dental cement was applied to secure and insulate the headmount. Postoperative buprenorphine (0.05 mg/kg s.c.) was administered for analgesia, and mice were monitored daily for 1 week, with meloxicam (5 mg/kg s.c.) administered once daily for 48 h post‐surgery. On recording days, mice were habituated to recording chambers 30 min before data acquisition. Pre‐amplifiers were connected to the headmount, allowing for unrestricted movement. Acute effects of antiseizure drugs on atypical absence seizures were evaluated using a vehicle‐controlled, within‐subjects crossover design. KI mice were randomized to treatment order using a Latin‐square design with a one‐week washout period between experiments. Ethosuximide (200 mg/kg), ulixacaltamide (60 mg/kg), CBD (100 mg/kg), or vehicle was administered (i.p.) 30 min before the light–dark transition, which is the period during which atypical absence seizures are most frequently observed. 12 EEG and EMG signals were acquired using Sirenia Acquisition software (Pinnacle Technology, v2.2.2). One‐hour epochs spanning the light–dark transition (17:30–18:30) were scored offline using Somnivore software by an experimenter blinded to treatment as previously described. 12 , 13 Atypical absence seizures were defined as diffuse, generalized SSW activity (3–7 Hz) accompanied by behavioral arrest. Spasms On postnatal day 16 (P16), male and female pups were separated from the home cage and individually video‐recorded for 10 min. Recordings were scored offline by a blinded observer for epileptic spasms, defined as brief (~1 s) flexion‐extension movements accompanied by behavioral arrest. 12 To assess sensitivity to ASMs, KI pups were randomly allocated to receive vigabatrin, CBD, or vehicle. From P13 to P15, pups received once‐daily i.p. injections of vigabatrin (100 mg/kg), CBD (100 mg/kg), or vehicle. Spasm burden was assessed on P16 as described above. Statistical methods Sex was included as a fixed effect in preliminary statistical analyses. No statistically significant main effect of sex or interaction effects were detected; data were therefore pooled across sex for all analyses. Parametric or non‐parametric tests were selected based on the assessment of residuals distribution. Statistical significance was set at p < 0.05 for all analyses. Atypical absence seizure incidence and duration were analyzed in 10 min bins using generalized linear mixed models (GLMM) in R 2024.4.2 as previously described. 19 Cumulative seizure incidence and duration over the 1 h recording period were analyzed using paired t‐tests in GraphPad Prism 10.6.1 (GraphPad Software, USA). Epileptic spasm incidence was analyzed in GraphPad Prism 10.6.1. Cumulative spasm counts over 10 min were compared between WT and KI mice using an unpaired t‐test. Treatment effects were analyzed using the Kruskal‐Wallis test with post‐hoc Dunn's multiple comparisons. One outlier was excluded from the vehicle group, as identified by the ROUT method (Q = 1%), indicating that the data point was 99% likely to be a statistical outlier. RESULTS Gabrb3 +/D120N mice exhibit spontaneous atypical absence seizures Gabrb3 +/D120N mice (n = 7) but not WT littermates (n = 6) exhibited spontaneous atypical absence seizures characterized by diffuse, low‐frequency (3–7 Hz) generalized slow spike‐waves on EEG, accompanied by behavioral arrest (Figure 1A). No WT (0/6) and all KI animals (7/7) displayed atypical absence seizures (mean ± SEM 7.4 ± 3.1; range 1–21) with cumulative duration 46.3 ± 22.7 s (range 4–152 s) at various timepoints over the 1 h epoch (Figure 1B–E). Gabrb3 +/ D120N knock‐in (KI) mice exhibited spontaneous atypical absence seizures. 1 h video‐EEG recordings spanning the light–dark transition were acquired from KI (n = 7: 3 male, 4 female) and wildtype (WT, n = 6: 3 male, 3 female) littermates and analyzed for atypical absence seizures. (A) Representative electroencephalographic (EEG) and electromyographic (EMG) traces of an atypical absence seizure in a Gabrb3 +/D120N mouse. Arrows indicate seizure onset and offset. (i) Expanded view of the underlined segment shows a characteristic ictal slow spike‐and‐wave EEG pattern. (B, C) Seizures with frequent incidence and (D, E) prolonged duration were observed in KI but not WT mice. Timecourse of seizure incidence and seizure duration were quantified in 10 min bins. Data are represented as mean ± SEM; overlaid circles indicate individual mouse data. Antiseizure drugs reduce atypical absence seizures in Gabrb3 +/D120N mice Ethosuximide (200 mg/kg) reduced total seizure incidence by 64% (vehicle 21.4 ± 4.7, range 5–45 vs. ethosuximide 7.7 ± 1.8, range 1–20; t(8) = 2.6, p = 0.03) and total seizure duration by 74% (vehicle 272.9 ± 62.5 s, range 48–544 s vs. ethosuximide 70.7 ± 16.9 s, range 20–184 s; t(8) = 3.2, p = 0.01) (Figure 2A,C). Timecourse analysis revealed a significant main effect of ethosuximide treatment on seizure incidence (χ 2[1] = 21.3, p = 0.02), with fewer seizures at 10 (p = 0.02), 20 (p = 0.003), 30 (p < 0.001), 40 (p < 0.001), and 50 min (p = 0.02) post‐administration (GLMM with Sidak correction; Figure 2B). There was no significant main effect of ethosuximide treatment on seizure duration (χ 2[1] = 2.5, p = 0.1; Figure 2D), nor a significant treatment‐by‐time interaction (χ 2[1] = 6, p = 0.1). Ethosuximide reduced atypical absence seizures in Gabrb3 +/D120N mice. Ethosuximide (200 mg/kg i.p.) reduced atypical absence (A) total seizure incidence and (B) seizure incidence at various timepoints 1 h post‐administration, as well as (C) total seizure duration, (D) with no significant timecourse effects on seizure duration 1 h post‐administration n (n = 9: 5 male, 4 female). Seizure number and duration were quantified in 10 min bins over a 1 h recording period and analyzed using a generalized linear mixed model with Sidak correction; cumulative totals were compared using paired t‐tests or Wilcoxon tests. Data are represented as mean ± SEM; overlaid circles indicate individual mouse data. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Ulixacaltamide (60 mg/kg) reduced total seizure incidence by 64% (vehicle 21.1 ± 2.8, range 5–44 vs. ulixacaltamide 7.5 ± 2.3, range 0–22; t(10) = 3.5, p = 0.006) and total seizure duration by 60% (vehicle 168.4 ± 25.55 s, range 48–544 s vs. ulixacaltamide 67.6 ± 21.4 s, range 0–200 s; t(10) = 3.2, p = 0.009) (Figure 3A,C). Timecourse analysis revealed a significant treatment effect of ulixacaltamide on seizure incidence (χ 2[1] = 6, p = 0.01), with fewer seizures at 10 (p < 0.001), 20 (p < 0.001), 30 (p < 0.001), 40 (p < 0.001), and 50 min (p = 0.003) post‐administration (Figure 3B). There was no significant main effect of ulixacaltamide treatment on seizure duration (χ 2[1] = 1.35, p = 0.2; Figure 3D), nor a significant treatment by time interaction (χ 2[1] = 0.4, p = 0.5). Ulixacaltamide reduced atypical absence seizures in Gabrb3 +/D120N mice. Ulixacaltamide (60 mg/kg i.p.) reduced atypical absence (A) total seizure incidence and (B) seizure incidence at various timepoints 1 h post‐administration, as well as (C) total seizure duration 1 h post‐administration, (D) with no significant timecourse effects on seizure duration (n = 11: 5 male, 6 female). Seizure number and duration were quantified in 10 min bins over a 1 h recording period and analyzed using a generalized linear mixed model with Sidak correction; cumulative totals were compared using paired t‐tests or Wilcoxon tests. Data represent mean ± SEM (**p < 0.01, ****p < 0.0001). CBD (100 mg/kg) reduced total seizure incidence by 46% (vehicle 24.4 ± 2.6, range 16–40 vs. CBD 13.2 ± 1.6, range 5–22; t(8) = 3.3, p = 0.01), and total seizure duration by 51% (vehicle 184.9 ± 22.9 s, range 88–308 s vs. CBD 90.2 ± 11.7 s, range 44–172 s; t(8) = 3.95, p = 0.004) (Figure 4A,C). Timecourse analysis revealed significant effects of CBD treatment on seizure incidence (χ 2[1] = 6.4, p = 0.01) and seizure duration (χ 2[1] = 4.1, p = 0.04) with CBD‐treated mice exhibiting significantly fewer seizures than vehicle controls at 30 (p = 0.01), 40 (p = 0.007), and 50 min (p = 0.05) and reduced seizure duration at 30 (p = 0.03) and 40 min (p = 0.02) post‐administration (Figure 4B,D). There were no significant treatment by time interaction effects for seizure incidence (χ 2[1] = 0.9, p = 0.35) or duration (χ 2[1] = 0.3, p = 0.6). Cannabidiol (CBD) reduced atypical absence seizures in Gabrb3 +/D120N mice. CBD (100 mg/kg) reduced atypical absence (A) total seizure incidence and (B) seizure incidence at various timepoints 1 h post‐administration, as well as (C) total seizure duration and (D) seizure duration at various timepoints 1 h post‐administration (n = 9: 3 male, 6 female). Seizure number and duration were quantified in 10 min bins over a 1 h recording period and analyzed using a generalized linear mixed model with Sidak correction; cumulative totals were compared using paired t‐tests or Wilcoxon tests. Data are represented as mean ± SEM; overlaid circles indicate individual mouse data (*p < 0.05, **p < 0.01). Neonatal Gabrb3 +/D120N mice exhibit spontaneous epileptic spasms that are ameliorated by ASMs A subset of WT mice (4/11) and all KI mice (10/10) exhibited epileptic spasms characterized by brief (~1 s) rapid flexion‐extension movements accompanied by behavioral arrest. Spasm frequency was significantly greater in KI mice (4.9 ± 1.9, range 4–9) than WT littermates (1 ± 1.4, range 0–4; t(19) = 5.4, p < 0.001; Figure 5A). Gabrb3 +/D120N mice had spontaneous epileptic spasms on postnatal day 16 (P16) that were reduced by cannabidiol (CBD) and vigabatrin (VGB). (A) Knock‐in (KI, n = 10: 5 male, 5 female) pups exhibited greater spasm incidence than wild‐type (WT, n = 11: 5 male, 6 female) littermates on P16. (B) Daily dosing with VGB (100 mg/kg i.p., n = 12; 6 male, 6 female) or CBD (100 mg/kg i.p., n = 13: 6 male, 7 female) from P13 to P15 reduced spasms on P16 compared to vehicle (n = 11: 5 male, 6 female). Spasms were quantified over 10 min recording period. KI and WT data were compared using the Mann–Whitney test; effects of treatments were analyzed using the Kruskal‐Wallis test with Dunn's correction. Data are represented as mean ± SEM; overlaid circles indicate individual mouse data. *p < 0.05, **p < 0.01, ****p < 0.0001. Daily dosing with vigabatrin (100 mg/kg) or CBD (100 mg/kg) from P13 to P15 reduced spasm burden on P16. Vigabatrin reduced spasm incidence by 41% (2.9 ± 0.3, range 2–5) and CBD reduced spasm incidence by 45% (2.7 ± 0.65, range 1–8) compared to vehicle (4.9 ± 0.3, range 3–6; Kruskal‐Wallis = 11.87, p = 0.003; Figure 5B). Notably, while CBD significantly reduced the mean spasm burden, we observed a bifurcation in treatment response: 54% (7/13) of the cohort demonstrated a near‐complete suppression of spasms (one spasm), while the remaining 46% (6/13) showed a modest reduction or no reduction (2–8 spasms). DISCUSSION There is an urgent need to develop novel therapies for DEEs, such as LGS, given high rates of pharmacoresistance. However, progress in preclinical drug discovery has been limited by a paucity of syndrome‐relevant animal models. 9 In this study, we evaluated a Gabrb3 +/D120N mouse model as a translational platform for LGS and DEE‐related drug screening. 12 We validated the seizure phenotype of this model, demonstrating that Gabrb3 +/D120N mice displayed robust atypical absence seizures that were reduced by ethosuximide and ulixacaltamide – agents established to suppress absence seizures in preclinical models. 16 , 20 Further, P16 mutants exhibited frequent epileptic spasms at rates significantly exceeding those of wild‐type littermates. We provide the first pharmacological validation of neonatal spasms in this model, demonstrating that vigabatrin, a clinical first‐line ASM for infantile spasms, reduced spasm burden in Gabrb3 +/D120N mice. 21 Notably, we show for the first time that the cannabinoid ASM CBD attenuates both atypical absence seizures and epileptic spasms in this model. These findings extend pharmacological validation of the Gabrb3 +/D120N mouse model for DEE‐related drug discovery and provide a benchmark for preclinical evaluation of novel cannabinoid therapies. Atypical absence seizures are generalized non‐motor seizures characterized by impaired awareness and SSW activity on EEG. 22 They occur in approximately 60% of individuals with LGS, often persisting into adulthood and contributing to neurocognitive impairment. 3 , 14 Consistent with previous reports, Gabrb3 +/D120N mice exhibited frequent spontaneous atypical absence seizures characterized by irregular low‐frequency (3–7 Hz) SSW activity, accompanied by behavioral arrest. 12 These features resemble other rodent models of atypical absence seizures, including the GABABR1a overexpression mouse and the Gabrb3 +/N328D mouse, and are comparable to the ≤2.5 Hz SSW patterns observed in clinical populations, accounting for species‐specific differences in oscillatory frequency. 1 , 23 , 24 Notably, seizure incidence peaked during the light–dark transition, mirroring the clinical association between atypical absence seizures and sleep transition in human LGS. 14 Initial characterization of this model by others demonstrated variable ASM responsiveness, with ethosuximide and clobazam, but not topiramate, reducing atypical absence seizures. 12 We replicated and extended these findings by demonstrating that ethosuximide, ulixacaltamide, and CBD significantly reduced the burden of atypical absence seizures. Ethosuximide is a first‐line ASM for treating absence seizures, but is used adjunctively in LGS based on expert consensus rather than clinical trial data. 15 , 25 While its primary mechanism involves Cav3 inhibition, it also modulates Cav1, sodium, and potassium conductances. 26 , 27 , 28 To isolate the effects of Cav3 inhibition in this model, we evaluated ulixacaltamide, a potent and selective Cav3 inhibitor in clinical development. Ulixacaltamide produced robust reductions in atypical absence seizures, consistent with reported efficacy in Genetic Absence Epilepsy Rats from Strasbourg (GAERS). 16 These findings support Cav3 channels as rational therapeutic targets for atypical absence seizures. Absence seizures are driven by aberrant Cav3 activity within corticothalamic and thalamic reticular neurons, with selective Cav3 blockade suppressing seizures in GAERS. 29 However, converging evidence suggests that atypical absence seizures involve neural circuitry beyond the thalamocortical network, including limbic and hippocampal regions. 23 , 30 This broader network involvement likely contributes to pharmacoresistance and may explain the partial efficacy of ASMs in this model. We report the first evaluation of CBD in the Gabrb3 +/D120N model. Approved as an adjunctive treatment for LGS based on reductions in drop seizures, clinical data suggest that CBD is also effective against atypical absence seizures. 31 , 32 Our results align with these data and complement the earlier finding that a synthetic cannabinoid, WIN 55,212–2, reduces seizures in this model. 12 Mechanistically, CBD lacks the CB1 receptor agonism of WIN 55,212–2, 33 instead likely acting via transient receptor potential (TRP) channel desensitization, G protein‐coupled receptor 55 (GPR55) antagonism, and GABAAR modulation. 34 , 35 , 36 Notably, CBD also inhibits Cav3 in vitro, suggesting a mechanistic convergence with ethosuximide and ulixacaltamide, though further research is required to determine its precise mechanism of action in the Gabrb3 +/D120N mouse model. 37 It must be acknowledged that we observed a noticeable increase in a typical absence seizure bouts and duration in vehicle‐treated mice in our pharmacological studies compared to untreated animals used in the initial characterization of the model. This discrepancy was consistent across independent experimental cohorts, suggesting that procedural stress associated with intraperitoneal injections may precipitate seizures in Gabrb3 +/D120N mice. Exposure to injection‐related stressors triggers the release of glucocorticoids, which increase neuronal excitability, lower seizure thresholds, and aggravate seizures in various rodent models. 38 , 39 Future studies may mitigate the effects of injection stress by habituating animals to sham injections before experiments or by utilizing alternative routes of administration. Infantile epileptic spasms are age‐dependent motor seizures characterized by brief (~1 s) flexion and/or extension of axial and limb muscles occurring in clusters, with onset occurring between 3 and 12 months of age. 6 , 40 An estimated 20–30% of LGS patients have a prior diagnosis of epileptic spasms, a trajectory associated with poor prognosis. 4 , 5 , 6 The pathological Gabrb3 variant modeled here was identified in a patient with a disease progression from infantile spasms to LGS. 10 We observed spontaneous spasms in mutant mice during a restricted neonatal window (P13–17), mirroring the clinical progression to multiple seizure types in a significant subset of DEE patients. 1 This model appears to recapitulate this disease progression, providing a promising opportunity to investigate pathological mechanisms and develop targeted therapies for this highly treatment‐resistant population. While other monogenic DEE models, such as the Gabrb3 +/N110D mouse, also exhibit spasms evolving into mixed seizures, atypical absence seizure burden is significantly higher in the present model (~445 events/24 h vs. <10 events/24 h in N110D mice), underscoring its utility for preclinical investigation. 12 , 41 We provide the first pharmacological validation of epileptic spasms in this model, demonstrating that both vigabatrin and CBD significantly reduce spasm burden. Vigabatrin is an irreversible GABA transaminase inhibitor used as a first‐line IESS therapy alongside the corticosteroid therapies, ACTH, and prednisone. 17 , 42 The partial efficacy of vigabatrin against spasms in Gabrb3 +/D120N mice aligns with the clinical observation that only approximately one‐third of IESS patients achieve long‐term remission on first‐line therapies due to high rates of pharmacoresistance and relapse. 21 , 43 Current treatments are further constrained by adverse effects, including cardiac and metabolic risks of corticosteroid therapies and vigabatrin‐induced irreversible retinal toxicity. 44 , 45 Our finding that CBD reduces infantile spasms in Gabrb3 +/D120N mice, coupled with its excellent safety profile, reinforces its potential as an IESS treatment. There is increasing interest in the use of CBD for epileptic spasms, with caregiver reports indicating high perceived efficacy and safety in IESS and LGS cohorts. 46 However, the limited clinical evidence remains mixed. Data from the CBD Expanded Access Program showed spasm reductions of 85–90%, while open‐label studies reported spasm reductions in 60–68% of treatment‐refractory patients, with 25–33% achieving spasm freedom. 47 , 48 , 49 However, a recent Phase II trial of CBD in a refractory IESS cohort reported an 89% non‐responder rate. 50 This clinical variability is reflected in the bimodal response of Gabrb3 +/D120N mice to CBD, with 54% of animals exhibiting near‐complete spasm suppression and comparable spasm incidence to wild‐type littermates, while 46% were non‐responsive. This variance in responders and non‐responders positions the Gabrb3 +/D120N mouse as a potential tool for investigating the neuromolecular determinants of CBD sensitivity. As in human LGS, no agent was completely efficacious in suppressing seizure activity. The partial efficacy of vigabatrin and CBD aligns with the clinical reality of DEEs, supporting the predictive validity of this model and establishing a baseline for the evaluation of novel compounds. Several limitations of this study warrant consideration. Principally, although multiple seizure types characterize LGS, we prioritized atypical absence seizures as the primary experimental endpoint due to their high frequency in this model and their well‐defined electroclinical signature. In clinical practice, however, atypical absence seizures are rarely treated in isolation, with broad‐spectrum reductions in total seizure burden the goal of management. 1 Tonic seizures, in particular, are considered a hallmark of LGS, yet they remain difficult to study in preclinical models. In Gabrb3 +/D120N mice, tonic seizures occur less frequently and are difficult to score accurately as they lack distinct electrographic characteristics. 12 , 24 To investigate broad‐spectrum antiseizure effects, future research may incorporate multi‐day vEEG recordings to capture less frequent seizure events. Additionally, due to the technical challenges of electrode implantation in neonatal mice, spasms were validated behaviourally without EEG confirmation. While our behavioral scoring aligns with prior preclinical research, in clinical practice, vEEG is essential to confirm spasms and identify hypsarrhythmia, an interictal pattern that contributes to neurocognitive decline. 21 Future research may explore emerging non‐invasive EEG technologies, such as NeuroGrid arrays, to determine if vigabatrin or CBD can resolve hypsarrhythmia in this model, as suggested by clinical reports. 48 , 51 Notably, we observed infrequent spasms in a subset of wild‐type mice. This observation aligns with previous reports of infrequent spasm‐like events in various mouse strains. 12 , 41 , 52 We speculate that such events may represent physiological motor activity akin to benign myoclonus observed in human infants. 53 Reliance on behavioral scoring introduces the possibility that reductions in spasm frequency could be confounded by drug‐induced sedation. To mitigate acute behavioral effects, drugs were injected after recording with a 24 h washout period. While CBD is non‐sedating at 100 mg/kg, we note that prolonged vigabatrin administration has been anecdotally associated with mild sedation in neonatal mice. 41 , 54 , 55 Although gross changes in exploratory behavior were not visually apparent in our blinded analyses, future research may incorporate quantitative locomotor assays to definitively rule out potential confounding effects of sedation. Finally, LGS has heterogeneous etiologies with distinct structural and genetic contributions, and frequently emerges as an evolving electroclinical phenotype within the broader epilepsy spectrum of several DEEs. 1 While the present model reflects a high‐impact GABRB3 variant, it represents only one phenotype within the broader DEE landscape. GABRB3 mutations cause a spectrum of monogenic DEEs, including LGS, IESS, Childhood Absence Epilepsy, and Angelman syndrome. 56 , 57 At present, mouse models of GABRB3 mutations provide the best available animal models for understanding pathophysiological mechanisms and developing new therapies. 9 , 12 , 24 However, additional environmental and genetic animal models are likely to be forthcoming as the field develops a better understanding of LGS etiology. Even within the sphere of genetic factors, which account for ~30–40% of LGS cases, a growing list of genes linked to LGS could be exploited to develop new models. 58 For example, variants in KCNB1, which encodes the inwardly rectifying potassium channel Kv2.1, have been identified in patients with drug‐resistant epilepsy. 59 Kcnb1 R306C/+ mice display an LGS‐like phenotype, with prolonged SSW discharges on EEG and infrequent spontaneous motor seizures. 59 The development of new animal models of LGS with distinct mechanisms, such as the Kcnb1 R306C/+ mouse model, will certainly help improve the translation of new therapeutic candidates for LGS. The ability of the Gabrb3 +/D120N model to recapitulate a developmental trajectory from IESS to LGS allows for longitudinal studies to assess whether successful treatment of neonatal spasms can prevent the emergence of LGS‐associated seizures and cognitive‐behavioral deficits. Given that Gabrb3 +/D120N mice display impaired learning and social interaction, future research may incorporate maze‐based memory tests and behavioral assays as endpoints for evaluating disease‐modifying effects of early interventions, such as CBD, which has a good safety profile in humans. 12 , 60 , 61 Conclusion The Gabrb3 +/D120N mouse model of LGS demonstrates robust, drug‐responsive atypical absence seizures and neonatal spasms that recapitulate the developmental trajectory of human LGS. The present findings support the model's predictive validity and its utility as a tool for identifying urgently needed therapies for DEEs. We report for the first time that the clinically approved cannabinoid ASM CBD can reduce both atypical absence seizures and infantile spasms in the Gabrb3 +/D120N mouse, providing a benchmark for future evaluation of cannabinoid therapies. AUTHOR CONTRIBUTIONS Jonathon C. Arnold and Thomas Harman conceived the current project. Thomas Harman contributed to data acquisition for animal experiments, data analysis and representation, and the drafting of the manuscript. Ka Lai Yip and Laura A. Sharman contributed to data acquisition and EEG analysis. Miguel Bedoya‐Pérez contributed data analysis of animal experiments. Jonathon C. Arnold, Thomas Harman, and Vaishali Janve contributed to experimental design. All authors reviewed the manuscript. Jonathon C. Arnold and Thomas Harman contributed to the final draft and manuscript submission. FUNDING INFORMATION This study was supported by the Lambert Initiative for Cannabinoid Therapeutics, a philanthropically funded center for medicinal cannabis research at the University of Sydney, and the Australian National Health and Medical Research Council. CONFLICT OF INTEREST STATEMENT J.C.A. has served as an expert witness in various medicolegal cases involving cannabis and has received consulting fees from the World Health Organization, Medical Cannabis Industry Australia, and Haleon (consumer health care subsidiary of GSK). He reports research grants from the Australian National Health and Medical Research Council and from the Lambert Initiative for Cannabinoid Therapeutics. J.C.A. holds a patent on novel cannabinoid therapies for epilepsy (WO2019071302). The remaining authors have no conflicts of interest. We confirm that we have read the Journal's position on ethical publication issues and affirm that this report is consistent with those guidelines. ETHICS STATEMENT We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.