Seizure control by ketogenic diet-associated medium chain fatty acids

Abstract

The medium chain triglyceride (MCT) ketogenic diet is used extensively for treating refractory childhood epilepsy. This diet increases the plasma levels of medium straight chain fatty acids. A role for these and related fatty acids in seizure control has not been established. We compared the potency of an established epilepsy treatment, Valproate (VPA), with a range of MCT diet-associated fatty acids (and related branched compounds), using in vitro seizure and in vivo epilepsy models, and assessed side effect potential in vitro for one aspect of teratogenicity, for liver toxicology and in vivo for sedation, and for a neuroprotective effect. We identify specific medium chain fatty acids (both prescribed in the MCT diet, and related compounds branched on the fourth carbon) that provide significantly enhanced in vitro seizure control compared to VPA. The activity of these compounds on seizure control is independent of histone deacetylase inhibitory activity (associated with the teratogenicity of VPA), and does not correlate with liver cell toxicity. In vivo, these compounds were more potent in epilepsy control (perforant pathway stimulation induced status epilepticus), showed less sedation and enhanced neuroprotection compared to VPA. Our data therefore implicates medium chain fatty acids in the mechanism of the MCT ketogenic diet, and highlights a related new family of compounds that are more potent than VPA in seizure control with a reduced potential for side effects. This article is part of the Special Issue entitled 'New Targets and Approaches to the Treatment of Epilepsy'.

Supplementary Fig. 1

Raw data for spike frequencies for all compounds, as described in Fig. 1.

Supplementary Fig. 2

The effect of medium chain fatty acid and related compounds at low concentration on the frequency of PTZ-induced epileptiform discharges in combined entorhinal cortex–hippocampus. (A) Structures of valproic acid (VPA), valnoctamide (VCD) and decanoic acid (DA). (B) The effect of VPA (n = 3), DA (n = 3) and VCD (n = 3) at 0.1 mM on frequency of PTZ-induced epileptiform discharges. The frequency of epileptiform discharges induced by PTZ plotted against time. (B) Comparison of the mean frequency of PTZ-induced epileptiform discharges. Graphs show means ± SEM. ** Indicates a significant difference at p < 0.01 compared to control.

Supplementary Fig. 3

Concentration-dependent in vitro seizure control activity for 4-methyloctanoic acid and nonanoic acid compared to VPA. PTZ-induced epileptiform discharges in combined entorhinal cortex–hippocampus was monitored at three compound concentrations. (A) The example of trace record of VPA, 4-MO and NA (B) and frequency of epileptiform discharges plotted against time (n = 5 for all). (C) Comparison of the mean frequency of PTZ-induced epileptiform discharges at different treatment concentrations, showing mean ± SEM. *p < 0.05 compared to control, **p < 0.01 compared to control.

Fig. 1

Structurally-specific medium chain fatty acids strongly reduce frequency of in vitro epileptiform activity. (A) A range of medium chain fatty acids were analysed in this study: Straight medium chain fatty acids octanoic (OA), nonanoic (NA) and decanoic (DA) acids contain 8, 9 and 10 carbon backbones respectively. Related structures analysed are derivates of octanoic acid branched at the second carbon (2-propyloctanoic acid (2-PO) and 2-butyloctanoic acid (2-BO)) and the fourth carbon (4-methyloctanoic acid (4-MO) and 4-ethyloctanoic acid (4-EO)); and two related structures, a heptanoic acid derivative (7 carbon backbone) branched on the second carbon (2-methylheptanoic acid (2-MH)), and an octanoic acid derivative branched on both the third and the seventh carbon (3,7-dimethyloctanoic acid (3,7-DO)). The frequency of epileptiform activity is plotted against time following (B) control (DMSO) and 3,7-DO (n = 3) gave no effect on epileptiform activity, whereas VPA and 2-MH (n = 3) treatment showed a weak effect. (C) The straight chain fatty acid OA showed no effect, whereas a strong effect was shown for NA and DA. (D) Octanoic acid derivatives branched on the second carbon, 2-PO, 2-BO; and (E) on the fourth carbon 4-MO, 4-EO are also highly active. (F) Comparison of the mean frequency of PTZ-induced burst discharges, averaged from 20 to 40 min post compound addition (data shown as means ± SEM). * and ** indicate a significant difference at p < 0.05 or p < 0.01 compared to control respectively; + and ++ indicate similar levels of significance compared to VPA. Data is provided for all compounds tested at 1 mM from at least five repeats unless indicated. Illustrative trace recordings plotted against time for all compounds are provided in Supplementary Fig. 1.

Fig. 2

Structurally-specific medium chain fatty acids show reduced inhibition of human histone deacetylase enzyme activity. Quantification of HDAC inhibition assay employing human nuclear extracts enzyme (from HeLa cells) as the source of HDAC activity shown as fitted dose–response curves for (A) VPA; (B) unbranched medium chain fatty acids (OA, NA, DA); (C) medium chain fatty acids branched on the second carbon, 2-BO and 2-PO; and (D) medium chain fatty acids branched on the fourth carbon, 4-MO and 4-EO, showing means ± SEM for four independent measurements at each of five concentrations. (E) Comparison of the HDAC inhibition with different treatments at a concentration of 10 mM, showing means ± SEM for four independent measurements for each compound. ** significant difference at p < 0.01 compared to control. ++ significant difference at p < 0.01 compared to VPA.

Fig. 3

Structurally-specific medium chain fatty acids show variable effects on human liver cell viability. Hepatotoxicity was evaluated by determining the mitochondrial function of human hepatocyte (Huh7) cells following 24 h treatment, measured by the mitochondrial conversion of MTT into blue formazan that decreases in direct proportion to cell viability. Data is provided for (A) VPA; (B) unbranched medium chain fatty acids (OA, NA, DA); (C) medium chain fatty acids branched on the second carbon, 2-BO and 2-PO; and (D) medium chain fatty acids branched on the fourth carbon, 4-MO and 4-EO. Fitted dose–response curves (fitting by three parameters, GraphPad) are based on 8 concentrations and with 4 independent measurements of each concentration. (E) Comparison of hepatotoxicity at a concentration of 1 mM, showing means ± SEM for four independent measurements for each compound. * or + indicate significant difference compared to control or VPA respectively, where one, two or three symbols indicate p < 0.5, p < 0.1, or p < 0.01 respectively.

Fig. 4

EEG recording spike frequency and amplitude and seizure behaviour following status epilepticus induction is decreased by 4-methyloctanoic and nonanoic acid compared to VPA. (A) Illustrative examples of EEG recordings from animals in self-sustaining status epilepticus (SSSE), induced by perforant pathway stimulation, with traces shown prior to compound administration (baseline) and following administration of control (DMSO), or VPA, 4-MO or NA (all at 400 mg/kg). (B) Time course of the effects on spontaneous spike frequency following administration of control (DMSO; n = 5), VPA, 4-MO or NA (n = 7 for all at 400 mg/kg) 10 min after stopping perforant path stimulation. (C) Time course of the effects on spontaneous spike amplitude following administration of control (DMSO; n = 5), VPA, 4-MO or NA (n = 7 for all at 400 mg/kg) 10 min after stopping perforant path stimulation. (D) Time course of the effects of compounds on behaviour seizure which was scored using the Racine score during SSES induced by perforant pathway stimulation.

Fig. 5

Sedation effects for 4-methyloctanoic and nonanoic acid compared to VPA. Time course of the effect on sedative score following administration of control (DMSO), VPA, 4-MO or NA at (A) 400 mg/kg or (B) 600 mg/kg. All results are obtained from five animals. Sedation score: 0, spontaneous movement; 1, intermittent spontaneous movement; 2, no spontaneous movement; 3, loss of auditory reflex; 4, loss of corneal reflex; 5, loss of response to tail pinch, with data presented as mean ± SEM and significance given by the Mann–Whitney test. *p < 0.05 compared to control; +p < 0.05 compared to VPA; **p < 0.01 compared to control; ++p < 0.01 compared to VPA.

Fig. 6

Neuroprotective effects in the hilus of the hippocampus following status epilepticus following nonanoic acid, 4-methyloctanoic acid and VPA treatment. Two months after induction of status epilepticus (following indicated treatments), hippocampal slices were prepared and neuronal cell loss was visualised (illustrated by diffuse dark staining) in the hilus (outlined in A). Animals (A) without status epilepticus (control, n = 5) or following status epilepticus that were treated with (B) vehicle only (DMSO) (n = 4); (C) VPA (n = 6); (D) 4-MO (n = 4); and (E) NA (n = 3) at 400 mg/kg. (F) Quantification of neuroprotective effect. Graph shows means ± SEM. *p < 0.05 compared to control, **p < 0.01 compared to control; +p < 0.05 compared to control. Scale bar = 20 μm. DG; dentate gyrus.