Regulation of oligodendrocyte metabolism and myelination by acetyl-L-carnitine in a mouse model of post-weaning social isolation Authors: Hyun-Jeong Yang, Youngja Hwang Park, Dalnim Kim, Ahreum Lee, Minkuk Park, Eugene Koh, Gakyung Lee Published: 2026 ABSTRACT Adolescent social stress can impair myelination and increase vulnerability to psychiatric symptoms. We investigated whether acetyl-L-carnitine (ALC), a metabolite linking energy and lipid metabolism, regulates oligodendrocyte (OL) myelination and rescues behavioral deficits in a post-weaning social isolation (PWSI) mouse model. In vitro, ALC uptake and conversion through the OCTN2/CrAT axis promoted myelin sheath expansion without affecting OL precursor proliferation or early differentiation. ALC activated ERK signaling, increased histone acetylation, promoted PPARγ nuclear translocation, and selectively enhanced mitochondrial respiration in mature OLs. In vivo, oral ALC supplementation restored social preference and medial prefrontal cortex MBP expression in PWSI mice, whereas cuprizone co-administration abolished these effects, suggesting that the therapeutic effects of ALC are closely associated with its impact on myelination. Lipidomic analysis of the corpus callosum showed that ALC restored PWSI-induced changes in fatty acid chain length and unsaturation. These findings identify ALC as a metabolic regulator that restores social isolation-induced myelin deficits by coordinating myelin protein synthesis, mitochondrial function, and myelin lipid remodeling. ALC may therefore provide a metabolic strategy for targeting myelin-related neurodevelopmental and psychiatric disorders. Introduction Adolescence is a critical developmental period marked by the rapid maturation of cognitive, emotional, and social functions [1]. It represents a window of heightened vulnerability during which mental disorders such as depression and schizophrenia often emerge [2]. Notably, the adolescent brain is highly sensitive to stress; exposure during this stage may contribute to enduring alterations in brain structure and function, with potential implications for later psychopathology [3]. The post-weaning social isolation (PWSI) model is an animal model used to investigate the effects of adolescent social stress. In this paradigm, animals are reared in isolation after weaning, thereby depriving them of normal social interactions, including play behavior [4]. PWSI disrupts various neurodevelopmental processes and has been utilized as an experimental model for psychiatric disorders such as depression and schizophrenia [4]. Studies have shown that social isolation leads to myelin deficits in the medial prefrontal cortex (mPFC), a key hub for social behavior [5, 6]. Indeed, when young mice are reared under PWSI conditions from postnatal day (P) 21 to P35, they exhibit reduced sociability and decreased myelin thickness in the mPFC, compared to normally group-housed controls, and these alterations persist even after subsequent re-exposure to a social environment [7]. While the mPFC is essential for social behavior [4, 6, 7], social functioning also critically depends on the biochemical integrity of the corpus callosum, the brain’s primary white matter tract. Emerging evidence suggests that alterations in lipid composition within the corpus callosum, which serves as the fundamental bridge for interhemispheric communication, are robustly linked to impaired social behavior [8–10]. Therefore, examining both the focal changes in the mPFC and the lipidomic landscape of the corpus callosum is essential for a comprehensive understanding of how social stress disrupts brain connectivity and myelin molecular architecture. During differentiation and myelin formation, oligodendrocytes (OLs) undergo intensive lipid synthesis and high energy expenditure. This high metabolic demand renders them particularly susceptible to mitochondrial dysfunction, metabolic imbalance, and oxidative stress [11]. Social isolation-induced overactivation of the hypothalamus-pituitary-adrenal (HPA) axis further exacerbates neuroinflammation and oxidative stress [12, 13], which may interfere with the physiological maturation and myelination of OLs. Despite these insights, research on social isolation has predominantly focused on neuronal alterations, leaving the mechanisms underlying impaired myelination and potential strategies for its restoration under-explored. Acetyl-L-carnitine (ALC) is a metabolite that crosses the blood-brain barrier (BBB) and is taken up into cells via the organic cation transporter OCTN2 (encoded by Slc22a5) [14]. Once inside the cell, ALC is converted into acetyl-CoA by carnitine acetyltransferase (CrAT), thereby linking it to mitochondrial energy metabolism, epigenetic regulation, and lipidogenesis [15]. Although the ALC transporter Slc22a5 is highly expressed in newly formed OLs [16], the functional role of ALC in these cells remains poorly understood. Based on existing evidence for the involvement of ALC in myelination [17, 18] and its multifaceted roles in enhancing energy metabolism, lipid synthesis, and epigenetic regulation—all of which are critical for OL maturation—we hypothesized that ALC directly modulates myelination by acting on OLs. Furthermore, we postulated that ALC supplementation would rescue the myelination deficits and sociability impairments induced by PWSI. The present study aimed to examine the direct impact of ALC on OL maturation and myelination in vitro. Furthermore, we investigated whether ALC supplementation could restore impaired myelination and social behavior in the PWSI model during the critical myelinogenic period (P21–P35). By analyzing alterations in myelin lipid composition, we sought to uncover the specific mechanisms through which ALC facilitates myelin protection and recovery. Materials and methods Animals All animal experiments were conducted in accordance with applicable laws and institutional guidelines and were approved by the Institutional Animal Care and Use Committee of the University of Brain Education. Mice were maintained in a specific pathogen-free facility under a 12-h light/dark cycle (lights on at 8:00 AM) with ad libitum access to chow and water. For in vivo experiments, C57BL/6N mice (Mus musculus) were purchased from Orient Bio Inc. (Seongnam, Korea), and C57BL6/J mice were obtained from an in-house colony. The comparison between regular-environment and social-isolation conditions was first performed using C57BL/6J mice and then replicated using C57BL/6N mice. All subsequent in vivo experiments were performed using C57BL/6N mice. For primary oligodendrocyte precursor cell (OPC) cultures, P0-P1 ICR mice (Mus musculus) were obtained from Orient Bio Inc. For the social isolation condition, mice were housed individually from P21 to P35. Mice in the regular environment condition were group-housed, with three to four mice per cage, during the same period. For drug administration, mice received a daily oral gavage of distilled water (DW), ALC hydrochloride (1.251 mmol/kg; 300 mg/kg; A6706, Sigma-Aldrich), or an equimolar dose of L-carnitine hydrochloride (1.251 mmol/kg; 247.4 mg/kg; C0283, Sigma-Aldrich) during the rearing period (P21-P35). To induce demyelination, mice were fed a diet containing 0.15% cuprizone (DooYeol Biotech, Seoul, Korea) [19], which was replaced every two days. Following behavioral testing, mice were euthanized for tissue collection. Both sexes were included in all experiments, except for gut metabolomics, which was performed exclusively in males. Preliminary analyses indicated no apparent sex-specific differences in the main behavioral and molecular outcomes (Supplementary Fig. S1); therefore, data from both sexes were pooled to increase statistical power. Sample sizes for individual experiments are reported in the corresponding figure legends, with detailed sex-specific sample sizes provided in Supplementary Table S1. Sample sizes were based on previous studies using similar experimental paradigms and were selected to detect biologically meaningful effects while minimizing animal use. Sociability test Mice were habituated for 5 min in an open-field arena (40 × 40 × 40 cm) containing two empty wire-mesh containers. For the sociability test, one container was replaced with a novel social stimulus, an age- and sex-matched mouse, while the other container remained empty. The test mouse was allowed to explore the arena for 5 min. Movements were recorded and analyzed using ANY-maze software (Stoelting Co., Wood Dale, IL, USA). The interaction zone was defined as the area within 2.5 cm of each container. The duration spent in the mouse zone, distance traveled, and number of head entries into the mouse zone were quantified. Antibodies and reagents Detailed information is provided in Supplementary Table S2. Quantitative real time PCR Total RNA was extracted from brain tissues or cultured cells using TRI reagent (Sigma-Aldrich, St. Louis, MO, USA) and reverse-transcribed into cDNA using SuperScript IV VILO Master Mix (Thermo Fisher Scientific, Waltham, MA, USA). Quantitative real-time PCR was performed using PowerUp SYBR Green Master Mix (Applied Biosystems, Austin, TX, USA). Primer sequences are provided in Supplementary Table S2. Relative mRNA expression levels were normalized to Actb or Ppia. Western blot Tissues or cells were lysed on ice using cold RIPA buffer (WSE-7420, ATTO, Tokyo, Japan) and centrifuged at 15,000 rpm for 15 min at 4 °C. Protein concentrations in the supernatants were quantified, and samples were mixed with sample buffer and boiled. Proteins were separated by SDS-PAGE and transferred to PVDF membranes. Membranes were blocked with EZBlock Chemi (AE-1475, ATTO) for 30 min and incubated overnight at 4 °C with primary antibodies. After washing, membranes were incubated with secondary antibodies for 1 h at room temperature. Protein signals were detected using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific) and captured with an Amersham Imager 600 (GE Healthcare, Chicago, IL, USA). Band intensities were quantified using ImageJ software (NIH, Bethesda, MD, USA). Original western blot images are provided in Supplementary Fig. S2. Primary OPC culture and cell lines Mixed glial cultures were prepared from the cortices of P0–P1 ICR mice and plated on poly-D-lysine (PDL)-coated flasks. Cultures were maintained in DMEM/F-12 (Gibco, Grand Island, NY, USA) supplemented with 10% FBS, 5% horse serum, and 1× penicillin-streptomycin (P/S). On DIV10, OPCs were isolated by shaking and purified using the differential adhesion method to remove astrocytes. Purified OPCs were seeded onto PDL-coated coverslips or nanofiber chamber slides (Merck, Darmstadt, Germany) and maintained in Sato medium consisting of DMEM supplemented with 1× B-27, 1× GlutaMAX, 1× P/S, 1% horse serum, 1× sodium pyruvate, 0.34 μg/ml T3 and 0.4 μg/ml T4) [20]. Medium was changed in every other day. The Oli-neu mouse oligodendroglial precursor cell line was kindly provided by Prof. Elior Peles at the Weizmann Institute of Science. Cells were maintained in DMEM supplemented with 10% FBS and 1× P/S and tested negative for mycoplasma contamination. The cell line was not recently authenticated by STR profiling because STR profiling is not routinely applied to mouse cell lines. For differentiation, the medium was replaced with Sato medium. For advanced differentiation, PD174265 (Sigma-Aldrich) was added at a final concentration of 1 μM. siRNA transfection and subcellular fractionation Cells were transfected with 10 nM siRNA (si-NC, si-Slc22a5, or si-Crat; Bioneer, Daejeon, Korea) using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA). For nuclear and cytoplasmic fractionation, transfected Oli-neu cells were treated with 100 μM ALC or vehicle for 24 h, and fractions were isolated using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific). Seahorse extracellular flux analysis The oxygen consumption rate (OCR) of Oli-neu cells transfected with 10 nM si-Slc22a5 or si-NC, with or without ALC treatment, was measured using the Seahorse XF HS Mini Analyzer and the XFp Cell Mito Stress Test Kit (Agilent, 103010-100). Transfected cells were seeded at a density of 5000–7000 cells per well, which was optimized through prior cell titration experiments. Cells were cultured either in standard growth medium (DMEM supplemented with 10% FBS) for undifferentiated Oli-neu cells or in differentiation medium (Sato medium supplemented with 1 μM PD174265) for differentiated OL-like cells. To examine the effect of ALC in differentiated OL-like cells, cells were incubated with 2 mM ALC in DMEM for 24 h (Fig. 3D-F). Twenty-four hours after seeding, cells were washed with pre-warmed Seahorse assay medium and incubated in a CO₂-free incubator at 37 °C for 1 h. The stress test was performed by measuring OCR at basal levels and after sequential injections of oligomycin, FCCP, and rotenone/antimycin A at final concentrations of 1.5, 0.5, and 0.5 μM, respectively, in all Seahorse experiments except those shown in Fig. 3G-I. In Fig. 3G-I, ALC was injected through Port A to a final concentration of 100 μM, and OCR was monitored for 2 h before subsequent injection of oligomycin, FCCP, and rotenone/antimycin A. Data were analyzed using the Seahorse XF Cell Mito Stress Test Report Generator (Agilent), which provided values for basal respiration, ATP production, maximal respiration, and other related parameters. OCR data were normalized to protein content per well after the assay. Elisa Mitochondrial biogenesis was assessed according to the manufacturer’s protocol (ab110217, Abcam). TNFα and IL6 levels were quantified using commercial kits (Mouse TNFα DuoSet ELISA, DY410, R&D Systems; Mouse IL6 DuoSet ELISA, DY406, R&D Systems). Immunohistochemistry and Immunocytochemistry For immunohistochemistry, mice were transcardially perfused with 4% PFA. Brains were cryoprotected in 30% sucrose and sectioned at 15 μm using a cryostat (Leica Biosystems, Nussloch, Germany). For immunocytochemistry, cultured cells were fixed with 4% PFA. Brain sections or cultured cells were blocked in 5% normal donkey serum containing 0.2% Triton X-100 and incubated overnight at 4 °C with primary antibodies, including rat anti-MBP (Millipore, Burlington, MA, USA) and rabbit anti-Olig2 (Abcam, Cambridge, UK). After washing in PBS, samples were incubated with fluorophore-conjugated secondary antibodies for 1 h at room temperature. Following a final wash, samples were mounted for visualization. Fluorescence was visualized using an ECLIPSE TS2 microscope (Nikon, Tokyo, Japan). Image analysis was performed using ImageJ software (NIH, Bethesda, MD, USA). Statistical analysis, reproducibility, and exclusion criteria The number of biological replicates or independent experiments is indicated in each figure legend. Exclusion criteria were pre-established before analysis. Animals or samples were excluded only in cases of technical failure, severe health issues unrelated to the experimental design, unsuccessful sample collection, or failure to meet assay-specific quality-control criteria. All animals that completed the experimental procedures and all samples passing quality control were included in the final analyses. Representative images and blots were selected from biological replicates or independent experiments that yielded similar results. Statistical analyses were performed using SigmaPlot 14.0 software. Data distribution and variance were assessed before selecting statistical tests. For two-group comparisons, Student’s t-test was used for normally distributed data, whereas the Wilcoxon rank-sum test was used for non-normally distributed data. For comparisons among more than two groups, one-way ANOVA followed by Holm–Sidak post hoc tests was used for normally distributed data, whereas Kruskal–Wallis one-way ANOVA on ranks followed by Tukey’s post hoc tests was used for non-parametric data. A value of p < 0.05 was considered statistically significant. Statistical significance and corresponding p values are indicated in the corresponding figures, figure legends, and Supplementary Information. A detailed description of the statistical analysis of omics data is available in the Supplementary Materials and Methods. Group variation is presented as mean ± SD or mean ± SEM, as indicated in the figure legends. Metabolomics and lipidomics profiling Detailed procedures for sample preparation, liquid chromatography-mass spectrometry (LC-MS) parameters, and data processing are provided in the Supplementary Materials and Methods. Randomization and blinding At weaning, mice were randomly assigned to experimental groups while balancing sex across groups. Investigators could not be fully blinded during animal housing and treatment administration because the rearing conditions and dietary/treatment procedures were apparent during the experiment. However, downstream quantitative analyses were performed using coded sample identifiers whenever applicable. Behavioral recordings and immunostaining images were analyzed using coded sample identifiers by investigators blinded to group allocation. Other instrument-based assays were analyzed from numerical outputs generated using standardized acquisition and preprocessing procedures. Results Post-weaning social isolation impairs sociability, reduces myelin gene expression in the medial prefrontal cortex, and depletes gut acetyl-L-carnitine levels To establish the PWSI model [7], mice were reared under regular environment or social isolation conditions from P21 to P35 (Fig. 1A). Social isolation impaired social motivation without affecting locomotor activity (Fig. 1B-E and Supplementary Fig. S1A-B) and reduced mPFC expression of myelination-related genes (Myrf, Mag, Mbp) [21] and myelin basic protein (MBP) (Fig. 1F-G, and Supplementary Fig. S1C). These behavioral and molecular deficits were consistent across both sexes (Supplementary Fig. S1D-M). To explore the systemic impact of SI on the gut-brain axis, we performed non-targeted metabolomic profiling of the cecum. Of the 8968 detected metabolites, 405 were significantly altered by social isolation (Supplementary Fig. S1N), with distinct metabolic signatures (Fig. 1H and Supplementary Fig. S1O). Pathway analysis identified disruptions in multiple biological pathways (Fig. 1I). Among the differentially abundant metabolites (Fig. 1J, Supplementary Fig. S1P-R, and Supplementary Table S3), ALC, a BBB-permeable metabolite implicated in myelination [14, 17, 18], was significantly reduced in mice subjected to social isolation. Given these findings, we focused our subsequent investigation on the functional effects of ALC on OL-lineage cells.Fig. 1Post-weaning social isolation impairs sociability, reduces myelin gene expression in the medial prefrontal cortex, and alters cecal metabolomic profiles.A Experimental scheme. After weaning at P21, mice were reared under regular-environment conditions (RE; 3–4 mice per cage) or post-weaning social isolation conditions (SI; 1 mouse per cage) until P35. Sociability test was then conducted within 3 days, followed by euthanasia and tissue collection. B–E Sociability test. B, C The two orange circles indicate the positions of wire-mesh cylindrical cages, one containing an age- and sex-matched stranger mouse (left) and the other empty (right). Heatmaps B and track plots C show movement patterns, with color gradients from blue to red representing increasing time spent in each area. D, E Quantitative analysis of sociability behavior. D Total distance traveled in the arena (m). E Number of head entries into the interaction zone. N = 8 and 7 mice for the RE and SI group, respectively D-E. F Relative mRNA levels of myelin-related genes in the medial prefrontal cortex (mPFC), normalized to Actb. N = 8 mice per group. Data are presented as mean ± SD D-F. G Myelin basic protein (MBP) expression levels in the mPFC. H–J Cecal metabolomics analysis. H Two-way hierarchical cluster analysis. Red and green bars indicate the RE and SI groups, respectively. Metabolite intensities are represented on a blue-to-red scale, indicating low to high abundance. I Pathway and functional analyses using the Mummichog algorithm showing significantly altered pathways under SI conditions. J Box-and-whisker plots of significantly altered predicted metabolites under SI conditions. Both sexes were used in comparable numbers for the sociability and mPFC analyses B–G, and no significant sex differences were observed (Supplementary Fig. S1). Cecal metabolomic analysis H–J was performed in male mice only. N = 6 mice per group (H-J). *p < 0.05, **p < 0.01; Student’s t-test. Post-weaning social isolation impairs sociability, reduces myelin gene expression in the medial prefrontal cortex, and alters cecal metabolomic profiles. A Experimental scheme. After weaning at P21, mice were reared under regular-environment conditions (RE; 3–4 mice per cage) or post-weaning social isolation conditions (SI; 1 mouse per cage) until P35. Sociability test was then conducted within 3 days, followed by euthanasia and tissue collection. B–E Sociability test. B, C The two orange circles indicate the positions of wire-mesh cylindrical cages, one containing an age- and sex-matched stranger mouse (left) and the other empty (right). Heatmaps B and track plots C show movement patterns, with color gradients from blue to red representing increasing time spent in each area. D, E Quantitative analysis of sociability behavior. D Total distance traveled in the arena (m). E Number of head entries into the interaction zone. N = 8 and 7 mice for the RE and SI group, respectively D-E. F Relative mRNA levels of myelin-related genes in the medial prefrontal cortex (mPFC), normalized to Actb. N = 8 mice per group. Data are presented as mean ± SD D-F. G Myelin basic protein (MBP) expression levels in the mPFC. H–J Cecal metabolomics analysis. H Two-way hierarchical cluster analysis. Red and green bars indicate the RE and SI groups, respectively. Metabolite intensities are represented on a blue-to-red scale, indicating low to high abundance. I Pathway and functional analyses using the Mummichog algorithm showing significantly altered pathways under SI conditions. J Box-and-whisker plots of significantly altered predicted metabolites under SI conditions. Both sexes were used in comparable numbers for the sociability and mPFC analyses B–G, and no significant sex differences were observed (Supplementary Fig. S1). Cecal metabolomic analysis H–J was performed in male mice only. N = 6 mice per group (H-J). *p < 0.05, **p < 0.01; Student’s t-test. Acetyl-L-carnitine enhances myelin sheath expansion ALC, an acetylated derivative of L-carnitine, is transported by OCTN2 (encoded by Slc22a5) and interconverted with L-carnitine via CrAT. Since Slc22a5 is highly expressed in newly formed OLs [16], we first confirmed that treatment with ALC or L-carnitine elevates the corresponding intracellular metabolite levels in Oli-neu cells (Fig. 2A, Supplementary Fig. S3, and Supplementary Tables S4–5). Notably, siRNA-mediated knockdown of Slc22a5 or Crat abolished the ALC- or L-carnitine-induced increase in MBP expression (Fig. 2B, and Supplementary Fig. S4A-D), demonstrating that MBP induction requires functional OCTN2 and CrAT.Fig. 2Acetyl-L-carnitine promotes MBP expression and oligodendrocyte membrane sheath expansion through OCTN2/CrAT-dependent and ERK-associated pathways.A LC-MS-based quantitative analysis for acetyl-L-carnitine (ALC) and L-carnitine (LC) levels in Oli-neu cells treated with vehicle, ALC, or LC for 24 h. B Western blot analysis of MBP expression in Oli-neu cells transfected with si-negative control (si-NC), si-Slc22a5, or si-Crat, followed by 24 h treatment with vehicle, ALC, or LC under advanced differentiation conditions. C-H Immunofluorescence analysis of primary OPC cultures subjected to transfection (TF) with si-NC, si-Slc22a5, or si-Crat at DIV1 and treated with vehicle, ALC, or LC in Sato medium starting at DIV2. Cells were cultured on PDL-coated coverslips C-F or 3D nanofiber scaffolds G-H, fixed at DIV7, and stained for Olig2 (red), MBP (green), and DAPI (blue). OPC proliferation (D; Olig2+ cells/DAPI+ cells), maturation (E; MBP+ cells/Olig2+ cells), membrane sheath formation on 2D coverslips (F; MBP+ membrane area per mature OL), and myelin-like area on 3D nanofibers (H; MBP+ myelin-like area per mature OL) were quantified. For D and E, N = 9, 11, 8, 8, 6, 7, 11, 7, and 7 fields of view per group, respectively; for F, N = 52, 53, 46, 66, 34, 31, 62, 47, and 40 cells per group, respectively. These data were obtained from three independent cultures across nine groups: si-NC/vehicle, si-NC/ALC, si-NC/LC, si-Slc22a5/vehicle, si-Slc22a5/ALC, si-Slc22a5/LC, si-Crat/vehicle, si-Crat/ALC, and si-Crat/LC. For H, N = 61, 29, 52, 92, 28, 52, 53, and 22 cells per group, respectively, obtained from three independent cultures across eight groups: si-NC/vehicle, si-NC/ALC, si-NC/LC, si-Slc22a5/vehicle, si-Slc22a5/ALC, si-Slc22a5/LC, si-Crat/vehicle, and si-Crat/ALC. I, J MBP protein expression in primary OPC cultures treated with vehicle or ALC from DIV1 to DIV4. I Representative western blot image of MBP. J Quantification of I. K Relative mRNA levels of genes related to OL differentiation, lipid metabolism, mitochondrial function, and carnitine transport in Oli-neu cells treated with the indicated concentrations of ALC for 24 h. L, M ALC-associated signaling in Oli-neu cells. L Western blot images of Oli-neu cells treated with ALC for 24 h with or without 2 h pretreatment with U0126 under advanced differentiation conditions. M Quantification of L. N, O Effect of Slc22a5 knockdown on ALC-associated signaling in Oli-neu cells. N Western blot images of cytosolic (“Cyt”) and nuclear (“Nuc”) fractions from si-NC or si-Slc22a5-transfected Oli-neu cells treated with vehicle or ALC. O Quantification of N. Unless otherwise indicated, ALC and LC were used at 100 μM, and U0126 at 10 μM. N = 6 independent cultures per group A and 3 independent cultures per group B, I-O. *p < 0.05, **p < 0.01, and ***p < 0.001. One-way ANOVA followed by Holm-Sidak post hoc test or Kruskal-Wallis test followed by Dunn’s post hoc test was performed within each siRNA group B-H or across ALC concentrations K, according to data distribution. Student’s t-test was used for indicated pairwise comparisons A, H, J, O. ###p < 0.001 indicates comparisons between si-NC and si-Slc22a5 or si-Crat by Student’s t-test F, H. Bars indicate mean ± SD. P Schematic model of ALC-mediated multi-modal regulation. Proposed framework illustrating the tri-modal action of ALC in driving oligodendrocyte maturation via myelin protein synthesis, lipid remodeling, and mitochondrial enhancement. For a comprehensive conceptual overview and detailed descriptions, see Supplementary Fig. S7. Acetyl-L-carnitine promotes MBP expression and oligodendrocyte membrane sheath expansion through OCTN2/CrAT-dependent and ERK-associated pathways. A LC-MS-based quantitative analysis for acetyl-L-carnitine (ALC) and L-carnitine (LC) levels in Oli-neu cells treated with vehicle, ALC, or LC for 24 h. B Western blot analysis of MBP expression in Oli-neu cells transfected with si-negative control (si-NC), si-Slc22a5, or si-Crat, followed by 24 h treatment with vehicle, ALC, or LC under advanced differentiation conditions. C-H Immunofluorescence analysis of primary OPC cultures subjected to transfection (TF) with si-NC, si-Slc22a5, or si-Crat at DIV1 and treated with vehicle, ALC, or LC in Sato medium starting at DIV2. Cells were cultured on PDL-coated coverslips C-F or 3D nanofiber scaffolds G-H, fixed at DIV7, and stained for Olig2 (red), MBP (green), and DAPI (blue). OPC proliferation (D; Olig2+ cells/DAPI+ cells), maturation (E; MBP+ cells/Olig2+ cells), membrane sheath formation on 2D coverslips (F; MBP+ membrane area per mature OL), and myelin-like area on 3D nanofibers (H; MBP+ myelin-like area per mature OL) were quantified. For D and E, N = 9, 11, 8, 8, 6, 7, 11, 7, and 7 fields of view per group, respectively; for F, N = 52, 53, 46, 66, 34, 31, 62, 47, and 40 cells per group, respectively. These data were obtained from three independent cultures across nine groups: si-NC/vehicle, si-NC/ALC, si-NC/LC, si-Slc22a5/vehicle, si-Slc22a5/ALC, si-Slc22a5/LC, si-Crat/vehicle, si-Crat/ALC, and si-Crat/LC. For H, N = 61, 29, 52, 92, 28, 52, 53, and 22 cells per group, respectively, obtained from three independent cultures across eight groups: si-NC/vehicle, si-NC/ALC, si-NC/LC, si-Slc22a5/vehicle, si-Slc22a5/ALC, si-Slc22a5/LC, si-Crat/vehicle, and si-Crat/ALC. I, J MBP protein expression in primary OPC cultures treated with vehicle or ALC from DIV1 to DIV4. I Representative western blot image of MBP. J Quantification of I. K Relative mRNA levels of genes related to OL differentiation, lipid metabolism, mitochondrial function, and carnitine transport in Oli-neu cells treated with the indicated concentrations of ALC for 24 h. L, M ALC-associated signaling in Oli-neu cells. L Western blot images of Oli-neu cells treated with ALC for 24 h with or without 2 h pretreatment with U0126 under advanced differentiation conditions. M Quantification of L. N, O Effect of Slc22a5 knockdown on ALC-associated signaling in Oli-neu cells. N Western blot images of cytosolic (“Cyt”) and nuclear (“Nuc”) fractions from si-NC or si-Slc22a5-transfected Oli-neu cells treated with vehicle or ALC. O Quantification of N. Unless otherwise indicated, ALC and LC were used at 100 μM, and U0126 at 10 μM. N = 6 independent cultures per group A and 3 independent cultures per group B, I-O. *p < 0.05, **p < 0.01, and ***p < 0.001. One-way ANOVA followed by Holm-Sidak post hoc test or Kruskal-Wallis test followed by Dunn’s post hoc test was performed within each siRNA group B-H or across ALC concentrations K, according to data distribution. Student’s t-test was used for indicated pairwise comparisons A, H, J, O. ###p < 0.001 indicates comparisons between si-NC and si-Slc22a5 or si-Crat by Student’s t-test F, H. Bars indicate mean ± SD. P Schematic model of ALC-mediated multi-modal regulation. Proposed framework illustrating the tri-modal action of ALC in driving oligodendrocyte maturation via myelin protein synthesis, lipid remodeling, and mitochondrial enhancement. For a comprehensive conceptual overview and detailed descriptions, see Supplementary Fig. S7. Next, we investigated the roles of ALC or L-carnitine in OPC proliferation, maturation, and membrane sheath formation using primary OPC cultures on coverslips (Fig. 2C-F) or 3D nanofibers (Fig. 2G-H). ALC did not alter OPC proliferation (Fig. 2D) or the proportion of differentiated OLs (Fig. 2E), but significantly increased the MBP-positive membrane area per OL (Fig. 2F). Slc22a5 knockdown reduced membrane sheath size and blunted the effect of ALC. In contrast to ALC, L-carnitine did not significantly increase sheath size (Fig. 2F), suggesting that the acetylated form of L-carnitine is required for this effect. Because 2D cultures do not fully recapitulate the geometric cues required for membrane wrapping, we further assessed myelin-like sheath formation using a 3D nanofiber assay. This system allowed us to validate whether the increased MBP+ area observed in 2D translates into an enhanced capacity for structural engagement with axon-mimetic cylindrical geometries. Similarly, on 3D nanofibers, ALC, but not L-carnitine, increased myelin-like area per OL (Fig. 2H). Notably, ALC responsiveness was partially retained on nanofibers even under knockdown conditions, raising the possibility that nanofiber-derived structural cues may partially compensate for reduced OCTN2 or CrAT function (Fig. 2H). Moreover, ALC increased MBP expression in primary cultures (Fig. 2I-J and Supplementary Fig. S4E). Expansion of the myelin sheath is driven by several synergistic factors, including the suppression of OL differentiation inhibitors, the activation of myelin-related transcription factors [22], enhanced mitochondrial efficiency [23], and active lipogenesis [24]. To investigate whether ALC modulates gene expression related to these pro-myelination pathways, we treated Oli-neu cells with ALC. ALC downregulated OL differentiation-inhibitory genes (Id2, Id4) [21] and upregulated genes associated with lipid metabolism (Srebf2, Pparg) [25] and mitochondrial function (Pparg, Ppargc1a) [23] (Fig. 2K). Of the pathways examined regarding OL differentiation and myelination [26], ALC was found to activate ERK signaling (Supplementary Fig. S4F). MEK1/2 inhibition with U0126 attenuated ALC-induced ERK phosphorylation and MBP expression (Fig. 2L-M), indicating that ALC promotes myelination via ERK-dependent mechanisms. Additionally, ALC increased PPARγ protein expression, independently of ERK signaling (Fig. 2L-M and Supplementary Fig. S4G-H). Given the reported link between ALC and histone acetylation [15, 27], we examined nuclear H3K27 acetylation (H3K27Ac). Consistently, ALC increased H3K27Ac levels, while this effect was abolished upon the knockdown of Slc22a5 (Fig. 2N-O). Given that histone acetylation-induced chromatin remodeling facilitates transcriptional accessibility [28], we next investigated the nuclear translocation of the aforementioned transcriptional modulators. ALC promoted the nuclear translocation of PPARγ, while having no effect on PGC1α or SIRT1; notably, this location was abrogated in Slc22a5 siRNA-treated cells (Fig. 2N, O, and Supplementary Fig. S4I-N). In summary, ALC acts directly on OL-lineage cells through OCTN2- and CrAT-dependent mechanisms. These effects involve ERK-dependent MBP upregulation, suppression of differentiation-inhibitory genes, activation of lipid- and mitochondrial metabolism-associated programs, PPARγ nuclear translocation, and increased H3K27 acetylation (Fig. 2P). Acetyl-L-carnitine enhances mitochondrial function in mature oligodendrocytes ALC upregulated mitochondria-related factors (PGC1ɑ, PPARγ) and promoted PPARγ nuclear translocation via Slc22a5 (Fig. 2K-O), suggesting its involvement in the mitochondrial function of OLs. To further investigate this, we assessed mitochondrial parameters, using the Seahorse XF system following ALC treatment or siRNA-mediated Slc22a5 knockdown. The effect of ALC on mitochondrial function depended on the stage of OL differentiation. In OPCs, ALC treatment did not alter any OCR-related parameters (Fig. 3A-C). In contrast, it markedly increased key respiratory measures, including basal respiration, proton leak, non-mitochondrial oxygen consumption, and ATP production, in differentiated OLs (Fig. 3D-F). Unlike the stage-specific effect of ALC, Slc22a5 knockdown consistently reduced mitochondrial function in both OPCs and differentiated OLs (Fig. 3G-L). Furthermore, in OPCs, L-carnitine treatment even led to a decrease in these parameters (Supplementary Fig. S5A-C), while Crat knockdown modestly reduced mitochondrial function (Supplementary Fig. S5D-F). In summary, ALC enhances mitochondrial function specifically in mature OLs, whereas its major transporter, OCTN2, is essential for maintaining mitochondrial function in both OPCs and mature OLs. Additionally, CrAT may contribute to mitochondrial function during the OPC stage.Fig. 3Acetyl-L-carnitine enhances stage-dependent mitochondrial respiration, whereas slc22a5 knockdown impairs mitochondrial function in Oli-neu cells.Seahorse Cell Mito Stress Test was performed in undifferentiated (‘OPC’, A-C, G-I) or differentiated (‘Differentiated OL’, D-F, J-L) Oli-neu cells transfected with siRNAs and treated with vehicle or ALC. Oxygen consumption rate (OCR) A, D, G, J and extracellular acidification rate (ECAR) B, E, H, K were monitored using the Seahorse XF system. OCR-derived mitochondrial parameters, including basal respiration, proton leak, maximal respiration, non-mitochondrial oxygen consumption, ATP production, spare respiratory capacity (%), and coupling efficiency, were calculated C, F, I, L. (A-C) Effect of ALC on mitochondrial function in undifferentiated Oli-neu cells. Cells transfected with si-negative control (si-NC) were treated with vehicle or ALC for 24 h before measurement. D–F Effect of ALC on mitochondrial function in differentiated Oli-neu cells. Cells transfected with si-NC were treated with vehicle or ALC for 24 h before measurement. G–I Effect of Slc22a5 knockdown on mitochondrial function in undifferentiated Oli-neu cells. Cells transfected with si-NC or si-Slc22a5 were monitored for real-time responses to ALC. J–L Effect of Slc22a5 knockdown on mitochondrial function in differentiated Oli-neu cells. Cells were transfected with si-NC or si-Slc22a5 and analyzed 2 days after transfection. N = 3 independent cultures per group. #p < 0.1; *p < 0.05; **p < 0.01, Student’s t-test. Bars indicate mean ± SD. Acetyl-L-carnitine enhances stage-dependent mitochondrial respiration, whereas slc22a5 knockdown impairs mitochondrial function in Oli-neu cells. Seahorse Cell Mito Stress Test was performed in undifferentiated (‘OPC’, A-C, G-I) or differentiated (‘Differentiated OL’, D-F, J-L) Oli-neu cells transfected with siRNAs and treated with vehicle or ALC. Oxygen consumption rate (OCR) A, D, G, J and extracellular acidification rate (ECAR) B, E, H, K were monitored using the Seahorse XF system. OCR-derived mitochondrial parameters, including basal respiration, proton leak, maximal respiration, non-mitochondrial oxygen consumption, ATP production, spare respiratory capacity (%), and coupling efficiency, were calculated C, F, I, L. (A-C) Effect of ALC on mitochondrial function in undifferentiated Oli-neu cells. Cells transfected with si-negative control (si-NC) were treated with vehicle or ALC for 24 h before measurement. D–F Effect of ALC on mitochondrial function in differentiated Oli-neu cells. Cells transfected with si-NC were treated with vehicle or ALC for 24 h before measurement. G–I Effect of Slc22a5 knockdown on mitochondrial function in undifferentiated Oli-neu cells. Cells transfected with si-NC or si-Slc22a5 were monitored for real-time responses to ALC. J–L Effect of Slc22a5 knockdown on mitochondrial function in differentiated Oli-neu cells. Cells were transfected with si-NC or si-Slc22a5 and analyzed 2 days after transfection. N = 3 independent cultures per group. #p < 0.1; *p < 0.05; **p < 0.01, Student’s t-test. Bars indicate mean ± SD. Acetyl-L-carnitine administration during post-weaning social isolation restores sociability and myelin basic protein expression in the medial prefrontal cortex Based on the pro-myelinating effects of ALC observed in vitro (Figs. 2–3), we investigated whether it could alleviate PWSI-induced social deficits and myelin loss in vivo. ALC treatment increased serum levels of both ALC and L-carnitine; however, only ALC was significantly elevated in the mPFC (Fig. 4A, Supplementary Fig. S6A-B, and Supplementary Tables S6–9). These results suggest that ALC accumulates more efficiently than L-carnitine in the brain under PWSI conditions.Fig. 4Acetyl-L-carnitine improves sociability and restores mPFC myelin expression in socially isolated mice.Mice were reared under social isolation (SI; 1 mouse per cage) from P21 to P35 and received daily oral gavage of distilled water (DW), acetyl-L-carnitine (ALC), or L-carnitine (LC) at equimolar doses (1.251 mmol/kg). The ALC + CPZ group was fed chow containing 0.15% cuprizone (CPZ) throughout the experimental period to disrupt myelination. After 2 weeks of SI, sociability testing was performed. Mice were then euthanized, and tissues were collected for biochemical and immunohistochemical analyses. A Liquid chromatography-mass spectrometry-based quantification of ALC and LC concentrations in serum (μM) and the mPFC (nmol/mg protein) after 2 weeks of treatment. N = 5 and 4 mice for the DW and ALC groups, respectively. B–D Sociability test results. B Representative behavioral heatmaps (upper) and track plots (lower). Orange circles indicate the wire-mesh cages, one containing a stranger mouse (upper right) or the other left empty (lower left). C Total distance traveled (m). D Number of entries into the mouse zone. N = 6 mice per group B-D. E Representative immunohistochemistry images of MBP (green) and DAPI (blue) in the mPFC. F,G Western blot analysis F and quantification G of MBP protein levels in the mPFC of SI mice treated with DW, ALC, LC, or ALC + CPZ. N = 3, 3, 3, and 5 mice for DW, ALC, LC, and ALC + CPZ groups, respectively F, G. Data are presented as mean ± SD. *p < 0.05; Student’s t-test for indicated comparisons. Both sexes were used in approximately equal numbers for analyses. Acetyl-L-carnitine improves sociability and restores mPFC myelin expression in socially isolated mice. Mice were reared under social isolation (SI; 1 mouse per cage) from P21 to P35 and received daily oral gavage of distilled water (DW), acetyl-L-carnitine (ALC), or L-carnitine (LC) at equimolar doses (1.251 mmol/kg). The ALC + CPZ group was fed chow containing 0.15% cuprizone (CPZ) throughout the experimental period to disrupt myelination. After 2 weeks of SI, sociability testing was performed. Mice were then euthanized, and tissues were collected for biochemical and immunohistochemical analyses. A Liquid chromatography-mass spectrometry-based quantification of ALC and LC concentrations in serum (μM) and the mPFC (nmol/mg protein) after 2 weeks of treatment. N = 5 and 4 mice for the DW and ALC groups, respectively. B–D Sociability test results. B Representative behavioral heatmaps (upper) and track plots (lower). Orange circles indicate the wire-mesh cages, one containing a stranger mouse (upper right) or the other left empty (lower left). C Total distance traveled (m). D Number of entries into the mouse zone. N = 6 mice per group B-D. E Representative immunohistochemistry images of MBP (green) and DAPI (blue) in the mPFC. F,G Western blot analysis F and quantification G of MBP protein levels in the mPFC of SI mice treated with DW, ALC, LC, or ALC + CPZ. N = 3, 3, 3, and 5 mice for DW, ALC, LC, and ALC + CPZ groups, respectively F, G. Data are presented as mean ± SD. *p < 0.05; Student’s t-test for indicated comparisons. Both sexes were used in approximately equal numbers for analyses. Next, we examined whether ALC rescues altered behaviors and mPFC hypomyelination in the PWSI model. Mice were orally administered with DW, ALC, or L-carnitine, concomitant with either a normal diet or a cuprizone-containing diet during the PWSI period. Cuprizone treatment, known to cause demyelination in the corpus callosum and cortex [29] and impair sociability [30], was used to test whether intact myelination is required for the behavioral or molecular effects of ALC. While locomotor activity was unaffected across groups (Fig. 4C), ALC significantly improved sociability, whereas L-carnitine showed no effect (Fig. 4D and Supplementary Fig. S6C). Notably, cuprizone co-administration abolished the behavioral rescue induced by ALC, suggesting that this effect is linked to its pro-myelinating properties. Consistent with the link between social behavior and mPFC myelination [7] (Fig. 1F-G), ALC treatment increased MBP expression, which was reversed by cuprizone (Fig. 4E-G). These findings suggest that ALC-induced recovery of social behavior is mediated, at least in part, by its pro-myelinating effects. Notably, in mice reared in a regular environment, ALC administration increased serum ALC levels but failed to enhance brain ALC concentrations or influence behavioral phenotypes (Supplementary Fig. S6D-F). In summary, administered ALC accumulates in the brain under PWSI conditions, thereby promoting myelin synthesis and restoring social behavior. These findings highlight ALC’s crucial role in maintaining brain homeostasis in response to pathological challenges. Acetyl-L-carnitine restores neural activity and myelination while attenuating neuroinflammation across multiple brain regions following social isolation PWSI reduces neural activity and myelination [7, 31], while promoting HPA axis activation and neuroinflammation across the mPFC, hippocampus, and brainstem [12, 32–34]. We therefore investigated whether ALC supplementation ameliorates these pathological features across multiple brain regions. In the mPFC, ALC restored social isolation-induced changes in the expression of genes related to neural activity (Fos), neurotransmission (Chat, Th), myelination (Myrf, Mbp, Mag), metabolism (Pparg), HPA axis (Nr3c1), and inflammation (Tnf) (Fig. 5A). Consistent with in vitro data, ALC increased MBP and SREBF2 protein levels and ERK phosphorylation in the mPFC (Fig. 5D, F). In the hippocampus, ALC significantly upregulated PPARγ expression, whereas increases in MBP and SREBF2 levels did not reach statistical significance (Fig. 5E, G). Given that the hippocampus is particularly vulnerable to isolation-induced neuroinflammation [32], we assessed inflammatory markers in this region. Hippocampal Tnf mRNA expression and TNFα protein levels tended to be higher in socially isolated mice and lower with ALC treatment, although these differences did not reach statistical significance (Fig. 5H, I). In the brainstem, ALC reversed the SI-induced expression changes associated with neural activity (Fos, Trpa1), synaptic function (Syp), glutamate metabolism (Glul), HPA axis signaling (Nr3c1), and neuroinflammation (Il6) (Fig. 5B). ALC also significantly increased brainstem glutamate levels, accompanied by recovery of Glul expression (Fig. 5C). In summary, ALC supplementation effectively rescues social isolation-induced impairments in neural activity and myelination, while attenuating neuroinflammation across extensive brain regions.Fig. 5Acetyl-L-carnitine modulates the expression of genes related to myelination, neural activity, and inflammation in the brains of socially isolated mice.Mice were assigned to three groups: regular environment with distilled water treatment (RE-DW), social isolation with distilled water treatment (SI-DW), and social isolation with acetyl-L-carnitine treatment (SI-ALC). RE mice were group-housed (3–4 mice per cage), whereas SI mice were housed individually (1 mouse per cage) from P21 to P35. Mice received daily oral gavage of distilled water (DW) or ALC (1.251 mmol/kg). After the 2-week rearing period, mice were euthanized, and tissues were collected for molecular and biochemical analyses. A, B RT-qPCR analysis of genes related to neuronal activity and neurotransmission, myelination, and oxidative stress/inflammation in the mPFC A and brainstem B. C Liquid chromatography–mass spectrometry-based quantification of glutamate and glutamine levels in the brainstem. D–G Western blot images D, E and protein quantification (F, G) for myelin markers (MBP, MAG), lipid/metabolic regulators (SREBF2, PPARγ, PGC1α, SIRT1), and ERK signaling in the mPFC D, F and hippocampus (HP) E, G. H, I Inflammatory markers in the hippocampus. H mRNA levels of Nfkb, Tnf, and Il6. I Cytokine concentrations of TNFα and IL6 measured by ELISA. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001; one-way ANOVA with Holm-Sidak post hoc test A, B, H, I or Student’s t-test C, F, G. For RE-DW, SI-DW, and SI-ALC groups, respectively, N = 3, 3, and 7 mice A–C, N = 3, 4, and 6 mice (H), and N = 3, 3, and 5 mice I. N = 3 mice per group D–G. Both sexes were used in approximately equal numbers for biochemical and molecular analyses. Acetyl-L-carnitine modulates the expression of genes related to myelination, neural activity, and inflammation in the brains of socially isolated mice. Mice were assigned to three groups: regular environment with distilled water treatment (RE-DW), social isolation with distilled water treatment (SI-DW), and social isolation with acetyl-L-carnitine treatment (SI-ALC). RE mice were group-housed (3–4 mice per cage), whereas SI mice were housed individually (1 mouse per cage) from P21 to P35. Mice received daily oral gavage of distilled water (DW) or ALC (1.251 mmol/kg). After the 2-week rearing period, mice were euthanized, and tissues were collected for molecular and biochemical analyses. A, B RT-qPCR analysis of genes related to neuronal activity and neurotransmission, myelination, and oxidative stress/inflammation in the mPFC A and brainstem B. C Liquid chromatography–mass spectrometry-based quantification of glutamate and glutamine levels in the brainstem. D–G Western blot images D, E and protein quantification (F, G) for myelin markers (MBP, MAG), lipid/metabolic regulators (SREBF2, PPARγ, PGC1α, SIRT1), and ERK signaling in the mPFC D, F and hippocampus (HP) E, G. H, I Inflammatory markers in the hippocampus. H mRNA levels of Nfkb, Tnf, and Il6. I Cytokine concentrations of TNFα and IL6 measured by ELISA. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001; one-way ANOVA with Holm-Sidak post hoc test A, B, H, I or Student’s t-test C, F, G. For RE-DW, SI-DW, and SI-ALC groups, respectively, N = 3, 3, and 7 mice A–C, N = 3, 4, and 6 mice (H), and N = 3, 3, and 5 mice I. N = 3 mice per group D–G. Both sexes were used in approximately equal numbers for biochemical and molecular analyses. Alterations in lipid profiles of the corpus callosum induced by post-weaning social isolation and acetyl-L-carnitine administration To investigate how PWSI alters lipid composition in myelin-enriched white matter, we performed lipidomic analysis of the corpus callosum (Fig. 6A). The corpus callosum was selected not only for its critical role in the interhemispheric integration of social information [35, 36], but also for its high myelin density. This approach enabled the detection of PWSI- and ALC-associated changes in corpus callosum lipid composition.Fig. 6Lipidomic profiling of the corpus callosum following post-weaning social isolation and acetyl-L-carnitine treatment.Mice were assigned to three groups: regular environment with distilled water treatment (RE-DW), social isolation with distilled water treatment (SI-DW), and social isolation with ALC treatment (SI-ALC). RE mice were group-housed (3–4 mice per cage), whereas SI mice were housed individually (1 mouse per cage) from P21 to P35. Mice received daily oral gavage of distilled water (DW) or ALC (1.251 mmol/kg). After the 2-week rearing period, mice were euthanized, and corpus callosum samples were collected for liquid chromatography‒Orbitrap mass spectrometry-based lipidomics. A Schematic workflow of the lipidomic analysis. B Total number of identified lipid species. C Principal component analysis (PCA) score plot derived from the lipidomics dataset. D–H Relative abundances of cholesterol D, fatty acyls E, glycerophospholipids F, sphingolipids G, and glycerolipids H in the corpus callosum of RE-DW, SI-DW, and SI-ALC mice. I–J Volcano plots illustrating significantly changed lipids between RE-DW and SI-DW I and between SI-DW and SI-ALC J, based on log2 fold change and p values. K Schematic overview of major lipid metabolic pathways related to lipid composition in myelin-enriched white matter. For RE-DW, SI-DW, and SI-ALC groups, respectively, N = 3, 4, and 7 mice. Both sexes were used in approximately equal numbers. *p < 0.05; Wilcoxon rank-sum test. Graphs show mean ± SEM of normalized abundance. Lipidomic profiling of the corpus callosum following post-weaning social isolation and acetyl-L-carnitine treatment. Mice were assigned to three groups: regular environment with distilled water treatment (RE-DW), social isolation with distilled water treatment (SI-DW), and social isolation with ALC treatment (SI-ALC). RE mice were group-housed (3–4 mice per cage), whereas SI mice were housed individually (1 mouse per cage) from P21 to P35. Mice received daily oral gavage of distilled water (DW) or ALC (1.251 mmol/kg). After the 2-week rearing period, mice were euthanized, and corpus callosum samples were collected for liquid chromatography‒Orbitrap mass spectrometry-based lipidomics. A Schematic workflow of the lipidomic analysis. B Total number of identified lipid species. C Principal component analysis (PCA) score plot derived from the lipidomics dataset. D–H Relative abundances of cholesterol D, fatty acyls E, glycerophospholipids F, sphingolipids G, and glycerolipids H in the corpus callosum of RE-DW, SI-DW, and SI-ALC mice. I–J Volcano plots illustrating significantly changed lipids between RE-DW and SI-DW I and between SI-DW and SI-ALC J, based on log2 fold change and p values. K Schematic overview of major lipid metabolic pathways related to lipid composition in myelin-enriched white matter. For RE-DW, SI-DW, and SI-ALC groups, respectively, N = 3, 4, and 7 mice. Both sexes were used in approximately equal numbers. *p < 0.05; Wilcoxon rank-sum test. Graphs show mean ± SEM of normalized abundance. A total of 389 lipid species, classified into four major classes and 26 subclasses, were identified (Fig. 6B and Supplementary Table S10). Principal component analysis revealed distinct clustering among groups, indicating group-dependent alterations in lipid profiles (Fig. 6C). Cholesterol, a key lipid comprising 46% of myelin lipids [37], showed a decreasing trend in the SI-DW group, though it was not statistically significant (Fig. 6D and Supplementary Table S11). Comparison of total lipid subclass abundance revealed a prominent decrease in phosphatidylethanolamine-plasmalogen (PE-P) in the social isolation group (Fig. 6E-H and Supplementary Table S12). Analysis of individual lipid species revealed that the levels of three triacylglycerol (TG) species were significantly higher in the socially isolated mice (SI-DW) compared to standard-housed controls (RE-DW). Conversely, nine glycerophospholipids, three ceramides, and one additional TG species were significantly lower in the socially isolated group (Fig. 6I and Supplementary Table S13). Notably, the reduced glycerophospholipids predominantly included species containing saturated fatty acids from the PE and phosphatidylcholine (PC) classes, such as PE (P-16:0/22:6), PE (P-18:0/16:0), PE (P-16:0/20:4), PE (P-16:0/16:0), PE (O-18:0/16:0), PC 20:1/22:6, PC 18:0/20:4, PE 18:0/22:4, and PE 18:0/20:1. In mice subjected to social isolation, ALC administration significantly reduced certain glycerophospholipids and glycerolipids, while elevating levels of Cer (d18:1/20:0) and TG 18:1/18:2/18:3 (Fig. 6J) compared to the DW-treated control group. These results suggest that PWSI reduced PE-P levels in corpus callosum lipids, particularly among species containing saturated fatty acids (Fig. 6K). Molecular shifts in corpus callosum myelin lipids induced by post-weaning social isolation are reversed by acetyl-L-carnitine We next analyzed acyl-chain features to characterize molecular shifts induced by social isolation and ALC administration. Compared with mice in a regular environment, those subjected to social isolation (both receiving DW) showed lipid fold changes in TG, PE, diacylglycerol (DG), PC, and phosphatidylserine (PS) species that negatively correlated with the degree of unsaturation and total carbon chain length, whereas ALC treatment reversed these trends (Fig. 7A–B, Supplementary Tables S14–15). This indicates that social isolation preferentially reduces longer and more unsaturated lipid species, whereas ALC shifts the lipid profile toward that of regular-environment controls. Within the PE-P and ether-linked phosphatidylethanolamine (PE-O) subclasses, species containing saturated or polyunsaturated fatty acids were significantly depleted by social isolation but partially recovered following ALC administration (Fig. 7C). The reduction in saturated fatty acid-containing species under social isolation was accompanied by the increased expression of desaturase genes, including Scd1 and Scd3, which were suppressed by ALC administration (Fig. 7D–F). In summary, ALC promotes the recovery of acyl-chain composition, including saturated and polyunsaturated fatty acid-containing species and chain-length distribution, thereby reversing PWSI-associated lipidomic shifts (Fig. 7G).Fig. 7Alterations in fatty-acyl composition of corpus callosum lipids induced by post-weaning social isolation and acetyl-L-carnitine administration.Mice were assigned to three groups: regular environment with distilled water treatment (RE-DW), social isolation with distilled water treatment (SI-DW), and social isolation with ALC treatment (SI-ALC). RE mice were group-housed (3–4 mice per cage), whereas SI mice were housed individually (1 mouse per cage) from P21 to P35. Mice received daily oral gavage of distilled water (DW) or ALC (1.251 mmol/kg). After the 2-week rearing period, mice were euthanized, and corpus callosum samples were collected for liquid chromatography‒Orbitrap mass spectrometry-based lipidomics. A Correlation between fold changes in triacylglycerol (TG), phosphatidylethanolamine (PE), diacylglycerol (DG), phosphatidylcholine (PC), and phosphatidylserine (PS) species and the degree of unsaturation in SI-DW vs. RE-DW (left) and SI-ALC vs. SI-DW (right). B Correlation between fold changes in TG and PE species and fatty-acyl carbon chain length in SI-DW vs. RE-DW (left) and SI-ALC vs. SI-DW (right). C Distribution of saturated fatty acid (SFA)-, monounsaturated fatty acid (MUFA)-, and polyunsaturated fatty acid (PUFA)-containing species in phosphatidylethanolamine-plasmalogen (PE-P) and ether-linked phosphatidylethanolamine (PE-O) lipids. D Schematic overview of fatty-acid metabolic pathways highlighting key enzymes involved in SFA, MUFA, and PUFA biosynthesis. E,F Relative mRNA expression of enzymes involved in fatty-acid desaturation E and elongation F in the corpus callosum. G Summary of changes in fatty-acyl composition within corpus callosum lipids following SI and ALC treatment. For RE-DW, SI-DW, and SI-ALC groups, respectively, N = 3, 4, and 7 mice. Both sexes were used in approximately equal numbers. *p < 0.05 (RE-DW vs. SI-DW), #p < 0.05 (SI-DW vs. SI-ALC); Wilcoxon rank-sum test. Data are presented as mean ± SEM of normalized abundance. Alterations in fatty-acyl composition of corpus callosum lipids induced by post-weaning social isolation and acetyl-L-carnitine administration. Mice were assigned to three groups: regular environment with distilled water treatment (RE-DW), social isolation with distilled water treatment (SI-DW), and social isolation with ALC treatment (SI-ALC). RE mice were group-housed (3–4 mice per cage), whereas SI mice were housed individually (1 mouse per cage) from P21 to P35. Mice received daily oral gavage of distilled water (DW) or ALC (1.251 mmol/kg). After the 2-week rearing period, mice were euthanized, and corpus callosum samples were collected for liquid chromatography‒Orbitrap mass spectrometry-based lipidomics. A Correlation between fold changes in triacylglycerol (TG), phosphatidylethanolamine (PE), diacylglycerol (DG), phosphatidylcholine (PC), and phosphatidylserine (PS) species and the degree of unsaturation in SI-DW vs. RE-DW (left) and SI-ALC vs. SI-DW (right). B Correlation between fold changes in TG and PE species and fatty-acyl carbon chain length in SI-DW vs. RE-DW (left) and SI-ALC vs. SI-DW (right). C Distribution of saturated fatty acid (SFA)-, monounsaturated fatty acid (MUFA)-, and polyunsaturated fatty acid (PUFA)-containing species in phosphatidylethanolamine-plasmalogen (PE-P) and ether-linked phosphatidylethanolamine (PE-O) lipids. D Schematic overview of fatty-acid metabolic pathways highlighting key enzymes involved in SFA, MUFA, and PUFA biosynthesis. E,F Relative mRNA expression of enzymes involved in fatty-acid desaturation E and elongation F in the corpus callosum. G Summary of changes in fatty-acyl composition within corpus callosum lipids following SI and ALC treatment. For RE-DW, SI-DW, and SI-ALC groups, respectively, N = 3, 4, and 7 mice. Both sexes were used in approximately equal numbers. *p < 0.05 (RE-DW vs. SI-DW), #p < 0.05 (SI-DW vs. SI-ALC); Wilcoxon rank-sum test. Data are presented as mean ± SEM of normalized abundance. Discussion This study provides evidence that ALC acts directly on OL-lineage cells, particularly during the sheath-forming stage, to promote myelin sheath expansion and mitigate PWSI-induced social and myelination deficits in vivo. We propose that ALC supports the sheath-forming stage of OL development through three complementary mechanisms: upregulation of myelin-related gene and protein expression, enhancement of mitochondrial bioenergetics, and remodeling of lipid species in myelin-enriched white matter (Supplementary Fig. S7). Together, these effects may help preserve the cellular and regional conditions required for myelin maintenance and remodeling under PWSI conditions. Our findings suggest that ALC acts primarily during the sheath-forming stage rather than during OPC proliferation or early maturation. ALC-derived acetyl-CoA has been reported to contribute to epigenetic regulation by supporting histone acetylation [38]. Consistent with this possibility, ALC treatment increased nuclear H3K27 acetylation in an Slc22a5-dependent manner (Fig. 2N, O). Previous studies have shown that ALC can modify histone acetylation by increasing acetyl-CoA availability [15], and H3K27Ac is closely associated with enhancer activation at major myelin genes, including MBP and PLP1 [27]. Furthermore, ALC increased PPARγ nuclear translocation (Fig. 2N, O), consistent with previous reports identifying PPARγ as a transcriptional regulator that promotes brain myelination [39]. In contrast, L-carnitine failed to reproduce the sheath-expanding effects of ALC (Fig. 2), suggesting that the acetylated form of L-carnitine may be important for OL membrane sheath expansion. In addition, ALC restored the expression of neural activity-related genes in vivo (Fig. 5), raising the possibility that recovery of activity-dependent cues may further support OL-mediated myelination [31]. In terms of intracellular signaling, the PI3K/Akt/mTOR pathway has been implicated in OPC differentiation [40], whereas ERK signaling is known to regulate myelin thickness and sheath growth after differentiation [41, 42]. In this study, ALC increased ERK phosphorylation without detectably affecting PI3K/Akt signaling among the pathways examined, and MEK/ERK inhibition attenuated ALC-induced MBP expression (Supplementary Fig. S4E-H and Fig. 2L-M). Although ALC reduced the expression of differentiation-inhibitory genes such as Id2 and Id4 (Fig. 2K), the most prominent phenotype in primary cultures was increased myelin sheath expansion rather than altered OPC proliferation or differentiation (Fig. 2D-F). The loss of ALC-induced sheath expansion under Slc22a5 knockdown conditions (Fig. 2) supports the importance of the OCTN2-dependent ALC metabolic pathway in OL sheath formation. Interestingly, ALC responsiveness was partially preserved in 3D nanofiber cultures under Slc22a5 or Crat knockdown conditions (Fig. 2H). This observation suggests that axon-mimetic mechanical or topographical cues may partially support sheath expansion when OCTN2/CrAT-dependent ALC metabolism is impaired. Because these pathways were not directly tested in the present study, future work will be required to determine whether mechanotransductive signaling contributes to the compensatory response observed in the nanofiber assay. OLs require substantial energy to synthesize and maintain the extensive membrane structures that form myelin sheaths. This process depends on coordinated mitochondrial function and lipid biosynthesis. During development, mature OLs undergo metabolic remodeling, including increased reliance on glycolysis [23], yet mitochondrial activity remains important for myelin maintenance and lipid biosynthesis. Previous studies have implicated ALC in the regulation of mitochondrial function in neurons and other neural contexts. Specifically, ALC treatment has been shown to exert neuroprotective effects against MDMA-induced neurotoxicity at the mitochondrial level [43]. Furthermore, in the acute phase of spinal cord injury, ALC preserves neural integrity and improves motor function by suppressing mitochondrial dysfunction [44]. However, whether ALC modulates mitochondrial function in OLs, another metabolically demanding cell population in the nervous system, has remained unclear. In our experiments, ALC did not alter OCR-related parameters in OPCs but enhanced mitochondrial respiration in mature OLs, suggesting that the metabolic response to ALC is stage-dependent. Conversely, Slc22a5 knockdown reduced mitochondrial function in both OPCs and mature OLs (Fig. 3), indicating that OCTN2 is required for basal metabolic maintenance across OL-lineage stages. This broader requirement may reflect the role of OCTN2 in transporting carnitine and acylcarnitine derivatives beyond ALC alone. Our lipidomic analysis indicates that PWSI alters corpus callosum lipid composition and that ALC supplementation partially shifts this profile toward that of regular-environment controls. Although ALC increased SREBF2 expression in cell and tissue analyses (Figs. 2 and 5), cholesterol levels in the corpus callosum were not significantly changed (Fig. 6D). This may reflect the relatively slow turnover of myelin cholesterol, which has been reported to be much slower than that of PE species [45]. Therefore, the 14-day treatment period may have been insufficient to detect significant changes in cholesterol abundance but sufficient to reveal changes in more dynamically regulated lipid subclasses such as PE-P (Fig. 6F). The depletion of PE-P species containing saturated and polyunsaturated fatty acyl chains, along with reductions in PE-O species containing saturated fatty acyl chains in the corpus callosum of PWSI mice, suggests that social isolation affects lipid features important for myelin stability and antioxidant capacity. PE-P is highly enriched in white matter PE and contributes to membrane organization, electrical insulation, and myelin wrapping [24]. A reduction in saturated fatty acid-containing PE-P species may weaken membrane packing and structural stability, whereas the reduction of polyunsaturated fatty acid-containing species may reflect vulnerability to oxidative stress associated with social isolation [37, 46, 47]. The ALC-associated recovery of these lipid species suggests that ALC may improve myelin quality by supporting membrane lipid remodeling and reducing susceptibility to oxidative damage. However, because oxidative damage was not directly measured in this lipidomics experiment, this interpretation should be tested in future studies using direct markers of lipid peroxidation and redox imbalance. In addition, social isolation altered acyl-chain features across several lipid classes, including the degree of unsaturation and total carbon chain length (Fig. 7). The decrease in the degree of unsaturation observed across major lipid classes (e.g., TG and PE) indicates altered myelin-associated lipid remodeling [48], while reduced chain length may affect membrane packing properties that are relevant to axonal insulation [24]. ALC partially normalized these acyl-chain parameters, supporting the view that ALC promotes adaptive remodeling of corpus callosum lipid composition under PWSI conditions. Our results demonstrate that ALC restores mPFC myelination markers while also normalizing lipid features in the corpus callosum. The mPFC is a key region for social behavior, whereas the corpus callosum supports interhemispheric communication required for coordinated sensory, cognitive, and social processing [35, 36]. Therefore, ALC may exert its behavioral effects through both local restoration of mPFC myelination and broader support of white matter lipid homeostasis. Although the present study does not directly measure long-range circuit efficiency, the recovery of corpus callosum lipid composition suggests a potential mechanism by which ALC may support interregional communication in socially isolated mice. Unlike L-carnitine, ALC administration increased mPFC ALC levels, enhanced myelin-related markers, and restored sociability under PWSI conditions (Fig. 4). Although ALC and L-carnitine share the transporter OCTN2 [14, 49], ALC has additional metabolic properties related to acetyl moiety. ALC has been associated with neuroprotective and antioxidant effects [50] and can regulate intracellular acetyl-CoA/CoA balance [51]. ALC-derived acetyl units can enter acetyl-CoA pools that support mitochondrial oxidative metabolism, lipid synthesis, histone acetylation, and acetylcholine synthesis [15, 52, 53]. In addition, ALC-derived carbon can be incorporated into glutamate, glutamine, and GABA through TCA-cycle–linked metabolism [52, 54–56]. These broader metabolic functions may explain why ALC, but not L-carnitine, showed robust pro-myelinating and behavioral effects in the present study. Notably, the effects of ALC were more pronounced under social isolation conditions than under regular-environment conditions. In the present study, ALC treatment increased mPFC ALC levels and restored social behavior in socially isolated mice (Fig. 4), whereas the same treatment increased serum ALC levels without elevating brain ALC levels or altering behavior in regular-environment mice (Supplementary Fig. S6D–F). These findings suggest that PWSI may create a metabolic or inflammatory state that favors ALC accumulation or action in the central nervous system. Although BBB permeability was not directly measured, social isolation-induced neuroinflammation or metabolic imbalance may alter transporter activity, including OCTN2-dependent transport, or modify functional barrier kinetics. In contrast, under regular-environment conditions, intact metabolic homeostasis may limit the additional brain accumulation or biological effect of exogenous ALC. This interpretation is consistent with reports that ALC supplementation under normal conditions does not necessarily increase myelin protein expression and may even reduce it [18], whereas ALC promotes remyelination in nerve injury or demyelination models [17, 57]. Thus, ALC supplementation may be most effective in pathological contexts characterized by increased metabolic demand or impaired myelin homeostasis. ALC may also promote myelination by improving the microenvironment surrounding OLs. OL-lineage cells are vulnerable to inflammation and oxidative stress because of their high metabolic demands and relatively limited antioxidant reserves [58]. Chronic glucocorticoid exposure can promote reactive oxygen species accumulation and inhibit OPC differentiation [59, 60]. In our study, ALC attenuated social isolation-induced inflammatory signals and normalized glucocorticoid receptor-related changes across multiple brain regions (Fig. 5). These findings suggest that ALC supports myelination not only through direct metabolic effects on OLs but also by reducing environmental stressors that impair OL function. Beyond its cell-autonomous effects on OLs, ALC’s in vivo efficacy appears to involve broader restoration of the pathological brain environment across multiple regions (Fig. 5). PWSI reduced neural activity-related markers and increased markers associated with HPA-axis activation and neuroinflammation [12, 32]. In the mPFC and brainstem, ALC restored the expression of neural activity and synaptic markers, including Fos, Chat, and Syp, which may provide activity-dependent cues that support myelination [31]. ALC also reduced pro-inflammatory markers across multiple regions, including Tnf expression in the mPFC, TNFα levels in the hippocampus, and Il6 expression in the brainstem, while normalizing Nr3c1 expression in the mPFC and brainstem (Fig. 5). Although we examined whether the pro-myelinating effects of ALC were associated with Bdnf [61], Bdnf mRNA levels in the mPFC did not reach statistical significance. Together, these findings suggest that recovery in our model may be driven by three complementary mechanisms: direct metabolic priming of OLs, restoration of activity-dependent neuronal inputs, and improvement of the neuroinflammatory and stress-related microenvironment. The convergence of these mechanisms may contribute to restored myelination and social behavior in socially isolated mice. Our behavioral analysis focused specifically on sociability, reflecting our primary interest in the functional relationship between mPFC myelination and social behavior. Although ALC restored social preference, its effects on other behavioral domains, such as anxiety-like behavior or cognitive function, remain to be determined. Future studies should examine whether ALC has broader therapeutic effects across multiple behavioral phenotypes and whether these effects are mediated by shared or region-specific myelination mechanisms. In summary, ALC rescues PWSI-induced social and myelination deficits through multifaceted mechanisms involving myelin protein expression, mitochondrial bioenergetic support, and remodeling of corpus callosum lipid composition. By supporting OL metabolism and improving the inflammatory and stress-related microenvironment, ALC may represent a potential therapeutic avenue for conditions involving impaired myelination and social dysfunction. Supplementary information Supplementary information Supplementary information