Antisense oligonucleotide therapy for KCNT1 encephalopathy

Abstract

Developmental and epileptic encephalopathies (DEEs) are characterized by pharmaco-resistant seizures with concomitant intellectual disability. Epilepsy of infancy with migrating focal seizures (EIMFS) is one of the most severe of these syndromes. De novo variants in ion channels, including gain-of-function variants in KCNT1, which encodes for sodium activated potassium channel protein KNa1.1, have been found to play a major role in the etiology of EIMFS. Here, we test a potential precision therapeutic approach in KCNT1-associated DEE using a gene-silencing antisense oligonucleotide (ASO) approach. We generated a mouse model carrying the KCNT1 p.P924L pathogenic variant; only the homozygous animals presented with the frequent, debilitating seizures and developmental compromise that are seen in patients. After a single intracerebroventricular bolus injection of a Kcnt1 gapmer ASO in symptomatic mice at postnatal day 40, seizure frequency was significantly reduced, behavioral abnormalities improved, and overall survival was extended compared with mice treated with a control ASO (nonhybridizing sequence). ASO administration at neonatal age was also well tolerated and effective in controlling seizures and extending the life span of treated animals. The data presented here provide proof of concept for ASO-based gene silencing as a promising therapeutic approach in KCNT1-associated epilepsies.

Keywords: Epilepsy; Gene therapy; Genetics; Mouse models; Neuroscience.

Figure 1. Phenotype of the Kcnt1 L/L mouse model.

(A) Difference in size at P21 of L/L mice (red arrow) compared with their L/+ and +/+ littermates. (B) Body weight is significantly reduced in L/L mice compared with their +/+ littermates at weaning age (Kruskal-Wallis test, followed by Dunn’s post hoc analysis; +/+ n = 15, L/+ n = 21, L/L n = 18). Data are presented in a box-and-whisker plot with maximal and minimal data points (whiskers) and median (line). (C) Life span is shortened in L/L mice with a median survival of 43 days (Kaplan-Meier curve, log-rank test P < 0.0001; +/+ n = 15, L/+ n = 15, and L/L n = 35). (D) Representative ECoG trace of ictal activity and interictal spikes in the L/L mice. Top: Acute interictal spikes. Bottom: Seizures can be preceded by an increase in frequency of acute interictal spikes. Spontaneous tonic-clonic seizures correlated with fast, high-amplitude signal, followed by electric suppression (black arrow). (E) Acute high-amplitude (>500 μV) spikes are present in the L/L mice, with a median of 1,470 spikes in 24 hours (P = 0.03, Kruskal-Wallis test, followed by Dunn’s post hoc analysis, n = 3 for each genotype). (F) Seizure frequency over 72 hours (+/+ n = 3, L/+ n = 3, L/L n = 6; median seizure frequency for L/L of 23 events. Kruskal-Wallis test, followed by Dunn’s post hoc analysis P = 0.024).

Figure 2. Behavioral profile of the Kcnt1 L/L mouse model.

(A) Representative images of minimal and maximal scores of nesting behavior. The left image corresponds to a score of 5 (+/+ mouse), while the right image exemplifies a score of 1 (L/L mouse). (B) Nesting behavior is impaired in L/L mice (+/+ n = 7, L/+ n = 7, L/L n = 8, Kruskal-Wallis test with Dunn’s post hoc analysis). (C) Total ambulatory distance explored in the locomotor cells test. L/L mice are more active compared with L/+ and +/+ (+/+ n = 12, L/+ n = 20, L/L n = 25, Kruskal-Wallis test with Dunn’s post hoc analysis). (D) L/L mice spend less time in the light compartment during the light/dark box test (+/+ n = 12, L/+ n = 20, L/L n = 17, Kruskal-Wallis test with Dunn’s post hoc analysis). (E) L/L mice display a preference for the open arms of the elevated plus maze (+/+ n = 10, L/+ n = 11, L/L n = 15, Kruskal-Wallis test with Dunn’s post hoc analysis). Data are presented in a box-and-whisker plot with maximal and minimal data points (whiskers) and median (line).

Figure 3. Kcnt1 ASO produces a dose-dependent knockdown of Kcnt1 mRNA in the mouse CNS.

Dose-response curves for Kcnt1 ASO in the brain cortex (A) and thoracic spinal cord (B) for +/+ mice injected at P40. The tissue was collected 2 weeks after injection and processed for mRNA quantification (n = 3 for each dose and PBS control, curves were fitted with the Motulsky regression). (C) Mouse cortex was collected 2 weeks after i.c.v. injection for mRNA quantification. The i.c.v. administration of Kcnt1 ASO reduced the levels of Kcnt1 mRNA (Kruskal-Wallis test P = 0.0027), without affecting the paralog gene Kcnt2 (Kruskal-Wallis test P = 0.4794); data are presented as bar plots with mean and SEM, untreated n = 4, Kcnt1 ASO n = 5, control ASO n = 3. (D) Western blot showing a reduction of Kcnt1 protein in WT mice left hemisphere 2 weeks after i.c.v. injection. (E) Average band signal for Kcnt1 protein. Data are presented as bar plots with mean and SD (n = 3 mice for each treatment condition, NS for all comparisons except PBS versus 500 μg, 1-way ANOVA with Dunnett’s multiple comparisons test). (F) Distribution of the Kcnt1 ASO in the mouse brain. Coronal brain sections of +/+ mice treated with Kcnt1 ASO 75 μg. Tissue was collected 2 weeks after i.c.v. injection and stained with an ASO antibody (red) and neuronal marker (NeuN; green) and counterstained with nuclear stain DAPI (blue). Kcnt1 ASO was found throughout the meninges, hippocampus, and cortical layers (n = 3 experiments). Scale bars represent 500 μm.

Figure 4. ASO-mediated knockdown of Kcnt1 at P40 markedly improves the disease phenotype of adult L/L mice.

(A) Experimental timeline for behavioral studies. (B) Kaplan-Meier curves show a dose-dependent improvement in survival of adult L/L mice treated with Kcnt1 ASO (P < 0.0001 for Kcnt1 ASO ED₅₀, ED₈₀, and 500 μg, log-rank test), while mice treated with control ASO showed a survival similar to that of untreated animals (P = 0.237, log-rank test, untreated n = 16, control ASO n = 16, Kcnt1 ASO ED₅₀ n = 13, ED₈₀ n = 11, 500 μg n = 11). (C) Acute spike frequency over 24 hours (untreated n = 4, Kcnt1 ASO ED₅₀ n = 6, ED₈₀ n = 7, 500 μg n = 8; 1-way ANOVA F[3, 21] = 7.978, P = 0.001). (D) Seizure frequency was significantly reduced after treatment with Kcnt1 ASO ED₈₀ and 500 μg. Although ED₅₀ did not reach statistical significance, a trend toward reduction was observed. Treatment with control ASO did not reduce the occurrence of seizures (control ASO n = 9; ED₅₀ n = 8; ED₈₀ n = 10, 500 μg n = 9; seizure frequency was compared using the nonparametric Wilcoxon matched pairs signed-rank test, with Pratt’s method for identical rows). (E) Representative images of nesting behavior of Kcnt1 ASO ED₈₀ (top) and control ASO–treated (bottom) L/L mice. (F) Nesting score of animals treated with Kcnt1 ASO showed a significant improvement compared with control ASO–treated animals (ED₅₀ vs. control P = 0.0002; ED₈₀ vs. control P = 0.0015; 500 μg vs. control P < 0.0001; untreated vs. control P > 0.9. Kruskal-Wallis test with Dunn’s post hoc analysis. Untreated n = 8, control ASO n = 12; ED₅₀ n = 13; ED₈₀ n = 10, 500 μg n = 12). Data are presented in a box-and-whisker plot with maximal and minimal data points (whiskers) and median (line).

Figure 5. ASO-mediated knockdown of Kcnt1 at P40 improves the behavioral phenotype of adult L/L mice.

(A) Total ambulatory distance was reduced in mice treated with Kcnt1 ASO 500 μg compared with control-treated mice (P = 0.0002) but not for mice treated with ED₅₀ or ED₈₀ (P = 0.775 and 0.839, respectively) (Kruskal-Wallis test, followed by Dunn’s multiple comparisons, +/+ n = 12, untreated n = 25, control ASO n = 13; ED₅₀ n = 13; ED₈₀ n = 10, 500 μg n = 11). (B) Time spent in the light compartment during the light/dark box test (Kruskal-Wallis test, followed by Dunn’s multiple comparisons P = 0.588; +/+ n = 12, L/L n = 17, control ASO n = 11, ED₅₀ n = 13, ED₈₀ n = 10, 500 μg n = 12). (C) Time spent in the open arms of the EPM (Kruskal-Wallis test, followed by Dunn’s post hoc analysis; +/+ n = 10; untreated n = 15; control ASO n = 9; ED₅₀ n = 13; ED₈₀ n = 10, 500 μg n = 12). (D) Time spent in the novel arm of the Y maze (Kruskal-Wallis test, followed by Dunn’s post hoc analysis P = 0.551; +/+ n = 8; untreated n = 10; ED₅₀ n = 12; ED₈₀ n = 10, 500 μg n = 10).

Figure 6. Neonatal administration of Kcnt1 ASO in the L/L mouse model.

(A) Dose-response curve for +/+ mice injected at P2 (n = 3–11 per dose). (B) Survival curve showing prolonged life span of Kcnt1 ASO–treated mice (P < 0.0001). Control ASO produced a small improvement (P = 0.003, log-rank test) (untreated n = 36; control ASO n = 23; Kcnt1 ASO 3.4 μg n = 17). (C) Weight at P40 (P = 0.016, +/+ n = 28, untreated n = 17, control ASO 50 μg n = 14, Kcnt1 ASO 3.4 μg n = 17). (D) Nesting score (P = 0.0001, untreated n = 8; control ASO 50 μg n = 9; Kcnt1 ASO 3.4 μg n = 15). (E) Seizure frequency (P < 0.0001, control ASO 50 μg n = 15, Kcnt1 ASO 3.4 μg n = 17). (F) Ambulatory distance (+/+ n = 12, untreated n = 25, Kcnt1 ASO 3.4 μg n = 17). (G) Time spent in open arms of EPM (+/+ n = 10, untreated n = 15, Kcnt1 ASO 3.4 μg n = 17). Mann-Whitney test (C–G). Data are presented in a box-and-whisker plot with maximal and minimal data points (whiskers) and median (line). (H) Redosing at P30 further reduced Kcnt1 mRNA (Kcnt1 ASO 3.4 μg n = 3; Kcnt1 ASO 3.4 μg + control ASO 500 μg n = 3; Kcnt1 ASO 3.4 μg + Kcnt1 ASO 35 μg n = 2; Kcnt1 ASO 3.4 μg + Kcnt1 ASO 75 μg n = 3). (I) Survival curve for L/L mice redosed with ED₈₀ at P167 (untreated n = 36; ED₈₀ reinjected n = 4). (J) Nesting score of L/L mice reinjected with ED₈₀ (n = 4); P < 0.0001 (1-way ANOVA with Dunnett’s multiple comparisons test) for P40 versus P165.