Treatment of epilepsy using a targeted p38γ kinase gene therapy

Abstract

Hyperphosphorylated microtubule-associated protein tau has been implicated in dementia, epilepsy, and other neurological disorders. In contrast, site-specific phosphorylation of tau at threonine 205 (T205) by the kinase p38γ was shown to disengage tau from toxic pathways, serving a neuroprotective function in Alzheimer's disease. Using a viral-mediated gene delivery approach in different mouse models of epilepsy, we show that p38γ activity-enhancing treatment reduces seizure susceptibility, restores neuronal firing patterns, reduces behavioral deficits, and ameliorates epilepsy-induced deaths. Furthermore, we show that p38γ-mediated phosphorylation of tau at T205 is essential for this protection in epilepsy, as a lack of this critical interaction reinstates pathological features and accelerates epilepsy in vivo. Hence, our work provides a scope to harness p38γ as a future therapy applicable to acute neurological conditions.

Fig. 1.. p38γ kinase treatment protects DS mice.

(A) Scn1a +/Δ mouse line generation and experimental strategy for p38γ therapeutic study. (B) Western blotting of the cortex (top) and immunostaining of the hippocampus (bottom) validated the transgene expression of p38γ^CA [hemagglutinin (HA) tagged] or control [green fluorescent protein (GFP)] in 6-month-old Scn1a (B6) mice (representative images from n = 4 to 6 mice per treatment). Glyceraldehyde-3 phosphate (GAPDH) served as a loading control. DAPI, 4′,6-diamidino-2-phenylindole. (C) Survival curve for treated Scn1a (B6) mice; Scn1a +/Δ.p38γ^CA (B6) mice show significantly improved survival compared with nontherapeutic control-administered mice [Scn1a +/Δ.control (B6)] over 200 days [n ≥ 11 per group; *P = 0.0113, ***P = 0.0006, and ****P ≤ 0.0001 (Mantel-Cox test)]. (D) Electroencephalographic (EEG) power spectral density profiles at low (left) and high (right) frequencies during wakeful periods in 5- to 6-month-old treated Scn1a (129) mice. Scn1a +/Δ.control (129) mice demonstrated significantly elevated power spectral density when compared with Scn1a +/+ mice in theta (4- to 12-Hz) and gamma (25- to 100-Hz) frequency bands [*P = 0.0156 (theta) and ****P ≤ 0.0001 (gamma)]. Aberrant network activity was ameliorated in Scn1a +/Δ.p38γ^CA (129) mice that exhibited theta power not significantly different from that of Scn1a +/+ mice of both treatment groups [**P = 0.002 (theta) and **P = 0.0016 (gamma); ns, not significant (ANOVA)], although gamma power remained moderately increased compared with Scn1a +/+.control mice (**P = 0.0015). Gamma power in Scn1a +/+.p38γ^CA was significantly altered compared to Scn1a +/+.control and Scn1a +/+.p38γ^CA mice (n = 3 to 5 mice per group; ****P ≤ 0.0001 and **P = 0.0041). AUC, area under the curve. (E) Open-field behavioral analysis. Compared with Scn1a +/+.control (129) mice, Scn1a +/Δ.control (129) mice traveled significantly further distances, with more circling, and more frequently in inner zone entries [****P ≤ 0.0001 (distance), ****P ≤ 0.0001 (circling), and **P = 0.0012 (zone)]. Behaviors of Scn1a +/Δ.p38γ^CA (129) mice were significantly normalized compared with Scn1a +/Δ.control (129) mice [**P = 0.0086 (distance), *P = 0.0164 (circling), and ns, P = 0.2012 (zone)] and not significantly different from that of Scn1a +/+ mice of both treatment groups [n = 9 to 15 mice per group; ns (ANOVA)].

Fig. 2.. TauT205 is a protective site in epilepsy.

(A) Western blotting of brains of 2-week-old Scn1a Δ/Δ (129) mice shows significantly increased and distinct tau phosphorylation at T205 (ptauT205), S202 + T205 (ptauAT8), and S262 (ptauS262) relative to normalized total tau levels and compared to Scn1a +/+ (129) and Scn1a +/Δ (129) mice. Similarly, individual Scn1a Δ/Δ (129) samples showed increased phosphorylated p38 MAPK levels (pp38 MAPK). Total tau levels and its phosphorylation at sites tauS202, tauS214, tauS422, and tauPHF1 (S396 + S404) as well as total p38γ levels were comparable between Scn1a +/+, Scn1a +/Δ, and Scn1a Δ/Δ (129) mice [n = 4 to 5 mice per genotype; ****P ≤ 0.0001; ns (ANOVA)]. (B) Schematic of the experimental design for p38γ therapeutic study before pilocarpine-mediated seizure induction. (C) Seizure severity and latency (top), survival (bottom left), and mean seizure severity (bottom right) after pilocarpine induction in 6-week-old tauT205T/T and tauT205A/A mice, following prior AAV treatment at 4 weeks of age. Wild-type B6 (tauT205T/T) mice pretreated with p38γ^CA (tauT205T/T.p38γ^CA) show significantly prolonged latency to develop more severe seizures (**P = 0.0049), improved survival (**P = 0.0054), and reduced mean seizure severity (**P = 0.0015) as compared to tauT205T/T.control mice. In contrast, tauT205A/A.p38γ^CA mice did not display significant differences in seizure latency, mean seizure severity, or survival compared to both tauT205T/T.control and tauT205A/A.control groups. Accordingly, tauT205A/A.p38γ^CA showed significantly reduced latency (***P = 0.0007), increased mean seizure severity (***P = 0.0003), and increased mortality (***P = 0.0006) compared to tauT205T/T.p38γ^CA mice. n = 15 to 20 mice per group (ANOVA and Mantel-Cox test, respectively).

Fig. 3.. p38γ^CA treatment after seizure induction is protective.

(A) Schematic of experimental strategy for p38γ therapeutic study following pilocarpine-mediated seizure induction. (B) Survival curve following pilocarpine induction shows improved DBA/2.p38γ^CA survival compared to DBA/2.control mice during a 4-month recording period following pilocarpine injection (day 0). n = 9 to 11 mice per group; *P = 0.0150 (Mantel-Cox test). (C) EEG power spectral density profiles during wake in DBA/2 mice 2 months following pilocarpine injection. DBA/2.control mice exhibited significantly increased high and low spectral power compared to untreated DBA/2 mice (theta: ***P = 0.0002; gamma: ****P ≤ 0.0001). DBA/2.p38γ^CA mice showed significantly reduced high and low spectral power, compared to DBA/2.control mice (theta: ***P = 0.0010; gamma: ****P ≤ 0.0001). DBA/2.p38γ^CA theta range was not significantly (ns) different from that of DBA/2.untreated mice but remained moderately increased in the gamma range (****P ≤ 0.0001). N = 4 mice per group (ANOVA).