How to balance the brain energy budget while spending glucose differently

The brain is an organ with high energy demands. Thus, while it represents only 2 % of the body weight, it contributes up to 20 % of its resting metabolism. Glucose is considered the almost exclusive blood-borne energy substrate utilized by the adult brain to fuel its activity. Most of the energy necessary for brain function is derived from the full oxidation of glucose. Recently, calculations of the energy requirements for the different components involved in brain activity have led to the conclusion that the vast majority of energy expenditure (81 %) is devoted to neuronal excitatory signalling (Attwell & Laughlin, 2001). When estimations were made on a cellular basis, it was concluded that no more than 5 % of energy usage can be attributed to glial function while the rest (comprising resting consumption + signalling costs) is accounted for by neurons. Although these calculations do not allow any prediction about the mechanims involved in energy production, their distribution and proportion among various cell types nor the possibility of metabolite exchange, the aforementioned conclusions about the prominent role of glucose oxidation seriously restrict the possibilities. As a logical consequence, one would predict that the proportion of glucose utilized by neurons should approach 95 % while the remaining 5 % should be enough to satisfy the modest glial energy needs (Fig. 1A).

Figure 1.

A, classical brain energy budget; B, new brain economy with lactate exchange.

In this issue of The Journal of Physiology, Vega et al. (2003) studied the uptake and distribution of locally applied deoxyglucose, glucose or lactate in a rat vagus nerve preparation. Using a particularly elegant method to track the radioactively labelled metabolites and to model their diffusion as well as their distribution between two compartments (axons and Schwann cells), they come to the unexpected conclusion that approximately 78 % of the labelled 2-deoxyglucose (and by extension of glucose utilization) occurs in Schwann cells, the counterparts of astrocytes in the peripheral nervous system. How can this be possible without violating the energy budget rules? One rule-breaking possibility could be that the energy-requiring processes and energy-producing pathways used in the peripheral nervous system are different from those in the central nervous system. This does not appear to be the case since it has long been known that active transport of ions necessary to re-establish the electrochemical gradients dissipated by action potential propagation is the predominant ATP-consuming process in nerves, also accounting for almost all oxygen consumption in this preparation (Ritchie, 1967).

A possible solution to this apparent paradox is offered by previous work suggesting the existence of a net lactate transfer from glial cells to neurons, known as the lactate shuttle hypothesis (Magistretti et al. 1999). Thus, it was postulated that in the central nervous system, astrocytes could respond to increased synaptic activity via glutamate uptake by increasing their rate of aerobic glycolysis, i.e. glucose consumption and lactate production (Pellerin & Magistretti 1994). After its release from astrocytes, lactate would be taken up by neurons via a specific monocarboxylate transporter and used as energy substrate. Mounting evidence has been provided indicating that lactate could constitute a significant energy substrate for neurons both in peripheral nerves and in the brain. Moreover, the existence of a lactate shuttle between cells within the same tissue is not unique to the nervous system since it has been already reported in other tissues (e.g. striated muscle), suggesting that it might be a rather universal concept (Brooks, 2002). Data presented by Véga et al. (2003) strongly indicate that a net transfer of energy substrate must be taking place between Schwann cells and axons, and points at lactate as the likely candidate.

What is the consequence of this alternative distribution of glucose utilization on the ratio of ATP production between neurons and glial cells? Glycolysis occurring in glial cells would provide 2 ATP and lead to the production of 2 lactate per glucose consumed. In return, oxidation of 2 lactate by the neuron (or nerve) could potentially provide up to 36 ATP. Interestingly enough, this distribution of ATP production respects the previously established energy budget with ≈5 % of the ATP being produced in the glial cell while the rest would be generated from lactate in the neuron (Fig. 1B). Of course, peripheral nerves and Schwann cells may represent an extreme case of metabolic compartmentation, like the retina where glial cells are almost exclusively glycolytic while neuronal cells are highly oxidative and glucose is predominantly taken up by glial cells while released lactate is used by neurons (Poitry-Yamate et al. 1995). It is likely, however, that even in the central nervous system a certain degree of such metabolic preference exists and that lactate exchange occurs to maintain an appropriate energy substrate supply to the energy-demanding neuron (see the recent article by Ainscow et al. 2002). Moreover, the proportion of glycolytically consumed glucose by glial cells and lactate produced for utilization by neurons may well vary according to the level of neuronal activity. A signal triggering such a metabolic response between astrocytes and neurons both in the central nervous system and in the retina has been identified as the excitatory neurotransmitter glutamate. Whether a similar signal is sent from axons to Schwann cells is not known although some candidates such as K+ could be proposed. The dogma that glucose is fully oxidized by the brain is still valid. However, new data obtained in experimental models affording a cellular resolution strongly suggest a contribution from glial cells via aerobic glycolysis that will provide lactate to meet the energy demands of nerve cells where the monocarboxylate would be fully oxidized.