Spatio-temporal heterogeneity in hippocampal metabolism in control and epilepsy conditions

The hippocampus’s dorsal and ventral parts are involved in different operative circuits, the functions of which vary in time during the night and day cycle. These functions are altered in epilepsy. Since energy production is tailored to function, we hypothesized that energy production would be space- and time-dependent in the hippocampus and that such an organizing principle would be modified in epilepsy. Using metabolic imaging and metabolite sensing ex vivo, we show that the ventral hippocampus favors aerobic glycolysis over oxidative phosphorylation as compared to the dorsal part in the morning in control mice. In the afternoon, aerobic glycolysis is decreased and oxidative phosphorylation increased. In the dorsal hippocampus, the metabolic activity varies less between these two times but is weaker than in the ventral. Thus, the energy metabolism is different along the dorsoventral axis and changes as a function of time in control mice. In an experimental model of epilepsy, we find a large alteration of such spatiotemporal organization. In addition to a general hypometabolic state, the dorsoventral difference disappears in the morning, when seizure probability is low. In the afternoon, when seizure probability is high, the aerobic glycolysis is enhanced in both parts, the increase being stronger in the ventral area. We suggest that energy metabolism is tailored to the functions performed by brain networks, which vary over time. In pathological conditions, the alterations of these general rules may contribute to network dysfunctions.

Energy production in brain cells is assumed to be optimized to perform tasks or activities in a brain region–specific manner (1–3). Simultaneously, neuronal activity should be adapted to minimize the energy expenditure required for its fueling (4, 5). Several metabolic pathways are available for energy production, in particular glycolysis and oxidative phosphorylation (1). Whether specific metabolic pathways are preferably recruited in different brain areas in a task-dependent manner is not known. The hippocampus is an ideal region to test this hypothesis. The dorsal hippocampus (DH) is involved in learning and memory associated with navigation, exploration, and locomotion, whereas the ventral hippocampus (VH) is involved in motivational and emotional behavior (6–8). These functions are supported by very distinct anatomical (9, 10), morphological (11–13), molecular (14–19), and electrophysiological (12, 13, 20–23) properties of hippocampal cells. The hippocampus structure is also highly heterogeneous at the gene level, from its dorsal to its ventral tip (24, 25), which may serve as a substrate for different functional networks related to cognition and emotion to emerge (7, 26, 27). Given the hippocampus’s heterogeneity, from structure to function, along its dorsoventral axis, our first hypothesis is that energy production is different between the VH and the DH.If energy production is tailored to a given structure–function relationship, we predicted that a change in energy production should accompany a change in the hippocampus’s functional state. Epilepsy is a particularly relevant situation to test this hypothesis. The different types of epilepsies are associated with numerous metabolic and bioenergetic alterations (28). Hypometabolism of epileptic regions is a common signature of mesial temporal lobe epilepsy (TLE) in humans and animal models, such as the pilocarpine model (29, 30). Importantly, in patients with TLE, only the temporal part (including the hippocampus) is epileptogenic. The temporal part corresponds to the ventral part in rodents, which has been identified as the epileptogenic region in the pilocarpine mode (31). Our second hypothesis is that any dorsoventral organization of energy metabolism found in control condition is altered in epilepsy.Finally, hippocampal functions demonstrate circadian regulation (32), in particular place cell properties (33) and long-term synaptic plasticity (34) as well as memory and learning processes (35, 36). Our third hypothesis is that energy production, specifically the respective contributions of oxidative phosphorylation and aerobic glycolysis (3), vary as a function of the time of the day in control and epilepsy.To test these three hypotheses, we used an ex vivo approach to evoke an energy-demanding electrophysiological activity standardized for the three independent variables considered here: time, region (DH/VH), and perturbation (control/epilepsy). As a first step toward a better understanding of the time regulation of hippocampal metabolism, we considered two time points during the night/day cycle: Zeitgeber 3 (ZT3) and Zeitgeber 8 (ZT8), as they correspond to low and high seizure probability in the TLE model used (37). We found that the control DH and VH have distinct time-dependent metabolic signatures regarding glycolysis and oxidative phosphorylation. In experimental epilepsy, there is no more dissociation between DH and VH at ZT3, but the regional difference reappears at ZT8.