Lactate is a potentially harmful substitute for brain glucose fuel: consequences for metabolic restoration of neurotransmission
Bibliographic record
Abstract
The metabolite lactate (L-lactate) can be generated and released by diverse brain cells, including neurons, astrocytes, and oligodendrocytes (Kann, 2023; Rae et al., 2024). Lactate production usually requires the degradation of glucose (D-glucose) – and glycogen in astrocytes – to pyruvate by glycolysis and subsequent conversion of pyruvate to lactate by the enzyme lactate dehydrogenase (Figure 1A; Dienel, 2019; Rae et al., 2024). Lactate production occurs when local glycolysis exceeds oxidative phosphorylation in mitochondria, a complex metabolic process that requires the tricarboxylic acid cycle for pyruvate utilization, the respiratory chain and molecular oxygen as an electron acceptor to finally enable adenosine-5′-triphosphate (ATP) generation (Dienel, 2019; Kann, 2023). Lactate production is generally favored during hypoxia (anaerobic glycolysis). Alternatively, lactate is generated in the presence of oxygen (aerobic glycolysis), for example, in astrocytes or in neuronal structures that miss mitochondria like a fraction of presynaptic terminals in pyramidal cell axons (Magistretti and Allaman, 2018; Kann, 2023). Aerobic glycolysis in neurons and glial cells usually results in some net lactate efflux from the active brain to the blood circulation (Dienel, 2019). By contrast to lactate, glucose is also required for essential processes such as build-up of glycogen and management of oxidative stress (Dienel, 2019; Kann, 2023).Figure 1: Neuronal energy metabolism and potential risks of lactate utilization.(A) Simplified scheme about glucose and lactate metabolism in neurons. Glucose crosses the neuronal cell membrane mainly through glucose transporters 3 and 4 (GLUT3, GLUT4) and is metabolized by glycolysis in the cytosol. Lactate is shuttled through MCT2, in symport with a proton. Depending on the metabolic conditions, lactate can be either produced (i) or fully oxidized (ii) in neurons. The complete metabolism of one glucose (six-carbon skeleton) and two lactate (three-carbon skeleton) molecules delivers about 32 ATP and 30 ATP, respectively. (B) Lactate can be provided by adjacent neurons and glial cells, including activated microglia. Lactate is elevated in blood plasma, for example, during intense physical exercise or lactate infusion. (C) Several risk factors in neurons need to be considered when lactate is used as an alternative fuel to glucose. (D) The risk factors can contribute to the occurrence of hyperexcitability (excitation-inhibition imbalance) or attenuation of activity in neural networks. Created using CorelDRAW® (Corel Corporation, Ottawa, Canada). ATP: Adenosine-5′-triphosphate; Glc: glucose; H+: proton; Lac: lactate; LDH: lactate dehydrogenase; Mito: mitochondria; O2: oxygen; Pyr: pyruvate.Lactate can be also taken up by neurons through monocarboxylic acid transporters (MCTs) and used as a supplemental brain fuel (Dienel, 2019; Rae et al., 2024). This lactate might originate from adjacent neurons, astrocytes, oligodendrocytes, or activated microglia (resident macrophages) (Figure 1B; Chausse et al., 2020; Kann, 2023). Moreover, lactate influx across the blood–brain barrier to the brain parenchyma occurs during physical exercise when working skeletal muscles release substantial amounts of lactate into the blood circulation (Rasmussen et al., 2010; Dienel, 2019). In addition, high levels of lactate exist in pathologic conditions such as sepsis, cerebral ischemia, seizures, and traumatic brain injury (Magistretti and Allaman, 2018; Kann, 2023). Lactate shuttling among blood plasma and brain parenchyma as well as among diverse brain cells is dependent on certain subtypes of MCTs that are permeable for lactate, pyruvate, and ketone bodies (Magistretti and Allaman, 2018; Rae et al., 2024). MCTs are symporters that transport a proton (H+) with each monocarboxylate along the concentration gradients (Dienel, 2019; Rae et al., 2024). Neuronal lactate uptake can potentially elicit intracellular acidification and concomitant disturbances in glycolysis and ion channel function, particularly in structures like presynaptic terminals (Munsch and Pape, 1999; Dienel, 2019; Hollnagel et al., 2020; Kann, 2023). Lactate can serve as a metabolic fuel for mitochondrial ATP generation in the utilizing cells. For this purpose, lactate is coupled to oxidative phosphorylation (see above), which initially requires conversion back to pyruvate by bidirectional lactate dehydrogenases. Metabolism of two lactate molecules in mitochondria delivers about 30 ATP, whereas one glucose molecule delivers two more ATP because of glycolysis (Figure 1A; Dienel, 2019). In neurons, most ATP is expended to power various ion pumps such as Na+/K+-ATPase and presynaptic vacuolar H+-ATPase for the establishment of neuronal excitability and synaptic transmission. Notably, the cellular consumption of oxygen rises linearly with complete lactate oxidation (3 O2 + 1 lactate → 3 CO2 + 3 H2O) (Kann, 2023). Sharp wave-ripple activity that is generated in neural networks of the hippocampus and features moderate energy demand, for example, associates with increased cerebral metabolic rate of oxygen by about 9% when glucose fuel (glucose only) is experimentally replaced by lactate fuel (lactate only) in slices of the rat hippocampus (Hollnagel et al., 2020). Lactate oxidation might also support sparing glucose for aerobic glycolysis in neurons. In addition, lactate may function as an intercellular signaling molecule that modulates neurotransmission, for example, through hydroxycarboxylic acid receptor 1 (Magistretti and Allaman, 2018; Cai et al., 2024; Rae et al., 2024). For more than 30 years, many experimental reports using dissociated primary cell cultures or the cortex in vivo have gained valuable insights into the role of lactate in the brain (Magistretti and Allaman, 2018; Dienel, 2019). These reports have limitations, however, because of the utilization of (i) cell cultures at different developmental stages, (ii) excessive glucose and/or lactate concentrations, (iii) antibiotics, and (iv) anesthetics that significantly attenuate neurotransmission and energy metabolism. These limitations have played a part in the vivid discussion about the role of lactate in neuronal energetics and signaling (Magistretti and Allaman, 2018; Dienel, 2019; Kann, 2023; Rae et al., 2024). Another set of reports explored the use of lactate in neurotransmission and neural network oscillations, mainly in slice preparations of the postnatal rodent hippocampus. Despite featuring oxygen and substrate concentration gradients from the surface to the core, hippocampal slices enable the induction of specific neural network activity states under defined experimental conditions without antibiotics and anesthetics (Kann, 2023). Among neural network phenomena, gamma oscillations (30–70 Hz) and sharp wave-ripples have been extensively investigated in hippocampal slices (Kann, 2023). Gamma oscillations appear in many cortex regions during perception, memory formation, and motor activity in vivo. They support timing of action potentials and synaptic plasticity. Hippocampal sharp wave-ripples (sharp waves overlaid by fast oscillations of > 180 Hz called “ripples”) appear during consummatory behavior and slow-wave sleep, for example, and they are thought to distribute compressed information to neocortical circuitries during memory consolidation. Gamma oscillations and sharp wave-ripples require mutual synaptic communication between excitatory pyramidal cells and inhibitory interneurons. Notably, gamma oscillations associate with much higher energy demand compared with sharp wave-ripples or network activity in the presence of anesthetic drugs (Hollnagel et al., 2020; Kann, 2023). The features of sharp wave-ripples present during energy substrate supply with glucose (glucose only) through the external recording solution were fully retained even one hour after exchange to lactate (lactate only) in hippocampal slices (Hollnagel et al., 2020). The intermittent nature of sharp wave-ripples (every 5 seconds) presumably permitted sufficient provision of ATP through diffusion to neuronal structures that depend on aerobic glycolysis (Dienel, 2019; Hollnagel et al., 2020; Kann, 2023). This observation reconciled some previous reports on lactate use that applied artificial electric stimuli at a low rate (every 60 seconds) to evoke synaptic neural population responses (Izumi et al., 1997; Kann, 2023). Likewise, lactate could serve as a glucose fuel alternative for neurons under anesthesia in vivo (Wyss et al., 2011). These findings indicate that lactate can substitute glucose during neural network activity states showing lower energy demand. Replacing glucose with lactate in the presence of gamma oscillations, however, elicited recurrent neural bursts in hippocampal slices (Figure 1C; Hollnagel et al., 2020). Bursts reflect hyperexcitability because of excitation-inhibition imbalance in neural networks and might facilitate the onset of epileptic seizures (Figure 1D). The bursts were widely prevented by raising the glucose fraction in energy substrate supply (lactate/glucose ratio of < 2:1). Likewise, calculations indicated that transport-saturating levels of lactate in the blood plasma can supply up to 60% of total oxidative energy metabolism, with glucose providing the rest (Dienel, 2019). In slice cultures, lactate (lactate only) attenuated gamma oscillations by reducing amplitude rather than changing frequency. Theta-gamma oscillations evoked by optogenetic tools also showed lower amplitudes in lactate (lactate only), whereas the frequency was resistant (Hollnagel et al., 2020). Similarly, replacing glucose with lactate partially attenuated synaptic population responses evoked by electric stimuli at a higher rate (every 2 seconds) (Cox and Bachelard, 1988). Single and paired patch-clamp recordings in identified glutamatergic pyramidal cells and fast-spiking, GABAergic interneurons revealed that lactate (lactate only) could attenuate neurotransmission by reducing transmitter release from presynaptic terminals, whereas action potential generation in axons was normal (Hollnagel et al., 2020). These findings indicate that lactate is only a partial substitute for glucose during neural network activity states showing high energy demand. Moreover, high lactate/glucose ratios are potentially harmful, possibly through the lack of some glycolytic ATP generation in certain presynaptic terminals. Various pathophysiological conditions and brain diseases have been reported to associate with increased lactate levels, albeit to a variable extent. Examples are exhaustive physical exercise, brain infection, seizures, Alzheimer’s disease, depression, and schizophrenia (Dienel, 2019; Zebhauser et al., 2022; Kann, 2023; Cai et al., 2024; Rae et al., 2024). Infusion of lactate (Na+-lactate) has been used, for example, in traumatic brain injury, insulin-induced hypoglycemia, and glucose transporter 1 deficiency syndrome for therapeutic intervention (Magistretti and Allaman, 2018; Kann, 2023; van Gemert et al., 2023) but also in some rodent models for experimental induction of anxiety-like responses (Magistretti and Allaman, 2018; Cai et al., 2024). The beneficial or harmful outcomes of increased lactate/glucose ratios on neurotransmission, neural network oscillations, cognition, and behavior, however, are critically dependent on the pathophysiological context. Medium lactate/glucose ratios can fuel neural network activities showing high energy demand such as cortical gamma oscillations when the oxygen supply is sufficient (Hollnagel et al., 2020). During intense physical exercise, for example, skeletal muscle activity increases lactate in blood plasma severalfold, which is taken up by the brain. This lactate can partially replace glucose fuel in neurotransmission and network oscillations (Rasmussen et al., 2010). Similarly, the infusion of lactate recovered cognitive tasks and stimulation-induced neural population responses, lowered the release of stress hormones, and attenuated autonomic symptoms during insulin-induced hypoglycemia in healthy subjects (Kann, 2023 and references therein). In patients with glucose transporter 1 deficiency syndrome, in which glucose transport across the blood–brain barrier is limited, the infusion of lactate associated with increases in serum lactate, sodium, and pH and partially suppressed epileptic discharges (van Gemert et al., 2023). High lactate/glucose ratios can fuel neural network activities showing lower energy demand like sharp wave-ripples or neural networks in anesthesia (Wyss et al., 2011; Hollnagel et al., 2020). In severe hypoglycemia, however, infusion of lactate was ineffective, whereas glucose rapidly reversed symptoms such as stupor and coma (Dienel, 2019 and references therein). Notably, high lactate/glucose ratios become risky when neural networks express activities with high energy demand (Figure 1C and D; Cox and Bachelard, 1988; Hollnagel et al., 2020). In addition, limited oxygen availability in mitochondria significantly disturbs lactate oxidation and thus ATP generation. Exhaustive physical exercise, for example, associates with moderate hypoglycemia and reduced cerebral oxygenation, and brain lactate uptake may even supersede glucose uptake (Rasmussen et al., 2010). In this condition, lactate becomes an important supplemental fuel but might also contribute to central fatigue (Dienel, 2019; Kann, 2023). Elevated levels of nitric oxide (NO) can also inhibit the respiratory chain in mitochondria (Chausse et al., 2020). NO production already occurs through neuronal NO synthase during neurotransmission in physiological conditions. In neuroinflammation, however, high NO levels can result from activated microglia that upregulate inducible NO synthase and switch to aerobic glycolysis with concomitant lactate release (Chausse et al., 2020; Rae et al., 2024). In this context, high lactate/glucose ratios might become harmful by eliciting hyperexcitability in neural networks (Hollnagel et al., 2020). Similarly, seizures and coma are well-known clinical sequelae in hypoglycemia, an adverse effect of insulin and sulphonylureas in the treatment of diabetes mellitus. Hyperexcitability might reflect differences in glucose and lactate use in excitatory principal cells and inhibitory interneurons (Kann, 2023). Conversely, these neuronal subtypes likely feature different susceptibility to high lactate/glucose ratios, local hypoxia, and NO levels. In depression and schizophrenia, the mechanisms underlying increased lactate levels are not well defined and require further systematic investigation. Moreover, the reported effects of lactate infusion on neurotransmission in cortical networks in both diseases are complex and difficult to interpret (Magistretti and Allaman, 2018; Cai et al., 2024). In Alzheimer’s disease, increased lactate levels in the cerebrospinal fluid were identified in earlier disease stages and might reflect the presence of inflammatory processes (Zebhauser et al., 2022). The above considerations of basic research and clinical medicine suggest that therapeutic infusion of sodium lactate in patients requires careful attention to the individual pathophysiological context as well as monitoring of lactate, glucose, electrolytes, and pH in the blood. Moreover, future experimental research may address (i) expression of lactate dehydrogenases and MCTs in neuronal subtypes and compartments, (ii) metabolic rates of lactate and glucose, including oxygen availability during different neural network activity states, and (iii) mitochondrial (dys)function in animal models of neurologic and psychiatric diseases. The authors thank Petra Schling (Heidelberg University Biochemistry Center, Germany) for the helpful discussion. C-Editors: Zhao M, Liu WJ, Qiu Y; T-Editor: Jia Y
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".