Abstract
In the central nervous system, lactate is formed under both aerobic and anaerobic conditions. The question is whether the generated lactate after neural excitation by glutamate is immediately metabolized. It has been hypothesized that lactate is formed after glutamate uptake in the glia, then released into the intercellular compartment and subsequently used by neurones (the so-called astrocyte-neuron lactate shuttle; Pellerin and Magistretti, 1994). In two recent reports (Schurr, 2006; Aubert et al, 2005), the question was addressed whether lactate could serve as a substrate for oxidative metabolism during nerve cell activation under aerobic conditions in vivo. Aubert et al (2005) designed a mathematical model to show that the time course of extracellular changes of lactate levels during enhanced neural activity could be interpreted as aerobic metabolism. Schurr (2006) summarized the arguments favouring the idea that in the functioning brain, lactate is the major end product of both aerobic and anaerobic glycolysis and that lactate subsequently serves as an oxidative substrate. For both reports, the results shown by Hu and Wilson (1997) are crucial. In that study, the perforant pathway of the rat was simulated for 5 secs once or repeatedly and the time course of extracellular lactate, oxygen, and glucose in the dentate gyrus of the rat brain was measured with rapidly responding biosensors. A single stimulation showed an initial decrease of all the analytes, followed by transient increases. After repeated stimulations, extracellular lactate was substantially increased, with a concomitant decrease of glucose and little, if any change of oxygen. This study and several others (e.g., De Bruin et al, 1990; Kuhr et al, 1988; Van der Kuil and Korf, 1991; Krugers et al, 1992) show that lactate is formed under conditions of enhanced neural activity which is—at least in part—associated with lower extracellular glucose. The time course of hippocampus lactate after a single stimulation (lasting 5 secs; Hu and Wilson, 1997) and detected with the biosensors is very similar to that seen with continuous flow analysis of microdialysates in rats, subjected to a single electroconvulsive stimulus with an intact entorhinal-hippocampal glutamatergic pathway (Krugers et al, 1992). Modelling shows that such a time course can best be described by a very rapid (nearly immediate) increase of intracellular lactate that is subsequently released into the extracellular compartment via a carrier-mediated process (Kuhr et al, 1988).
To appreciate the results of in vivo studies, a few cautionary remarks should be made. It should be realized that a lactate level is the net result of appearance and disappearance, which is a composite of the rates of lactate influx and efflux across the blood brain barrier, lactate influx and efflux across the membranes of brain cells, lactate formation from glucose and glycogen, lactate oxidation in any brain cells, diffusion to and from the site of the sensor, and also the size of extracellular space in brain and blood lactate concentration. The same factors also determine glucose and oxygen levels, with the obvious differences in transport capacity and diffusion capability. Changes in extracellular space are emphasized below as an example, but the other factors also need careful consideration. Both electroconvulsive shock (used by Kuhr et al, 1988) and electrical stimulation (applied by Hu and Wilson, 1997) of cerebral neuronal pathways must be seen as an uncommon (and perhaps even an artificial) condition, as in an intact brain it is unlikely that all neurons of a pathway fire are synchronous and at the same rate. It is also of interest to mention here, that Hu and Wilson (1997) did not observe changes in oxygen, lactate, and glucose at low stimulation for periods shorter than 5 secs. Apparently, the observed metabolic effects seen by electroconvulsive shock or electrical stimulation for 5 secs are not necessarily representative of those in a normally functioning brain.
In my opinion, the results of Hu and Wilson (1997) do not give convincing arguments to support the idea that a substantial portion of lactate formed after neural activity is immediately used as an aerobic substrate. So, it is questionable whether these experiments support the so-called astrocyte-neuron lactate shuttle hypothesis.
(1) The initial decrease of lactate after a single stimulation is very small as compared with the subsequent increase. The Hu and Wilson study (Figure 1; same as Figure 1 in the Aubert paper) shows that the area under the curve of the decrease in the levels of lactate immediately after a single stimulus is only approximately 1% of that of the subsequent rise of the lactate. The percentage (relative) decrease of lactate is somewhat smaller than that of glucose and oxygen, but this can easily be understood because of the rapid release of lactate into the extracellular compartment, as previously been demonstrated (Kuhr et al, 1988). Yet, this minor decrease has served as an argument to support aerobic metabolism of lactate.
(2) The initial decrease of lactate seen after a single stimulation becomes larger after repeated stimulation, when the levels of lactate are enhanced. The area under the curve of the decrease is increased more than 5 × . So, if the early decrease is attributed to oxidative metabolism then metabolism of lactate is apparently substantially increased per stimulus. Surprisingly, there is not less glucose metabolized, as the concomitant early dips in glucose remain the same, independent of the changes in lactate. Also, there is no change in the oxygen signal, as would be expected if the early drop of lactate is due to an increased oxidation concomitant with unaltered glucose consumption.
(3) Figure 2 of the Hu and Wilson study (1997) shows that virtually all the fluctuations of the recordings of glucose and oxygen coincide within seconds. So, after a single stimulation of 5 secs between 12 and 25 secs, transient decreases are seen of lactate, oxygen, and glucose, followed by increases again in precisely the same time frame and an apparent normalization of the analytes with a maximum after 35 secs. The relative changes are in the order of 5% for oxygen and glucose and of 12% for lactate. This observation points to a common mechanism, rather than to sequential processing of the energy substrates. This analysis is applicable after the second stimulus onwards and refers to the subsequent stimuli: the lactate changes follow glucose and oxygen in the first stimulus, but lactate starts then to rise to levels of approximately 170% of the baseline.
The question is what kind of common mechanism is possible. Essential for quantitative interpretation of the biosensor recordings is that the diffusion of analytes remains constant in the extracellular compartment. In studies on ischemia, it is clear that the shrinkage of the extracellular space, as assessed with impedance measurements, affects lactate diffusion to the dialysis probe. Shrinkage of the extracellular compartment, as the consequence of neuronal activity and associated swelling, has also been seen both in vivo and in vitro preparations.
Two recent examples are given.
In vitro, rapid increases in cellular volume as after a short stimulation of 2 secs are found (see Holthoff and Witte (1996)). There was a rapid increase of extracellular potassium (within 2 secs), that was back at the resting levels within 25 secs. The levels of an extracellular marker (tetramethylannoniumions), and also light scattering were similarly increased within 2 secs and normalized after 20 to 40 secs after the stimulation. The extracellular space shrinks by approximately 30% within 2 secs. In vivo, related observations have been reported. Changes in the water apparent diffusion coefficient during neuronal activation reflect transient microstructural changes of the neurons or glia during activation. Darquie et al (2001) showed transient decreases in the diffusion of water in the human occipital cortex during visual stimulation. Such decreases coincided with an increase in the blood oxygenation level-dependent signal and was initiated at approximately 15 secs and reached a minimum approximately 45 secs after the onset of the visual stimulus. The study of Darquie and coworkers matches with the study of Mangia et al (2003), measuring lactate with time-resolved proton magnetic resonance spectroscopy after brief visual stimulation (event-related design). They observed an even more rapid decrease of lactate (within 5 secs) and a normalization with 12 secs: a time frame also close to the in vitro observations (Holthoff and Witte, 1996). In this experiment, the stimulus was given repeatedly, so there must be consistently increased levels of lactate during the experiment. The Mangia study may point to a rapid extrusion of lactate during or after increased neural activity.
I conclude that it is unlikely that the initial dip in lactate as reported in the Hu and Wilson study is due to oxidative metabolism. It is not here to state that lactate cannot be used in the oxidative pathway. In fact, Hu and Wilson (1997) reported a slow transient decrease in the levels of extracellular glucose after stimulation concomitant with a longer lasting increase in lactate levels and constant oxygen levels. This may indeed point to the possible utilization of lactate instead of glucose as an oxidative substrate 10 mins after the onset of stimulation onward. Anyway, the possible oxidative lactate metabolism is a relatively slow pathway.
The most consistent interpretation of the mentioned studies is that most glucose used through the glycolysis proceeds directly through the oxidative pathway, and that some glucose is converted to lactate, in particular during high neuronal activity. Very similar time frames of extracellular lactate were seen in the muscle, an organ that should inspire us, ʻbrainy investigatorsʼ to think (Schurr, 2006). In this organ, the cellular export of lactate during exercise or after stimulation of the muscle starts to rise after approximately 30 secs to 1 min (e.g., De Boer et al, 1991).
In vivo estimates of the turnover of lactate and glucose in the freely moving rat suggest that a minor proportion of these substrates traffic through the extracellular space (Rhemrev-Boom et al, 2005). It is likely that a substantial proportion of cerebrally formed lactate disappears (un-metabolized) via the blood circulation (Leegsma-Vogt et al, 2003). This may also explain the often observed discrepancy between glucose and oxygen uptake and metabolism in brain during activation (Leegsma-Vogt et al, 2004). The general concept of brain metabolism is perhaps that we should not intend to localize sequential energy-related processes in a few cells. Rather, the cellular heterogeneity of the brain should be to taken into consideration to explain parallel metabolic pathways.