Hippocampal Bioenergetics and Metabolic Profiling Identifies Fatty Acid Oxidation as a Potential Therapeutic Target in Traumatic Brain Injury
Di Zhou, Mengxuan Shi, Mitchell D. Kilgore, Yuwen Xiu, Yingjie Wang, Danni Wang, Thin Yadanar Sein, Charles Vidoudez, Amin Iskender, Gaby A. Moyano, Lauren M. Dumont, Yinghua Jiang, Prasad V. G. Katakam, Bo Ning, Aaron S. Dumont, Xiaoying Wang, Jia Fan, Ning Liu
TL;DR
This study explores how fatty acid oxidation could be a new treatment target for traumatic brain injury by examining metabolic changes in the hippocampus.
Contribution
The study identifies fatty acid oxidation as a novel therapeutic target for TBI through bioenergetic and metabolic profiling.
Findings
Mitochondrial oxidative phosphorylation in the dentate gyrus was reduced acutely after TBI but recovered by day 7.
Fatty acid oxidation was elevated by day 7 post-TBI, indicating a compensatory metabolic adaptation.
Sodium octanoate administration improved mitochondrial respiration and reduced neurodegeneration after TBI.
Abstract
Traumatic brain injury (TBI) causes lasting neurological impairments, particularly learning and memory deficits associated with hippocampal damage. Emerging evidence suggests that hippocampal vulnerability may be linked to bioenergetic dysfunction, though its role remains poorly defined. A deeper understanding of post-TBI metabolic disturbances and their association with pathological outcomes could reveal novel therapeutic targets. In this study, we conducted functional bioenergetic assessments and multi-omics analyses on hippocampal slices using a mouse controlled cortical impact model of TBI. Seahorse analysis revealed a significant reduction in mitochondrial oxidative phosphorylation in dentate gyrus (DG) slices at day 1 (acute phase), which recovered by day 7 (subacute phase) post-TBI. Metabolomic profiling revealed acute impairments in purine nucleotide, glucose, amino acid, and…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsTraumatic Brain Injury and Neurovascular Disturbances · Traumatic Brain Injury Research · S100 Proteins and Annexins
Introduction
Traumatic brain injury (TBI) is a leading cause of mortality and long-term disability worldwide [1], with no effective pharmacological therapies currently available to limit its consequential neurological deficits [2]. Although early neuroprotective strategies targeting neuronal cell death have shown promise in preclinical models, they have largely failed to translate into improved functional outcomes in clinical settings [3], due in part to an incomplete understanding of the underlying molecular mechanisms.
While TBI commonly results in focal cortical damage, it also induces diffuse injury across various brain regions, most notably the hippocampus, a structure essential for learning and memory [4–6]. The hippocampus is particularly vulnerable to TBI-related insults, including metabolic disruption, structural damage, and functional impairment [7, 8]. Increasing evidence suggests that cognitive dysfunction following TBI is closely linked to secondary phase alterations in hippocampal function [9, 10]. In particular, persistent inflammation and neurodegeneration in the hippocampus significantly contribute to long-term cognitive deficits after TBI [5, 11], underscoring the need for targeted interventions to mitigate this region's vulnerability.
Cerebral bioenergetics dysfunction is well-documented in patients with moderate-to-severe TBI and correlates strongly with clinical outcomes [12, 13]. Experimental models consistently demonstrate that TBI disrupts bioenergetics and mitochondrial function in the hippocampus [14–18], which exhibits vulnerability comparable to that of the cortex [19–21]. This dysfunction is associated with deficits in hippocampal-dependent learning and memory [16]. Notably, hippocampal bioenergetic impairment post-TBI is a dynamic process, initiating within minutes to hours after injury, persisting for days, and with severity that correlates closely with the extent of the injury [16, 19, 22]. However, critical knowledge gaps remain regarding the molecular insights on the temporal bioenergetic and metabolic alterations in the hippocampus after moderate-to-severe TBI, particularly those involving mitochondrial respiration, ATP production, and the metabolism of glucose, fatty acids, and amino acids. Furthermore, whether early bioenergetic impairment directly contributes to subsequent neuroinflammation and neurodegeneration remains unclear. It is also unknown whether restoring bioenergetic function can be leveraged to develop effective adjunctive therapies for TBI. Therefore, a comprehensive, region-specific understanding of these processes is essential for identifying novel therapeutic targets and guiding the development of effective strategies in this setting.
In this study, we employed the controlled cortical impact (CCI) model, a well-established approach for modeling moderate-to-severe TBI. This model induces significant cortical tissue loss while sparing the hippocampus and results in reproducible spatial learning and memory deficits [9]. We conducted comprehensive analyses of hippocampal bioenergetics using Seahorse analyzers and metabolic profiles through multiple powerful methodologies for elucidating molecular mechanisms including metabolomics, isotope tracing, and proteomics [23]. Our findings reveal profound bioenergetic dysregulation in acute phases post-TBI, with marked disruptions in purine nucleotide, glucose, amino acid, and fatty acid metabolism. Most of these pathways exhibited partial to full recovery by the subacute phase. Notably, enhancing fatty acid oxidation (FAO) via sodium octanoate administration restored mitochondrial bioenergetics, reduced microglial activation, and mitigated hippocampal neurodegeneration after TBI. These results underscore the therapeutic potential of targeting bioenergetic dysfunction and FAO to alleviate TBI-induced neurological damage.
Material and Methods
Animals
All animal experiments were conducted in accordance with the guidelines of the National Institutes of Health (NIH) and approved by the Institutional Animal Care and Use Committee (IACUC) of Tulane University. Male C57BL/6 J mice (3 months old; Stock No.: 000664) were obtained from Jackson Laboratory. Mice were randomly assigned to experimental groups, and all treatments were administered in a blinded manner. The animals were used for two experimental cohorts, including an untreated group and a sodium octanoate-treated group, as illustrated in Fig. 1, which outlines the experimental timeline, injury groups, and treatment strategy. The study design adhered to the ARRIVE guidelines. Animals were excluded based on the following predefined criteria: (1) signs of severe distress, including inability to eat or drink, poor grooming (ruffled fur), or marked reduction in spontaneous activity; (2) body weight loss exceeding 20%; (3) postoperative complications such as incision infection or excessive bleeding; and (4) seizures or other unexpected illnesses or injuries following surgery.Fig. 1. The experimental design and schematic diagram of the experimental timeline: A untreated group, and B group treated with sodium octanoate
Controlled Cortical Impact Models of Traumatic Brain Injury
Controlled cortical impact (CCI) models were generated as we described previously, with minor modifications [24–27]. This model exhibits cortical tissue loss and partial tissue loss in the hippocampal CA1 region, while the hippocampal DG area remains intact. Briefly, mice were anesthetized with 2.5% isoflurane (Anaquest, Memphis, TN, USA) in 70% N_2_O/30% O_2_ and positioned in a stereotactic frame with a gas anesthesia mask delivering 2% isoflurane via a Fluotec 3 Vaporizer (Colonial Medical Supply, Londonderry, NH, USA). The scalp was reflected to expose the skull, and a 5-mm craniectomy was performed over the left cerebral hemisphere (3 mm lateral to the sagittal suture, equidistant between lambda and bregma) using a microdrill (Braintree Scientific, #MD-1200) and a 5-mm trephine bit (Fine Science Tools, #18004-50). A moderate-to-severe cortical injury was induced using a pneumatic CCI device (Precision Systems and Instrumentation, TBI-0310) with a 3-mm flat-tipped impactor (velocity: 4.6 m/s; depth: 0.65 mm; dwell time: 500 ms). The scalp was closed with 4–0 silk sutures (Ethicon, #50–118–0796), and mice were monitored during recovery. Each surgical procedure lasted approximately 8–10 min. Sham-operated mice underwent craniectomy and suture without cortical impact. All mice were maintained on a heating pad after the surgical procedure until they regained consciousness. To minimize technical variability, the TBI-day 7 and Sham-day 7 groups were initiated 6 days before the TBI-day 1 and Sham-day 1 groups, allowing all mice to be sacrificed and sampled simultaneously.
Drug Administration
Sodium octanoate (Santa Cruz, #sc-212944) was administered as a single intraperitoneal (i.p.) injection at 100 mg/kg body weight in sterile 1 × PBS, 60 min post-CCI. Control mice received an equal volume of vehicle (PBS; Santa Cruz Biotechnology, #SC-281692) at the same time point.
Preparation of Oxygenated Artificial Cerebrospinal Fluid
Artificial cerebrospinal fluid (aCSF) was prepared as described previously [28], with minor modifications. To mimic physiological cerebrospinal fluid, the aCSF was composed of the following components in distilled water: 120 mM NaCl, 3.5 mM KCl, 1.3 mM CaCl_2_, 1 mM MgCl_2_ hexahydrate, 0.4 mM KH_2_PO_4_, 5 mM HEPES, 10 mM glucose. On the day of the experiment, 1 mM sodium pyruvate and 2 mM L-glutamine were added, and the pH was adjusted to 7.4 using 10 M NaOH. The aCSF was oxygenated by bubbling oxygen (95%) for approximately 15 min prior to contact with the slice tissue.
Euthanasia and Tissue Collection
For hippocampal slice preparation, mice were anesthetized at 1 or 7 days post-TBI using 3% isoflurane inhalation for 3 min, until unresponsive to noxious stimuli and the absence of reflexes was confirmed. Mice were then euthanized via decapitation using a guillotine, and brains were rapidly removed and immersed in ice-cold, oxygenated artificial cerebrospinal fluid (aCSF).
For immunostaining, at 48 h post-injury, mice were anesthetized with 2.5% isoflurane inhalation and transcardially perfused with ice-cold 1 × PBS, followed by 4% paraformaldehyde (PFA) in PBS. Brains were harvested and post-fixed in 4% PFA for 48 h, then either cryoprotected in 30% sucrose for 48 h prior to embedding in OCT compound (Thermo Fisher Scientific, #23–730-571) and stored at −20 °C, or processed for formalin-fixed, paraffin-embedded (FFPE) tissue sectioning.
Preparation of Brain Hippocampal Slices
Hippocampal slices are widely used for assessing mitochondrial bioenergetics using Seahorse assays [29, 30] and for metabolic analyses [31, 32]. Given the high sensitivity of brain sections to oxygen and glucose deprivation, we adapted a method commonly used in slice electrophysiology to preserve tissue viability during hippocampal slice preparation [29, 33]. Coronal brain sections (200 µm in thickness) were collected from consistent bregma levels (−1.46 to −2.46 mm) using a McIlwain tissue chopper (The Vibratome Company, O'Fallon, MO). Sections containing the hippocampus were transferred to a chamber containing room-temperature, oxygenated aCSF. Biopsy punches of the hippocampal DG region from each mouse were collected using a 0.75-mm stainless steel biopsy punch (World Precision Instrument, FL, USA). The entire procedure, from brain collection to tissue punch collection, was completed within approximately 45 min post-sacrifice. This timeframe is unlikely to induce significant metabolic alterations, as prior research indicates that TCA cycle intermediates (citrate, fumarate, and succinate) and fatty acids remain stable in brain slices during ex vivo incubation periods ranging from 35 min to 2.5 h [31]. For metabolomics analysis, three DG punches from each mouse were pooled and snap-frozen in liquid nitrogen as a single sample. For U-^13^C-Octanoate tracing, DG brain slices (3 slices per mouse, 200 µm thickness, 0.75 mm diameter) were incubated in aCSF containing 2 mM sodium octanoate (^13^C_8_, 99%; Cambridge Isotope Laboratories, CLM-9617-PK) for 30 min. After incubation, slices were washed with cold PBS and snap-frozen with liquid nitrogen. For Seahorse analysis, individual punches were placed on the mesh side of an XF islet capture microplate (Agilent, Cat. No. 101122-100) with the section facing down. Each well was immediately filled with aCSF, avoiding air bubbles, and the brain punches were gently centered in each well using a pipette fitted with a 200 µL tip.
Seahorse Analysis of Brain Hippocampal Slices
Seahorse analysis of brain hippocampal slices was performed using a standard protocol as described previously [29, 30]. One day prior to the assay, the sensor cartridge was hydrated by adding 1 mL of calibration buffer and incubating overnight at 37 ℃ in a non-CO_2_ incubator. Next, the microplate containing brain slices was incubated at 37 °C in the non-CO_2_ incubator for 40 min. During this time, the drug injection ports of the sensor cartridge were loaded with 10× stock solutions (prepared in aCSF): 316 µM oligomycin (Port A, 70 µL), 100 µM FCCP (Port B, 75 µL), and 100 µM Antimycin A/50 µM Rotenone (Port C, 80 µL). The cartridge was then loaded into the XFe24 analyzer for calibration for approximately 20 min. Following calibration, the sensor cartridge was transferred to the assay plate containing the brain slices. The mitochondrial stress test protocol included 3–5 measurement cycles for each injection step: baseline, oligomycin, FCCP, and Rotenone/Antimycin A. Each cycle consisted of a 3-min mix, 3-min wait, and 2-min measurement period.
High-Performance Liquid Chromatography and High-Resolution Mass Spectrometry and Tandem Mass Spectrometry (LC-MS/MS)
Metabolites were extracted from hippocampal brain slices using a methanol-chloroform phase separation, as we described previously [25, 27]. Briefly, tissue was homogenized in 2 mL of methanol containing 0.1 µL Metabolomics Amino Acid Mix (Cambridge Isotope Labs Incorporated, MSK-A2-1.2) and 1 nmol 13:0 Coenzyme A (Avanti Research, 870713P-5 mg), followed by centrifugation at 4 °C, 3000 rpm, 20 min. The upper aqueous phase was collected, evaporated under nitrogen, and analyzed by LC-MS/MS at the Harvard Center for Mass Spectrometry [25, 27]. For analysis, 5 µL of each sample was injected onto an IDX mass spectrometer (Thermo Scientific) equipped with a Zic-pHILIC column (150 × 2.1 mm, 5 µm, 40 °C, Sigma Aldrich). The mobile phases were as follows: (A) 20 mM ammonium carbonate with 0.1% ammonium hydroxide in water, and (B) 97% acetonitrile in water. The LC gradient began at 99% B, ramped to 40% B over 17 min, then to 0% B over 10 min, held for 5 min, returned to 99% B in 4 min, and re-equilibrated for 11 min at a flow rate of 0.15 mL/min. Data were acquired in polarity-switching mode at 120,000 resolution (m/z 65–1000; AGC target 1e5). MS1 files were used for both untargeted and targeted metabolomics. MS/MS spectra were acquired from pooled samples using AcquireX DeepScan (Thermo Scientific) in both positive and negative ion modes over four iterative cycles, incorporating blank exclusion and pooled sample inclusion lists. For targeted analysis, a 1 µM metabolite standard mix was run post-sample to verify retention times. Metabolite levels were normalized to internal controls. Untargeted data were processed using Compound Discoverer 3.1 (Thermo Fisher Scientific) and annotated using the mzCloud spectral library.
Metabolic Profiling Analysis
Untargeted LC-MS/MS metabolomics for global metabolic profiling and pathway enrichment analysis was performed as we described previously [27]. Processed metabolomics datasets are provided in Supplemental File 1. Multivariate statistical analyses, including heatmap visualization, partial least squares–discriminant analysis (PLS-DA), variable importance in projection (VIP) scoring, and random forest analysis, were performed using MetaboAnalyst 6.0 (http://www.metaboanalyst.ca). PLS-DA was used to visualize group separation and global metabolic shifts, with metabolites ranked by VIP scores. Metabolites with VIP scores > 1 were considered influential features. Random forest analysis was applied as a complementary, non-parametric method to identify discriminatory metabolites. Pathway enrichment analysis was performed using the MetaboAnalyst pathway analysis module, mapping metabolites to KEGG pathways with significance determined by pathway impact and adjusted p-values. For metabolomics analyses, data were log-transformed and auto-scaled prior to statistical testing to reduce heteroscedasticity and approximate normality, as confirmed by the Shapiro–Wilk test. Raw abundance values are shown in figures to preserve biological scale. Targeted metabolites analysis was subsequently performed to quantitatively assess predefined metabolites involved in key bioenergetic pathways, including glycolysis, the TCA cycle, amino acid metabolism, and fatty acid metabolism.
Proteomics Analysis
Proteomics data from our previous study [26] were reanalyzed for in-depth evaluation. During data preprocessing, proteins with zero expression across all samples were excluded. Overlaps in protein expression among sample groups were visualized using the VennDiagram (v1.7.3) and ComplexUpset (v1.3.3) R packages. Differential protein expression across regions and groups was assessed using univariate analysis of variance (ANOVA), with a two-tailed P value < 0.05 considered statistically significant. Gene Ontology (GO) enrichment analysis of differentially expressed proteins was performed using clusterProfiler (v4.12.0) and GO.db (v3.19.1), and semantic similarity among GO terms was evaluated using GOSemSim (v2.30.0). Additional data visualizations were generated using ggplot2 (v3.5.1), ComplexHeatmap (v2.20.0), and dendextend (v1.18.1). All analyses were conducted in R (v4.3.3) within RStudio (2024.04.1 Build 748).
Immunofluorescence
Immunofluorescence staining was performed as previously described [27]. Briefly, frozen brain tissues embedded in OCT were sectioned into 20-µm slices using a Leica CM1950 cryostat (Leica Biosystems, Nussloch, Germany). Free-floating sections were washed three times in 1 × PBS (10 min each), followed by incubation in blocking solution containing 5% donkey serum (Sigma Aldrich, #D9663), 0.3% Triton X-100 (Sigma Aldrich, #X100-500ML), and 3% BSA in 1 × PBS for 1 h at room temperature. Sections were then incubated overnight at 4 °C with primary antibodies diluted in the same blocking solution. After three additional PBS washes, sections were incubated for 1 h at room temperature with the appropriate secondary antibodies diluted in 1 × PBS. Following another three PBS washes, sections were mounted on slides and coverslipped using VECTASHIELD Antifade mounting medium with DAPI (Vector Labs, #H-1200-10). Images were captured using an Olympus IX-83 confocal laser scanning microscope (Olympus Life Sciences, Tokyo, Japan). The following antibodies were used: Rabbit anti-Iba1 (Fujifilm Irvine Scientific Inc. Nc9288364, 1:200), Rabbit anti-Hadhb (Proteintech, 29091-1-AP, 1:200), Rabbit anti-Decr1 (ABclonal, A13014, 1:100), Mouse anti-GFAP (Cell Signaling, #3670, 1:200), Mouse anti-NeuN (Millipore Sigma, MAB377, 1:200), Donkey anti-rabbit Alexa Fluor 488 (Thermofisher, A21206, 1:1000), and Donkey anti-mouse Alexa Fluor 594 (Thermofisher, A-21203, 1:1000).
Neurodegeneration Histology
To assess treatment-associated differences in neurodegeneration within the dentate gyrus (DG) of the hippocampus at day 3 post-injury, both terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and Fluoro-Jade C (FJC) staining were performed. Formalin-fixed paraffin-embedded (FFPE) brain samples were sectioned at 5 µm using a Microm HM 340 E microtome (Microm International, Walldorf, Germany) and mounted onto slides. Deparaffinization was carried out with four 10-min xylene washes, followed by rehydration through a graded ethanol series (2 × 100%, 2 × 95%, and 1 × 70% ethanol, 3 min each). Slides were then rinsed in 1 × PBS for 5 min.
For TUNEL staining, rehydration was followed by heat-induced epitope retrieval via submersion in boiling 1 mM trisodium citrate buffer (pH = 6.0) for 10 min. Slides were cooled for 15 min, rinsed three times in PBS, and incubated in a blocking solution for 1 h. Sections were then incubated overnight at 4 °C with mouse anti-NeuN primary antibody (clone A60, Millipore, MAB377, 1:200) diluted in incubation buffer. The next day, slides were washed (3 × in PBS) and incubated with donkey anti-mouse Alexa Fluor 555 secondary antibody (Invitrogen, A31570, 1:500) for 60 min at room temperature. Following another PBS wash (3 ×), slides were incubated in the TUNEL reaction mixture (In Situ Cell Death Detection Kit, Fluorescein; Roche, #11684795910) for 1 h at room temperature in the dark. After a final PBS wash, slides were coverslipped with VECTASHIELD Antifade Mounting Medium with DAPI (Vector Laboratories, H-1200-10) and imaged using an Olympus IX-83 Confocal Laser Scanning Microscope (Olympus Life Sciences, Tokyo, Japan).
For Fluoro-Jade C (FJC) staining, rehydration was followed by immersion in diH2O for two minutes. FJC staining was achieved using the FJC Ready-to-Dilute Staining Kit for detecting degenerating neurons (Biosensis, TR-100-FJ) following the manufacturer’s recommended protocol for FFPE sections. After staining, slides were air dried at 60 °C for 20 min and coverslipped using DPX mounting medium (Sigma-Aldrich, #1.00579). Imaging was performed with a Nikon Eclipse Ts2R Inverted Fluorescence Microscope (Nikon Corporation, Tokyo, Japan).
Quantitative Analysis of Iba1⁺, TUNEL⁺, and FJC⁺ Cells
For quantitative analysis of Iba1⁺ microglia, TUNEL⁺ apoptotic cells, and Fluoro-Jade C (FJC)⁺ degenerating neurons, coronal brain sections at the level of the CCI lesion (approximately bregma − 2.0 mm) were stained for the respective markers. Whole coronal sections were imaged, and all analyses were performed by a blinded investigator using ImageJ software (NIH). For each section, the average background fluorescence intensity was calculated, and pixels exceeding two standard deviations above background were considered positive for the corresponding antigen. The DG was manually delineated, and positively labeled cells within this region were quantified using ImageJ’s “Analyze Particles” function. Cell counts were subsequently normalized to the two-dimensional area of the DG within each coronal section. For microglial quantification, whole-section images were first masked using the DAPI channel to ensure that only cells with nuclei in the plane of section (i.e., DAPI⁺/Iba1⁺ cells) were included, thereby excluding processes from out-of-plane microglia. Quantification then proceeded using the same approach as described for TUNEL and FJC analyses.
Statistical Analysis
Sample sizes were determined through an a priori power analysis based on pilot data and prior literature [27], conducted using G*Power 3.1 with an assumed large effect size (Cohen’s f = 0.8), α = 0.05, and power (1 − β) = 0.8. Statistical analyses were performed using GraphPad Prism 10. Data normality was assessed using the Shapiro–Wilk test. Comparisons between two groups were conducted using unpaired Student’s t-tests, while comparisons among more than two groups were performed using one-way ANOVA followed by Tukey–Kramer post hoc tests to correct for multiple comparisons. For analyses involving multiple time points, one-way ANOVA was used to focus on a priori planned comparisons within each time point rather than testing injury × time interactions. Data are presented as mean ± standard deviation (SD), with statistical significance defined as P < 0.05.
Results
Seahorse analysis reveals bioenergetics impairment in the hippocampus during the acute phase following TBI
To evaluate hippocampal bioenergetics during the acute and subacute phases after TBI, we performed Seahorse extracellular flux analysis using the XFe24 analyzer on acute mice brain tissue punches, adapted from a previously published method [30] (Fig. 2A). Tissue punches were collected from the DG region at days 1 and 7 post-sham or CCI (Fig. 2B). Analysis of the bioenergetic profiles of the hippocampal DG with Seahorse OCR analysis (Fig. 2C) revealed no significant differences between Sham-day 1 and Sham-day 7 groups (Fig. 2D). Therefore, to reduce animal use and technical variability, a single sham group was used in comparison with both the 1-day and 7-day post-TBI samples in subsequent experiments. Notably, mitochondrial basal respiration, maximal respiration, and ATP production were significantly reduced at day 1 post-TBI, which was not sustained by day 7 (Fig. 2D). Collectively, these findings demonstrate that hippocampal mitochondrial bioenergetics is markedly impaired during the acute, but not in the subacute phases, following TBI.Fig. 2. The effect of TBI on mitochondrial respiration in the hippocampal slices.** A** Brain slices (200 µm) were collected with stainless steel biopsy punches from the hippocampus of sham (day 1), sham (day 7), TBI (day 1) and TBI (day 7) mice with McIlwain Tissue Chopper and placed into 24 wells Seahorse plate, which was filled with 700 µl of aCSF, followed by OCR measurement in Seahorse. The graph is created with BioRender (https://www.biorender.com/). B Representative coronal section indicating the dentate gyrus (DG; arrow) of the mouse hippocampus used for Seahorse analysis. C Seahorse oxygen consumption rate (OCR) analysis in DG slices under basal conditions and following sequential injections of oligomycin (Oligo), FCCP, Rotenone/antimycin (AA/RO). D Quantification of mitochondrial OCR parameters, including Basal respiration, Maximal respiration, and ATP production. Data are presented as mean ± SD (n = 5–8 mice per group; each point represents one mouse brain slice). Statistical comparisons were made by one-way ANOVA with Tukey’s post hoc test; *p < 0.05, **p < 0.01, ***p < 0.001
Untargeted metabolomic profiling identifies a significant reduction in metabolites involved in energy production and purine metabolism in the hippocampus during the acute phase post-TBI
To gain a comprehensive understanding of the metabolic profile in the hippocampal DG at days 1 and 7 post-TBI, we performed untargeted metabolomic analyses using LC-MS/MS. Untargeted metabolomics data (Supplemental File 1) were analyzed with MetaboAnalyst 6.0. Partial Least Squares-Discriminant Analysis (PLS-DA) revealed distinct metabolic shifts among hippocampal DG slices from Sham, TBI-1d, and TBI-7d groups (Fig. 3A). Heatmap analysis showed a greater number of downregulated metabolites on day 1 post-TBI compared to Sham controls (Fig. 3B). Volcano plot analysis further demonstrated that DG slices from TBI-1d exhibited significant metabolite reductions, while TBI-7d samples displayed both increases and decreases in metabolite levels. Compared to TBI-1d, the TBI-7d group showed broader recovery of metabolite levels (Supplemental Fig. 1). Pathway analysis (TBI-1d vs Sham, TBI-7d vs Sham, and TBI-7d vs TBI-1d) identified purine metabolism as one of the most significantly impacted pathways (Supplemental Fig. 1). Furthermore, Random Forest plot analysis indicated that metabolites involved in energy production and purine metabolism, including Adenosine diphosphate (ADP), Guanosine diphosphate (GDP), and Adenosine, were significantly reduced at day 1 or day 7 post-TBI (Fig. 3C). Further analysis revealed that energy-associated metabolites, including adenosine monophosphate (AMP) and ADP, were significantly decreased at day 1, but not at day 7 post-TBI (Fig. 3D). Guanosine monophosphate (GMP) and GDP levels were also significantly reduced at both time points, though to a lesser extent at day 7 post-TBI (Fig. 3D). Notably, glucose, the key brain fuel source of the brain, was also significantly reduced at day 1, but not at day 7 post-TBI (Fig. 3D). Collectively, these findings indicate that TBI induces acute metabolic deficits in energy production and purine metabolism within the hippocampus, most prominently during the early post-injury phase.Fig. 3. Untargeted metabolomic profiling of hippocampal DG following traumatic brain injury.** A** Partial least squares–discriminant analysis (PLS-DA) scores plot showing clear separation of Sham (black), TBI-1d (red), and TBI-7d (blue) groups. B Hierarchical clustering heatmap of differentially abundant metabolites across the three conditions (rows: top 50 features; columns: individual biological replicates). C Pathway impact analysis from untargeted data, ranking the top altered metabolic pathways by Random Forest mean decrease accuracy. D Quantitative changes in select metabolites identified by untargeted LC-MS/MS, including the metabolites involved in energy and purine nucleotides metabolism, such as adenosine monophosphate (AMP), adenosine diphosphate (ADP), guanosine monophosphate (GMP), guanosine diphosphate (GDP), and key brain fuels, including glucose, glutamine. Data are expressed as mean ± SD (n = 3–4 per group). Normality was assessed using the Shapiro-Wilk test. To satisfy the assumptions of parametric tests, metabolomics data were log-transformed prior to analysis. Statistical comparisons were performed by one-way ANOVA followed by Tukey’s post hoc test; *p < 0.05, **p < 0.01, ***p < 0.001
Targeted metabolites analysis revealed significant dysregulation of glycolysis, the glutamine-α-KG axis, and medium-chain fatty acid metabolism in the hippocampus during the acute phase post-TBI
To assess metabolic pathway alterations, we performed targeted metabolite analysis focusing on glycolysis, the TCA cycle, amino acid and fatty acid metabolism (Fig. 4A). In this hippocampal DG at day 1 post-TBI, upstream glycolytic intermediates (glucose-phosphate and fructose-phosphate) were significantly elevated, whereas lactate, a downstream glycolytic end product, was markedly reduced and restored by day 7 post-TBI (Fig. 4B). Analysis of TCA cycle intermediates showed a significant reduction in α-ketoglutarate (α-KG) at day 1 post-TBI, with other TCA metabolites remaining unchanged at both day 1 and day 7 (Fig. 4C). Regarding amino acids, glutamate levels were significantly decreased at day 1 but not day 7 post-TBI, whereas other amino acids showed no significant changes (Fig. 4D). In fatty acid metabolism, long-chain fatty acids (C10–C16) remained stable across all time points (Fig. 4E). However, we observed a significant reduction in medium-chain fatty acid (C8, octanoate) levels at day 1 post-TBI, but not at day 7 (Fig. 4E). Short-chain fatty acids (C6) followed a similar but nonsignificant trend (Fig. 4E). Collectively, these results indicate acute dysregulation of glycolysis, the glutamine-α-KG axis, and medium-chain fatty acid metabolism in the hippocampus at day 1 post-TBI, which largely resolves by day 7.Fig. 4. Targeted metabolite analysis of hippocampal DG following traumatic brain injury.** A** Schematic overview of central metabolic pathways analyzed. B Targeted LC-MS/MS quantification of glycolytic intermediates, including glucose-phosphate, fructose-phosphate, 3-phosphoglycerate, phosphoenolpyruvate, pyruvate, and lactate. C Targeted LC-MS/MS quantification of TCA cycle metabolites, including citrate, aconitate, α-ketoglutarate (α-KG), succinate, fumarate, and malate. D Targeted LC-MS/MS quantification of amino acids, including glutamate, aspartate, alanine, leucine, lysine, and serine. E Targeted LC-MS/MS quantification of fatty acids, including C6, C8 (octanoate), C10, C12, and C16 fatty acids. Data are expressed as mean ± SD. (n = 3–4 mice per group). Statistical comparisons were performed by one-way ANOVA with Tukey’s post hoc test; *p < 0.05, **p < 0.01, ***p < 0.001
Medium-chain fatty acid isotope tracing revealed enhanced FAO flux through the TCA cycle in the hippocampus during the subacute phase post-TBI
Given the transient reduction in octanoate in the hippocampus at day 1 post-TBI, we examined the metabolic flux of C8 fatty acid through the TCA cycle using a U-^13^C-octanoate tracer (Fig. 5A). At day 7, but not day 1 post-TBI, we observed decreased M + 0 and increased M + 1 isotopologues of C06 and C04 fatty acids (Fig. 5B). Similarly, acetoacetate, derived from acetyl-CoA, showed reduced M + 0 and increased M + 1 and M + 4 labeling at day 7 (Fig. 5B). These data suggest a significant elevation of fatty acid β-oxidation in the hippocampus at the subacute phase post-TBI. We next analyzed U-^13^C-octanoate-derived TCA cycle intermediates (α-KG, succinate, malate) and aspartate. At day 1 post-TBI, α-KG and succinate displayed increased M + 0 and reduced M + 1 labeling. In contrast, at day 7 post-TBI, there is a reduction in M + 0 and an elevation in M + 1 or M + 5 labeling (Fig. 5C). These data revealed that fatty acid-derived TCA cycle oxidative metabolism, located at the upstream of succinate, is inhibited at day 1 post-TBI but elevated at day 7 post-TBI. Moreover, malate and aspartate exhibited persistently decreased M + 0 labeling, and aspartate showed a persistent elevation of M + 1 labeling at both time points (Fig. 5C), suggesting sustained FAO contribution downstream of succinate. Collectively, these results demonstrate that FAO-driven TCA flux is markedly elevated during the subacute phase following TBI.Fig. 5. The effect of TBI on the oxidative metabolism of fatty acids in the hippocampus post-TBI.** A** Schematic presentation of the contribution of U-^13^C-Octanoate-derived carbons to TCA cycle intermediates and amino acids; ^13^C carbons are depicted in blue and ^12^C carbons in black. B Mass isotopomer distributions of short-chain fatty acids (C06 and C04) and acetoacetate. C Mass isotopomer distributions of TCA cycle intermediates, including α-ketoglutarate, succinate, malate, and associated amino acids aspartate. n = 3. Data are expressed as mean ± SD, *p < 0.05, **p < 0.01, and ***p < 0.001. One-way ANOVA followed by Tukey’s post hoc analysis
Spatial proteomics uncovers dysregulation of bioenergetics-related proteins involved in purine nucleotide metabolism and FAO in the hippocampus following TBI
We recently performed region-specific spatial proteomics to profile general protein abundance and global pathway enrichment across hippocampal subregions, including DG1, DG2, pyramidal layer (PL), and stratum moleculare (SM), during the acute (day 1) and subacute (day 7) phases post-TBI [26]. In this study, we reanalyzed the raw dataset with a refined focus on GO enrichment of differentially expressed proteins to uncover the molecular mechanisms driving dynamic changes in bioenergetics and metabolic profiles in the hippocampal region following TBI. In total, 3890 proteins were detected, with 3638 proteins co-expressed across all 4 hippocampal regions. Unique proteins identified in each subregion included 7 in DG1, 2 in DG2, 4 in PL, and 5 in SM (Supplemental Fig. 2A). We further performed differentially expressed protein analysis across various comparison groups (Sham vs. TBI day 1; Sham vs. TBI day 7; TBI day 7 vs. TBI day 1) within each hippocampal subregion. This analysis revealed 693 differentially expressed proteins in DG1, 587 in DG2, 675 in PL, and 379 in SM (Supplemental Fig. 2B-C).
To gain functional insights, we conducted GO enrichment analysis on the differentially expressed proteins and visualized the enrichment results across the four hippocampal regions (Fig. 6A). We selected the top five upregulated and downregulated pathways in each comparison group and performed GO term semantic similarity clustering to better understand pathway alterations over time (Fig. 6B). With respect to metabolic processes, we found that FAO in the DG1, DG2, and PL regions was significantly upregulated at day 7 post-TBI compared to day 1, indicating the potential enhancement of FAO during the subacute phase (Fig. 6B). The FAO process includes a total of 117 proteins, of which 41 were detected in our proteomics dataset (Fig. 6C). Notably, key proteins involved in this process, including Acadl, Acadvl, Acsl4, Decr1, and Hadhb, were prominently affected (Fig. 6D). Immunostaining validation of the FAO proteins revealed significantly elevated levels of Hadhb and Decr1 in the DG at day 7 post-TBI, compared to sham controls and day 1 post-TBI. Furthermore, we found that Hadhb was primarily expressed in astrocytes (GFAP^+^ cells), and Decr1 was partially localized to astrocytes (GFAP^+^ cells) (Fig. 6E).Fig. 6. Proteomic profiling reveals significant alterations in hippocampal bioenergeticpathways, particularly FAO, following TBI.** A** Bubble plot of GO Biological Process enrichment for differentially expressed proteins in DG1, DG2, PL, and SM regions. For each comparison (Dpi 1 vs Sham; Dpi 7 vs Sham; Dpi 7 vs Dpi 1), the top five upregulated and downregulated terms are shown. Bubble size ∝ –log_10_(p.adjust); color indicates direction (red = upregulated, blue = downregulated). B GO term semantic similarity clustering for significantly altered proteins across regions and comparisons. C Heatmaps of log_2_ fold-changes in FAO-related proteins for each comparison. Columns represent comparisons; shading indicates fold-change direction (red = up, blue = down); asterisks denote statistical significance (*p < 0.05; **p < 0.01; ***p < 0.001). Bar plots below summarize the numbers of significantly up- and downregulated proteins per comparison. D Boxplots showing normalized abundance of key FAO proteins in Sham (green), Dpi 1 (red), and Dpi 7 (blue) groups. *p < 0.05, **p < 0.01, ***p < 0.001 by one-way ANOVA with Tukey’s post hoc test. E Representative immunofluorescence images of Hadhb or Decr1 (green) co-stained with GFAP (red; astrocyte marker) in the hippocampal DG of the ipsilateral hemisphere at the indicated time points post-TBI
In addition, we found that the purine nucleotide triphosphate metabolic process was downregulated in the hippocampus as early as 1 day post-TBI (Fig. 6B). The purine nucleoside triphosphate metabolic process comprised 169 proteins, 98 of which were detected in our dataset (Supplementary Fig. 3A). Differential analysis indicated a general downregulation of this pathway post-injury; however, temporal dynamics were observed. Notably, critical proteins in the DG1 area, including Adk, Ak2, Guk1, Nudt2, and Stoml2, were significantly reduced at day 1 but exhibited varying degrees of recovery toward control levels by day 7 (Supplementary Fig. 3B).
To further investigate alterations in energy metabolism, we also analyzed proteins associated with the glycolytic pathway, although it was not among the top five enriched pathways. The glycolytic pathway comprises 104 annotated proteins, of which 36 were detected in our proteomic dataset (Supplementary Fig. 3C). In contrast to the pronounced changes observed in FAO and purine nucleotide triphosphate metabolism, key glycolytic proteins exhibited relatively modest alterations following TBI (Supplementary Fig. 3D).
Sodium octanoate administration partially restores mitochondrial bioenergetics in the hippocampus following TBI
Given that octanoate (C8 fatty acids) levels are reduced on day 1 and restored by day 7 post-TBI, and that octanoate-derived FAO, as well as the proteins associated with FAO, are markedly upregulated on day 7, we hypothesized that enhancing FAO through exogenous octanoate supplementation could improve hippocampal bioenergetics and ultimately neurological outcomes following TBI. Notably, octanoate is a blood–brain barrier (BBB) permeable medium-chain fatty acid, comprising up to 13% of the circulating free fatty acid pool in humans and contributing approximately 20% to total brain oxidative energy production [34]. Moreover, it has been well-characterized that sodium octanoate administration dose-proportionally elevates brain octanoate levels [35]. Based on this, we administered sodium octanoate to CCI mice and assessed hippocampal bioenergetics on day 1 post-TBI using the Seahorse analyzer (Fig. 7A). We observed that TBI significantly reduced both basal and maximal respiration in hippocampal slices. While maximal respiration did not significantly recover, sodium octanoate treatment partially restored basal mitochondrial respiration (Fig. 7B). These findings suggest that sodium octanoate confers a modest improvement in mitochondrial bioenergetics following TBI.Fig. 7. The effect of sodium octanoate administration on bioenergetics in the hippocampus post-TBI.** A** Seahorse oxygen consumption rate (OCR) analysis in DG slices from Sham, TBI, and TBI + sodium octanoate mice under basal conditions and following sequential injections of Oligomycin (Oligo), FCCP, Rotenone/antimycin (AA/RO). n = 4–7 mice per group. B Quantification of mitochondrial OCR parameters, including Basal respiration, Maximal respiration, and ATP production. Data are expressed as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001. One-way ANOVA followed by Tukey’s post hoc analysis
Sodium octanoate administration reduces microglia counts and neurodegeneration in the hippocampus following TBI
To determine whether sodium octanoate attenuates microglial accumulation after TBI, the Iba1 + cells in the hippocampal regions were quantified (Fig. 8A). The results suggest that sodium octanoate significantly reduces TBI-increased microglial numbers within the hippocampal DG at 48 h post-injury (Fig. 8B). To test whether the improvement of hippocampal bioenergetics by sodium octanoate can contribute to reduced acute brain damage, neuronal cell death in the hippocampal area was assessed. TUNEL staining (Fig. 8C) revealed a marked reduction in TUNEL-positive neurons within the DG of sodium octanoate-treated mice compared to controls at day 3 post-injury (Figs. 8D), indicating that sodium octanoate treatment reduced acute neuronal apoptosis in this hippocampal subregion. This finding was corroborated by FJC staining (Fig. 8E), in which treatment with sodium octanoate resulted in a significant reduction in the degenerating neurons in the DG at day 3 post-injury, independent of cell death mechanism (Figs. 8F). Together, these results suggest that sodium octanoate administration following TBI mitigates hippocampal bioenergetic dysfunction, reduces microglial accumulations, and promotes neuronal survival in this vulnerable brain region.Fig. 8. The effect of sodium octanoate administration on neuroinflammation and neurodegenerationin the hippocampus post-TBI.** A**, B Representative immunofluorescent staining (A) and quantitation analysis (B) of microglia (Iba1 + positive cells) in the perilesional hippocampal DG area of the ipsilateral brain of Sham, TBI, and TBI + sodium octanoate mice at 48 h after TBI. Scale bar, 50 μm. n = 4. C, D TUNEL staining (C) and quantitation analysis (D) of apoptotic cells in the perilesional hippocampal DG area of the ipsilateral brain of Sham, TBI, and TBI + sodium octanoate mice at 3 days post-TBI. Scale bar, 200 µm. n = 4. E, F Fluorojade C staining (E) and quantitation analysis (F) of neurodegenerative cells in the perilesional hippocampal DG area of the ipsilateral brain of Sham, TBI, and TBI + sodium octanoate mice at 3 days post-TBI. Scale bar, 200 µm. n = 4. Data are expressed as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001. One-way ANOVA followed by Tukey’s post hoc analysis
Discussion
In this study, we identified a dynamic pattern of hippocampal bioenergetic remodeling following TBI that is strongly dependent on the injury phase. First, Seahorse bioenergetics assays revealed a significant decline in mitochondrial function at day 1 post-TBI, which was not sustained by day 7. Untargeted metabolomics showed broad reductions in metabolites related to energy production and purine metabolism, while targeted metabolite analyses indicated acute dysregulation of glycolysis, suppression of the glutamine–α-KG axis and medium-chain fatty acid metabolism at day 1. Notably, these metabolic abnormalities largely resolved by day 7 post-TBI. U-^13^C-octanoate isotope tracing demonstrated enhanced FAO flux through the TCA cycle during the subacute phase, reflecting a metabolic shift toward FAO-driven energy production. Spatial proteomics further supported this temporal distinction, revealing significant suppression of purine nucleotide metabolism-related proteins at day 1 and elevated FAO-related protein expression by day 7. Third, octanoate supplementation partially restored mitochondrial respiration and reduced microglial accumulation and neurodegeneration in the hippocampus. Together, these findings (summarized in Fig. 9) highlight the dynamic bioenergetic adaptations of the hippocampus after TBI and support targeting FAO enhancement as a promising strategy to mitigate acute trauma-induced cerebral metabolic failure and secondary injury.Fig. 9. Proposed working model.** A** Schematic summary illustrating hippocampal metabolic and bioenergetic reprogramming during the acute and subacute phases following TBI. The hippocampal DG exhibits acute dysregulation of glycolysis, inhibition of the glutamine-α-KG axis and medium-chain fatty acid metabolism, and suppression of OXPHOS and ATP production, which are partially restored during the subacute phase post-TBI. B Conceptual model proposing that therapeutic enhancement of FAO may restore bioenergetic homeostasis and prevent neurodegeneration after TBI. The graph is created with BioRender (https://www.biorender.com/)
Our Seahorse assay analysis of hippocampal brain slices from a moderate-to-severe TBI mouse model revealed significant bioenergetic impairment at the acute (day 1), but not subacute (day 7), post-injury phase. A key advantage of this approach is the ability to measure mitochondrial respiration parameters in spatially defined brain subregions, although glycolytic capacity or reserve cannot be accurately assessed [29]. To our knowledge, hippocampal slice respiration has not previously been evaluated in moderate-to-severe TBI. One study using a mild fluid percussion injury (mFPI) model in rats reported no effect on mitochondrial respiration in CA1, reduced respiration in CA3, and elevated respiration in DG at 24 h post-injury [29]. These discrepancies may stem from differences in TBI severity and experimental models. Supporting this, in vivo proton magnetic resonance spectroscopy (^1^H-MRS) has shown that moderate, but not mild TBI, induces a transient reduction in N-acetylaspartate (NAA), a marker of neuronal viability and mitochondrial metabolism, within 4–24 h post-injury, followed by a delayed increase in lactate by day 5 in the rat hippocampus [22]. In line with our observation, another study using a CCI model in mice demonstrated significant mitochondrial impairment as early as 30 min post-injury, with persistent reductions in respiratory capacity up to 72 h. Specifically, complex I- and II-driven state III (ATP-linked) respiration was decreased by 79% and 81%, respectively, in the hippocampus [19]. Together, these findings suggest that acute hippocampal bioenergetic dysfunction is severity-dependent and temporally restricted, with partial recovery or metabolic adaptation emerging during the subacute phase. Accordingly, therapeutic strategies targeting mitochondrial metabolism may need to be phase-specific and sustained to achieve optimal benefit.
Further metabolic analyses suggest that impaired purine nucleotide metabolism may contribute to the reduction in hippocampal bioenergetics observed on day 1 post-TBI. Specifically, we detected reductions in all measured nucleotides (ADP, AMP, GDP, and GMP) in hippocampal DG slices post-TBI. While this appears inconsistent with earlier reports of increased AMP and ADP following ATP hydrolysis after TBI [36], such discrepancies may reflect variation in injury severity and the timing of analysis. For example, increased AMP has been observed at 15–40 min post-injury, coinciding with early ATP depletion [36], whereas by 6 h post-TBI, ADP and AMP levels, along with ATP, are significantly reduced [37]. Similar reductions in ADP and AMP (∼20%) have been reported in the injured hippocampus 24 h after hypoxia/ischemia [38], which is a significant pathophysiology of severe TBI [39]. These findings suggest that although AMP and ADP may transiently accumulate, they ultimately decline post-TBI. Our proteomic analysis further supports this conclusion, revealing marked downregulation of purine nucleoside triphosphate metabolism–related proteins in hippocampal subregions at day 1 post-TBI [26]. This suggests that bioenergetic failure and nucleotide depletion may be largely driven by loss of ATP metabolism–associated enzymes. However, we cannot rule out additional contributing factors, including impaired adenine nucleotide translocator function [40], enhanced nucleotide degradation and consumption [41], and possible neuronal loss. These mechanisms warrant further investigation.
Moreover, reduced glucose levels may also contribute to bioenergetics deficits in hippocampal slices post-TBI. Despite the decrease in glucose, upstream glycolytic intermediates such as glucose-phosphate and fructose-phosphate were elevated at day 1, potentially reflecting a compensatory increase in glycolytic flux. However, lactate, the endpoint of glycolysis, was significantly reduced at day 1 post-TBI, suggesting a downstream bottleneck in glycolysis, possibly due to impaired NAD⁺ regeneration or feedback inhibition of key enzymes. Hyperglycolysis is a well-documented acute response to TBI [42], typically occurring within minutes and lasting for hours, depending on the injury’s type and severity [43]. This is often followed by a phase of metabolic depression, characterized by reduced glucose metabolism and OXPHOS [42, 44, 45]. In severe TBI, early post-ischemic conditions, marked by reduced cerebral blood flow and oxygen deprivation, can suppress glucose uptake, utilization, and oxidation, impairing both glycolysis and OXPHOS [46]. Thus, the combined glycolytic and OXPHOS deficits observed at day 1 likely reflect both injury severity and the timing of metabolic evolution. Although most TCA cycle intermediates were unchanged, α-KG and glutamate were selectively reduced at day 1 post-TBI. Given that glutamate can be converted to α-KG via GDH, these findings suggest impaired anaplerotic flux into the TCA cycle. Consistently, previous studies have reported diminished GDH activity in microglia [27] and in mitochondria isolated from the ipsilateral cortex and striatum [47] during the acute phase of TBI. Interestingly, our spatial proteomics [26] revealed increased GDH expression in the hippocampal DG2 area on day 1 post-TBI, possibly reflecting a compensatory response to decreased GDH enzymatic activity. These results suggest that restoration of GDH function may serve as a viable strategy to enhance hippocampal bioenergetics during the acute phase following TBI.
The temporal enhancement of FAO may provide a mechanistic explanation for why hippocampal bioenergetic and metabolic disturbances are pronounced during the acute phase but are not sustained into the subacute phase following TBI. Although long-chain fatty acid levels remain unchanged in hippocampal slices following TBI, levels of the medium-chain fatty acid C08 (octanoate) are significantly reduced at day 1 post-TBI and largely recovered by day 7. This suggests an acute impairment in FAO, leading to a transient depletion of medium- or short-chain fatty acids in hippocampal DG slices post-TBI. Consistent with this interpretation, U-^13^C-octanoate isotope tracing demonstrated enhanced FAO flux in the hippocampal DG at day 7 post-TBI, and spatial proteomics revealed substantial recovery of FAO-related pathways by this time point. Together, these findings highlight dynamic, phase-dependent alterations in FAO-linked mitochondrial bioenergetics following TBI. Medium-chain fatty acids are essential energy substrates, contributing up to 20% of the brain’s energy supply [34]. In circulation, fatty acids are typically bound to serum albumin, with free fatty acid concentrations estimated at 444 µM, approximately 74% of albumin levels (600 µM) [48]. Notably, FAO is enhanced in the brain under conditions of impaired glucose metabolism, such as ischemic stroke [49] and AD [50]. Similarly, prior studies have shown that following TBI, the brain tends to increase FAO, particularly in the subacute phase after the initial injury [51]. Therefore, this compensatory enhancement of FAO may represent the brain’s adaptive mechanism to utilize alternative fuel sources to support energy metabolism and promote functional recovery.
While treatments with fatty acids omega-3 [52] or fatty acids transporter acetyl L-carnitine [53] have shown beneficial effects in TBI, their mechanisms are primarily attributed to anti-inflammation and neuroprotection. However, whether directly enhancing FAO in the TBI brain confers metabolic and functional benefits remains to be investigated. Octanoate, a C8 medium-chain fatty acid, is known to cross the BBB [54] and is readily taken up by cells to undergo mitochondrial FAO for energy production [55]. Notably, octanoate supplementation has been shown to improve brain energy metabolism and cognitive performance in patients with mild cognitive impairment and AD [56, 57]. In this proof-of-concept study, we tested whether enhancing FAO during the acute phase of TBI could restore mitochondrial bioenergetics. Our results demonstrate that sodium octanoate supplementation partially improves mitochondrial bioenergetics, reduces microglial accumulation, and attenuates neurodegeneration in the hippocampus following TBI. However, in the treated group, mitochondrial bioenergetics was assessed only at day 1 post-TBI, despite having established baseline bioenergetic profiles at both day 1 and day 7 in the untreated cohorts. The lack of day 7 data in the treated group limits our ability to assess the duration and progression of sodium octanoate’s effects. Future studies incorporating longitudinal assessments will be essential to determine whether the observed benefits are sustained or transient. Interestingly, astrocytes are considered the primary site for octanoate metabolism in the brain [34]. They metabolize octanoate to produce ketone bodies [58] and promote GABA synthesis in neurons by increasing glutamine supply [48]. Notably, we found that FAO-associated proteins (Hadhb and Decr1) are mainly expressed in astrocytes. Thus, octanoate may enhance glial FAO and promote neuron-astrocyte metabolic coupling, ultimately reducing neuronal death and secondary brain damage after TBI. While this mechanism cannot yet be directly validated in vivo, it will be explored in future studies using neuron-astrocyte co-culture systems.
We acknowledge several limitations of this study. First, the bioenergetic and metabolic profile of the hippocampus in females following TBI has not been investigated. Previous studies have shown that sex influences hippocampal gene expression after TBI, with evidence suggesting that females may exhibit distinct metabolic responses[6]. These observations warrant further investigation. Second, ex vivo tissue handling, including slicing and incubation, can influence metabolite levels and tissue viability, even in sham controls. Although Seahorse extracellular flux analysis has been widely and successfully applied to acute brain slices, including hippocampal DG regions of approximately 200 µm [29, 30, 59], hippocampal tissue remains highly sensitive to transient oxygen deprivation. However, as all groups were processed identically, the relative differences, particularly in mitochondrial metabolism, TCA cycle intermediates, and FAO, remain biologically meaningful. Future in vivo metabolic analyses may help refine baseline measurements. Third, although Seahorse analyses showed no significant differences between day 1 and day 7 sham groups, time-matched controls would provide a more rigorous reference, especially for metabolomics and proteomics data, where subtle temporal effects could exist. Moreover, although Sham animals are standard comparators in TBI models, we observed a tendency of decline in mitochondrial basal respiration from day 1 to day 7 post-Sham surgery. This raises uncertainty as to whether the apparent recovery at Day 7 post-TBI reflects true physiological improvement or normalization relative to an unstable Sham baseline. Future studies that include a naïve (non-surgical) control will be important for accurately isolating TBI-specific metabolic changes. Fourth, although our metabolomics and proteomics data indicate changes in the hippocampal DG area post-TBI, they do not delineate metabolic dysregulation at the level of specific cell types, such as neurons, astrocytes, or microglia. Future studies employing single-cell analysis techniques could provide more detailed insights into cell-type-specific metabolic alterations. Lastly, we acknowledge that normalizing Seahorse or metabolomics data to protein concentration would improve the rigor of analyses using hippocampal slices. In our study, the CCI model did not cause significant tissue loss in the hippocampal DG, and TUNEL staining revealed limited cell death in this region, supporting the reliability of our quantification. However, we recognize that potential variability due to subtle cell loss cannot be completely ruled out. In future studies, incorporating protein-based normalization will help strengthen the validity of our findings.
In summary, our study reveals a dynamic remodeling of hippocampal bioenergetic and metabolic pathways following TBI and suggests that targeting early bioenergetic deficits, particularly through enhancement of FAO, may represent a promising therapeutic strategy. Future studies aimed at defining the temporal reprogramming of metabolic pathways, including purine, fatty acids, and lipids, and energy metabolism, as well as their cell-type-specific interactions within the hippocampus, will be critical for identifying new opportunities to modulate recovery and improve outcomes after TBI.
Supplementary Information
Below is the link to the electronic supplementary material.ESM 1(DOCX 1.55 MB)ESM 2(CSV 24.4 KB)