Ketone Fuel Keeps Mouse Brain Tissue Running When Glucose Is Gone

What has remained genuinely unresolved is whether ketones can, on their own, keep brain tissue working when glucose is entirely absent. In living animals that question is nearly impossible to answer, because even in prolonged starvation the body manufactures glucose through gluconeogenesis and keeps supplying it to the brain.

The study, published as a Brief Report in the journal Cellular and Molecular Neurobiology, examined two linked questions: whether mitochondrial respiration persists after hours of glucose-free, BHB-rich incubation, and whether the tissue retains enough energetic capacity to survive spreading depolarization, arguably the most energetically demanding event that occurs in the brain. Spreading depolarization, often abbreviated SD, is a self-propagating wave of near-complete breakdown of the transmembrane ion gradients that neurons painstakingly maintain. Restoring those gradients afterward requires substantial ATP-dependent ion pumping, which is precisely why the researchers chose it as a severe stress test for fuel-deprived tissue.

The experimental design was elegantly simple. Horizontal slices of mouse neocortex, 400 micrometers thick, were prepared from male and female C57BL/6 mice aged three to nine weeks and randomly assigned to artificial cerebrospinal fluid containing either 10 millimolar glucose, 10 millimolar racemic BHB, or 10 millimolar sucrose. The sucrose condition served as a no-exogenous-fuel control: brain tissue lacks the enzymatic machinery to split sucrose into metabolizable monosaccharides, so sucrose preserved solution osmolality without supplying any energy. All three solutions were osmolality-matched, with pH held between 7.3 and 7.4, and the BHB and sucrose solutions were made by substituting glucose with equimolar sodium β-hydroxybutyrate or sucrose.

For the metabolic measurements, slices were maintained for four hours at 37 degrees Celsius in their respective solutions, then homogenized and loaded into an Oroboros O2k high-resolution respirometer. The researchers ran a substrate-uncoupler-inhibitor titration protocol, stepping through non-phosphorylating leak respiration supported by glutamate and malate, ADP-driven complex I-linked oxidative phosphorylation, total oxidative phosphorylation capacity after adding succinate, maximal uncoupled electron transfer capacity induced with the uncoupler CCCP, and complex II-only electron transfer after rotenone inhibition. Cytochrome c was added to check mitochondrial outer membrane integrity, and preparations showing more than ten percent stimulation were excluded. Residual oxygen consumption was subtracted to isolate true mitochondrial flux.

The result was consistent across every respiratory state measured: oxygen flux was significantly lower in BHB-incubated tissue than in glucose-incubated tissue. Complex I-linked leak respiration was reduced, as were complex I-supported and complex I-and-II-supported oxidative phosphorylation capacity, maximal electron transfer capacity, and complex II-linked electron transfer capacity. Tissue incubated with sucrose alone, relying purely on endogenous energy reserves, showed a further reduction in oxygen flux across all coupling-control states compared with glucose. In other words, mitochondria kept working after four hours of exclusive BHB fuel, but they breathed more slowly than their glucose-fed counterparts.

The authors are careful about what that reduction does and does not mean. Different catabolic substrates can support similar power outputs at different respiration rates because of differences in coupling efficiency, proton leak, ATP-to-oxygen stoichiometry, and the thermodynamic state of ATP hydrolysis. Ketone metabolism has been proposed to increase mitochondrial coupling efficiency and ATP flux, and lower oxygen consumption following BHB incubation might reflect a more efficient metabolic strategy. But the researchers explicitly note that these mechanisms were not tested in this study, and the lower flux could equally reflect a genuine reduction in ATP-generative capacity, an adaptation to reduced energetic demand, or an acute suppressive effect of BHB on respiratory capacity.

The second, and arguably more striking, set of experiments tested whether BHB-fed tissue could withstand repeated spreading depolarizations. Slices were held in an interface chamber at 34 to 37 degrees Celsius, perfused continuously with their respective solutions, and monitored with intrinsic optical signal imaging, which captures the light-scattering changes as the depolarization wave sweeps across the neocortex. High-potassium artificial cerebrospinal fluid, with potassium raised to 26 millimolar, was perfused for up to seven minutes to trigger an SD, followed by ten minutes of recovery in physiological potassium solution before the cycle was repeated.

Every slice in all three groups generated a first spreading depolarization, demonstrating that endogenous energy reserves alone were sufficient for one initial energetic challenge. The decisive difference appeared at the second induction. Slices supplied with glucose or BHB underwent and recovered from a second SD, while every sucrose-treated slice, cut off from any exogenous fuel, failed to generate one. In statistical terms, the group difference in second-cycle SD occurrence was highly significant, with pairwise Fisher’s exact tests after Bonferroni correction showing glucose versus sucrose and BHB versus sucrose both at p = 0.0006. During the first cycle, latency to SD and propagation speed did not differ significantly among groups, and all groups returned to baseline reflectance, consistent with reversal of the cellular swelling that accompanies SD.

The second cycle revealed subtler differences. BHB-treated slices showed a significantly longer latency to SD onset and a slower propagation speed than glucose-treated slices. Both measures are generally interpreted as markers of reduced susceptibility to spreading depolarization, since propagation depends on the wavefront’s ability to recruit adjacent neurons into the regenerative wave. Peak amplitude of the SD-associated optical signal was significantly lower in both BHB- and sucrose-incubated slices than in glucose-incubated slices during the first cycle, while the time to fifty percent recovery from peak did not differ among groups. Supplementary experiments with a lower, 2 millimolar BHB concentration showed that most slices still generated an initial SD and a subset retained capacity for a second, hinting at a dose-dependent picture the authors did not fully explore.

The findings carry real weight for conditions where ketogenic metabolism matters clinically, including drug-resistant epilepsy and GLUT1 deficiency syndrome, and for neurological disorders where ketogenic strategies are increasingly being considered. Yet the authors are unusually candid about the limits of the ex vivo approach. The preparation does not reproduce the vascular, hormonal, transport, and systemic adaptations of in vivo ketosis; the substrate concentrations do not directly replicate the living brain’s metabolic environment; racemic BHB contains D- and L-forms that are metabolized differently; the sodium counterion in Na-BHB modestly raised extracellular sodium without a counterion-matched control; sample sizes for some optical analyses were modest; and investigators were not blinded during acquisition or analysis, a potential source of observer bias in the segmentation pipeline. No direct assay of tissue viability was performed. What the study establishes is narrower but firm: for several hours, in the complete absence of exogenous glucose, β-hydroxybutyrate supplied enough energy for mouse neocortical tissue to generate and recover from repeated spreading depolarizations, a feat that endogenous reserves alone could not match. That is a clear demonstration of cerebral fuel flexibility, and a challenge to the assumption that exogenous glucose is always indispensable for the brain’s most demanding work.

Subject of Research: Whether the ketone body β-hydroxybutyrate can sustain energetically demanding neural function in brain tissue deprived of exogenous glucose

Article Title: β-Hydroxybutyrate Maintains Energetically Demanding Neural Functions During Exogenous Glucose Deprivation

Article References: Ricks, R. L., Nevers, D. S., Poulos, T., Shafer, T. L., Dunford, C., Hansen, J. M., Reynolds, P. R., Bikman, B. T., & Parrish, R. R. (2026). β-Hydroxybutyrate Maintains Energetically Demanding Neural Functions During Exogenous Glucose Deprivation. Cellular and Molecular Neurobiology, 46(1), Article 147. https://doi.org/10.1007/s10571-026-01834-3

Image Credits: AI Generated

DOI: 10.1007/s10571-026-01834-3

Keywords: ketone bodies, β-hydroxybutyrate, glucose deprivation, mitochondrial respiration, spreading depolarization, brain slices, cerebral metabolism, neocortex, ketogenic diet, ex vivo, bioenergetics, mice