Abstract The accumulation of lipids in microglia/macrophage-induced inflammation exacerbation represents a pivotal factor contributing to secondary injury following spinal cord injury (SCI). N-Lactoyl-Phenylalanine (L-P), a metabolic byproduct of exercise, exhibits the capacity to regulate carbohydrate and lipid metabolism and may serve as a potential regulator of lipid metabolism in microglia/macrophage. This study investigates the role of L-P in modulating lipid homeostasis in microglia/macrophage and its therapeutic implications for SCI recovery. By establishing a mouse model of SCI, we confirmed that L-P administration markedly altered lipid metabolism in microglia/macrophage. This metabolic reprogramming was mediated through the activation of the AMPK-PGC1α-PPARγ signaling pathway, which plays a crucial role in regulating cellular energy metabolism and inflammatory responses. Our findings demonstrate that L-P treatment enhances the lipid metabolic capacity of microglia/macrophage, thereby attenuating neuroinflammation and promoting tissue repair after injury. Moreover, the polarization of microglia/macrophage shifts toward the anti-inflammatory M2 phenotype, providing substantial support for the regenerative process of the injured spinal cord. Functional analysis revealed that mice treated with L-P exhibited significantly improved motor function compared to the control group. Collectively, these results underscore the therapeutic potential of L-P in SCI and suggest its utility as a metabolic intervention strategy by modulating microglia/macrophage lipid metabolism to accelerate recovery. Graphical Abstract [38]graphic file with name 12974_2025_3495_Figa_HTML.jpg Supplementary Information The online version contains supplementary material available at 10.1186/s12974-025-03495-3. Introduction Spinal cord injury (SCI) is a profoundly disabling neurological condition that frequently leads to permanent functional impairments and a limited regenerative capacity [[39]1]. After SCI, complex pathological cascades occur, including neuroinflammation, oxidative stress, and metabolic disturbances, which aggravate tissue damage and impede recovery. Among these pathological processes, dysregulated lipid metabolism has attracted increasing attention due to its significant role in neuroinflammation and central nervous system (CNS) repair [[40]2]. Microglia and macrophage play a crucial role in mediating post-SCI inflammation and tissue remodeling [[41]3]. Axonal demyelination and Wallerian degeneration induced after SCI can lead to the formation of a large number of myelin debris. After phagocytosing these debris, microglia/macrophage will accumulate lipid droplets intracellularly and undergo foaming, have defects in phagocytosis, generate high levels of reactive oxygen species, and secrete a large amount of pro-inflammatory cytokines, thereby exacerbating the death of neurons and the formation of scars [[42]4]. However, the molecular mechanisms connecting lipid metabolism to microglia/macrophage function and subsequent tissue repair remain insufficiently understood. The AMP-activated protein kinase (AMPK)-peroxisome proliferator-activated receptor γ coactivator 1α (PGC1α)-peroxisome proliferator-activated receptor γ (PPARγ) axis has emerged as a critical regulatory pathway in cellular energy homeostasis and lipid metabolism [[43]5]. AMPK is a central metabolic sensor that becomes activated under energy stress conditions and plays a key role in maintaining metabolic balance by promoting catabolic pathways and inhibiting anabolic processes [[44]6]. Downstream of AMPK, PGC1α acts as a transcriptional coactivator that regulates mitochondrial biogenesis and oxidative metabolism, while PPARγ, a nuclear receptor, governs lipid uptake and storage. Activation of the AMPK-PGC1α-PPARγ axis has been associated with modulating inflammation and metabolic reprogramming in various cell types, including microglia/macrophage [[45]7, [46]8]. Thus, targeting this axis represents a promising strategy for improving metabolic and inflammatory outcomes in SCI. Recent studies have emphasized the therapeutic potential of small molecules that regulate metabolic pathways to promote CNS repair [[47]9, [48]10]. One such molecule is N-Lactoyl-Phenylalanine (L-P), a naturally occurring metabolite derived from lactate and phenylalanine, is a blood-derived signal that can inhibit appetite and obesity. This signal can be stimulated by exercise. Long-term use of L-P can reduce fat and weight, and improve glucose balance [[49]11, [50]12]. AMPK is a crucial molecule for exercise-induced metabolic adaptation, and L-P is a metabolic signaling molecule generated by exercise. It is possible that the two have a synergistic role in maintaining systemic energy balance. L-P has been implicated in regulating metabolic processes in peripheral tissues, but its role in CNS injury and repair remains largely unexplored. Given the central role of lipid metabolism in microglia/macrophage function and SCI pathophysiology, L-P may constitute a novel therapeutic agent for enhancing recovery after SCI. In this study, we examined the effects of L-P on microglia/macrophage lipid metabolism and its subsequent influence on SCI recovery. Specifically, we hypothesized that L-P facilitates metabolic reprogramming in microglia/macrophage by activating the AMPK-PGC1α-PPARγ axis, leading to enhanced lipid metabolism and reduced neuroinflammation. To test this hypothesis, we utilized a mouse model of SCI and conducted a series of molecular, cellular, and behavioral analyses to evaluate the therapeutic potential of L-P. Our findings demonstrate that L-P significantly improves functional recovery in mice with SCI, accompanied by enhanced lipid metabolism and reduced pro-inflammatory microglia/macrophage activation. These results not only clarify a mechanism by which L-P promotes CNS repair but also provide a strong basis for targeting microglia/macrophage metabolism in the treatment of SCI. Materials and methods Reagents and antibodies N-Lactoyl-Phenylalanine (HY-150012, purity:99.82%) and BODIPY 493/503 (HY-W090090) was bought from MCE (NJ, USA). The Oil red O staining kit was bought from Solarbio (G1261, Beijing, China). BV2 cell was bought from Ubigene (YC-C035, Guangzhou, China). Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum (FBS) were provided by Gibco (CA, USA). Antibodies NF200 (ab4680), GFAP (ab7260), ACE tubulin (ab24610), Iba1 (ab178846) and Iba1 (ab283319) were provided by Abcam (MA, USA). Antibodies pAMPK (44-1150G) and AMPK (PA5-36045) were provided by Invitrogen (CA, USA). Antibodies pan-Klac (PTM-1401RM) was provided by PTM bio (Hangzhou, China). Antibodies SREBP1 (14088-1-AP), ACOX1 (10957-1-AP), PLIN2 (15294-1-AP), CD163 (16646-1-AP), ARG1 (16001-1-AP), PGC1α (66369-1-Ig), Ly6c (FITC-65140) and PPARγ (16643-1-AP) were provided by Proteintech (Wuhan, China). Antibodies CD36 (DF13262) was provided by Affinity (Suzhou, China). Antibodies CD86 (sc-19617) was provided by Santa Cruz (California, USA). Antibodies IL1β (516288), IL10 (502171) and β actin (380624) were provided by Zenbio (Chengdu, China). Antibody TYR tubulin (MAB1864-1) was provided by millipore (Burlington, USA). Donkey anti-rabbit IgG 488 (ab150073); Donkey anti-rabbit IgG 555 (ab150062); Donkey anti-mouse IgG 488 (ab150105); Donkey anti-mouse IgG 555 (ab150110); Goat anti-rabbit IgG HRP (ab205718); Goat anti-mouse IgG HRP (ab6789) were provided by Abcam (MA, USA). An enhanced chemiluminescence (ECL) kit was provided by Zenbio (Chengdu, China). Animal models C57BL/6 mice approximately eight weeks old (weighing 20–22 g, male and female) were supplied by Charles River, Shanghai, China. AMPK^flox/flox mice (C57BL/6JGpt-Prkaa1^em1Cflox/Gpt, Strain NO. T007182), Cx3cr1^iCre mice (C57BL/6JGpt-Cx3cr1^em1Cin(iCre)/Gpt, Strain NO. T006768) and Tmem119^CreERT (Tmem119-P2A-CreERT2, Strain NO. T050766) mice were supplied by GemPharmatech. R26-CAG-LSL-tdTomato mice (Cat. NO. NM-KI-225042) were purchased from Shanghai Model Organisms Center, Inc. The animals were acclimatized for 7 days before the experiment under a controlled room temperature ranging between 23 °C and 25 °C, with unrestricted access to water and food. The protocol for the use and care of the animals was approved by the Animal Care and Use Committee of Wenzhou Medical University and strictly adhered to the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health (Ethics No. Wydw2024–0539). C57BL/6 mice were anesthetized with 1% pentobarbital. After the spinous and lamina processes of T9-T10 were removed, the spinal cord was exposed. The spinal cord at T9 was contused using the Reward Spinal Cord Impactor (Reward, Shenzhen, China. parameters: 1.5 m/s, 0.6 mm). Cefazolin sodium (0.9%) was injected into the peritoneal cavity of the mice twice daily, and artificial assisted urination was carried out every morning and evening. Animal groups The C57BL/6 mice were divided into 6 groups: Sham group (in which mice had their vertebral laminae pried open but without spinal cord injury, n = 24), SCI group (with spinal cord contusion injury, n = 24), Exercise group (Mice with spinal cord contusion were exercised on a running wheel for 30 min each day. Specifically, the protocol involved starting from the second day after surgery, when the mice regained consciousness from anesthesia and the ability to move freely, and then continuously running on the wheel for 30 min each day, n = 12), L-P group (After the establishment of the spinal cord contusion model, a dose of 20 mg/kg of N-Lactoyl-Phenylalanine was administered via the tail vein every day for a week. The first administration was immediately carried out after the SCI model was established and the wound was sutured. n = 18). 10 mg/kg L-P group (After the establishment of the spinal cord contusion model, a dose of 10 mg/kg of N-Lactoyl-Phenylalanine was administered via the tail vein every day for a week, n = 6). 30 mg/kg L-P group (After the establishment of the spinal cord contusion model, a dose of 30 mg/kg of N-Lactoyl-Phenylalanine was administered via the tail vein every day for a week, n = 6). The AMPK^flox/flox; Tmem119^CreERT−/− and AMPK^flox/flox; Tmem119^CreERT+/− mice were divided into 2 groups: AMPK^flox/flox; Tmem119^CreERT−/− group (AMPK^flox/flox; Tmem119^CreERT−/− mice after the establishment of the spinal cord contusion model, a dose of 20 mg/kg of N-Lactoyl-Phenylalanine was administered via the tail vein every day for a week, n = 6). AMPK^flox/flox; Tmem119^CreERT+/− group (AMPK^flox/flox; Tmem119^CreERT+/− mice after the establishment of the spinal cord contusion model, a dose of 20 mg/kg of N-Lactoyl-Phenylalanine was administered via the tail vein every day for a week, n = 6). The Cx3cr1^iCre+/−; tdTomato^+/− mice and Tmem119^CreERT+/−; tdTomato^+/− mice were divided into 2 groups: SCI group (with spinal cord contusion injury, n = 15), L-P group (After the establishment of the spinal cord contusion model, a dose of 20 mg/kg of N-Lactoyl-Phenylalanine was administered via the tail vein every day for a week, n = 15). Extraction of plasma from exercising mice The mice in the Exercise group were euthanized after exercise, and 1 mL of whole blood was obtained from the heart and injected into an anticoagulant tube. The collected anticoagulant whole blood was placed at 4 °C and centrifuged at 3000 rpm for 10 min. Subsequently, the supernatant was carefully aspirated using a pipette to acquire the plasma. The determination of Lac-Phe After extracting the spinal cord proteins from the SCI group and the exercise group with NP40 lysis buffer, the lactylated molecules were bound with K-lac antibody. The antibody bound with the protein was adsorbed by immunomagnetic beads, and then the supernatant was discarded. The bound protein was eluted with elution buffer, and finally the content of Lac-Phe was detected with a phenylalanine detection kit (ab241000). Oil red O staining Frozen sections were fixed in 10% formalin for 30 min, washed twice with distilled water (3 min each), and immersed in 60% isopropanol for 20–30 s. Sections were stained with modified Oil Red O for 10–15 min, briefly washed with 60% isopropanol (5–10 s), and rinsed in distilled water for 1 min. Mayer’s hematoxylin re-stained nuclei for 1–2 min, followed by 10 min of tap water rinsing for bluing. After brief cleaning in distilled water, moisture was removed with filter paper, and sections were mounted using glycerin gelatin. Cell culture The BV2 cell was cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin. Incubation of cells was carried out at 37℃ in a humidified incubator with a CO[2] concentration of 5%. Drug administration to cell The cells were divided into 6 groups: Control group (cells were not treated), NaOl group (stimulated with 100 µM sodium oleate), Exercise group (stimulated with 100 µM sodium oleate and 100 µL blood plasma of exercising mice), L-P group (stimulated with 100 µM sodium oleate and 100 µM N-Lactoyl-Phenylalanine, ), L-P + Compound C group (stimulated with 100 µM sodium oleate, 100 µM N-Lactoyl-Phenylalanine, and 2 µM of Compound C) and L-P + SR-18,292 group (stimulated with 100 µM sodium oleate, 100 µM N-Lactoyl-Phenylalanine, and 20 µM of SR-18292). BODIPY staining 1 mg of BODIPY 493/503 is dissolved in 382µL of DMSO to acquire a 10mM stock solution. The stock solution is diluted at a ratio of 1:1000 and added to adherent cells from which the culture medium has been removed, followed by incubation for 30 min. The cells are washed twice with the culture medium for 5 min each time. Fluorescence microscopy is employed to observe the cells. Regarding the flow-sorted cells, following fluorescence quenching by light exposure, the fluorescence intensity was first compared with that of unstained cells to confirm the effectiveness of fluorescence quenching. Subsequently, BODIPY dye was added for staining. The fluorescence intensity was then measured using a microplate reader at an excitation wavelength of 488 nm and an emission wavelength range of 500–520 nm. Network Pharmacological analysis First, we systematically retrieved and screened therapeutic targets related to spinal cord injury (SCI) using the GeneCards and OMIM databases [[51]13]. Subsequently, potential biological targets of the L-P were identified through the high-precision prediction tool SwissTargetPrediction [[52]14]. Next, by constructing a Venn diagram, we identified the intersection of these two sets of targets, referred to as common targets. Based on these common targets, a protein-protein interaction (PPI) network was constructed using the STRING database to explore potential connections between them [[53]15]. To further understand the functional attributes of these targets and their roles in biological pathways, we conducted GO analysis and KEGG pathway enrichment analysis using the DAVID database [[54]16]. The results of the data analysis were visualized using R Studio to present our research findings clearly and intuitively. Transcriptome sequencing RNA isolation and library Preparation Total RNA was extracted using the TRIzol reagent (Invitrogen, CA, USA) following the manufacturer’s protocol. The purity and quantification of RNA were determined via the NanoDrop 2000 spectrophotometer (Thermo Scientific, USA). The integrity of RNA was evaluated by the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Subsequently, the libraries were built with the VAHTS Universal V6 RNA-seq Library Prep Kit in accordance with the manufacturer’s instructions. The transcriptome sequencing and analysis were performed by OE Biotech Co., Ltd. (Shanghai, China). RNA sequencing and differentially expressed genes analysis The libraries were sequenced on an llumina Novaseq 6000 platform and 150 bp paired-end reads were generated. About 51.33 M raw reads for each sample were generated. Raw reads of fastq format were firstly processed using fastp [[55]17–[56]25] and the low quality reads were removed to obtain the clean reads. Then about 48.05 M clean reads for each sample were retained for subsequent analyses. The clean reads were mapped to the house mouse genome using HISAT2 [[57]18]. FPKM [[58]19] of each gene was calculated and the read counts of each gene were obtained by HTSeq-count [[59]26]. PCA analysis were performed using R (v 3.2.0) to evaluate the biological duplication of samples. Based on the hypergeometric distribution, enrichment analysis of GO [[60]27] pathways for DEGs was respectively carried out to screen the significant enriched terms by using R (v 3.2.0). R (v 3.2.0) was employed to draw the column diagram, the chord diagram and bubble diagram of the significant enrichment terms. Molecular docking study Ligands’ 2D structures were retrieved from PubChem and converted to 3D using the LigPrep module. Diverse conformations were generated with the OPLS-2005 force field, selecting the lowest-energy conformer for molecular docking [[61]28]. PGC1α (PDB ID: 3CS8) and AMPK (PDB ID: 4QFR) from the Protein Data Bank were prepared using Schrodinger Maestro’s protein preparation wizard, with protonation states assigned, structure restrained, and partial energy minimized to a 0.3 Å RMSD limit [[62]29]. Docking simulations were conducted using Schrodinger’s Glide module, where each bioactive molecule was docked into the designated binding site of the grid. The ideal ligand–protein interactions were identified by selecting the lowest-energy binding posture for each docking experiment. Flow cytometry detection Immediately after the animals were sacrificed, the spinal cords were removed and placed on ice. The dura mater was stripped off, and the spinal cords were cut into 1 mm³ pieces with ophthalmic scissors. They were then placed in pre-cooled DPBS. 2 ml of 0.25% trypsin digestion solution was added to each spinal cord, and gently pipetted. The digestion time was 5 min. After the cells passed through the cell sieve, 2 ml of DMEM medium containing 10% FBS was added to terminate the digestion. The obtained cell suspension was resuspended in DPBS and sorted for tdTomato positive cells using a flow cytometer (Beckman Coulter, Moflo Astrios EQ). The sorted positive cells were collected for WB detection. As for the sorting of Ly6c + and tdTomato positive cells, we first incubated the sorted Cx3cr1 tdTomato cells with Ly6c FITC antibody (10 µg/mL) for 1 h, and then sorted them through flow cytometry. Enzyme-linked immunosorbent assay In this study, enzyme-linked immunosorbent assay (ELISA) was used to quantitatively detect the concentration of specific biomarkers in the samples. Briefly, 96-well plates were coated with 100 µL of capture antibody diluted in buffer and incubated overnight at 4 °C. After blocking with 1% BSA in PBS for 1 h at room temperature, 100 µL of sample or standard (prepared according to the manufacturer’s instructions) was added to each well. The plates were incubated at room temperature for 2 h and then washed with PBS containing 0.05% Tween 20. Biotinylated detection antibody was added to each well and incubated for 1 h. After washing, 100 µL of streptavidin-HRP was added and incubated for 30 min. All detections were repeated and the results were compared with the standard curve. Western blot Tissues and cells are lysed using a buffer (RIPA) with protease and phosphatase inhibitors to prevent degradation. Protein concentration is measured by the BCA method. Depending on target protein size, polyacrylamide gels adapted to the concentration are prepared. Samples are mixed with loading buffer, boiled, and subjected to electrophoresis at 120 V. Separated proteins are transferred to a PVDF membrane via wet transfer. The membrane is blocked with 5% skim milk to prevent non-specific binding. Primary antibody incubation occurs overnight at 4 °C, followed by three TBST washes. A secondary antibody is incubated at room temperature for 1 h, then washed with TBST. Finally, an ECL substrate enables a chemiluminescent reaction, and signals are captured and analyzed with an imaging system. Immunofluorescence After the paraffin sections were dewaxed and hydrated, they were rinsed thoroughly with PBS (phosphate-buffered saline) three times for 5 min each to remove any remaining residues and ensure proper preparation of the tissue. To retrieve antigens, the sections were subjected to heat-induced epitope retrieval by boiling them in citrate buffer. This process helps restore antigenicity that might have been masked during the fixation process. After cooling to room temperature, the sections were again rinsed with PBS three times for 5 min each to remove residual citrate buffer. To minimize non-specific binding, the samples were blocked using a 5% bovine serum albumin (BSA) solution. Next, the sections were incubated with the primary antibody at 4 °C overnight, ensuring specific binding to target antigens. Afterward, the sections were washed with PBS three times and incubated with the secondary antibody at room temperature for 1 h. Finally, after three additional PBS washes, the sections were mounted with an anti-fluorescence quenching agent containing DAPI for nuclear staining. H&E and Nissl staining The tissue 28 days after SCI were fixed with 4% paraformaldehyde, followed by gradient dehydration and paraffin embedding. Subsequently, the embedded tissues were sectioned into 5 μm - thick slices and mounted onto glass slides. For H&E staining, after dewaxing and rehydration, the sections were stained with hematoxylin for 6 min to stain the nuclei. Then, they were differentiated with 1% hydrochloric acid - ethanol solution and blue - toned with tap water. After that, eosin staining was performed for 2 min. Finally, the sections were dehydrated, made transparent, and mounted with neutral resin. For Nissl staining, following dewaxing and rehydration, the sections were stained with tar purple staining solution at 56 °C for 1 h. Subsequently, differentiation, dehydration, transparency treatment were carried out, and then the sections were mounted with neutral resin. Analysis of animal behavior At 28 days post SCI, mice were placed in a 50 cm-long plastic track to analyze overground locomotion. Anatomical landmarks (iliac crest, hip, knee, ankle, last toe, L1, and L4 vertebrae) were tracked, and at least six gait cycles were recorded per mouse at 60 fps using an iPhone. Analysis focused on the central 30 cm of the track, where mice walked 6–10 gaits. Errors from fast movement or moderate camera speed were noted. If hindlimbs were immobile, forelimbs determined the gait. Videos meeting criteria (continuous forward movement without head rotation) were analyzed with DeepLabCut to track joints [[63]30–[64]33]. Kinematic data, including angular excursions and limb swings, were processed in MATLAB, and analyses followed methods from Alluin et al. [[65]31]. and Zorner et al. [[66]30, [67]34]. Data collection and analysis were performed independently. BMS score One day before the experiment, the mice were placed in the test field to adapt to the environment. During the test, the mice were gently placed in the center of the field and their free movement was observed. The focus was on the movement of the hind limbs: joint activities (ankle, knee, hip), gait coordination, trunk stability, and foot placement on the ground, etc. The tests were conducted and recorded on the 1st, 7th, 14th, and 28th days after spinal cord injury. The experiment and data recording were carried out using a double-blind method. Statistical analysis Generalized linear mixed models were used to analyze the BMS scores. Statistical significance between two experimental groups was assessed using a two-sample t-test. To compare data across more than two groups, analysis of variance (ANOVA) followed by Tukey’s post hoc tests was performed. Statistical analyses were conducted using Excel and GraphPad Prism 9. Results are presented as mean ± standard deviation (SD), and a p-value of < 0.05 was considered statistically significant. Results Exercise for SCI mice can ameliorate the accumulation of lipid droplets in the injured area and inhibit inflammation Firstly, we employed Oil Red O staining to examine the accumulation of lipid droplets in the spinal cord. The data results indicated that a large number of lipid droplets accumulated in the injury area after SCI. Nevertheless, after exercising the SCI mice, the accumulation of lipid droplets in the injury area was significantly mitigated (Fig. [68]1A, F). Subsequently, we utilized Iba1 and PLIN2 immunofluorescence co-localization and discovered that a considerable amount of lipid droplets were located in Iba + cells after SCI. After exercising the SCI mice, the lipid droplets in Iba + cells were significantly reduced (Fig. [69]1B-C, G-I). Then, we extracted the serum from the exercised mice and found that the serum of exercised mice could significantly decrease the lipid overload of Iba + cells induced by sodium oleate in vitro. The detection of M1/M2-related indicators of Iba + cells indicated that exercise could facilitate the M2 transformation of Iba + cells (Fig. [70]1D-E, J-M). Subsequently, we employed the BMS score to assess the behavioral recovery of mice in each group. The experimental outcomes indicated that the motor function of mice in the exercise group achieved a more pronounced recovery (Sfigure [71]1). To sum up, the metabolites after exercise can effectively promote lipid metabolism in Iba + cells and promote the recovery of motor function in SCI mice. Fig. 1. [72]Fig. 1 [73]Open in a new tab Exercise accelerates lipid metabolism in Iba + cells and inhibits inflammation. (A) Oli Red O staining of each group, scale bar: 1000 μm. (B) Immunofluorescence expression of Iba1 (green), PLIN2 (red) and DAPI (blue) in each group, scale bar: 50 μm. (C) Immunofluorescence expression of BODIPY 493/503 (green) and DAPI (blue) in each group, scale bar: 50 μm. (D) Immunofluorescence expression of CD86 (green), Iba1 (red) and DAPI (blue) in each group, scale bar: 50 μm. (E) Immunofluorescence expression of Iba1 (green), CD163 (red) and DAPI (blue) in each group, scale bar: 50 μm. (F) Statistical analysis of Lipid droplet area ratio in each group. (G, I, J, L) Statistical analysis of PLIN2, BODIPY 493/503, CD86 and CD163 expression in each group. (H, K, M) Statistical analysis of PLIN2, CD86 and CD163 expression in Iba1 + cell in each group. * represents P < 0.05. ** represents P < 0.01 vs. the SCI group. n.s. stands for no statistical significance. The data are represented as the mean ± SD (n = 6) The rationale for selecting N-Lactoyl-Phenylalanine as a representative metabolite for exercise and the network pharmacology prediction analysis In recent years, the study of lactylation has been one of the hotspots. The accumulation of a large amount of lactate due to anaerobic metabolism after exercise provides a foundation for lactylation [[74]35]. The results of pan-Klac detection reveal that the lactylation level in the spinal cord of mice after exercise is significantly increased (Fig. [75]2A-B). N-Lactoyl-Phenylalanine (L-P) is a lactylation exercise metabolite and has emerged as an important role in the field of metabolic regulation in recent years [[76]11]. We detected the content of L-P in the lactoylated molecules. The experimental results demonstrated that the L-P in the exercise group was significantly higher than that in the SCI group (Fig. [77]2C). Subsequently, we used network pharmacology to predict the possible interacting proteins and pathways of L-P. The results show that it is highly enriched in various biological responses such as lysosomes and inflammatory responses (Fig. [78]2D-G). This is conducive to enhancing the lipid metabolic capacity of Iba + cells and promoting the repair of SCI. Fig. 2. [79]Fig. 2 [80]Open in a new tab The Basis for Selecting L-P and Network Pharmacology Analysis of L-P. (A) Western blotting showing the expression of pan-Klac in each group. (B) Quantitative analysis of pan-Klac protein expression. (C) Quantitative analysis of pan-Klac protein expression. (D) Common targets for both SCI and L-P. (E) The Protein-Protein Interaction Network of L-P and SCI. (F) GO enrichment analysis of the candidate targets for L-P and SCI. (G) KEGG enrichment analysis of the candidate targets for L-P and SCI. ** represents P < 0.01 vs. the SCI group. The data are represented as the mean ± SD (n = 6) N-Lactoyl-Phenylalanine can ameliorate the accumulation of lipid droplets in the injured area and inhibit inflammation Firstly, we employed the BMS score to initially screen the optimal treatment concentration of L-P. The data results indicated that the therapeutic efficacy of L-P essentially peaked at a dosage of 20 mg/kg. When the dosage was increased to 30 mg/kg, no significant alteration in the therapeutic effect was observed (Sfigure [81]2). Subsequently, to verify whether L-P possesses the ability to promote lipid metabolism in the spinal cord, we examined and analyzed the lipid accumulation in the injury area following L-P administration. The results of Oil Red O staining indicated that L-P could significantly decrease lipid accumulation in the injury area (Fig. [82]3A, F). In view of the significant alterations in the number of Iba1 + cells following treatment, we performed a quantitative analysis of microglia in each group. On this foundation, and with the quantification of the expression of various indicators as a premise, we augmented the statistical analysis of the expression levels of these indicators in Iba1 + cells of each group (Sfigure [83]3). The co-localization staining of Iba1 and PLIN2, as well as the BODIPY staining results in vitro, demonstrated that lipid droplets in Iba + cells were significantly eliminated after L-P treatment (Fig. [84]3B-C, G-I). Finally, we detected the polarization status of Iba + cells. The results of co-labeling CD86 and CD163 with Iba1 suggested that after regulating lipid metabolism in Iba + cells, L-P could effectively reverse Iba + cells polarized to M1 due to SCI and promote their transformation to M2 (Fig. [85]3D-E, J-M). To sum up, L-P can effectively eliminate the accumulated lipids in Iba + cells and facilitate their transformation to M2. Fig. 3. [86]Fig. 3 [87]Open in a new tab L-P accelerates lipid metabolism in Iba + cells and inhibits inflammation. (A) Oli Red O staining of each group, scale bar: 1000 μm. (B) Immunofluorescence expression of Iba1 (green), PLIN2 (red) and DAPI (blue) in each group, scale bar: 50 μm. (C) Immunofluorescence expression of BODIPY 493/503 (green) and DAPI (blue) in each group, scale bar: 50 μm. (D) Immunofluorescence expression of CD86 (green), Iba1 (red) and DAPI (blue) in each group, scale bar: 50 μm. (E) Immunofluorescence expression of Iba1 (green), CD163 (red) and DAPI (blue) in each group, scale bar: 50 μm. (F) Statistical analysis of Lipid droplet area ratio in each group. (G, I, J, L) Statistical analysis of PLIN2, BODIPY 493/503, CD86 and CD163 expression in each group. (H, K, M) Statistical analysis of PLIN2, CD86 and CD163 expression in Iba1 + cell in each group. * represents P < 0.05. ** represents P < 0.01 vs. the SCI group. n.s. stands for no statistical significance. The data are represented as the mean ± SD (n = 6) The regulatory pathways of N-Lactoyl-Phenylalanine screened based on transcriptome To determine the functional pathway of L-P, we initially conducted transcriptome sequencing for the SCI group and the L-P group. The results of GO enrichment analysis indicated that the long-chain fatty acid metabolism-related pathway in the L-P group was significantly downregulated, and the representative genes CD36 and PLIN2 were also significantly downregulated (Fig. [88]4A-B). CD36, as a key protein for the transport of fatty acids into cells, is closely associated with lipid accumulation [[89]36]. Meanwhile, PLIN2 is a lipid protein in cells, and its expression increases with the accumulation of lipid droplets [[90]37]. The expression of these two genes is closely related to the regulation of PPARγ [[91]38]. Consequently, we examined the expression of PGC1α, PPARγ, and CD36 in the spinal cord. The research findings demonstrated that the PGC1α/PPARγ pathway was activated after L-P treatment in the spinal cord, with significantly elevated expressions of PGC1α and PPARγ. Moreover, CD36, as a downstream molecule of PPARγ, had its expression significantly suppressed. Further analyses revealed that as the expressions of PGC1α and PPARγ in Iba + cells were upregulated, the expression of CD36 decreased accordingly (Fig. [92]4C-K). To sum up, L-P might regulate lipid metabolism in Iba + cells via the PGC1α/PPARγ pathway. Fig. 4. [93]Fig. 4 [94]Open in a new tab The regulatory pathways of N-Lactoyl-Phenylalanine screened based on transcriptome. (A) Gene Ontology (GO) enrichment analysis of transcriptome sequencing. (B) GO enrichment analysis and string diagrams. (C) Immunofluorescence expression of PGC1α (green), Iba1 (red) and DAPI (blue) in each group, scale bar: 50 μm. (D) Statistical analysis of PGC1α expression in each group. (E) Statistical analysis of PGC1α expression in Iba1 + cell in each group. (F) Immunofluorescence expression of Iba1 (green), PPARγ (red) and DAPI (blue) in each group, scale bar: 50 μm. (G) Statistical analysis of PPARγ expression in each group. (H) Statistical analysis of PPARγ expression in Iba1 + cell in each group. (I) Immunofluorescence expression of Iba1 (green), CD36 (red) and DAPI (blue) in each group, scale bar: 50 μm. (J) Statistical analysis of CD36 expression in each group. (K) Statistical analysis of CD36 expression in Iba1 + cell in each group. * represents P < 0.05. ** represents P < 0.01 vs. the SCI group. n.s. stands for no statistical significance. The data are represented as the mean ± SD (n = 6) N-Lactoyl-Phenylalanine regulates the lipid metabolism of Iba + cells via the AMPK/PGC1α/PPARγ pathway We initially acquired the binding poses and interactions of L-P and PGC1α through Schrodinger Maestro. The minimum binding energy of the interaction was − 4.483 kcal/mol, suggesting a relatively weak binding capacity between them. Which factors mediate the activation of the PGC1α/PPARγ pathway still requires further exploration (Fig. [95]5A-B). AMPK, as a crucial molecule for metabolic regulation, is an important upstream signal of PGC1α [[96]39]. Through Schrodinger Maestro, we obtained the binding poses and interactions of L-P and AMPKα1. The minimum binding energy of the interaction was − 5.883 kcal/mol, indicating a highly stable binding between them. According to the docking study, the interaction between L-P and AMPK-related proteins is mediated by strong electrostatic and hydrophobic effects, which encompass numerous potential active sites (ALA156 and ASP157) (Fig. [97]5C-D). Subsequently, WB verification demonstrated that L-P could effectively promote the phosphorylation of AMPK, thereby activating the AMPK/PGC1α/PPARγ pathway, downregulating PLIN2, and accelerating lipid metabolism (Fig. [98]5E-I). The data of inflammation-related indicators also revealed that the pro-inflammatory-related signals were significantly suppressed after L-P treatment, while the anti-inflammatory signals were significantly upregulated, and Iba + cells transformed to the M2 type (Fig. [99]5E, J-M). To sum up, The AMPK/PGC1α/PPARγ pathway is likely to be the major pathway through which L-P regulates the lipid metabolism of Iba + cells. Fig. 5. [100]Fig. 5 [101]Open in a new tab N-Lactoyl-Phenylalanine regulates the lipid metabolism of Iba + cells via the AMPK/PGC1α/PPARγ pathway. (A) The space-filling models and ribbon models of L-P in complex with PGC1α. (B) The 2D binding models between L-P and PGC1α complex. (C) The space-filling models and ribbon models of L-P in complex with AMPK. (D) The 2D binding models between L-P and AMPK complex. (E) Western blotting showing the expression of pAMPK, AMPK, PGC1α, PAPRγ, PLIN2, CD86, IL 1β, Arg1 and IL 10 in each group. (F-M) Quantitative analysis of pAMPK/AMPK, PGC1α, PAPRγ, PLIN2, CD86, IL 1β, Arg1 and IL 10 protein expression. * represents P < 0.05. ** represents P < 0.01 vs. the SCI group. The data are represented as the mean ± SD (n = 6) The regulation of the lipid metabolism of microglia/macrophage by N-Lactoyl-Phenylalanine through the AMPK/PGC1α/PPARγ pathway was verified by inhibiting key targets We utilized conditional knockout mice of AMPK in microglia to validate the in vivo pathway conditions. When AMPK in microglia was knocked out, the lipid metabolism-promoting effect of L-P was significantly inhibited. The results of Oil Red O staining and co-localization of Iba1 and PLIN2 demonstrated that a considerable amount of lipid accumulation still emerged in the AMPK^flox/flox; Tmem119^CreERT+/− group under the influence of L-P (Fig. [102]6A-B, F-G). The fluorescence co-localization detection results of the related pathways also indicated that PGC1α and PPARγ, which facilitate lipid metabolism, were significantly down-regulated in the AMPK^flox/flox; Tmem119^CreERT+/− group, while the expression level of CD36 significantly rose (Fig. [103]6C-E, H-J). Subsequently, to verify the specific cell types and lipid metabolism-related pathways involved in the action of L-P, we conducted further validation using mice with monocyte/macrophage & microglia lineage tracing genes and microglia lineage tracing genes. First, through immunofluorescence staining, we found that the Cx3cr1 fluorescent mice could completely include all lipid droplets, while the Tmem119 fluorescent mice had some omissions. Moreover, in both types of mice, L-P could significantly reduce lipid expression in positive cells (Fig. [104]7A-B). Next, we used flow cytometry to sort the tdTomato-positive cells from each group. The experimental results showed that the number of positive cells collected by flow cytometry in both types of mice was significantly reduced after L-P treatment (Fig. [105]7C-D). Then, we performed WB detection on the collected cells to analyze lipid metabolism indicators such as lipid generation, phagocytosis, storage, and β-oxidation, as well as inflammatory-related indicators in positive cells. The experimental results indicated that L-P could inhibit lipid generation, phagocytosis, and storage, but no significant effect was found on lipid β-oxidation. Additionally, L-P had similar therapeutic effects on monocytes/macrophages and microglia (Fig. [106]7E). To more precisely demonstrate that L-P also has lipid-regulating and anti-inflammatory effects on blood-derived macrophages, we further sorted Cx3cr1 tdTomato positive cells by flow cytometry, labeled them with the blood-derived macrophage-specific antibody Ly6c FITC to obtain the target cells (Sfigure [107]4 A), and then conducted lipid and inflammatory detection on Ly6c+ & tdTomato + cells. The experimental results indicated that L-P could effectively reduce lipid droplets in blood-derived macrophages and inhibit inflammatory expression (Sfigure [108]4B-D). Finally, we established a model of lipid overload in BV2 cells induced by NaOl and inhibited the AMPK/PGC1α/PPARγ pathway using the AMPK inhibitor Compound C and the PGC1α inhibitor SR-18,292. The WB results indicated that L-P could effectively enhance the expression of pAMPK in BV2 cells and activate the AMPK/PGC1α/PPARγ pathway, downregulating SREBP1, CD36 and PLIN2. When Compound C and PGC1α suppressed the expressions of pAMPK and PGC1α, the PPARγ signal was significantly inhibited, and the protein expression levels of SREBP1, CD36 and PLIN2 were upregulated. Furthermore, the β-oxidation-related indicator ACOX1 has not shown any significant changes in these regulatory processes. (Fig. [109]8A-H). Subsequently, upon analyzing the inflammation-related indicators, it was revealed that pro-inflammatory indicators such as CD86 and IL-1β were significantly inhibited following the action of L-P, while their expressions increased after the inhibition of pAMPK and PGC1α. Anti-inflammatory indicators such as ARG1 and IL-10 were significantly elevated by L-P, while their expressions were downregulated after the inhibition of pAMPK and PGC1α (Fig. [110]8A, I-L). Furthermore, we examined the expression of lipid droplets in BV2 cells of each group. The BODIPY staining results demonstrated that the lipid droplet clearance effect of L-P was significantly inhibited by Compound C and SR-18,292 (Fig. [111]8M-N). To sum up, L-P regulates the lipid metabolism of microglia/macrophage via the AMPK/PGC1α/PPARγ pathway. Fig. 6. [112]Fig. 6 [113]Open in a new tab Conditional knockout mice of AMPK in microglia verified that L-P regulates spinal cord lipid metabolism via the AMPK/PGC1α/PPARγ pathway. (A) Oli Red O staining of each group, scale bar: 1000 μm. (B) Immunofluorescence expression of Iba1 (green), PLIN2 (red) and DAPI (blue) in each group, scale bar: 50 μm. (C) Immunofluorescence expression of PGC1α (green), Iba1 (red) and DAPI (blue) in each group, scale bar: 50 μm. (D) Immunofluorescence expression of Iba1 (green), PPARγ (red) and DAPI (blue) in each group, scale bar: 50 μm. (E) Immunofluorescence expression of Iba1 (green), CD36 (red) and DAPI (blue) in each group, scale bar: 50 μm. (F) Statistical analysis of Lipid droplet area ratio in each group. (G, H, I, J) Statistical analysis of PLIN2, PGC1α, PPARγ and CD36 expression in Iba1 + cell in each group. ** represents P < 0.01 vs. the SCI group. The data are represented as the mean ± SD (n = 6) Fig. 7. [114]Fig. 7 [115]Open in a new tab Identification of the lipid metabolism effects of L-P on microglia and macrophages based on lineage tracing gene mice. (A) Immunofluorescence expression of PLIN2 (green), tdTomato (red, tracer mononuclear/macrophage cells) and DAPI (blue) in each group, scale bar: 50 μm. (B) Immunofluorescence expression of PLIN2 (green), tdTomato (red, tracer microglia) and DAPI (blue) in each group, scale bar: 50 μm. (C) Flow sort tdTomato-positive mononuclear/macrophage cells from each group. (D) Flow sort tdTomato-positive microglia from each group. (E) Western blotting showing the expression of PAPRγ, SREBP1, CD36, PLIN2, CD86, IL 1β, Arg1 and IL 10 in each group. ** represents P < 0.01 vs. the L-P group. The data are represented as the mean ± SD (n = 6) Fig. 8. [116]Fig. 8 [117]Open in a new tab The inhibition of AMPK and PGC1α was used to study the role of the AMPK/PGC1α/PPARγ pathway in lipid metabolism and inflammation. (A) Western blotting showing the expression of pAMPK, AMPK, PGC1α, PAPRγ, SREBP1, CD36, PLIN2, CD86, IL 1β, Arg1 and IL 10 in each group. (B-L) Quantitative analysis of pAMPK/AMPK, PGC1α, PAPRγ, SREBP1, CD36, PLIN2, CD86, IL 1β, Arg1 and IL 10 protein expression. (M) Statistical analysis of BODIPY493/503 expression in each group. (N) Immunofluorescence expression of BODIPY493/503 (green) and DPAI (blue) in each group, scale bar: 50 μm. * represents P < 0.05. ** represents P < 0.01 vs. the L-P group. The data are represented as the mean ± SD (n = 6) N-Lactoyl-Phenylalanine promotes the recovery in SCI mice We evaluated the recovery of spinal cord tissue in each group of mice through H&E staining and Nissl staining. The H&E staining results showed that the tissue of the L-P group mice recovered significantly compared to the SCI group, with a significant reduction in the area of tissue damage (Fig. [118]9A, E). The Nissl staining results indicated that more neurons survived at the edge of the injury area in the L-P group mice (Fig. [119]9B, F). Axons are the key structures for transmitting neural signals. Through GFAP and NF200 immunofluorescence staining, we found that the glial scar in the L-P group mice was looser than that in the SCI group, allowing more axons to penetrate the injury area (Fig. [120]9C, G). ACE tubulin and TYR tubulin are respectively the polymerized microtubules and depolymerized microtubules of axons. The co-staining results showed that L-P treatment could significantly promote microtubule stability and enhance the stability of axons (Fig. [121]9D, H). Finally, we examined the behavioral recovery of the mice. By marking the active joints of the mice and using the machine learning module for big data behavioral simulation analysis to process the movement videos of the mice, the visualized movement patterns of the hind limbs and the movements of each joint were obtained. The data results indicated that after L-P treatment, the hind limbs of the mice regained a certain degree of recyclable movement ability, and the movement trajectories of the joints were partially close to those of the Sham group (Fig. [122]10A). Through further detailed data analysis, it was concluded that the support capacity of the hind limbs of the mice in the L-P group was effectively improved, and the hind limbs could be lifted off the ground to a certain extent for cyclic movement (Fig. [123]10B-E). To sum up, L-P promotes the recovery in SCI mice by repairing tissues and protecting axons. Fig. 9. [124]Fig. 9 [125]Open in a new tab L-P promotes tissue repair and axonal protection in SCI mice. (A) H&E staining of each group, scale bar: 1000 μm. (B) Nissl staining of each group, scale bar: 1000 μm. (C) Immunofluorescence expression of NF200 (green), GFAP (red) and DAPI (blue) in each group, scale bar: 1000 μm. (D) Immunofluorescence expression of ACE tubulin (green), TYR tubulin (red) and DAPI (blue) in each group, scale bar: 1000 μm. (E) Statistical analysis of damage area in each group. (F) Statistical analysis of number of neurons in each group. (G) Statistical analysis of axonal penetration rate in each group. (F) Statistical analysis of ACE tubulin/TYR tubulin in each group. ** represents P < 0.01 vs. the SCI group. The data are represented as the mean ± SD (n = 6) Fig. 10. [126]Fig. 10 [127]Open in a new tab Figure 10. L-P promotes the recovery of motor function in SCI mice. (A) At 28 days after SCI, the hindlimb movement of mice was analyzed by visual bar graph and the joint movement trajectory was detected visually. (B) Maximum height of toes off the ground. (C) Maximum height of hips above the ground. (D) Maximum muscle strength. (E) Motor cycle recording. ** represents P < 0.01 vs. the SCI group. The data are represented as the mean ± SD (n = 6) Discussion The lipid accumulation in microglia/macrophage is one of the crucial factors contributing to the deterioration of the inflammatory microenvironment and the obstacles to repair following SCI [[128]40]. Post-SCI, the excessive activation of microglia/macrophage results in abnormal metabolic status, manifested as the accumulation of cholesterol and lipid droplets [[129]41]. Such lipid metabolic disorder not only keeps microglia/macrophage persistently in a pro-inflammatory state (M1 phenotype), secreting a large number of pro-inflammatory factors and intensifying neuroinflammation. Additionally, lipid accumulation hinders the transformation of microglia/macrophage to an anti-inflammatory and repair phenotype (M2 phenotype), restricting the processes of neural regeneration and vascular repair [[130]42]. In recent years, exercise intervention has been regarded as having a promoting effect on SCI repair, and its mechanism is associated with the release and regulatory role of exercise metabolites. Exercise metabolites are active substances secreted by tissues such as muscles and liver during exercise, including lactate, ketone bodies, N-lactoyl amino acids, etc [[131]43–[132]45]. These metabolites play significant roles in regulating energy metabolism, inflammatory responses, and cellular functions. Among them, ketogenesis-related treatments have undergone a series of studies in the direction of neuron protection [[133]46], but the regulation of exercise metabolites on the injured spinal cord microenvironment has not been explicitly reported. A stable spinal cord microenvironment is a hotbed for regeneration, and lipid-accumulated microglia/macrophage are the key factors hindering the remodeling of the spinal cord microenvironment [[134]47]. Therefore, in this study, based on the characteristics that exercise can promote lipid metabolism in microglia/macrophage and upregulate lactylation, we screened out the lactoylated exercise metabolite N-Lactoyl-Phenylalanine (L-P) for research. Based on the content difference of L-P detected in the exercise group and the SCI group, as well as the possible action pathways of L-P on SCI analyzed by network pharmacology, we hypothesized that L-P could exert therapeutic effects by regulating the lipid metabolism of microglia/macrophage, and further verified it through methods such as Oil Red O staining and immunofluorescence co-localization. The research findings indicate that L-P indeed can accelerate the lipid metabolism of Iba + cells after SCI. Subsequently, we examined the M1/M2 phenotype of Iba + cells and discovered that the effect of L-P could effectively facilitate the transformation of Iba + cells to M2. To verify the pathway through which L-P accelerates the lipid metabolism of Iba + cells, we conducted transcriptome sequencing and identified that CD36 and PLIN2, which are significantly downregulated downstream of the long-chain fatty acid metabolism signal, are closely associated with lipid accumulation and devouring, and located downstream of the PPARγ signal. Further studies suggest that the PGC1α/PPARγ signal might be the upstream regulator of CD36 and PLIN2 [[135]48, [136]49]. To confirm that this signaling pathway is the predominant functional pathway, we initially employed molecular docking to investigate the docking between L-P and PGC1α, yet the docking results were unsatisfactory. AMPK, as a key molecule in metabolic regulation, is an important upstream signal of PGC1α [[137]6]. Therefore, we examined the molecular docking between L-P and AMPKα1, and the results demonstrated a stable binding between them. Subsequently, the results of Western blot (WB) detection revealed that the AMPK/PGC1α/PPARγ pathway activated by L-P could effectively suppress the expression of PLIN2 and modulate the transformation of microglia/macrophages to M2. Subsequently, we used microglial AMPK conditional knockout mice to verify the upstream of L-P action and utilized lineage tracing mice to verify the cell types of L-P action and lipid metabolism regulation. The experimental results indicated that AMPK is upstream of L-P action and L-P can act on microglia and macrophages. It can regulate the lipid metabolism of microglia/macrophages by modulating lipid synthesis, storage and phagocytosis, thereby inhibiting inflammation. Then, we utilized AMPK inhibitor Compound C and PGC1α inhibitor SR-18,292 to inhibit the key targets of the pathway, and the results revealed that the regulatory effect of L-P was significantly inhibited, and the lipid metabolism rate in BV2 cells were significantly decreased. Finally, we evaluated the recovery of tissue architecture and motor function in mice treated with L-P, and the results indicated that L-P could promote the recovery in SCI mice by repairing tissues and protecting axons. At present, numerous metabolic regulators have been utilized in studies of SCI, including AMPK activators (such as metformin) and PPARγ agonists (such as rosiglitazone) [[138]50, [139]51]. Metformin regulates energy metabolism by activating AMPK and has extensive downstream effects, but it may lead to gastrointestinal discomfort or the risk of lactic acidosis [[140]52]. Rosiglitazone affects adipocyte differentiation and insulin sensitivity by activating PPARγ, but it may also cause side effects such as weight gain and an increased risk of cardiovascular events [[141]53]. In contrast to these synthetic drugs, L-P exists naturally in the body and has lower immunogenicity and toxic side effects, making it a more appropriate potential candidate drug for the treatment of SCI. In conclusion, L-P regulated the lipid metabolism of microglia/macrophage by activating the AMPK-PGC1α-PPARγ axis, significantly facilitating the repair of SCI mice. This finding not only deepens our knowledge of the relationship between the metabolism and function of microglia/macrophage but also offers a novel metabolic regulatory strategy for the treatment of SCI. Electronic supplementary material Below is the link to the electronic supplementary material. [142]Supplementary Material 1^ (1.5MB, pdf) [143]Supplementary Material 2^ (400KB, docx) Acknowledgements