Abstracts

Rationale: About a quarter of patients with mitochondrial disease develop a form of epilepsy¹. This epilepsy has a poor prognosis and is resistant to antiepileptic drugs (AEDs)². We developed an in vitro model of ‘mitochondrial epilepsy’ with predictive validity; epileptic activity is resistant to the majority of conventional AEDs. One AED capable of abolishing activity was pentobarbital. Whilst able to modulate GABA_A receptors it also blocks AMPA receptors (AMPARs). We examined the therapeutic benefit of inhibiting AMPARs in an in vitro model of mitochondrial epilepsy and in a pilot study in patients with mitochondrial epilepsy. Methods: Epileptiform discharges in rat hippocampal slices were recorded as extracellular local field potentials. Events were elicited by the co-application of fluorocitrate, rotenone and KCN. For pharmacological studies, drugs were applied to circulating artificial cerebrospinal fluid (aCSF). Data on the efficacy of Perampanel in patients with mitochondrial epilepsy was obtained from the MRC Mitochondrial Disease patient cohort. Results: We assessed the contribution of glutamatergic synapses using antagonists for NMDA and AMPA receptors. D-AP5 (50µM) had no effect on the frequency of epileptiform discharges (9.42 ± 1.05 burst/min baseline vs. 8.34 ± 2.18 burst/min drug, p>0.05, n=5). In contrast, NBQX (20µM) significantly reduced the frequency of discharges (12.84 ± 3.37 vs. 0.02 ± 0.02, p<0.05, n=5). Given this finding, we examined AMPAR inhibitors such as Decanoic acid (C10) and Perampanel. Application of C10, a direct inhibitor of AMPARs, at 100µM (15.26 ± 4.22 vs. 12.48 ± 4.34) and 300µM (6.82 ± 2.39) reduced the frequency of discharges, however this was not significant (p>0.05, n=5). Concentrations of 1mM (1.86 ± 0.91) and 2mM (0.38 ± 0.38) caused a significant (p<0.05, n=5) decrease in the burst count. This reduction was dose-dependent (IC_­50 value of 230.6µM). Perampanel, a clinically approved AED, when applied at 0.1µM, reduced the frequency of epileptiform discharges (13.86 ± 2.74 vs 11.00 ± 2.58) but this was not significant (p>0.05, n=5). Increasing the concentration to 0.3µM (8.76 ± 2.56), 1µM (5.94 ± 2.37), 3µM (2.58 ± 1.30), and 10µM (0.92 ± 0.59) caused a significant reduction (p<0.05, n=5). This effect was also dose-dependent (IC₅₀ value; 0.42µM). On the basis of these preclinical findings, 7 patients (m.3243A>G (n=3), m.5543T>C (n=1) and recessive POLG mutations (n=3)) with mitochondrial epilepsy and metabolic strokes were prospectively treated with Perampanel as an adjunctive therapy (median: 8mg; range: 2-10 mg). Seizure improvement was noted in 1 patient with the m.5543T>C mutation and 2 patients with recessive POLG mutations. Perampanel was withdrawn in 1 patient (2 mg) due to aggression and the drug’s efficacy was unclear in the remaining 3 patients. Conclusions: In an in vitro model of mitochondrial epilepsy, AMPARs antagonists such as C10 and Perampanel reduce epileptiform activity. Translating to the clinic, we observed that there is some benefit in terms of seizure control with Perampanel in patients with mitochondrial epilepsy.References:(1) Whittaker RG et al. Annals of Neurology. 78(6):949-957.(2) Bindoff LA and Engelsen BA. Epilepsia. 53:92-97. Funding: Eisai Ltd