Among many other hypotheses, PUFA (polyunsaturated fatty acids) have been proposed to underlie the anticonvulsant effect of the KD (Cunnane et al., 2002). Fraser et al. (2003) found that a 4:1 KD in children increased serum lipids, specifically linoleic acid, arachidonic acid, and docosahexaenoic acid (DHA), by 1.3- to 2.6-fold. Moreover, KD-fed rats showed slightly higher brain levels of arachidonic acid and DHA (Taha et al., 2005). DHA or another n-3 long chain PUFA, eicosapentaenoic acid (EPA), modestly increased the seizure threshold in the a rat cortical stimulation model (Voskuyl et al., 1998). Also, both PUFAs showed inhibitory actions in mouse hippocampal slices (Xiao & Li, 1999). Therefore, it is of considerable interest to determine to what extent PUFAs contribute to the anticonvulsant effect of the KD. Although we have not directly investigated this hypothesis, our recent studies shed some light on the effects of the long chain PUFA, EPA, and DHA, added to a standard diet versus the KD. Similar mouse seizure models were used to compare the effect of these oral treatments (Samala et al., 2008; Willis et al., in revision). Surprisingly, there was no anticonvulsant effect of either DHA or EPA after 4 weeks of feeding in any of the acute seizure models investigated, including the 6 Hz model, in which the KD was anticonvulsant. This lack of effect was paralleled by an increase in plasma esterified DHA by 1.7-fold with both EPA and DHA feeding, which is similar to the DHA increase found in children on the KD (Fraser et al., 2003). These data therefore suggest that the individual long chain PUFA, EPA, or DHA, only have limited anticonvulsant activities. Further, it is unlikely that blood DHA increases are the single underlying cause of the KD's anticonvulsant effect in the 6 Hz model. However, KD-induced concomitant increases of the enzyme inducer acetone may alter PUFA metabolism, which may contribute to the KD's anticonvulsant effect.
Lastly, it is important to realize that the KD influences specific PUFA levels differentially and feeding DHA or EPA alone does not mimic exactly the PUFA levels found after KD feeding. For example, the KD in humans elevated blood levels of several PUFA, including arachidonic acid (Fraser et al., 2003), and inhibits sodium currents and synaptic transmission (Fraser et al., 1993). In contrast, blood arachidonic acid levels were significantly decreased by EPA or DHA feeding in our mouse study and there is so far no information on PUFA levels after KD-feeding in the mouse. Therefore, to determine to what extent PUFA contribute to the KD's anticonvulsant effect will require future studies on PUFA levels in KD-fed mice and subsequent careful mimicking of the KD-induced changes in PUFA levels.