Abstract
Martínez-François JR, Fernández-Agüera MC, Nathwani N, Lahmann C, Burnham VL, Danial NN, Yellen G. eLife 2018;7:e32721.
Brain metabolism can profoundly influence neuronal excitability. Mice with genetic deletion or alteration of Bad (
Commentary
Medical students often struggle to memorize the “canonical” pharmacological mechanisms of anticonvulsant drugs. Although only a single answer is deemed “correct” on the students’ multiple-choice exams, the anticonvulsant benefits of many of these drugs often result from their actions on multiple receptors. Since small-molecule anti-seizure drugs act on several targets, we would expect that treatments that alter the entire body's metabolic state, as occurs with the high-fat, low-carbohydrate ketogenic diet (KD) would work through multiple biochemical pathways. Indeed, decades of research have revealed multiple putative mechanisms of KD action (1). Despite this complexity, there is great interest in elucidating the major mechanistic KD pathways that could be directly targeted to achieve a better-tolerated “ketogenic diet in a pill” (2) to provide relief to the many patients who find the KD cumbersome and unpalatable (3).
One postulated therapeutic mediator of the KD is the ATP-sensitive potassium channel (KATP), an ion channel that couples cellular energetics to electrical activity. At low intracellular ATP concentrations, KATP channels open and conduct outward K+ currents to produce more-negative resting membrane potentials and thus less-excitable neurons; higher levels of ATP close KATP channels and the resting membrane potential becomes more positive (4). Ex vivo brain slice experiments performed in conditions that mimic KD hypoglycemic and ketotic states (i.e., with low extracellular glucose or in the presence of exogenous ketone bodies) found that KATP channels were activated in these conditions (5, 6). Interestingly, the authors of the current paper had previously identified another cellular energetics pathway that modulates KATP currents in normal extracellular glucose concentrations and without exogenous ketone bodies. They found that that both unconditional deletion and phosphorylation site mutation of BCL-2 Agonist of Cell Death (BAD), a protein that modulates mitochondrial glucose metabolism in heathy cells (7)—as well performing its better-known role in mediating apoptosis in damaged cells—increased ketone body concentration and KATP currents in vitro in the presence of typical electrophysiological extracellular solutions (8). Moreover, in vivo, the loss of BAD function reduced pharmacologically evoked seizures, a protective effect that was lost in KATP deletion mice (8).
Although the previous BAD deletion/mutation experiments were performed without explicitly changing glucose or ketone concentrations, one might wonder if the unconditional manipulation of BAD (a protein involved in glucose metabolism in many tissues) might indirectly produce its effects on KATP activity apart from the brain regions of interest. To test this possibility, the authors of the current paper virally delivered recombinant BAD into the hippocampi of BAD−/− mice. Hippocampal slice patch-clamp recordings of dentate granule neurons performed in the presence of typical glucose concentrations revealed that only neurons reconstituted with recombinant BAD (but not untransformed neurons) exhibited ATP-sensitive increases in KATP activity, a result demonstrating regarding the effects of BAD deletion were cell-autonomous.
While it seems intuitive that the previous finding that unconditional BAD deletion reduces seizures resulted from increased KATP currents, there was no direct evidence that the increased KATP activity reduced neuronal excitability. Therefore, here, the authors directly tested the effects of BAD deletion on KATP-dependent changes in neuronal firing in hippocampal slices from wild-type and BAD−/− mice. Importantly, they used perforated-patch recordings, a technique that makes electrical contact with the cell without altering its intracellular composition including its internal ATP concentration. Compared with wild-type neurons, BAD−/− neurons exhibited significantly fewer action potentials and were therefore less excitable. Pharmacological blockade of the KATP channel eliminated the inhibitory effect in BAD−/− neurons but did not increase the excitability of wild-type neurons, results that suggest that the KATP channel mediates the inhibitory effects if BAD deletion.
Finally, the investigators demonstrated that BAD−/− deletion would reduce seizure-like activity in hippocampal slices (specifically, picrotoxin-evoked spiking of calcium-associated GCaMP6 fluorescence), a result consistent with their previous in vivo experiment that showed resistance of BAD−/− mice to kainic acid-evoked seizures (8). Perhaps more interesting, the current study also found that conditional deletion of BAD within the dentate gyrus was sufficient to produce the full inhibitory effect on the ex vivo seizure-like-activity. Therefore, BAD deletion-mediates upregulation of KATP activity within a key node of an epilepsy network that may be sufficient to prevent seizures.
The results of this paper expanded the authors’ prior work (8) and provided some important new information concerning the links among BAD, KATP, and seizures. We now know that BAD deletion enhances KATP currents within an individual neuron (as opposed to requiring an organism- or brain-wide effect), an important finding given that KD- or BAD-mediated changes in metabolic state could certainly have a multitude of global effects. This research also further clarified the anticonvulsant mechanism of BAD deletion by revealing the enhancement of KATP currents and, at least for this seizure-like model, the sufficient localization of the effect to the dentate gyrus.
The demonstration that conditional deletion of BAD within the dentate gyrus sufficiently inhibits seizure-like activity in hippocampal slices suggested intriguing translational/therapeutic possibilities. Potentially, direct microinfusion of BAD or KATP modulators to key epileptic networks nodes could prevent seizures without producing the side effects associated with systemic administration. However, because the KD is often prescribed for patients with extra-limbic seizures, including generalized or multifocal seizures in patients with epileptic encephalopathy syndromes (9), we need to first understand the effects of BAD and KATP manipulations within these other seizure networks as well as the hippocampus.
Additional work also needs to be done to understand the mechanistic relationships among alterations of systemic metabolism, as occurs in the KD, BAD activity, and KATP currents. The molecular underpinnings that mediate the effects of BAD on glucose homeostasis have been best-elucidated in hepatocytes and pancreatic β cells and these pathways are undoubtedly affected with the altered serum glucose concentrations that occur with the KD (7). Apart from the liver and pancreas, phosphorylation of Ser155 of neuronal BAD, the subject of this study, causes the neuronal mitochondria to catabolize glucose rather than ketone bodies (7, 8). However, we do not know if systemic hypoglycemia or ketosis affects BAD in directing this switch in mitochondrial fuel. In addition, we also do not know if BAD signaling is needed for extracellular low-glucose or ketone bodies to enhance KATP currents (5, 6).
Determining the interactions between the KD and neuronal BAD will not only provide mechanistic insights into the coupling of neuronal energetics and excitability but will also be necessary to translate discoveries into potential therapies. Does neuronal BAD inhibition or direct KATP enhancement contribute only a small portion to the total anti-seizure effects of the KD or closely match KD efficacy without requiring systemic hypoglycemia and ketosis? With the latter possibility, epilepsy patients may truly realize some of the benefits of the promised “ketogenic diet in a pill,” Moreover, countless medical students will have the “opportunity” to memorize yet one more pharmacological mechanism.