Skip to main content
NCBI home page
Cellular and Molecular Neurobiology logoLink to Cellular and Molecular Neurobiology
. 2026 Oct 2;46:147. doi: 10.1007/s10571-026-01834-3

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

Rachel L Ricks 1, Dallin S Nevers 2, Tyler Poulos 1,2, T Luke Shafer 2, Clayton Dunford 1, Jason M Hansen 1, Paul R Reynolds 1, Benjamin T Bikman 1, R Ryley Parrish 1,2,✉
PMCID: PMC13633234  PMID: 42825834

Abstract

Ketone bodies are a major source of cerebral energy during fasting and the ketogenic diet, but whether they can sustain brain tissue metabolism when glucose is absent remains uncertain. This question is difficult to resolve in vivo because circulating glucose is maintained, even during starvation, through endogenous production. We therefore used an ex vivo brain preparation to examine the metabolic capacity of tissue following incubation with β-hydroxybutyrate (BHB) as the only exogenous fuel, isolating brain tissue from its primary endogenous glucose sources. Mitochondrial respiration was monitored following 4-hour incubation in glucose-free, BHB-rich artificial cerebrospinal fluid, and tissue resilience was tested by inducing spreading depolarization, a severe energetic challenge that requires rapid restoration of ionic and metabolic homeostasis. Following acute reliance on exogenous BHB, in addition to endogenous energy reserves, mitochondrial electron transfer system function persisted, with lower O2 flux than following incubation with exogenous glucose. BHB-supported tissue maintained sufficient energetic capacity to generate and recover from repeated spreading depolarizations. In contrast, tissue relying solely on endogenous energy reserves failed to undergo repeated spreading depolarizations, suggesting that exogenous BHB provided the needed short-term energetic support beyond that available from endogenous substrates in the complete absence of exogenous glucose. The results broaden our understanding of cerebral fuel flexibility and provide evidence that BHB can support highly energetic neural function in the absence of exogenous glucose.

Graphical Abstract

graphic file with name 10571_2026_1834_Figa_HTML.webp

Open in a new tab

Cerebral fuel flexibility: glucose and ketones. Mitochondrial respiration continued with reduced O2 flux following β-hydroxybutyrate (BHB) incubation in the absence of exogenous glucose. BHB-supported brain slices generated and recovered from repeated spreading depolarizations (SD), whereas slices relying on endogenous energy reserves alone did not sustain repeated SDs. Together, these findings demonstrate that ketone-supported brain slices can sustain energetically demanding neural function beyond that supported by endogenous fuel reserves alone. ADP, adenosine diphosphate; ATP, adenosine triphosphate; Cyt c, cytochrome c; TCA, tricarboxylic acid. Created in BioRender. Parrish, R. (2026) https://BioRender.com/egdxr0j

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10571-026-01834-3.

Keywords: Energetics, Spreading depolarization, Spreading depression, Brain metabolism, Ketones

Introduction

The mammalian brain is classically viewed as a glucose-dependent organ, with high energetic demand and limited energy stores. Therefore, under normal physiological conditions, it is believed that glucose is the predominant substrate supporting cerebral ATP production (Dienel 2019; Mergenthaler et al. 2013). However, during prolonged fasting, ketogenic diet adherence, or other states of reduced carbohydrate availability, circulating ketone bodies increase and become an important alternative fuel for the brain, a state known as ketosis (Bolaños and Magistretti 2025; Chowdhury et al. 2014; García-Rodríguez and Giménez-Cassina 2021). β-Hydroxybutyrate (BHB), the major circulating ketone body, crosses the blood–brain barrier and is oxidized by neural cells to support mitochondrial energy production (Achanta and Rae 2017; Cahill and Veech 2003).

Although ketone bodies are widely recognized as glucose-sparing substrates, reducing cerebral glucose utilization in vivo (Jiang et al. 2011), it remains unclear whether BHB can sustain brain tissue energetics in the absence of glucose. Defining how brain tissue responds to glucose deprivation can aid in understanding therapeutic ketosis, particularly in conditions where it is widely used, such as drug-resistant epilepsy and GLUT1 deficiency syndrome (Bough and Rho 2007; Daley et al. 2026; Qiao et al. 2024; Williams et al. 2024), as well as other neurological conditions for which ketogenic strategies are increasingly being considered (Jensen et al. 2020; Luong et al. 2025). Direct experimental tests of brain function in which exogenous glucose is removed and replaced by BHB can help clearly define the limits of cerebral metabolic flexibility. In intact mammals, glucose remains available during prolonged starvation through gluconeogenesis (Rui 2014), and glucose transport is reduced rather than eliminated in GLUT1 deficiency syndrome (Klepper et al. 2020), making it difficult to isolate the brain’s dependence on glucose versus alternative fuels in vivo. However, this question can be answered using ex vivo brain preparations, in which brain tissue can be isolated from systemic sources of glucose, while retaining only locally available endogenous energy reserves.

In this study, we quantified brain tissue mitochondrial respiration following a period during which BHB was provided as the sole exogenous energetic substrate, along with cerebral ability to undergo and recover from multiple spreading depolarizations (SD), arguably the most energetically demanding event that occurs in the brain (Lindquist 2024). Of note, tissue relying solely on endogenous energy reserves failed to generate a second SD, suggesting that BHB provided support beyond that provided by any remaining endogenous fuel substrates. The mitochondria of brain slices incubated with exogenous BHB for 4 h demonstrated lower O2 flux across respiratory states compared to tissue that had been incubated with exogenous glucose. The ability of brain tissue to generate repeated SDs in the absence of exogenous glucose challenges the idea that exogenous glucose is necessary to sustain energetically demanding brain function and provides evidence for the capacity of BHB to serve as an alternative cerebral fuel beyond what is provided by endogenous brain reserves.

Materials and Methods

Animal Care and Ethical Approval

All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) and conducted in accordance with institutional guidelines. Experiments were performed using male and female C57BL/6 mice aged 3–9 weeks. Animals were housed in individually ventilated cages under a 12-hour light/dark cycle with ad libitum access to food (LabDiet 5001) and hyperchlorinated reverse osmosis-filtered water.

Brain Slice Preparation

Mice were anesthetized with isoflurane and decapitated. Brains were immediately removed and immersed in ice-cold carbogenated (95% O2/5% CO2) cutting solution containing (in mM): 126 NaCl, 26 NaHCO3, 3.5 KCl, 3 MgCl2, 1.26 NaH2PO4, and 10 glucose. Horizontal brain slices (400 μm) were prepared using a VT1200 vibratome (Leica Microsystems) and further dissected into neocortical sections as previously described (Withers et al. 2025). Brain slices were then quickly transferred to a continuously carbogenated artificial cerebrospinal fluid (aCSF) bath containing (in mM): 126 NaCl, 26 NaHCO3, 3.5 KCl, 2 CaCl2, 1 MgCl2, 1.26 NaH2PO4, and 10 glucose. The prepared neocortical brain slices were then randomly assigned to aCSF containing either 10 mM glucose (Fisher Chemical D16), 10 mM racemic β-hydroxybutyrate (BHB; Sigma 54965), or 10 mM sucrose (Fisher Chemical S5) for experimentation. A concentration of 10 mM glucose was selected based on conventional ex vivo brain slice preparations, in which elevated bath substrate concentrations relative to in vivo concentrations compensate for diffusion-limited delivery in the absence of functional microvasculature, with an equimolar concentration of BHB used for comparison. The sucrose condition served as a no-exogenous-fuel control because brain tissue lacks the enzymatic capacity to hydrolyze sucrose into metabolizable monosaccharides. Thus, sucrose maintained solution osmolality without providing an exogenous energetic substrate, allowing assessment of the extent to which endogenous energy substrates (Andersen and Schousboe 2023; Bolaños and Magistretti 2025; Brown and Ransom 2007; Kumar et al. 2025; Yip et al. 2016) could support the measured outcomes. Experimental solutions containing BHB or sucrose were made by substituting glucose with equimolar sodium β-hydroxybutyrate or sucrose, respectively. Osmolality (mean ± SD) was 303.0 ± 1.1, 306.4 ± 5.6, and 300.7 ± 9.1 mOsm/kg for glucose, BHB, and sucrose aCSF, respectively. Corresponding high-K+ solutions measured 304.3 ± 4.5, 308.8 ± 0.5, and 308.3 ± 2.1 mOsm/kg, respectively. Solution pH was maintained at 7.3–7.4.

Brain slices used for high-resolution respirometry were maintained in a submersion chamber at 37 °C for 4 h prior to experimentation. Brain slices used for SD experiments were maintained in an interface holding chamber at room temperature for 3–8 h in their respective solutions prior to experimentation.

High-Resolution Respirometry

Following the 4-hour incubation in aCSF containing 10 mM glucose or 10 mM BHB, brain slices were weighed using an analytical balance (Mettler Toledo AL104). Slices were then transferred to 1.7 mL microcentrifuge tubes containing ice-cold MiR05 respiration medium containing (in mM): 110 sucrose, 60 lactobionic acid, 20 taurine, 20 HEPES, 10 KH2PO4, 3 MgCl2, and 0.5 EGTA, supplemented with 1 g/L BSA. Tissue was mechanically homogenized by repeated aspiration with a pipette. The resulting homogenate was transferred to an Oroboros O2k high-resolution respirometer (Oroboros, Innsbruck, Austria) containing MiR05 maintained at 37 °C to evaluate mitochondrial respiratory function.

Following chamber loading, dissolved O2 concentration was elevated to 400 µM to establish baseline O2 availability. Mitochondrial respiration was then assessed using a substrate-uncoupler-inhibitor titration (SUIT) protocol. Each successive titration was performed after a stable O2 flux plateau was recorded.

Non-phosphorylating leak respiration (LEAK) was measured following additions of glutamate (G; 10 mM; Sigma P2256) and malate (M; 2 mM; Sigma M1000). Complex I-supported oxidative phosphorylation (OXPHOS) capacity was subsequently assessed by addition of ADP (D; 10 mM; Sigma 117105). Total OXPHOS capacity supported by complexes I and II was determined following addition of succinate (S; 20 mM; Sigma S2378). Maximal electron transfer (ET) capacity was measured by stepwise titration of the uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP, U; 0.1 µM increments; Sigma C2759) until maximum O2 flux was achieved. Complex II-supported ET capacity was then determined following inhibition of complex I with rotenone (Rot; 0.5 µM; Sigma R8875). Finally, to determine experimental O2 baseline flux, residual oxygen consumption (ROX) was measured following inhibition of complex III with antimycin A (Ama; 2.5 µM; Sigma A8674). To prevent oxygen limitation, supplemental O2 was added throughout the assay to maintain dissolved O2 concentration above 250 µM.

Mitochondrial outer membrane integrity was assessed by addition of cytochrome c (Cyt c; 10 µM; Sigma C7752), with preparations exhibiting > 10% stimulation of O2 flux excluded from analysis. Air calibration of the polarographic O2 sensors was performed immediately prior to each SUIT protocol, and zero-O2 calibrations were performed on non-experimental days by dithionite addition. O2 flux values were corrected for instrumental background, O2 diffusion, sensor O2 consumption, and ROX to isolate mitochondrial O2 flux.

Intrinsic Optical Signal Recording and Analysis

SD was induced using high-K+ aCSF, prepared by substituting NaCl with KCl to achieve a final K+ concentration of 26 mM (Andrew et al. 2017) while maintaining physiological osmolarity. This modification was applied to each experimental condition (glucose, BHB, and sucrose).

During experiments, brain slices were maintained in an interface system at 34–37 °C and continuously perfused (3.5 mL/min) with their respective carbogenated aCSF using a peristaltic perfusion system (PPS2 Multi Channel Systems). Following a 10-minute equilibration period, high-K+ (26 mM) aCSF was perfused for a total of 7 min or until SD occurred. Physiological-K+ aCSF (3.5 mM) was then restored for 10 min before repeating the SD induction protocol one additional time.

SD events were monitored using intrinsic optical signal (IOS) imaging. IOS images were acquired using a Lumenera INFINITY8-2 M microscope camera (Teledyne Vision Solutions, Thousand Oaks, CA) mounted on an AmScope SM-6 Series zoom trinocular stereo microscope (AmScope, Irvine, CA). Camera exposure and gain settings were held constant throughout each recording.

Slices were illuminated using a broad-field battery-powered LED light source (Harbor Freight, Calabasas, CA). Images were acquired at 1 Hz with a resolution of 1608 × 1104 pixels using INFINITY ANALYZE 7 software and stored as uncompressed TIFF sequences. Spatial calibration was performed for each slice using a reference standard of known dimensions.

Image sequences were analyzed using a custom Python-based (Python Software Foundation, Beaverton, OR) analysis pipeline developed for SD identification and metric quantification. SD events were manually defined as start-to-end frame ranges and opened in a dedicated interface for preprocessing, segmentation, mask correction, ROI definition, spatial calibration, and metric export.

For each SD event, frames were preprocessed by applying Gaussian smoothing (σ = 0.5 pixels), subtracting a pixel-wise median baseline computed from the 20–30 frames preceding the SD event, and normalizing intensities to the global 1st–99th percentile range to enhance contrast of the SD-associated IOS. The neocortex was manually delineated as the region of interest (ROI). SD segmentation was performed using the SAM2.1 video predictor with the Hiera Base+ checkpoint (sam2.1_hiera_base_plus; Meta AI). Initial masks were generated from user-provided point prompts and brush annotations, then propagated through the selected event frame range. Propagated masks were manually reviewed and corrected before analysis.

Quantitative metrics were derived from the final binary masks. For each frame, the primary SD region was identified as the largest external contour and smoothed by retaining the 25 lowest Fourier harmonics. Propagation speed was calculated from frame-to-frame outward displacement of the SD boundary and converted to µm/s using the spatial calibration factor and image acquisition rate. Outward displacement values were thresholded to exclude small movements below 2 pixels, and the remaining displacements were averaged to obtain the mean boundary advance for the frame transition. When multiple disconnected SD regions were present, propagation speeds were calculated separately for each lineage-aware object and combined as an area-weighted mean. Frames with missing or invalid masks were excluded from speed calculations.

To account for slow reflectance changes associated with solution exchange during high-K+ treatment and washout, SD recovery analyses were performed on baseline-corrected IOS signals. For each event, relative reflectance change was calculated as ΔR/R₀ = (R(t) − R₀)/R₀, where R₀ was defined as the mean reflectance of the 30 frames preceding K+ application (baseline reflectance). A recruited-region trace was calculated from the cumulative SD mask within the neocortical ROI, while a reference trace was calculated from non-recruited ROI pixels. A piecewise baseline model representing the solution-associated reflectance component was generated from the reference trace and aligned to the recruited-region trace over the interval spanning SD onset to high-K+ washout. The SD-associated IOS signal was then isolated by subtracting this baseline model from the recruited-region trace.

Peak SD-associated IOS amplitude was defined as the largest signed residual deflection within 100 s of SD onset, located to subframe resolution by parabolic interpolation. Recovery half-time (T50) was defined as the interval from peak amplitude to the first time the residual signal returned to 50% of peak amplitude relative to the corrected baseline, assessed up to 250 s past the peak. Events were excluded from recovery analyses if no detectable SD signal was present (peak amplitude < 0.5% ΔR/R₀), if solution washout occurred before T50, or if technical issues compromised data quality.

Statistical Analysis and Study Limitations

Data are presented as mean ± standard error of the mean (SEM) or median and interquartile range, as appropriate. No differences were observed between male and female mice, and therefore data were presented as a single group. Parametric data assessed by Shapiro–Wilk testing were analyzed by unpaired two-tailed Welch’s t test or one-way ANOVA with Tukey’s multiple-comparisons test, as appropriate; nonparametric data were analyzed by unpaired two-tailed Mann–Whitney U test or Kruskal–Wallis test. Contingency analyses were conducted using two-tailed Fisher’s exact tests, with pairwise comparisons adjusted using Bonferroni correction where applicable. All statistical analyses were performed using GraphPad Prism version 11.0 (GraphPad Software, San Diego, CA). IOS imaging data were analyzed using a custom-built Python graphical user interface as described above, as well as custom Python scripts. Sample sizes differed among metrics because progressively more stringent image-quality requirements were necessary for advanced IOS analyses, with propagation speed and residual ΔR/R₀ measurements requiring higher signal quality and greater stability than SD occurrence or latency measurements.

We acknowledge several limitations in this study. Sample sizes were modest for some IOS analyses, potentially limiting statistical power, and investigators were not blinded during data acquisition or analysis. Because IOS segmentation incorporated automated mask propagation followed by manual review and correction, the lack of blinding represents a potential source of observer bias. Additional considerations include the use of racemic BHB, comprising approximately equal concentrations of D- and L-BHB, which differ in their metabolism, and substrate concentrations that do not directly replicate the in vivo cerebral metabolic environment. Finally, while the ex vivo preparation enables controlled assessment of tissue-level responses to substrate availability, it does not reproduce the vascular, hormonal, transport, and systemic adaptations associated with in vivo ketosis.

Results

Basal Mitochondrial Function is Maintained During Exogenous Glucose Deprivation

Following 4-hour maintenance of mouse neocortical tissue in aCSF containing either 10 mM BHB, 10 mM glucose, or 10 mM sucrose, mitochondrial respiratory function was assessed by high-resolution respirometry. Across all respiratory states examined, O2 flux (JO2) was significantly lower following BHB incubation than following glucose incubation (Fig. 1b–f). Complex I-linked LEAK respiration, representing non-phosphorylating JO2 associated with proton leak, proton slip, cation cycling, and electron leak at high chemiosmotic potential, was reduced following acute BHB incubation (Fig. 1b). Following ADP-mediated recruitment of ATP synthase, JO2 remained lower during both complex I-supported (Fig. 1c) and complex I and II-supported oxidative phosphorylation (OXPHOS) (Fig. 1d). Similarly, JO2 associated with maximal uncoupled electron transfer (ET) capacity was lower following BHB incubation when supported by complexes I and II (Fig. 1e) or by complex II alone (Fig. 1f). In addition, tissue incubated without an exogenous fuel substrate (sucrose aCSF), relying on endogenous energy reserves, exhibited a further reduction in mitochondrial JO2 across all coupling-control states assessed compared to tissue incubated with glucose (Fig. S1). Collectively, these data demonstrate that mitochondrial respiration can be maintained under conditions of exogenous glucose deprivation, with a respiratory profile characterized by lower JO2 across respiratory states.

Fig. 1.

Fig. 1

Open in a new tab

Mitochondrial oxygen flux (JO2) in mouse neocortex is reduced following 4-hour incubation in 10 mM β-hydroxybutyrate (BHB) aCSF compared to 10 mM glucose aCSF. a Representative JO2 traces during a substrate-uncoupler-inhibitor titration (SUIT) protocol depicting change in mitochondrial JO2 over time per unit tissue mass (pmol·s− 1·mg− 1). b Complex I-linked LEAK respiration, representing JO2 supported by electron entry through complex I in the absence of oxidative phosphorylation (OXPHOS) following addition of glutamate and malate (GM). c Complex I-linked OXPHOS capacity, representing maximal JO2 supported by electron entry through complex I coupled to ATP production following addition of ADP. d Total OXPHOS capacity, representing maximal JO2 from the additive contributions of complexes I and II electron supply pathways converging at the Q-junction coupled to ATP production following addition of succinate (S). e Total electron transfer (ET) capacity, the noncoupled state of maximum mitochondrial respiration following stepwise addition of the uncoupler carbonyl cyanide m-chlorophenyl hydrazone (CCCP; U). f Complex II-linked ET capacity, representing maximal JO2 supported by electron entry through complex II in the noncoupled state following addition of rotenone (Rot). Residual oxygen consumption (ROX), determined by antimycin A (Ama) addition at the end of the SUIT protocol, was subtracted from JO2 values to isolate mitochondrial JO2 (baseline correction). Background-corrected data are presented to remove the contribution of O2 backdiffusion and O2 consumption by the polarographic O2 sensor. Data are presented as mean ± SEM (unpaired two-tailed Welch’s t test, n = 12). *p < 0.05, **p < 0.01. Created in BioRender. Parrish, R. (2026) https://BioRender.com/kit9f5b

BHB Can Supply the Energy Required for Highly Energetically Demanding Brain Events

Neocortical brain slices incubated in either glucose, BHB, or sucrose aCSF were exposed to a high-K+ bath to induce SDs (Fig. 2a). A sucrose aCSF group without glucose or BHB provided a control to account for the contribution of endogenous energy reserves. All slices from the three groups underwent a first SD (Fig. 2bi); however, only the BHB and glucose groups were able to undergo and recover from a second SD (Fig. 2bii).

Fig. 2.

Fig. 2

Open in a new tab

Neocortical brain slices supplied with β-hydroxybutyrate (BHB) in the absence of exogenous glucose can undergo and recover from energetically demanding spreading depolarizations (SD). a Experimental paradigm for sequential SD induction. b SD occurrence upon high-K+ (26 mM) exposure in media containing 10 mM glucose, 10 mM BHB, or 10 mM sucrose (no-exogenous-fuel condition) during bi cycle 1 (overall two-tailed Fisher’s exact test, p = 1.0000; n = 8 per group) and bii cycle 2 (overall two-tailed Fisher’s exact test, p < 0.0001; pairwise Fisher’s exact tests with Bonferroni correction; n = 8 per group; glucose vs. sucrose, p = 0.0006; BHB vs. sucrose, p = 0.0006). c Latency to SD following 26 mM high K+ exposure for ci cycle 1 (Kruskal–Wallis test, p = 0.5628; n = 8 per group) and cii cycle 2 (two-tailed Mann–Whitney U test, p = 0.027; n = 8 per group). d Representative intrinsic optical signal (IOS) images of SD propagating across the neocortex displayed at 25-second intervals. Red lines indicate the SD border. e Mean propagation speed of each SD event for ei cycle 1 (Kruskal–Wallis test, p = 0.1842; n = 8 for glucose and BHB, n = 7 for sucrose) and eii cycle 2 (two-tailed Welch’s t test, p = 0.0448; n = 8 per group). f Mean residual change in reflectance (ΔR/R₀) for the area recruited by SD over time for cycle 1 (n = 5 for glucose, n = 4 for BHB, n = 6 for sucrose). The solid line represents the mean for each group, and the shaded region represents ± SEM. g Peak amplitude of SD-associated IOS relative to baseline reflectance prior to SD (ΔR/R₀) for gi cycle 1 (one-way ANOVA, p = 0.0035, with Tukey’s multiple-comparisons test; n = 5 for glucose, n = 4 for BHB, n = 6 for sucrose; glucose vs. BHB, p = 0.0278; glucose vs. sucrose, p = 0.0032) and gii cycle 2 (two-tailed Welch’s t test, p = 0.9299; n = 6 for glucose, n = 5 for BHB). h Time from peak amplitude to return to 50% of the signed peak amplitude relative to baseline (T50) following SD for hi cycle 1 (Kruskal–Wallis test, p = 0.8597; n = 5 for glucose, n = 4 for BHB, n = 6 for sucrose) and hii cycle 2 (two-tailed Mann–Whitney U test, p = 0.1775; n = 6 for glucose, n = 5 for BHB). Boxplots display the median and interquartile range, with whiskers extending to the minimum and maximum values. *p < 0.05, **p < 0.01, ***p < 0.001. Created in BioRender. Parrish, R. (2026) https://BioRender.com/mqjiuxe

During the first high-K+ cycle, latency to SD and SD propagation speed did not vary significantly between groups (Fig. 2ci, ei). However, following a second SD induction, BHB-treated slices showed a significantly longer latency and slower rate of SD propagation than glucose-treated slices (Fig. 2cii, eii).

IOS analysis further revealed that all treatment groups returned to baseline reflectance following the first SD (Fig. 2f), consistent with reversal of cellular swelling associated with SD. ΔR/R₀ peak amplitude was significantly lower in both BHB- and sucrose-incubated slices compared with glucose-incubated slices (Fig. 2g). Time to 50% recovery from peak amplitude (T50) did not differ among groups (Fig. 2h).

Additional high-K+ bath experiments using 2 mM BHB are presented in Fig. S2. Most slices generated an initial SD and a subset retained the capacity to support a second SD, with differences in propagation speed but no significant post hoc differences in latency.

Collectively, these data demonstrate that BHB can support the induction of and recovery from repeated spreading depolarizations in the absence of exogenous glucose, beyond the capacity supported by endogenous energy reserves alone.

Discussion

Glucose has classically been viewed as a requisite energetic substrate for neural tissue (Dienel 2019). Neurons, however, demonstrate metabolic flexibility in their capacity to utilize a variety of substrates, including lactate (Yip et al. 2016; Pellerin and Magistretti 2004), ketone bodies (Achanta et al. 2017; Andersen et al. 2023; Jiang et al. 2011; Pan et al. 2002), and lipids (Greda et al. 2025; Kumar et al. 2025; Pavlowsky et al. 2025; Saber et al. 2025), with substrate uptake primarily facilitated by glucose transporters (GLUTs) and monocarboxylate transporters (MCTs) (Pellerin et al. 2005; Koepsell 2020). The present study examined whether BHB can support brain tissue metabolism during a period of exogenous glucose deprivation, beyond that sustained by endogenous energy reserves alone. Following incubation with BHB, mitochondrial JO2 was reduced across all respiratory states measured by high-resolution respirometry relative to glucose-incubated tissue. Despite this reduction, BHB-supported tissue demonstrated the ability to generate and recover from repeated SDs. Repeated SD induction also revealed prolonged latency to SD onset and slower propagation in BHB-treated tissue relative to glucose-treated tissue. Together, these findings suggest that BHB can sustain brain function for several hours in the absence of exogenous glucose and may reduce tissue susceptibility to repeated SD.

A notable finding was the reduction in mitochondrial JO2 associated with mitochondrial leak respiration, oxidative phosphorylation capacity, and electron transfer capacity following incubation in BHB compared with glucose. Tissue previously incubated without an exogenous fuel substrate exhibited still lower JO2 across coupling-control states, demonstrating that endogenous energy reserves alone can also sustain measurable mitochondrial respiration. Different catabolic substrates can exhibit different rates of respiration supporting similar power outputs due to differences in coupling efficiency, proton leak, ATP/O2 stoichiometry, and the thermodynamic state of ATP hydrolysis (ΔGATP) (Gnaiger 2020; Schmidt et al. 2021). While ketone body metabolism has been proposed to alter mitochondrial bioenergetics by increasing mitochondrial coupling efficiency and ATP flux (Holstein et al. 2025), increasing ΔGATP (Veech et al. 2001; Veech 2004), reducing free radical damage (Achanta and Rae 2017; Jarrett et al. 2008; Veech et al. 2001), and improving bioenergetic resilience during heightened energetic demand (Llorente-Folch et al. 2023), these effects were not tested in this study. It is possible that lower mitochondrial JO2 following BHB incubation reflects a more efficient metabolic strategy (Saito et al. 2022; Veech et al. 2001). In addition, lower mitochondrial JO2 could reflect a reduction in ATP generative capacity. Potential mechanisms include mitochondrial adaptation to reduced energetic demand, as ketone bodies can reduce neuronal excitability (Ma et al. 2007; Qiao et al. 2024; Zhang et al. 2026), or an acute suppressive effect of BHB on mitochondrial respiratory capacity.

SD was selected as a brief, reproducible model of extreme energetic demand because the profound disruption of transmembrane ion gradients requires substantial ATP-dependent ion transport to restore ionic homeostasis (Lindquist 2024; Ricks et al. 2026; Schoknecht et al. 2025). Slices without an exogenous fuel source (sucrose condition) generated an SD, demonstrating that endogenous energy reserves were sufficient to support an initial energetic challenge. However, these slices could not generate a second SD, whereas slices supplied with exogenous glucose or BHB generated and recovered from repeated SDs. Given that recovery from SD represents one of the most significant energetic challenges encountered by brain tissue, these findings indicate that BHB provided metabolic support beyond that sustained by endogenous energy reserves alone.

Initial latency from high-K+ insult to SD induction and average propagation speed of the SD wavefront were similar between glucose- and BHB-treated slices. Interestingly, for repeated SDs, BHB-treated tissue exhibited prolonged latency to SD onset and slower average SD propagation. Propagation of the SD depends on the ability of the SD wavefront to recruit adjacent neurons into the regenerative wave. Consequently, both increased latency and reduced propagation velocity are generally interpreted as markers of reduced SD susceptibility (Aiba and Noebels 2021; Dreier 2011; Jansen et al. 2024; Vitale et al. 2025). Because energy deprivation is a common SD induction paradigm (Ricks et al. 2025), BHB’s impact on SD kinetics may be brought about by altered energy availability during these extreme brain perturbations and/or by direct modulation of neuronal ion channels (Ma et al. 2007; Manville et al. 2020; Pflanz et al. 2019; Zhang et al. 2026).

One notable study limitation was the use of Na-BHB, which modestly increased extracellular [Na+] relative to glucose aCSF. Although solution osmolalities were similar and the predicted electrophysiological effect of the additional Na+ was small, a counterion-matched control was not included; therefore, a potential Na+-dependent contribution to the observed effects cannot be excluded.

Overall, these data demonstrate that β-hydroxybutyrate can support energetically demanding neural activity, assessed by induction of repeated spreading depolarizations in the short-term absence of exogenous glucose, suggesting substantial cerebral metabolic flexibility. No direct assay of tissue viability was performed in this study; therefore, future studies should determine how prolonged reliance on BHB affects processes beyond energy production, including biosynthetic pathways, redox homeostasis beyond an acute glutathione redox assessment (Fig. S3), and overall tissue health.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (369.9KB, docx)

Author Contributions

All authors contributed to the study conception and design. Material preparation, data collection, and analyses were performed by R.L.R., D.S.N., T.P., T.L.S., C.D., and J.M.H. IOS imaging analysis software was designed by C.D. The first draft of the manuscript was written by R.L.R., D.S.N., T.P., T.L.S., C.D., and R.R.P., and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This research was supported by internal funding from Brigham Young University.

Data Availability

The raw datasets generated during the current study are not publicly available due to the lack of a standard repository for this type of data but are available from the corresponding author upon reasonable request.

Declarations

Competing Interests

B.T.B. is an advisor for Unicity International and receives royalties for his role with them, along with royalties from the sale of a book about insulin resistance. B.T.B., P.R.R., and R.R.P. are advisors for Ketone Labs. All authors declare there are no additional conflicts of interest.

Ethical Approval

All animal experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (8th ed.) and were approved by the Brigham Young University Institutional Animal Care and Use Committee.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. Achanta LB, Rae CD (2017) β-Hydroxybutyrate in the Brain: One Molecule, Multiple Mechanisms. Neurochem Res 42(1):35–49. 10.1007/s11064-016-2099-2 [DOI] [PubMed] [Google Scholar]
  2. Achanta LB, Rowlands BD, Thomas DS, Housley GD, Rae CD (2017) β-Hydroxybutyrate Boosts Mitochondrial and Neuronal Metabolism but is not Preferred Over Glucose Under Activated Conditions. Neurochem Res 42(6):1710–1723. 10.1007/s11064-017-2228-6 [DOI] [PubMed] [Google Scholar]
  3. Aiba I, Noebels JL (2021) Kcnq2/Kv7.2 controls the threshold and bi-hemispheric symmetry of cortical spreading depolarization. Brain 144(9):2863–2878. 10.1093/brain/awab141 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Andersen JV, Schousboe A (2023) Milestone Review: Metabolic dynamics of glutamate and GABA mediated neurotransmission - The essential roles of astrocytes. J Neurochem 166(2):109–137. 10.1111/jnc.15811 [DOI] [PubMed] [Google Scholar]
  5. Andersen JV, Westi EW, Neal ES, Aldana BI, Borges K (2023) β-Hydroxybutyrate and medium-chain fatty acids are metabolized by different cell types in mouse cerebral cortex slices. Neurochem Res 48(1):54–61. 10.1007/s11064-022-03726-6 [DOI] [PubMed] [Google Scholar]
  6. Andrew RD, Hsieh YT, Brisson CD (2017) Spreading depolarization triggered by elevated potassium is weak or absent in the rodent lower brain. J Cereb Blood Flow Metab 37(5):1735–1747. 10.1177/0271678X16657344 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bolaños JP, Magistretti PJ (2025) The neuron–astrocyte metabolic unit as a cornerstone of brain energy metabolism in health and disease. Nat Metabolism. 10.1038/s42255-025-01404-9 [DOI] [PubMed] [Google Scholar]
  8. Bough KJ, Rho JM (2007) Anticonvulsant mechanisms of the ketogenic diet. Epilepsia 48(1):43–58. 10.1111/j.1528-1167.2007.00915.x [DOI] [PubMed] [Google Scholar]
  9. Brown AM, Ransom BR (2007) Astrocyte glycogen and brain energy metabolism. Glia 55(12):1263–1271. 10.1002/glia.20557 [DOI] [PubMed] [Google Scholar]
  10. Cahill GF Jr., Veech RL (2003) Ketoacids? Good medicine? Trans Am Clin Climatol Assoc 114:149–161 discussion 162–163 [PMC free article] [PubMed] [Google Scholar]
  11. Chowdhury GM, Jiang L, Rothman DL, Behar KL (2014) The Contribution of Ketone Bodies to Basal and Activity-Dependent Neuronal Oxidation in Vivo. J Cereb Blood Flow Metabolism 34(7):1233–1242. 10.1038/jcbfm.2014.77 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Daley SF, Masood W, Annamaraju P, Khan Suheb MZ (2026) The Ketogenic Diet: Clinical Applications, Evidence-based Indications, and Implementation. In: StatPearls [Internet]. StatPearls Publishing, Treasure Island (FL) [PubMed]
  13. Dienel GA (2019) Brain Glucose Metabolism: Integration of Energetics with Function. Physiol Rev 99(1):949–1045. 10.1152/physrev.00062.2017 [DOI] [PubMed] [Google Scholar]
  14. Dreier JP (2011) The role of spreading depression, spreading depolarization and spreading ischemia in neurological disease. Nat Med 17(4):439–447. 10.1038/nm.2333 [DOI] [PubMed] [Google Scholar]
  15. García-Rodríguez D, Giménez-Cassina A (2021) Ketone bodies in the brain beyond fuel metabolism: from excitability to gene expression and cell signaling. Front Mol Neurosci. 10.3389/fnmol.2021.732120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gnaiger E (2020) Mitochondrial pathways and respiratory control. An introduction to OXPHOS analysis. Bioenerg Commun. 5 edn. 10.26124/bec:2020-0002 [DOI] [Google Scholar]
  17. Greda AK, Gomes JP, Schmidt-Krueger V, Zurawska-Plaksej E, Fritsche-Guenther R, Rudolph IM, Telugu NS, Cömert C, Kirwan J, Kunz S, Rothe M, Johannsen M, Diecke S, Bross P, Willnow TE (2025) Interaction of sortilin with apolipoprotein E3 enables neurons to use long-chain fatty acids as alternative metabolic fuel. Nat Metab 7(11):2346–2365. 10.1038/s42255-025-01389-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Holstein DM, Saliba A, Lozano D, Kim J, Sharma K, Lechleiter JD (2025) β-hydroxybutyrate enhances brain metabolism in normoglycemia and hyperglycemia, providing cerebroprotection in a mouse stroke model. J Cereb Blood Flow Metab 45(8):1493–1506. 10.1177/0271678x251334222 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Jansen NA, Linnenbank C, Schenke M, Voskuyl RA, Jorge MS, Krivoshein G, Breukel C, Linssen MM, Claassens JWC, Brouwers C, van Heiningen SH, Heuck A, Lykke-Hartmann K, Tolner EA, van den Maagdenberg AMJM (2024) Spontaneous spreading depolarizations originate subcortically in a novel mouse model of familial hemiplegic migraine type 2. Neurobiol Dis 202:106714. 10.1016/j.nbd.2024.106714 [DOI] [PubMed] [Google Scholar]
  20. Jarrett SG, Milder JB, Liang LP, Patel M (2008) The ketogenic diet increases mitochondrial glutathione levels. J Neurochem 106(3):1044–1051. 10.1111/j.1471-4159.2008.05460.x [DOI] [PubMed] [Google Scholar]
  21. Jensen NJ, Wodschow HZ, Nilsson M, Rungby J (2020) Effects of ketone bodies on brain metabolism and function in neurodegenerative diseases. Int J Mol Sci. 10.3390/ijms21228767 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Jiang L, Mason GF, Rothman DL, de Graaf RA, Behar KL (2011) Cortical substrate oxidation during hyperketonemia in the fasted anesthetized rat in vivo. J Cereb Blood Flow Metab 31(12):2313–2323. 10.1038/jcbfm.2011.91 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Klepper J, Akman C, Armeno M, Auvin S, Cervenka M, Cross HJ, De Giorgis V, Della Marina A, Engelstad K, Heussinger N, Kossoff EH, Leen WG, Leiendecker B, Monani UR, Oguni H, Neal E, Pascual JM, Pearson TS, Pons R, Scheffer IE, Veggiotti P, Willemsen M, Zuberi SM, De Vivo DC (2020) Glut1 deficiency syndrome (Glut1DS): state of the art in 2020 and recommendations of the international Glut1DS study group. Epilepsia Open 5(3):354–365. 10.1002/epi4.12414 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Koepsell H (2020) Glucose transporters in brain in health and disease. Pflügers Archiv - Eur J Physiol 472(9):1299–1343. 10.1007/s00424-020-02441-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kumar M, Wu Y, Knapp J, Pontius CL, Park D, Witte RE, McAllister R, Gupta K, Rajagopalan KN, De Camilli P, Ryan TA (2025) Triglycerides are an important fuel reserve for synapse function in the brain. Nat Metab 7(7):1392–1403. 10.1038/s42255-025-01321-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lindquist BE (2024) Spreading depolarizations pose critical energy challenges in acute brain injury. J Neurochem 168(5):868–887. 10.1111/jnc.15966 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Llorente-Folch I, Düssmann H, Watters O, Connolly NMC, Prehn JHM (2023) Ketone body β-hydroxybutyrate (BHB) preserves mitochondrial bioenergetics. Sci Rep 13(1):19664. 10.1038/s41598-023-46776-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Luong T, Svart M, Gormsen LC, Søndergaard E (2025) Ketogenic diet in the management of disease. Curr Opin Clin Nutr Metab Care 28(6):477–482. 10.1097/mco.0000000000001158 [DOI] [PubMed] [Google Scholar]
  29. Ma W, Berg J, Yellen G (2007) Ketogenic diet metabolites reduce firing in central neurons by opening K(ATP) channels. J Neurosci 27(14):3618–3625. 10.1523/jneurosci.0132-07.2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Manville RW, Papanikolaou M, Abbott GW (2020) M-Channel Activation Contributes to the Anticonvulsant Action of the Ketone Body β-Hydroxybutyrate. J Pharmacol Exp Ther 372(2):148–156. 10.1124/jpet.119.263350 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Mergenthaler P, Lindauer U, Dienel GA, Meisel A (2013) Sugar for the brain: the role of glucose in physiological and pathological brain function. Trends Neurosci 36(10):587–597. 10.1016/j.tins.2013.07.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Pan JW, de Graaf RA, Petersen KF, Shulman GI, Hetherington HP, Rothman DL (2002) [2,4-13C2]-beta-Hydroxybutyrate metabolism in human brain. J Cereb Blood Flow Metab 22(7):890–898. 10.1097/00004647-200207000-00014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pavlowsky A, Silva B, Basu R, Correia Delecourt A, Geny D, Danglot L, Plaçais PY, Preat T (2025) Neuronal fatty acid oxidation fuels memory after intensive learning in Drosophila. Nat Metab 7(12):2438–2450. 10.1038/s42255-025-01416-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Pellerin L, Magistretti PJ (2004) Neuroenergetics: calling upon astrocytes to satisfy hungry neurons. Neuroscientist 10(1):53–62. 10.1177/1073858403260159 [DOI] [PubMed] [Google Scholar]
  35. Pellerin L, Bergersen LH, Halestrap AP, Pierre K (2005) Cellular and subcellular distribution of monocarboxylate transporters in cultured brain cells and in the adult brain. J Neurosci Res 79(1–2):55–64. 10.1002/jnr.20307 [DOI] [PubMed] [Google Scholar]
  36. Pflanz NC, Daszkowski AW, James KA, Mihic SJ (2019) Ketone body modulation of ligand-gated ion channels. Neuropharmacology 148:21–30. 10.1016/j.neuropharm.2018.12.013 [DOI] [PubMed] [Google Scholar]
  37. Qiao Y-N, Li L, Hu S-H, Yang Y-X, Ma Z-Z, Huang L, An Y-P, Yuan Y-Y, Lin Y, Xu W, Li Y, Lin P-C, Cao J, Zhao J-Y, Zhao S-M (2024) Ketogenic diet-produced β-hydroxybutyric acid accumulates brain GABA and increases GABA/glutamate ratio to inhibit epilepsy. Cell Discov 10(1):17. 10.1038/s41421-023-00636-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Ricks R, MacKenzie-Gray Scott C, Trevelyan AJ, Parrish RR (2025) Mechanistic insight into the initiation of spreading depolarizations: Is it really all about potassium? J Physiol 603(11):3269–3273. 10.1113/JP288809 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ricks R, Bawuah Afran-Okese K, Withers PC, Shafer TL, Roque I, Poulos T, Norby JH, Maxfield R, Parrish RR (2026) Cellular and network mechanisms of spreading depolarization. Adv Physiol Educ 50(3):929–941. 10.1152/advan.00201.2025 [DOI] [PubMed] [Google Scholar]
  40. Rui L (2014) Energy metabolism in the liver. Compr Physiol 4(1):177–197. 10.1002/cphy.c130024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Saber SH, Yak N, Yong XLH, Bong YT, Leeson H, Dai CY, Binder T, Lu S, Purushothaman R, Lenaerts AS, Almeida-Souza L, Koludarova L, Er S, Hlushchuk I, Gaudin A, Singh S, Nyman TA, Harmer JR, Zuryn S, Wolvetang E, Talbo GH, Airavaara M, Battersby BJ, van Waardenberg AJ, Anggono V, Balistreri G, Joensuu M (2025) DDHD2 provides a flux of saturated fatty acids for neuronal energy and function. Nat Metab 7(10):2117–2141. 10.1038/s42255-025-01367-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Saito ER, Warren CE, Hanegan CM, Larsen JG, du Randt JD, Cannon M, Saito JY, Campbell RJ, Kemberling CM, Miller GS, Edwards JG, Bikman BT (2022) A novel ketone-supplemented diet improves recognition memory and hippocampal mitochondrial efficiency in healthy adult mice. Metabolites. 10.3390/metabo12111019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Schmidt CA, Fisher-Wellman KH, Neufer PD (2021) From OCR and ECAR to energy: perspectives on the design and interpretation of bioenergetics studies. J Biol Chem. 10.1016/j.jbc.2021.101140 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Schoknecht K, Baeza-Lehnert F, Hirrlinger J, Dreier JP, Eilers J (2025) Spreading depolarizations exhaust neuronal ATP in a model of cerebral ischemia. Proc Natl Acad Sci U S A 122(19):e2415358122. 10.1073/pnas.2415358122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Veech RL (2004) The therapeutic implications of ketone bodies: the effects of ketone bodies in pathological conditions: ketosis, ketogenic diet, redox states, insulin resistance, and mitochondrial metabolism. Prostaglandins Leukot Essent Fatty Acids 70(3):309–319. 10.1016/j.plefa.2003.09.007 [DOI] [PubMed] [Google Scholar]
  46. Veech RL, Chance B, Kashiwaya Y, Lardy HA, Cahill GF Jr (2001) Ketone bodies, potential therapeutic uses. IUBMB Life 51(4):241–247. 10.1080/152165401753311780 [DOI] [PubMed] [Google Scholar]
  47. Vitale M, Tottene A, Zadeh MZ, Pietrobon D (2025) Mechanisms underlying the increased susceptibility to initiation of cortical spreading depression in a genetic mouse model of migraine. J Headache Pain 26(1):256. 10.1186/s10194-025-02139-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Williams RSB, Boison D, Masino SA, Rho JM (2024) Mechanisms of Ketogenic Diet Action. In: Noebels JL, Avoli M, Rogawski MA, Vezzani A, Delgado-Escueta AV (eds) Jasper’s Basic Mechanisms of the Epilepsies. 5th edn., New York, pp 1635–1666. 10.1093/med/9780197549469.003.0079 [DOI]
  49. Withers PC, Jones A, Afran-Okese KB, Calder B, Morrill HJ, Shafer TL, Nevers DS, Norby JH, Acosta R, Bikman BT, Suli A, Parrish RR (2025) Pharmacological inhibition of all known major inward cationic currents does not block the induction of spreading depolarizations. Front Cell Neurosci. 10.3389/fncel.2025.1668329 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Yip J, Geng X, Shen J, Ding Y (2016) Cerebral Gluconeogenesis and Diseases. Front Pharmacol 7:521. 10.3389/fphar.2016.00521 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zhang G, Wang XR, Su B, Li T, Liu H, Xu X (2026) A–type potassium channels mediate the inhibitory effect of β–hydroxybutyrate on CA1 pyramidal neurons in an immature mouse model of kainic acid–induced status epilepticus. Epilepsy Res 226:107841. 10.1016/j.eplepsyres.2026.107841 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (369.9KB, docx)

Data Availability Statement

The raw datasets generated during the current study are not publicly available due to the lack of a standard repository for this type of data but are available from the corresponding author upon reasonable request.

Articles from Cellular and Molecular Neurobiology are provided here courtesy of Springer

RESOURCES