The ketogenic diet (KD) in its various forms has been demonstrated to be effective at reducing seizures in individuals with drug-resistant epilepsy (Neal et al. 2009). Despite the effectiveness of the diet, the exact mechanism(s) of action in the context of seizure control are not known. However, a number of mechanisms have been proposed that include alterations in antioxidant status, changes in brain neurotransmitter levels, such as GABA, and the metabolic consequences of the diet on cellular energy metabolism (Rho and Sankar 2008). With regards to the latter, it is of note that mitochondrial dysfunction, at the level of the respiratory chain, has been documented in animal model systems of status epilepticus and in post mortem brain samples obtained from patients who had epilepsy (Kunz et al. 2000; Cock et al. 2002). Furthermore, mitochondrial biogenesis has been reported in the brain of rats fed the KD leading to the suggestion that this change has the capacity to elevate the seizure threshold (Bough et al. 2006).
A number of versions of the KD exist including the use of a medium-chain triglyceride (MCT)-based diet (Kossoff et al. 2009). This diet has been demonstrated to be as effective as the classical KD and can permit a greater proportion of carbohydrate and protein to be consumed when compared with the classical KD. However, compliance and tolerability of the diet is still an issue (Bahassan and Jan 2006). There is therefore a need for refined formulations that can replicate the efficacy of the diet and release patients and families from its current constraints. Critical to the development of such a treatment is better understanding of the biochemical changes that occur as a consequence of the KD.
Patients on the KD diet have raised plasma levels of ketone bodies (β-hydroxybutyrate and acetoacetate), but these correlate poorly with seizure control, suggesting the likely involvement of other factors with regards to conveying therapeutic benefit (Likhodii et al. 2000; Thavendiranathan et al. 2000). In addition to ketones, the medium-chain fatty acids octanoic (C8) and decanoic (C10) acid are also raised in the plasma of patients who are on the MCT-based diet (Haidukewych et al. 1982). However, the contribution of these fatty acids to efficacy of the KD has not been extensively studied. In view of the suggestion that increased cellular mitochondrial content may be beneficially associated with KDs, we have in this study investigated the effects of C8 and C10 on neuronal mitochondrial content with an aim to provide further insight into the mechanism whereby the MCT KD may exert its beneficial effects.
- Top of page
- Acknowledgements and conflict of interest disclosure
The MCT KD is widely and successfully used for the treatment of pharmaco-resistant epilepsy. Despite this, the biochemical mechanisms responsible for its effectiveness are poorly understood. Seizure control does not appear to correlate strongly with ketosis. However, plasma levels of the medium-chain fatty acids, C8 and C10, are also elevated (C8: 104–859 μM; C10: 87–552 μM) in patients on this form of the KD (Haidukewych et al. 1982), and we sought here to determine the relative efficacy of these molecules in increasing mitochondrial function.
Among the potential mechanisms whereby KDs convey their efficacy is via improvement of brain energy metabolism and mitochondrial function. Brain mitochondrial proliferation has been demonstrated following exposure to the KD (Bough et al. 2006). This observation may be particularly relevant given the loss of respiratory chain enzyme activity associated with seizure activity (Kunz et al. 2000; Cock et al. 2002). In view of these observations, and the efficacy of the MCT-based KD, we studied the effects of C8 and C10 on neuronal cell mitochondrial content. Using a concentration (250 μM) of these fatty acids comparable to that reported in the plasma of patients on a MCT-based KD (Haidukewych et al. 1982), we observed, for C10 but not C8, a significant increase in citrate synthase activity after 6 days of exposure. Citrate synthase is exclusively located in the mitochondrion and can be utilized as a marker enzyme for this organelle. Alterations in citrate synthase activity can therefore be used to reflect the mitochondrial enrichment of a cellular/tissue homogenate (Selak et al. 2000; Bernier et al. 2002; Itkonen et al. 2013). Consequently, the data reported here suggest an increase in the mitochondrial content of the SH-SY5Y cell homogenates. The SH-SY5Y neuronal cell line is widely used in neurological research and in the study of mitochondrial function (Wang et al. 2012; Hong et al. 2013). However, in view of the fact that this is a cell line, a degree of caution is required when interpreting our results. Our observation that C10 can also elicit increased citrate synthase activity in cultured skin fibroblasts suggests that one of the key findings of this study is not restricted to SH-SY5Y cells.
Mitochondrial proliferation and increase in citrate synthase activity may also occur as a consequence of decreased respiratory chain activity (Bernier et al. 2002). However, this is unlikely to be the situation here as there was no evidence of impaired respiratory chain enzyme activity following treatment with C10. Moreover, complex I activity was significantly increased. It could also be argued that the effect of C10 upon citrate synthase activity/mitochondrial content occurs as a consequence of the increased availability of an energy substrate. Our finding that C8, BHB and AcAC were unable to elicit a response with regards to citrate synthase suggests that our observation is specific to C10 and not related to provision of energy substrates. The lack of response with the ketone bodies also implies that BHB and AcAC are unlikely to be responsible for the mitochondrial proliferation associated with the KD.
Consistent with our citrate synthase data, the electron microscopy findings indicate an increase in mitochondrial number following C10 treatment. However, the mitochondria appear smaller and more electron dense when compared with the untreated cells. Although the number of mitochondria in the C10-treated cells had increased, their smaller morphology may mean that, overall, the proportion of mitochondria occupying the cytosol has not increased. It remains to be demonstrated whether the appearance of the mitochondria in the C10-treated cells reflect alterations in metabolic activity and/or “new” mitochondria formed as a consequence of an increased biogenesis.
Loss of respiratory chain enzyme activity, in particular complex I, is associated with seizure activity (Chuang et al. 2012). In contrast to the other components of the respiratory chain and decanoyl CoA dehydrogenase activity, complex I activity was significantly elevated by treatment with C10. Although the mechanism for this increase in complex I activity has not been addressed here, it is of interest to note that activity of this complex can, in neuronal cells, exert significant control over ATP synthesis (Davey et al. 1998). Whether this specific increase in complex I activity leads to an increase in cellular oxygen consumption and ATP production will be the subject of further studies.
In addition to key functions that include ATP production, calcium homeostasis and regulation of apoptosis, mitochondria are a source of reactive oxygen species. Consequently, mitochondrial proliferation could lead to an increase in oxidative stress. However, improved antioxidant status has been reported in rats fed a KD (Ziegler et al. 2003). In this study, we determined the cellular content of GSH and found it to be unaffected by the C10 treatment. Neuronal GSH content has previously been shown to be decreased by exposure to oxidizing species (Gegg et al. 2003). In view of this, the preservation of GSH status observed here indicates that significant cellular oxidative stress may not occur in the C10-treated cells. A contributing factor to minimizing oxidative stress may be the increased catalase activity seen in the C10-treated cells. Catalase catalyses the decomposition of hydrogen peroxide into oxygen plus water thereby minimizing cellular oxidative stress (Meilhac et al. 2000). Increased generation of reactive oxygen/nitrogen species is also associated with loss of cell viability and an increase in LDH release (Bolaños et al. 1995). Our finding that LDH release was comparable for control and C10-treated cells implies that the cells were not markedly stressed by exposure to this fatty acid.
To explore further the potential mechanism responsible for the above findings, we considered the involvement of PPARγ. Recently, C10, but not C8, has been shown to act as a ligand and activator of PPARγ (Malapaka et al. 2012). Furthermore, stimulation of PPARγ, using pioglitazone or rosiglitazone, has been shown to promote mitochondrial biogenesis in SH-SY5Y cells (Miglio et al. 2009). Our finding that the PPARγ antagonist BADGE prevented the C10-mediated increase in citrate synthase implicates PPARγ activation in our study and in addition explains the lack of effect of C8. Alterations in gene expression may also explain the finding that exposure to C10 for a number of days is required before significant effects upon mitochondrial function can be observed. In addition, our observation that removal of C10 from the cell culture medium led to a return of citrate synthase to control levels implies that persistent exposure to the fatty acid and/or PPARγ activation is required to maintain an increased mitochondrial content. The relative specificity of PPARγ towards C10, when compared with C8, also provides an explanation of our finding that only a very high ratio of C10 to C8 is capable of eliciting an increase in citrate synthase. PPARγ activation has also been shown to cause increased catalase activity, thereby providing a further explanation of our findings relating to this antioxidant enzyme (Diano et al. 2011).
Concerning PPARγ and mitochondrial function, activation, using rosiglitazone as an agonist, has been shown to increase complex I activity in a seizure model (Chuang et al. 2012). This observation is also consistent with our finding relating to the effects of C10 upon complex I activity discussed above. While C10 has been shown to occupy a novel binding site on PPARγ, it is also capable of binding, with less affinity, to PPARα and PPARβ/δ (Malapaka et al. 2012). Thus, it is possible that downstream metabolic events beyond PPARγ may occur as a consequence of C10 exposure.
Plasma levels of C8 and C10 are elevated in patients adhering to the MCT form of the KD. However, there are a number of metabolic fates with regards to these fatty acids, e.g. β-oxidation and ω-oxidation. This latter process generates the dicarboxylic acid form of C10, sebacic acid, which has been shown to exhibit biological activity including improving glucose tolerance (Membrez et al. 2010). In contrast to such studies, sebacic acid in our experimental system showed no effect which again suggests the specificity of C10 with regards to causing mitochondrial proliferation.
In conclusion, we have considered the role the medium-chain fatty acids C8 and C10 may play with regards to explaining the efficacy of the MCT KD. In this study, we have focussed upon mitochondrial involvement. Our finding that C10 has the potential to increase cellular mitochondrial number raises the possibility that simplified/more easily managed C10-based forms of the KD diet could be developed. Support for exploring the use of C10 further also comes from recent work that has demonstrated that C10, but again not C8, may acutely control seizure activity (Chang et al. 2012). Although this acute response would suggest a different mechanism to that being considered in our study, it does point to the fact that there may be many biological targets for C10, activation of which may be beneficial with regards to seizure control.