Imaging brain aerobic glycolysis as a marker of synaptic plasticity

Functional brain imaging techniques such as positron emission tomography (PET) or functional magnetic resonance imaging (fMRI) provide a unique opportunity to study the brain at work. These techniques detect metabolic and vascular signals that are coupled to changes in neuronal activity, thus affording the possibility to localize brain areas, changing their level of activity during particular behavioral modalities (1). They also inform about basal and activated metabolic states, particularly PET, which can monitor with appropriate tracers the metabolic rates for glucose and oxygen, as well as blood flow. With fMRI, in addition to activity changes, it is possible to identify the degree of functional and anatomical connectivity existing between a given brain region, taken as a seed, and other brain areas.

In a study by Shannon et al. (2) published in PNAS, the use of PET and fMRI imaging techniques has not only been extended to monitor metabolic and connectivity changes during a task, as such imaging studies usually explore, but also to identify with these techniques the metabolic correlates of learning and plasticity associated with a complex visuomotor task in healthy human subjects. The underlying hypothesis that provided the rationale to use PET and fMRI to study plasticity and learning stemmed from a series of experiments previously carried out by the same group that identified the occurrence of aerobic glycolysis, a particular metabolic pathway of glucose, in conditions of high synaptic plasticity and remodeling (3). Aerobic glycolysis occurs when glucose utilization exceeds oxygen consumption, resulting in the production of lactate from glucose despite the presence of adequate oxygen concentrations. It is also known as the Warburg effect, and is a metabolic hallmark of cancer cells (4).

Previous work by Marc Raichle and his colleagues has shown, using PET to image metabolic parameters, that during early stages of development, aerobic glycolysis levels are highest and present throughout the brain, with a peak at 10 y of age. This metabolic profile correlates with a high level of expression for genes involved in synaptic plasticity, growth, and remodeling (3). Interestingly, in adulthood, aerobic glycolysis becomes restricted to certain brain areas such as the superior and medial frontal gyrus, the posterior cingulate cortex, the dorsolateral prefrontal cortex, and the precuneus, where aerobic glycolysis accounts for 25% of glucose utilization, whereas in other areas such as the cerebellum, aerobic glycolysis is barely detectable (5). Interestingly, these areas with high aerobic glycolysis are the sites of intense expression for plasticity genes. Based on these observations (3, 5), the authors hypothesized that an association existed between aerobic glycolysis and plasticity.

In PNAS, Shannon et al. (2) demonstrate that aerobic glycolysis is indeed enhanced in brain areas that undergo plasticity during a learning task. The learning paradigm consisted of an out-and-back reaching task whereby the subject was requested to connect, using a stylus, a center circle to one of eight equally spaced peripheral circles on a screen. One group served as a control (C), whereas for a second rotation (R) group, the task was perturbed by covertly and gradually rotating the mapping between the stylus and the display screen, thus imposing a learning condition.

In a first set of experiments using fMRI, Shannon et al. (2) show that this complex visuomotor learning task results in the specific activation of Brodmann area 44 (BA44), an area generally mobilized by complex motor tasks. Having identified the area that is activated during this complex learning task, the authors went on to explore the metabolic profile of this area before and after the task, using PET. They monitored blood flow, glucose utilization, and oxygen consumption. A remarkable observation was that in the R group, which had to undergo adaptation, increased glucose utilization accompanied by a decrease in oxygen consumption was observed after the task. Such a metabolic profile represents the signature of aerobic glycolysis. Interestingly in the C group, an opposite metabolic profile was observed, with an increase in blood flow and in oxygen consumption.

Posttask Aerobic Glycolysis

Previous reports have indicated an increase in aerobic glycolysis during task performance (6), and, indeed, in the study by Shannon et al. (2), several motor and visual areas were activated during the visuomotor task. However, only BA44 showed a sustained posttask increase in aerobic glycolysis, indicating that this metabolic behavior is associated with learning-induced processes rather than simply providing additional substrates to match increased task-dependent energy demands.

Aerobic Glycolysis and Glia

The question then arises of the cellular processes that underlie such a metabolic profile in relation to plasticity and learning. A distinctive feature of aerobic glycolysis is that it results in the formation of lactate, despite adequate levels of available oxygen. Aerobic glycolysis triggered by glutamate uptake into astrocytes and resulting in lactate release has been proposed as a mechanism to couple neuronal activity to glucose utilization (1). This process, known as the astrocyte-neuron lactate shuttle, provides a mechanism to deliver lactate as an energy substrate to meet the energetic demands of activated neurons. More recently, lactate has been shown to be more than a metabolic substrate and to play a key role in plasticity and learning (7, 8). Thus, blocking the transfer of lactate from astrocytes to neurons impairs memory consolidation (7). Molecular analysis of this action of lactate indicates that the monocarboxylate acts as a signal for plasticity by inducing the expression of a variety of plasticity genes, such as Arc, Zif 268, and BDNF (9). In the context of the experiments reported by Shannon et al. (2), it is likely that the lactate produced by the sustained aerobic glycolysis provides a signal for plasticity occurring in BA44. This effect, along with the known role of aerobic glycolysis in providing molecular building blocks for biosynthesis (10, 11), may converge to support synaptic remodeling related to learning.

Shannon et al. (2) provide an additional thought-provoking hypothesis, also involving the role of glia, to explain the sustained aerobic glycolysis that occurs in BA44 following learning associated with the complex visuomotor task. They point to a possible role of activated microglia. Indeed, recent evidence indicates that microglia, in addition to their well-established role in inflammation and immune competence, can play a physiological role in synaptic remodeling during development and plasticity (12, 13). They base this hypothesis on a metabolic consideration that is fully consistent with the aerobic glycolysis observed in BA44. Indeed, microglia, like several other immune-competent cells, shift their energy metabolism from an oxidative one to a predominantly glycolytic one when they become activated (14, 15). Thus, it may well be that the observed increase in aerobic glycolysis associated with plasticity events in BA44 also reflects, in part, a physiologically activated state of microglia that contributes to synaptic remodeling.

The article by Shannon et al. (2) provides further evidence that energy metabolism and synaptic function are tightly coupled not only during activation, a phenomenon that has provided the physiological basis for functional imaging techniques, but also during the periods of intense synaptic remodeling associated with learning that follow activation. They also provide strong evidence for the use of aerobic glycolysis as detected by PET as a marker of synaptic plasticity. Because aerobic glycolysis is a metabolic pathway mainly localized in glia cells, both astrocytes and microglia, these results provide additional justification to reconsider the role of glia not only in providing energy support to neurons but also as active players in synaptic plasticity and learning.