Utilizing an acute hyperthermia-induced seizure test and pharmacokinetic studies to establish optimal dosing regimens in a mouse model of Dravet syndrome

Abstract

Objective:

The current standard of care for Dravet syndrome (DS) includes polytherapy after inadequate seizure control with one or more monotherapy approaches. Treatment guidelines are often based on expert opinions, and finding an optimal balance between seizure control and adverse drug effects can be challenging. This study utilizes the efficacy and pharmacokinetic assessment of a second-line treatment regimen that combines clobazam and sodium valproate with an add-on drug as a proof-of-principle approach to establish an effective therapeutic regimen in a DS mouse model.

Methods:

We evaluated the efficacy of add-on therapies stiripentol, cannabidiol, lorcaserin, or fenfluramine to clobazam and sodium valproate against hyperthermia-induced seizures in Scn1a^A1783V/WT mice. Clobazam, N-desmethyl clobazam (an active metabolite of clobazam), sodium valproate, stiripentol, and cannabidiol concentrations were quantified in plasma and brain using liquid chromatography-tandem mass spectrometry for the combinations deemed effective against hyperthermia-induced seizures. The concentration data were used to calculate pharmacokinetic parameters via non-compartmental analysis in Phoenix WinNonLin.

Results:

Higher doses of stiripentol or cannabidiol, in combination with clobazam and sodium valproate, were effective against hyperthermia-induced seizures in Scn1a^A1783V/WT mice. In Scn1a^WT/WT mice, brain clobazam and N-desmethyl clobazam concentrations were higher in the triple-drug combinations than in the clobazam monotherapy. Stiripentol and cannabidiol brain concentrations were greater in the triple-drug therapy than when given alone.

Significance:

A polypharmacy strategy may be a practical preclinical approach to identifying efficacious compounds for DS. The drug-drug interactions between compounds used in this study may explain the potentiated efficacy of some polytherapies.

Graphical Abstract

Introduction

Dravet syndrome (DS) is a rare but catastrophic infant-onset genetic epilepsy.^1,2 Over 900 distinct mutations in the Scn1a gene have been identified.^3,4 DS develops in the first year of life in approximately 1 per 30,000 live births.⁵ Seizure onset is typically induced by fever, which worsens and manifests as spontaneous recurrent generalized tonic-clonic seizures, sometimes resulting in status epilepticus. DS also includes developmental delays, cognitive impairment, and an increased risk for sudden unexpected death in epilepsy.

Antiseizure medications (ASMs) recommended as first- and second-line treatment regimens for DS patients have predominantly relied on expert opinions and an insufficient number of retrospective, limited sample-sized, open-label studies.⁶ Clinical response to conventional ASMs has been disappointing, with many DS patients remaining pharmacoresistant to available treatment options.^7,8 The recommended treatment algorithm includes sodium valproate (VPA) and clobazam (CLB) as first-line options. In typical clinical practice, the failure of one first-line drug to provide optimal seizure control leads to adding the other agent. Most pediatric DS patients use a second-line drug concomitantly with VPA and CLB and as such, most clinical trials of adjunctive therapies are given in combination with these drugs.⁹ Several epileptologists have suggested stiripentol (STP) and cannabidiol (CBD) as the best second-line agents to be used in combination with VPA and CLB, with fenfluramine (FFA) emerging as a promising option.^10–12

Despite the advances in the search for effective therapies for DS, including recommended combinatorial strategies, it is difficult to find an equilibrium between seizure control, adverse drug effects, and the burden of polypharmacy in patients with DS.^13,14 The choice of the drug combination and selected tolerable doses frequently rely on the suggested treatment regimen and experience in the clinical setting.¹⁵ Establishing an optimal dosing regimen for DS requires understanding the relationships between dose, exposure, and effect (i.e., the pharmacokinetics (PK) and pharmacodynamics (PD)) for commonly used ASMs alone and in combination with other ASMs.¹⁶ Preclinical studies are valuable for comparing PK with efficacy and tolerability and can inform clinical approaches. Thus, integrating PK information in an appropriate preclinical model can improve the development of polypharmacy strategies in the clinical setting.¹

We have previously described a mouse model of DS that offers advantages for screening novel compounds using a polypharmacy approach.¹⁷ The present study uses this previously described mouse model to determine PK parameters and test the efficacy of clinically relevant second-line drug combination therapies. These data increase our understanding of PK and PD drug-drug interactions (DDI) of add-on combination therapies in the treatment of DS and generate essential data for optimizing the sub-chronic or chronic dosing regimens in a mouse model of DS before embarking on costly clinical trials for investigational compounds that will be ultimately tested in children who are already taking combinations of ASMs. Additionally, the combined use of efficacy testing and PK monitoring sets the stage for the discovery of future investigational compounds that might be efficacious in this difficult-to-treat pediatric epileptic encephalopathy.

Materials and methods

Animals

The Institutional Animal Care and Use Committee of the University of Utah approved all animal care and experimental procedures. Animal experiments were conducted per Animal Research: Reporting of In Vivo Experiments guidelines (https://www.nc3rs.org.uk/arrive-guidelines). We generated experimental animals by breeding a floxed stop male Scn1a^A1783V (B6(Cg)-Scn1atm1.1Dsf/J, Jax #026133) with a Sox2-cre (B6. Cg-Edil3 ^{Tg(Sox2−cre)1Amc}/J) female mouse to produce both heterozygous (Scn1a^A1783V/WT) and wild-type offspring. Both female and male heterozygous and age-matched wild-type littermates were used for experiments. Mice were group‐housed in a specific pathogen‐free mouse facility under standard laboratory conditions (14‐h light/10‐h dark) and had access to food and water ad libitum, except during hyperthermia-induced seizure experiments and were transferred to the experimental room approximately 1 hour before testing. A detailed list of common data elements was recorded, and detailed case report forms were utilized to confirm all aspects of each study and are available upon request.

Hyperthermia-induced seizure test

Hyperthermia-induced seizure experiments were conducted in Scn1a^A1783V/WT mice (P28-P42) to evaluate the temperature at which the mice seized. A rectal probe was inserted 10 min before the time-to-peak effect (TPE) of the test drug(s) and hyperthermia-induced seizure test. Before the test started, mice acclimated to the temperature probe for 5 min in a glass chamber. Mice were placed under a heat lamp, and the core temperature was gradually elevated by 1°C every 2 min in the chamber until a generalized seizure was observed or the temperature reached 42.5°C.^18–20 If no seizure occurred, the mouse was considered seizure‐free. A neonatal mouse/rat rectal probe (Braintree Scientific, Inc, Braintree, MA) coupled to a TCAT-2LV controller (Physitemp Instruments, Inc) was used to monitor the core body temperature. Male and female Scn1a^A1783V/WT mice were randomly assigned to treatment groups before the test, and observers were blinded to the treatment. A mouse was removed from the study if it had a behavioral seizure between drug administration and testing. The temperature at which mice seized was recorded.

Materials

CLB and VPA were purchased as reference powders from Sigma-Aldrich. Lorcaserin (LCS) and CBD were purchased from Cayman Chemical (Ann Arbor, MI). FFA was obtained from Axon Medchem (Groningen, Netherlands). STP was purchased from Toronto Research Chemicals (Toronto, Canada).

Drug administration

All administered drugs were prepared as 0.5% methylcellulose (Sigma) suspensions. All drugs [CLB (5 mg/kg), VPA (75 mg/kg), STP (30 mg/kg, 70 mg/kg, 100 mg/kg, 130 mg/kg), CBD (70 mg/kg, 100 mg/kg, 135 mg/kg, 150 mg/kg), FFA (5 mg/kg, 10 mg/kg, 20 mg/kg, 30 mg/kg, LCS (5 mg/kg, 7 mg/kg, 10 mg/kg, 15 mg/kg], were administered at 0.01 mL/g volume and tested based on their time of peak effect (TPE) as observed in other models of epilepsy (TPE_{CLB, STP, CBD} = 1 hr, TPE_{FFA, LCS} = 0.5 hr, TPE_VPA = 0.25 hr) (Table S1).^17,21–22 Each drug was intraperitoneally administered separately before the hyperthermia-induced seizure test. All mice used in the study were monitored for sedation via physical observation in all treatment groups to assess drug safety. Additionally, mice treated with CBD were assessed for hypothermia, a commonly described effect of CBD, via rectal probe, as such an effect could obviate hyperthermia-induced seizures.

Sample collection

Blood samples were collected from age-matched Scn1a^WT/WT and Scn1a^A1783V/WT mice (P28-P42) at various time points: 0.08, 0.17, 0.25, 0.5, 1, 2, 4, and 8-hr post-dose. Mice used in these studies were not subjected to hyperthermia studies. Blood was collected via decapitation into tubes with K₂EDTA (Becton, Dickson and Company, NJ) as an anticoagulant. Blood was centrifuged (3500xg for 10 min at 4°C) to isolate plasma. Brains were rapidly removed following decapitation and weighed. Whole brains were homogenized in ultrapure water using a probe sonicator to obtain a target brain homogenate concentration of 200 mg/mL. Plasma and brain homogenate samples were stored at −80°C before analysis.

Bioanalysis

CLB, N-CLB, and VPA

CLB was diluted to 1.0 mg/mL in 50:50 methanol:water (MeOH:H₂O), while VPA was diluted to 10 mg/mL in 50:50 MeOH:H₂O. NCLB and VPA-d₆ were purchased as 1.0 mg/mL stock solutions in MeOH, while CLB-¹³C₆ and NCLB-¹³C₆ were purchased as 0.1 mg/mL stock solutions in acetonitrile, all from Sigma Aldrich. All mouse plasma samples were diluted 20-fold before extraction, with human plasma with EDTA anticoagulant used as a matrix for calibrators and quality control (QC) samples. Mouse brain homogenate was analyzed without dilution, using mouse brain homogenate as the control matrix for calibrators and QC samples.

A 100 μL aliquot of the 20-fold diluted plasma or undiluted brain homogenate was assayed using a protein precipitation approach. A 50 μL volume of internal standard (400 pg/μL VPA-d₆, 40 pg/μL CLB-¹³C_6, and NCLB-¹³C₆ in 10:90 MeOH:H₂O) was added to each sample aliquot, followed by 1 mL of cold acetonitrile. The samples were vortex-mixed and centrifuged to separate the supernatant from the protein pellet. The supernatant was then transferred to a fresh microcentrifuge tube and centrifuged again to ensure the removal of all solid particulates from the sample. The supernatant was transferred into a polypropylene tube and dried under house air (15 psi) in a TurboVap set (40°C). Samples were then reconstituted with 200 μL of 10:90 MeOH:H₂O, centrifuged, vortex mixed, and transferred to autosampler vials for analysis.

A 10 μL sample volume was injected for analysis on a Waters Acquity UPLC with a Quattro Premier XE UHPLC-MS/MS system. Chromatography was performed using a Phenomenex (Torrance, CA) Luna Omega Polar C18, 1.6 μm (2.1 × 50 mm), and gradient elution was maintained at a 300 μL/min flow rate. Mobile phases consisted of (A) 10mM ammonium acetate (pH 5.5) and (B) methanol. Analytes were monitored using the following mass transitions (collision energy, CE): 287.1→245.1 (NCLB, CE=20V); 293.1→251.1 (NCLB-¹³C₆, CE=20V); 301.1→259.1 (CLB, CE=20V); 307.1→265.1 (CLB-¹³C₆, CE=20V); 143.1→143.1 (VPA, CE=5V); and 149.1→149.1 (VPA-d₆, CE=5V). CLB, NCLB, and their internal standards (IS) utilized positive electrospray ionization (ESI), whereas VPA and its internal standard (IS) used negative ESI. The calibration curves utilized 1/x² weighted linear regression between 4–200 ng/mL for CLB and NCLB and 0.2–10 μg/mL for VPA.

STP

STP and STP-d₉ were purchased as reference powders from Toronto Research Chemicals (Toronto, Canada). STP was diluted to 1.0 mg/mL in 90:10 MeOH:H₂O. Mouse plasma samples were diluted 100-fold into 50:50 MeOH:H₂O before analysis. A 200 μL aliquot of the 100-fold diluted sample was analyzed with calibrators, and QC was prepared in 1% mouse plasma in 50:50 MeOH:H₂O. Mouse brain homogenate was diluted 10-fold into 50:50 MeOH:H₂O before analysis, and a 200 μL aliquot was assayed against calibrators and QCs prepared in 10% mouse brain homogenate in 50:50 MeOH:H₂O.

STP was extracted using a basified liquid-liquid extraction procedure. A 50 μL volume of 40 pg/μL STP-d₉ in 50:50 MeOH:H₂O was added to each sample. The sample pH was then adjusted with 0.5 mL of 5% (v:v) ammonium hydroxide in type-1 water and extracted with 2.0 mL of methyl tert-butyl ether (MTBE). The sample was mixed well and centrifuged to separate the organic and aqueous layers. The sample was frozen at −80°C for 30 min, and the organic layer was decanted into a fresh tube. The organic layer was dried under house air (15 psi) in a TurboVap set (40°C). The sample was then reconstituted with 200 μL of 50:50 MeOH:H₂O and transferred to an autosampler vial for analysis.

A 20 μL sample volume was injected onto a Waters Acquity UPLC and autosampler coupled with a ThermoScientific TSQ Quantum Access MS/MS. Chromatography utilized a Phenomenex Luna Omega PS C18 3μm (2.1 × 100mm) column, and gradient elution was maintained at a 300 μL/min flow rate. Mobile phases comprised (A) 0.1% formic acid in type-1 water and (B) methanol. STP and its IS were ionized using positive ESI, and the mass transitions (CE) 217.2→187.2 (11V) and 226.2→196.2 (11V) were monitored for STP and STP-d₉, respectively. The assay’s dynamic range was 1–1000 ng/mL, fitting with a 1/x weighted linear regression.

CBD and 7-OH-CBD

CBD and 7-OH-CBD were purchased as 1.0 mg/mL stock solutions in MeOH from Sigma-Aldrich, while CBD-d₃ and 7-OH-CBD-d₃ were purchased as 0.1 mg/mL stock solutions in MeOH. Plasma and brain homogenate were diluted before analysis 10-fold into human plasma with EDTA anticoagulant. Calibrators and QC samples were also prepared in human plasma. A 200 μL aliquot of the 10-fold diluted sample was analyzed following a method previously validated by the Center for Human Toxicology.

Samples were extracted using a liquid-liquid extraction under acidic conditions in polypropylene tubes. After addition of 50 μL of 20 pg/μL CBD-d₃ and 7-OH-CBD-d₃ in 50:50 MeOH: H₂O to each sample, the sample was acidified using 0.5 mL of 2 mM ammonium formate (pH 3.5) and extracted using 2.0 mL of 5:1 hexane:MTBE. The sample was mixed and centrifuged. The sample was frozen at −80°C for 30 min, and the organic layer was decanted into a fresh tube. The organic layer was dried under house air (15 psi) in a TurboVap set (40°C). The sample was then reconstituted with 200 μL of 50:50 MeOH:H₂O and transferred to an autosampler vial for analysis.

A 30 μL sample volume was injected onto a ThermoScientific Accela LC pump, and an autosampler interfaced with a ThermoScientific TSQ Vantage MS/MS. Chromatography used a Phenomenex Luna Omega PS C18, 3 μm (2.1 × 100 mm) column, and gradient elution maintained at a 250 μL/min flow rate. Mobile phases consisted of (A) 1 mM ammonium formate (pH 3.5) and (B) methanol. Positive ESI was used for CBD, 7OH-CBD, and their IS, monitoring the mass transitions (CE): 315.2→193.2 (20V, CBD), 318.2→196.2 (20V, CBD-d₃), 331.2→313.2 (10V, 7OH-CBD), and 334.2→316.2 (10V, 7OH-CBD-d₃). Calibration curves for both analytes are linearly regressed with 1/x² weighting between the concentrations of 0.5–100 ng/mL.

PK analysis

Before PK analyses, concentrations were averaged at each time point for each combination of drug regimen, tissue, and sample time. Concentrations below the assay limit of quantitation were excluded from the analysis. PK analyses were conducted in Phoenix WinNonLin v8.3, using the linear trapezoidal linear interpolation option calculation method in the non-compartmental analysis module. The software was allowed to determine the best-fit line for the elimination slope for all calculations. The parameters of interest for this analysis were the area under the concentration-time curve from zero to infinity (AUC_0-inf), maximum concentration (C_max), and half-life (t_1/2) for CLB, N-CLB, STP, CBD, and 7-OH-CBD.

Statistical analysis

Statistical analyses were conducted using a Log-rank (Mantel-Cox) test for the hyperthermia-induced test. A power analysis used α=0.05 and the hazard ratio=0.20. Significance was defined as a p<0.05. All analyses were conducted with GraphPad Prism 9.0.

Results

Higher doses of STP add-on to CLB+VPA were effective against hyperthermia-induced seizures in Scn1a^A1783V/WT mice.

We previously demonstrated that the combination of CLB (5 mg/kg) + VPA (75 mg/kg) + STP (100 mg/kg) was protective against hyperthermia-induced seizures; however, STP (100 mg/kg) alone, CLB (5 mg/kg) + VPA (75 mg/kg) without STP, and CLB (5 mg/kg) + VPA (75 mg/kg) + STP (30 mg/kg) were found to be ineffective.¹⁷ To better understand the STP dose required for seizure protection in the setting of triple-drug therapy, the current work evaluated hyperthermia-induced seizure protection observed from escalating STP doses (30, 70, 100, and 130 mg/kg) used in combination with CLB (5 mg/kg) and VPA (75 mg/kg) in our Scn1a^A1783V/WT mouse model of DS. Scn1a^A1783V/WT mice treated with 100 mg/kg (*p=0.047) and 130 mg/kg STP (**p=0.004) in addition to CLB (5 mg/kg) + VPA (75 mg/kg) had significantly higher temperature thresholds for seizure onset in comparison to animals treated with CLB (5 mg/kg) + VPA (75 mg/kg) alone (Figure 1). In contrast, lower doses of STP (30 mg/kg and 70 mg/kg) added to CLB (5 mg/kg) + VPA (75 mg/kg) did not significantly change temperature thresholds compared to CLB (5 mg/kg) + VPA (75 mg/kg). Seizure protection profiles of control groups from each study (CLB (5 mg/kg) + VPA (75 mg/kg) showed no statistically significant differences (Figure S1). Signs of sedation were observed in only the CLB+VPA+STP (130 mg/kg) mice post-treatment, suggesting potential behavioral impairment. Combined, these results indicate that of the evaluated STP doses, adding 100mg/kg STP to CLB (5 mg/kg) + VPA (75 mg/kg) is required for seizure protection (as measured by an increase in seizure temperature threshold) without inducing sedation.

Figure 1:

Seizure protection profile of add-on therapy of STP to CLB and VPA in Scn1a^A1783V/WT mice. (A) STP (30 mg/kg) + CLB (5 mg/kg) + VPA (75 mg/kg) did not significantly affect the temperature at which mice seized (CLB + VPA + STP: n = 9, CLB + VPA: n = 9 p = .7973) (B) STP (70 mg/kg) + CLB (5 mg/kg) + VPA (75 mg/kg) failed to over any significant protection against hyperthermia-induced seizures in mice (CLB + VPA + STP: n = 10, CLB + VPA: n = 10 p = .7743) (C) Add-on STP (100 mg/kg) to CLB (5 mg/kg) + VPA (75 mg/kg) significantly increased the median seizing temperature threshold at which mice seize (CLB + VPA + STP: n = 10, CLB + VPA: n = 10) (D) Add-on STP (130 mg/kg) to CLB (5 mg/kg) + VPA (75 mg/kg) significantly increased the median seizing temperature threshold at which mice seize (CLB + VPA + STP: n = 10, CLB + VPA: n = 10). *p < .05, **p < 0.005; Log-rank (Mantel-Cox). Figures 1A and 1C were modified and published with permission from Pernici et al.¹⁷

Plasma and brain PK profiles demonstrate potential drug-drug interactions between STP, CLB, and VPA.

After identifying the optimal STP dose (100 mg/kg) in the setting of triple-drug therapy with CLB (5 mg/kg) + VPA (75 mg/kg), we then sought to compare and contrast the PK of all three compounds when administered alone and in the setting of multi-drug therapy. Exposures (As determined by AUC_o-inf of CLB, and to a lesser extent its active metabolite N-CLB, were substantially higher within the triple-drug therapy arm when compared to treatment with either CLB alone (5 mg/kg) or CLB (5 mg/kg) + VPA (75 mg/kg) (Figures 2 and S2). CLB C_max in the brain was significantly higher in the triple-drug therapy cohort compared to both the CLB alone (p<0.0001) and CLB + VPA (p=0.0003) cohorts. AUC_0-inf of VPA in the brain was ~2-fold higher in both the CLB + VPA and CLB + VPA + STP cohorts relative to the VPA alone cohort (Table 1). However, sampling limitations prevented statistical testing of this observation. Finally, STP AUC_0-inf in the brain was approximately 3-fold higher in the CLB + VPA + STP cohort compared to treatment with STP alone; however, C_max values were not different between these two groups (p=0.64). CLB stand-alone PK data in Scn1a^A1783V/WT mice mirrored CLB PK data in Scn1a^WT/WT, demonstrating a comparable PK profile between Scn1a^A1783V/WT and Scn1a^WT/WT mice (Table S2). In summary, the PK of CLB, its active metabolite N-CLB, VPA, and STP differ between mono- and multi-drug therapy, supporting a critical need for evaluating PK in both settings.

Figure 2:

Concentration–time curves of (A) CLB (5 mg/kg), (B) N-CLB, (C) VPA (75 mg/kg), and (D) STP (100 mg/kg) in the brain of Scn1a^WT/WT mice following intraperitoneal administration. Brain samples were obtained 0–8 hr post-dose. Each data point is the mean brain concentration value for four mice. Missing data points before the 8-hr time in each curve denote that the analyte of interest was found to be below the limit of quantitation for the assay.

Table 1:

PK parameters following intraperitoneal administration of CLB, VPA, STP, and CBD in plasma and brain of Scn1a^WT/WT mice

Treatment type	Analytes	Plasma			Brain			Brain/Plasma ratio
		PK Parameters
		AUC0-inf (μg*hr/mL)	Cmax (μg/mL)	t1/2 (hr)	AUC0-inf (μg*hr/g)	Cmax (μg/g)	t1/2 (hr)	AUC0-inf	Cmax
Single Drug	CLB	0.648	0.418	1.86	0.749	0.446	2.46	1.15	1.07
	N-CLB	5.53	0.77	5.29	11.6	0.966	10.4	2.1	1.25
	VPA	129	155	0.397	12.3	23.6	0.28	0.1	0.15
	STP	53.3	34.3	0.868	5.56	4.39	0.863	0.1	0.13
	CBD	4.21	1.39	2.25	5.43	1.42	2.74	1.29	1.02
	7-OH-CBD	3.3	0.67	3.32	2.71	0.62	3.25	0.82	0.93
Double (CLB+VPA)	CLB	0.462	0.54	1.53	0.637	0.76	0.904	1.38	1.41
	N-CLB	3.61	0.567	2.61	5.13	0.788	3.3	1.42	1.39
	VPA	224	243	0.623	25.1	42.5	0.469	0.11	0.17
Triple (CLB+VPA+STP)	CLB	5.64	1.56	1.51	7.69	2.12	1.34	1.36	1.36
	N-CLB	7.13	0.88	3.83	16.3	1.22	11.2	2.29	1.14
	VPA	209	166	0.67	27.1	35.5	0.274	0.13	0.21
	STP	103	30.7	1.51	15.1	5.34	1.24	0.14	0.17
Triple (CLB+VPA+CBD)	CLB	2.05	0.579	2.11	2.87	0.858	1.69	1.4	1.48
	N-CLB	14.8	0.76	13	23.2	0.91	17.1	1.56	1.2
	VPA	295	233	1.55	36.1	42.1	1.19	0.12	0.18
	CBD	6.29	1.24	2.78	12.9	1.33	5.84	2.05	1.07
	7-OH-CBD	5.86	0.72	4.28	5.97	0.622	5.99	1.02	0.86

Abbreviations: AUC, area under the concentration–time curve; CLB, clobazam; C_max, maximum concentration; CBD, cannabidiol; n/a, not available; N-CLB, N-desmethylclobazam; PK, pharmacokinetic; 7-OH-CBD, 7-hydroxycannabidiol; STP, stiripentol; t_1/2, half-life; VPA, sodium valproate. Note: Brain/plasma AUC and C_max ratio assumes brain density is equal to the density of water (1 g/ml).⁵⁵

Higher doses of CBD add-on to CLB and VPA were effective against hyperthermia-induced seizures in Scn1a^A1783V/WT mice.

Previous experiments in our lab demonstrated that CBD (135 mg/kg) alone and CLB (5 mg/kg) + VPA (75 mg/kg) without CBD failed to increase temperature thresholds in Scn1a^A1783V/WT mice.¹⁷ To determine the CBD dose required for seizure protection in the setting of triple-drug therapy, we assessed the hyperthermia-induced seizure protection observed from escalating CBD doses (70, 100, 135, and 150 mg/kg) used in combination with CLB (5 mg/kg) and VPA (75 mg/kg) in our Scn1a^A1783V/WT mice. Scn1a^A1783V/WT mice treated with 135 mg/kg (*p=0.027) and 150 mg/kg CBD (*p=0.01) in addition to CLB (5 mg/kg) + VPA (75 mg/kg) had significantly higher temperature thresholds for seizure onset in comparison to animals treated with CLB (5 mg/kg) + VPA (75 mg/kg) alone (Figure 3). In contrast, lower doses of CBD (70 mg/kg and 100 mg/kg) added to CLB (5 mg/kg) + VPA (75mg/kg) did not significantly change temperature thresholds compared to CLB (5 mg/kg) + VPA (75 mg/kg). Seizure protection profiles of control groups from each study (CLB (5 mg/kg) + VPA (75 mg/kg) showed no statistically significant differences (Figure S3). All mice treated with CBD experienced a reduction in post-injection core body temperatures compared to baseline pre-injection core body temperature readings, potentially influencing the observed temperature threshold associated with seizures. No signs of sedation were observed in all mice treated with CBD.

Figure 3:

Antiseizure efficacy of CBD add-on to CLB and VPA in Scn1a^A1783V/WT mice. (A) CBD (70 mg/kg) + CLB (5 mg/kg) + VPA (75 mg/kg) did not significantly protect mice against hyperthermia-induced seizures (CLB + VPA + CBD: n = 10, CLB + VPA: n = 10 p = .7973) (B) CBD (100 mg/kg) + CLB (5 mg/kg) + VPA (75 mg/kg) failed to over any significant protection against hyperthermia-induced seizures in mice (CLB + VPA + CBD: n = 10, CLB + VPA: n = 10 p = .8746) (C) Add-on CBD (135 mg/kg) to CLB (5 mg/kg) + VPA (75 mg/kg) offered protection against hyperthermia-induced seizures in mice seize (CLB + VPA + CBD: n = 9, CLB + VPA: n = 10) (D) Add-on CBD (150 mg/kg) to CLB (5 mg/kg) + VPA (75 mg/kg) offered protection against hyperthermia-induced seizures in mice seize (CLB + VPA + CBD: n = 10, CLB + VPA: n = 10) *p < .05; Log-rank (Mantel-Cox).

Plasma and brain PK profiles demonstrate potential DDIs between CBD, CLB, and VPA.

Treatment with CLB (5 mg/kg) + VPA (75 mg/kg) + CBD (135 mg/kg) was effective against hyperthermia-induced seizures in Scn1a^A1783V/WT mice. The PK parameters of CBD treatment with CLB, N-CLB, and VPA in the brains and plasma of mice are displayed in Figures 4 and S4. CBD exposures in the triple-drug therapy were increased when compared to mice treated with CBD alone. Similarly, CLB exposures were increased in mice co-treated with CBD and VPA compared to those treated with CLB-only. Total N-CLB exposures in triple-drug-treated mice were observed to increase compared to N-CLB exposures in CLB+VPA or CLB monotherapy-treated mice. CBD administration resulted in an increase in the brain exposure of VPA in CBD+CLB+VPA compared to CLB+VPA or VPA-alone treatments (Table 1). The observed trend of an increase in N-CLB, VPA, and CBD exposures in the triple-drug therapy compared to N-CLB, VPA, and CBD exposures in the CLB+VPA or CLB or VPA and CBD monotherapies suggests potential DDI between CLB, N-CLB, CBD, and VPA.

Figure 4:

Concentration–time curves of (A) CLB (5 mg/kg), (B) N-CLB, (C) VPA (75 mg/kg), and (D) CBD (135 mg/kg) in the brain of Scn1a^WT/WT mice following intraperitoneal administration. Brain samples were obtained 0–8 hr post-dose. Each data point is the mean brain concentration value for four mice. Missing data points before the 8-hr time in each curve denote that the analyte was found to be below the limit of quantitation for the assay.

FFA and LCS treatment as adjunctive therapies to CLB and VPA were not effective against hyperthermia-induced seizures in Scn1a^A1783V/WT mice

In addition to STP and CBD as adjunct drugs, other FDA-approved and investigational add-on compounds, such as FFA and LCS, that have shown promise in the treatment of seizures in DS were evaluated. In a previous study, FFA (25 mg/kg) and LCS (10 mg/kg) were insufficient to prevent hyperthermia-induced seizures in the Scn1a^A1783V/WT mice when administered alone.¹⁷ Therefore, we evaluated the effect of FFA or LCS administration in combination with CLB+VPA against hyperthermia-induced seizures in DS mice. At all studied doses of FFA (5, 10, 20, and 30 mg/kg) administered in combination with CLB (5 mg/kg) + VPA (75 mg/kg), the combination treatment failed to confer protection from hyperthermia-induced seizures in the Scn1a^A1783V/WT mice. Similarly, there were no significant differences in the seizing temperatures when increasing doses of LCS (5, 7, 10, and 15 mg/kg) were combined with CLB (5 mg/kg) + VPA (75 mg/kg) compared to the control-treated group of CLB (5 mg/kg)+VPA (75 mg/kg) (p>0.05) (Table 2). Since CLB+VPA+FFA and CLB+VPA+LCS were ineffective against hyperthermia-induced seizures, we did not follow up with PK studies to evaluate potential DDI between the compounds.

Table 2:

Summary of the efficacy of STP, CBD, FFA, and LCS as add-on therapies to CLB and VPA against hyperthermia-induced seizures following intraperitoneal administration in Scn1a^A1783A/WT mice

STP add-on
Drug treatment	Seizing temp (°C)	Control treatment	Seizing temp (°C)	Statistical significance (Log-rank)
CLB (5mg/kg) + VPA (75 mg/kg) + STP (30 mg/kg)	39.8 (n=9)	CLB (5 mg/kg) + VPA (75 mg/kg)	39.7 (n=9)	p=0.797
CLB (5mg/kg) + VPA (75 mg/kg) + STP (70 mg/kg)	39.5 (n=9)		39.7 (n=9)	p=0.797
CLB (5mg/kg) + VPA (75 mg/kg) + STP (100 mg/kg)	40.4 (n=10)		39.5 (n=10)	*p=0.047
CLB (5mg/kg) + VPA (75 mg/kg) + STP (130 mg/kg)	40.2(n=9)		39.15 (n=10)	**p=0.004
CBD add-on
Drug treatment	Seizing temp (°C)	Control treatment	Seizing temp (°C)	Statistical significance (Log-rank)
CLB (5mg/kg) + VPA (75 mg/kg) + CBD (70 mg/kg)	39.0 (n=10)	CLB (5 mg/kg) + VPA (75 mg/kg)	39.5 (n=10)	p=0.186
CLB (5mg/kg) + VPA (75 mg/kg) + CBD (100 mg/kg)	39.5 (n=10)		39.6 (n=9)	p=0.875
CLB (5mg/kg) + VPA (75 mg/kg) + CBD (135 mg/kg)	40.7 (n=9)		39.85 (n=10)	*p=0.027
CLB (5mg/kg) + VPA (75 mg/kg) + CBD (150 mg/kg)	39.7 (n=10)		39.95 (n=10)	*p=0.01
FFA add-on
Drug treatment	Seizing temp (°C)	Control treatment	Seizing temp (°C)	Statistical significance (Log-rank)
CLB (5mg/kg) + VPA (75 mg/kg) + FFA(5 mg/kg)	39.4 (n=10)	CLB (5 mg/kg) + VPA (75 mg/kg)	39.4 (n=10)	p=0.787
CLB (5mg/kg) + VPA (75 mg/kg) + FFA (10 mg/kg)	39.4 (n=10)		39.0 (n=10)	p=0.334
CLB (5mg/kg) + VPA (75 mg/kg) + FFA (20 mg/kg)	39.0 (n=10)		38.4 (n=10)	p=0.982
CLB (5mg/kg) + VPA (75 mg/kg) + FFA (30 mg/kg)	39.2 (n=10)		39.1 (n=10)	p=0.345
LCS add-on
Drug treatment	Seizing temp (°C)	Control treatment	Seizing temp (°C)	Statistical significance (Log-rank)
CLB (5mg/kg) + VPA (75 mg/kg) + LCS (5 mg/kg)	39.6 (n=10)	CLB (5 mg/kg) + VPA (75 mg/kg)	39.9 (n=10)	p=0.346
CLB (5mg/kg) + VPA (75 mg/kg) + LCS (7 mg/kg)	39.5 (n=10)		39.7 (n=9)	p=0.513
CLB (5mg/kg) + VPA (75 mg/kg) + LCS (10 mg/kg)	38.3 (n=10)		39.7 (n=9)	p=0.939
CLB (5mg/kg) + VPA (75 mg/kg) + LCS (15 mg/kg)	38.9 (n=10)		39.4 (n=10)	p=0.133

Abbreviations: CBD, cannabidiol, CLB, clobazam; FFA, fenfluramine; LCS, lorcaserin; STP, stiripentol, VPA, sodium valproate.

Discussion

Although newer drugs have been approved for treating patients with DS, pharmacoresistance remains common in DS.²³ Thus, there is a critical need to introduce novel drugs and better treatment strategies for DS. DS is usually treated with multiple ASMs since monotherapy is generally inadequate.²⁴ Using etiologically relevant and well-characterized preclinical models that include comprehensive PK profiling is helpful for informing future clinical studies and identifying novel compounds that may aid in controlling seizures in the drug refractory DS patient. Furthermore, comparison to drug exposures in the human therapeutic range is essential to demonstrate that preclinical models have predictive validity to ensure the translation of ASMs into preclinical success.²⁵ The present study used single-administration intraperitoneal dosing and a polytherapy paradigm that demonstrated efficacy against hyperthermia-induced seizures as a vital proof-of-concept preclinical strategy to identify efficacy and pharmacokinetics of polytherapy in a DS mouse model. As these are known therapeutic approaches for treatment, this work demonstrates that a drug screening approach utilizing combination therapies and PK analysis is essential to determine whether a preclinical model has a good face, construct, and predictive validity. These are key attributes of preclinical studies if efficacious therapies are to be discovered for persons with refractory DS.

Following the polytherapy regimens, changes in some PK parameters such as AUC, C_max, and half-life occurred, suggesting potential DDI between either STP or CBD with CLB and VPA that could contribute to the antiseizure effects observed in the triple-drug treatment studies. The PK study in this work represents a framework to inform dosing in subchronic administration studies for combination therapies for DS, a priority of the epilepsy research field. To that end, we recently completed subchronic dosing studies that utilized the PK data from this manuscript to predict doses in triple drug therapy (CLB + VPA + STP) that would achieve concentrations for each compound within the targeted therapeutic window. To date, concentration data from this subchronic dosing study fall within the target therapeutic window (data not shown), supporting the extrapolation of PK data from the wild-type mice in this study to guide therapeutic dosing in our Scn1a^A1783V/WT mouse model. This approach is similar to exploratory Phase 1 clinical trials, which seek to define the PK of a novel drug in healthy volunteers prior to studies in volunteers with the targeted disease state.²⁶ Furthermore, these data demonstrate PK studies as an integral part of ASM discovery and development work. PK profiles of ASMs are critical to guide appropriate efficacy studies in subchronic recurrent spontaneous rodent seizure models yet are often lacking in preclinical assessments. Additionally, this work provides a framework for future studies designed to identify novel therapies for patients who do not respond to either CLB + VPA + STP or CBD.^8,14

In the present study, adding STP (100 mg/kg and 130 mg/kg) to CLB (5 mg/kg) + VPA (75mg/kg) demonstrated efficacy against hyperthermia-induced seizures in DS mice. Reports on the clinical success of STP as an add-on therapy have demonstrated mixed efficacy when combined with CLB + VPA. For example, Inoue et al.²⁷ reported that 13 out of 24 patients receiving 50 mg/kg/day (maximum) STP in addition to CLB (mean: 0.36 mg/kg/day) + VPA (mean: 26.9 mg/kg/day) had their seizure frequency decreased by ≥ 50%. Also, Chiron et al.¹¹ reported that 71% of patients were responders in a randomized placebo-controlled syndrome-dedicated trial when STP (50 mg/kg/day) was added to CLB (0.5 mg/kg/day maximum) + VPA (30 mg/kg/day maximum). It is as yet unclear if these clinically studied doses achieve the exposures necessary for seizure protection across the entire studied population, supporting the need to evaluate PK and optimal dosing strategies to improve DS treatment.

Our work shows changes in some key PK parameters of STP when STP is coadministered with CLB and VPA, suggesting potential DDIs between STP, CLB, N-CLB, and VPA. Several experiments also suggest that the antiseizure effects of STP are potentiated by the interaction with other ASMs, such as CLB, via DDIs.^28–29 Similarly, prior clinical studies have reported that STP-mediated inhibition of CYP2C19 can elevate N-CLB concentration in patients co-treated with CLB.^11,30 The present study observed greater total brain STP exposure when STP is combined with CLB and VPA compared to treatment with STP alone. Figure 2D suggests this finding is likely due to sustained concentrations of STP near C_max when co-administered with CLB + VPA, resulting in an increase in total STP absorption. CLB has been reported to increase STP concentration by 37%; however, the potential mechanism responsible for the increase of STP levels by CLB is not fully known but reported to be mediated via CYP 450 enzyme activity.³¹ Additionally, VPA inhibits CYP3A4 activity, an enzyme that mediates the first-pass effect of STP, increasing STP exposure.³² Overall, we found that plasma and brain STP exposures in mice increase when co-administered with CLB and VPA, consistent with human studies. This finding confers face and predictive validity to our multiple therapy screening approach for identifying novel investigational compounds that may serve to control seizures in patients with refractory DS better.

CBD (135 mg/kg and 150 mg/kg) add-on to CLB and VPA significantly elevated the temperature threshold for seizure onset in Scn1a^A1783V/WT mice. Several studies have reported that CBD (100–200 mg/kg) demonstrates an antiseizure effect when given alone or co-administered with CLB.^20,33

Although CBD proved efficacious in the present study, systemic and network meta-analysis comparative efficacy studies of STP, CBD, and FFA as first-line add-on therapies for seizures in DS showed that CBD demonstrated the least probability of reducing seizure frequency by ≥ 50%^34–35. Moreover, a randomized open-label extension trial by Iannone et al.³⁶ showed that the percentage of patients with at least a 50% reduction in seizure frequency following CBD+CLB+VPA treatment was 40.2%, suggesting that not all patients with DS have their seizures adequately suppressed by CBD. Potential DDI, coupled with differences in the Scn1a gene mutation types and independent PD effects of CLB and CBD, may contribute to the observed varying efficacy outcomes when CBD is given alone or as an adjunct therapy with CLB.^20,37

Our PK study demonstrated alterations in some key PK parameters of CBD when combined with CLB and VPA, suggesting potential DDIs between CBD, CLB, N-CLB, and VPA. Notably, several clinical PK studies on concomitant administration of CBD and CLB had reported potential bi-directional DDI between these compounds, resulting in clinically significant changes in N-CLB concentrations most likely mediated by inhibition of CYP2C19.^38–39 We observed a two-fold increase in brain exposure to CBD in the triple-drug treatment group of CBD+CLB+VPA compared to CBD monotherapy. The higher AUC values observed for CBD in combination therapy could be associated with reduced CBD clearance due to DDI with CLB.⁴⁰ CLB extensively inhibits CYP2D6, while VPA is a mild inhibitor of both CYP2C19 and CYP3A4, which are reported to metabolize CBD.^41–43 This coordinated inhibition of CYP450 enzymes relevant to CBD metabolism by CLB and VPA is expected to slow CBD clearance, increasing CBD half-life and total exposure in triple-drug therapy. Moreover, CBD is a substrate of breast cancer resistance protein (BCRP), an efflux transporter at the blood-brain barrier.⁴⁴ VPA inhibits the transport of BCRP substrates such as CBD, thus reducing the efflux of CBD at the blood-brain barrier, increasing CBD penetration into brain tissue.⁴⁵ Our study demonstrated changes in CBD brain PK profiles when CBD is added to CLB and VPA, thus suggesting an alteration in the dosage regimen of CBD when given concurrently with CLB and VPA for sub-chronic studies.

The PK of CLB and N-CLB differed between mice treated with CLB alone compared to CLB in combination with other drugs, suggesting potential DDI between CLB and VPA, STP, and CBD similar to what has been reported in humans.^20,46 Increased absorption of CLB was observed when CLB was co-administered with VPA + STP or CBD compared to CLB+VPA treatment, resulting in increased CLB exposure and C_max in the brain. CLB is a P-gp substrate,⁴³ while STP and CBD are P-gp inhibitors, thus reducing the efflux of CLB from the blood-brain barrier.^47–49 Our data indicate higher N-CLB (>2-fold increase) concentrations in triple-drug therapies compared to CLB+VPA treatment. Additionally, adding CBD to CLB+VPA resulted in higher CLB and N-CLB exposures than adding STP to CLB+VPA, suggesting CBD has a greater impact on CLB disposition than STP. The observed DDI increases the exposure of both CLB and its active metabolite, N-CLB, in the brain, likely improving the overall observed efficacy in the triple-drug treatment groups. These findings fit with what is known in humans and further demonstrate that our novel drug screening platform has good face and predictive validity for developing new efficacious therapies for DS.

The present study also evaluated the efficacy of FFA for treating seizures associated with DS as an adjunctive drug. At all doses tested, FFA failed to protect against hyperthermia-induced seizures when added to CLB and VPA. Preclinical and clinical studies have reported the successful use of FFA as an add-on treatment for DS, but there has been incomplete efficacy across some study populations.^50–52 Systematic reviews and meta-analyses comparing the efficacy and safety of adjunctive ASMs for DS reported that FFA is as effective as STP.^34–35 Similar to our findings with STP, a PK study of FFA might help determine appropriate doses to achieve therapeutic plasma and brain levels. Like FFA, LCS failed to protect against seizures when added to CLB and VPA in Scn1a^A1783V/WT mice. This suggests that our model may be a drug-resistant DS model for screening novel compounds for DS.

There are limitations of the present study. Each time point of the PK studies is pooled data from four wild-type mice rather than longitudinal samples collected from the same animal over time. This approach prevents the determination of inter-animal variability in PK and statistical testing of key PK parameters, such as AUC_0-inf. However, pooling can be used for comprehensive data analyses and exploratory PK studies for drug-dosing designs.^53–54 The PK work in this study used Scn1a^WT/WT rather than Scn1a^A1783V/WT mice. This approach is similar to phase I human studies, which seek to define the PK of a novel drug in healthy volunteers before advancing the study to patients with the targeted disease state. We subsequently compared the CLB PK profile in Scn1a^A1783V/WT mice to age-matched WT mice; this study demonstrated that the CLB PK profile in Scn1a^A1783V/WT mice mirrored that of WT mice (Table S2). Furthermore, the initial PK data from wild-type mice in the present work were used to successfully guide dosing in a subchronic triple-drug therapy study in Scn1a^A1783V/WT mice that achieved targeted therapeutic concentrations (data not shown), supporting the translation of PK data between wild-type and DS mice. This further demonstrates our study’s translatability and suggests comparable PK profiles between Scn1a^WT/WT and Scn1a^A1783V/WT mice. Also, we did not perfuse brain tissue with saline before homogenization; thus, residual blood in the sample may influence the measured brain concentrations. Moreover, the additional volume in the intraperitoneal space may alter perfusion in triple-drug studies compared to a single drug. Using a preclinical model allowed the evaluation of brain concentrations, which is not practical in clinical settings. Future work is therefore needed to understand the relationship between the clinically accessible plasma concentrations and drug concentrations in brain tissue, likely through the use of a physiologically based PK model. Despite these limitations, the work reported here demonstrates the importance of defining drug PK in the setting of multi-drug therapy.

While new adjunctive therapies have been approved for the person with DS, not all patients have their seizures adequately controlled with either older-generation ASMs or new ASMs used in combination with CLB + VPA. Thus, a continued unmet clinical need exists to identify compounds that may be efficacious in treating DS, often in combination with other ASMs. To find effective therapies for DS, investigational compounds may also need to be screened against spontaneous seizures. Thus, PK data from this study will guide the design of appropriate sub-chronic dosage regimens for evaluating the efficacy of STP or CBD when co-administered with CLB and VPA against spontaneous seizures and describe a framework for future studies to identify appropriate doses of CLB, VPA, and a novel drug in our DS mouse model. In conclusion, the results from these studies describe effective doses of STP and CBD when added to CLB+VPA, which are supported by an understanding of the underlying PK of these doses. These findings support the need for future studies of existing and investigational ASM to include PK studies in both single and multi-drug therapy to better translate preclinical findings to effective clinical care.

Supplementary Material

Supinfo1

Figure S1: Seizure protection profile of various CLB (5 mg/kg) + VPA (75 mg/kg) control groups compared to multiple doses of STP add-on to CLB + VPA in Scn1a^A1783V/WT mice. p = 0.3558; Log-rank (Mantel-Cox).

Supinfo2

Figure S2: Concentration–time curves of (A) CLB (5 mg/kg), (B) N-CLB, (C) VPA (75 mg/kg), and (D) STP (100 mg/kg) in the plasma of Scn1a^WT/WT mice following intraperitoneal administration. Plasma samples were obtained 0–8hr post-dose. Each data point is the mean plasma concentration value for four mice. Missing data points before the 8-hr time in each curve denote that the analyte of interest was found to be below the limit of quantitation for the assay.

Supinfo3

Figure S3: Seizure protection profile of the various CLB (5 mg/kg) + VPA (75 mg/kg) control groups compared to multiple doses of CBD add-on to CLB + VPA in Scn1a^A1783V/WT mice. p = 0.0604; Log-rank (Mantel-Cox).

Supinfo4

Figure S4: Concentration–time curves of (A) CLB (5 mg/kg), (B) N-CLB, (C) VPA (75 mg/kg), and (D) CBD (135 mg/kg) in the plasma of Scn1a^WT/WT mice following intraperitoneal administration. Plasma samples were obtained 0–8hr post-dose. Each data point is the mean plasma concentration value for four mice. Missing data points before the 8-hr time in each curve denote that the analyte of interest was found to be below the limit of quantitation for the assay.

Key Points.

• The hyperthermia-induced seizure assay in Scn1a^A1783V/WT mice mimics febrile seizures in clinical settings.

• The triple-drug therapy may be a valuable preclinical strategy for screening investigational compounds for DS.

• Enhance our understanding of drug-drug interactions of combination therapies in treating DS.

• The study designed a triple-drug paradigm that mimics clinical standards for identifying effective therapies.

Data availability

Data from the study would be made available to other investigators for replication and re-use upon reasonable request.