Eighty-one female wild-type C57BL/6J mice (8-week-old, 20–30 g) were procured from Nantong University’s Animal Center (Nantong, Jiangsu, China). Specimens were maintained under controlled temperature/humidity with 12-hour photoperiods. Breeding and experimental procedures occurred at Nantong’s Animal Experiment Center under regulatory compliance. All protocols adhered to ARRIVE (2.0) guidelines with ethical approval from Nantong University’s Animal Experiment Ethics Committee (No. S20250318-004). A comprehensive chronological flowchart delineated the experimental design (Fig. 1B).

Anesthesia was induced in the mice via i.p. injection of ketamine (K2753, Sigma-Aldrich, St. Louis, MO, USA) (80 mg/kg). The spinal cord on the T10 vertebrae was struck with a percussion device (68099, RWD Life Science, Shenzhen, China) to produce spinal cord contusion. After impact, the mice still exhibited tail-flick reflexes, and spinal cord hemorrhage was observed in the central region. The mice were placed on a heating pad for recovery. Post-recovery, bilateral hindlimbs showed no ankle joint movement, corresponding to a Basso Mouse Scale (BMS) score of 0, indicating establishment of a successful SCI model. In the sham group, mice underwent laminectomy only, wherein the vertebral laminae were completely removed, followed by sequential suturing of the muscle layer, fascial layer, and skin. For treatment, mice in the SCI group received i.p. injections of POL (HY-N0033, MedChemExpress, Woodbridge, NJ, USA) at doses of 15 mg/kg and 30 mg/kg immediately after successful model establishment and after resuscitation. Subsequent injections were administered every 24 h for 7 consecutive days. Sham and SCI groups received i.p. injection of an equivalent volume of saline daily. To prevent urinary retention, SCI mice underwent daily assisted bladder expression until recovery of spontaneous voiding function.

Motor function recovery in SCI mice was assessed using the BMS, hindlimb reflex score, and footprint analysis. To minimize subjective bias, all evaluations were performed by two independent evaluators, blind to group identification, who were familiar with the experimental protocol but not directly involved in the study. (1) BMS score: hindlimb motor function was assessed using the BMS on days 1, 3, 7, 14, 21, and 28 after SCI (days after injury = dpi) [24]; (2) hindlimb reflex score: At 1, 3, 7, 14, 21, and 28 dpi, mic Subsequentlye were suspended at a height of approximately 30 cm for 14 sec. Hindlimb motor function was assessed using the following scoring criteria: 0, Normal; 1, Incomplete hindlimb extension; 2, Hindlimb clasping; 3, Hindlimb paralysis; (3) Footprint analysis: at 28 dpi, mice were tested for gait analysis using dye-marked limbs (red: forelimbs; blue: hindlimbs) on a paper-lined runway. Quantitative assessment of locomotor recovery included measurements of stride length and stride width from the footprints.

After being deeply anesthetized by intraperitoneal injection of ketamine at a dose of 80 mg/kg, mice were euthanized by carbon dioxide asphyxiation to ensure death prior to tissue collection. Subsequently, the abdomen was incised and the diaphragm was punctured to expose the heart. Cold phosphate-buffered saline (PBS, ST476, Beyotime, Shanghai, China) and 4% polyformaldehyde (P6148, Sigma-Aldrich) solution were sequentially injected into the heart. After the perfusion was completed, the damaged central spinal cord segment (1.5 cm in length) was removed and immediately immersed in 4% polyformaldehyde solution for preservation. Then, dehydration treatment was carried out and sections were made using a paraffin sectioning machine (Kedee KD-202A, Hisure, Jinhua, Zhejiang, China). Thin sections of 5 micrometers were cut longitudinally. For the immunofluorescence experiment, the preparation process of the tissue samples was the same as described above. Subsequently, the tissue was stained using the instructions provided by the Nissl staining kit (G1036, Servicebio, Wuhan, Hubei, China) and observed under a microscope (DM2500, Leica, Berlin, Germany). The specific parameters are as follows: Bright field, 4$\times$, exposure time 20 ms.

Hematoxylin-eosin (H&E) staining was executed employing an H&E staining kit (G1005, Servicebio). Paraffin-embedded spinal cord sections underwent xylene dewaxing (5 min), graduated ethanol rehydration, hematoxylin immersion (5 min), eosin counterstaining (2 min), and acid alcohol differentiation (10 sec) prior to dehydration and mounting. Section evaluation utilized a DM2500 microscope (Leica) under bright-field illumination (4$\times$ magnification; 10 ms exposure).

Luxol Fast blue (LFB) staining was performed using the LFB staining kit (G1092, Servicebio). For LFB myelin staining, sections were incubated in 0.1% Luxol Fast Blue for 2 h, differentiated in lithium carbonate for 10 sec until clear gray-white matter contrast was achieved, with microscope (DM2500, Leica) validation of myelin integrity (bright blue signal). The specific parameters are as follows: Bright field, 4$\times$, exposure time 50 ms.

On the 7th day after SCI, the mice in the SCI and SCI+POL (30 mg/kg) groups were euthanized, and their spinal cords removed. RNA was extracted from the spinal cords using TRIzol reagent (EF0131, YiFeiXue, Nanjing, Jiangsu, China). The cDNA library construction and RNA sequencing were carried out by GeneChem (Shanghai, China). Seven days post-SCI, euthanized SCI and SCI+POL (30 mg/kg) cohort mice underwent spinal cord extraction. RNA isolation employed TRIzol reagent (YiFeiXue). cDNA library preparation and RNA sequencing were executed by GeneChem. Total RNA (1 µg/sample) underwent library construction using the NEBNext® Ultra RNA Library Prep Kit (E7530, New England Biolabs, Ipswich, MA, USA). Poly(A)+ mRNA enrichment utilized oligo(dT)-immobilized magnetic beads, followed by fragmentation in NEBNext® First Strand Buffer (94 °C, 15 min). First-strand cDNA synthesis deployed random hexamers with M-MuLV Reverse Transcriptase; second-strand synthesis employed DNA Polymerase I. After end repair and 3^′ adenylation, NEBNext® adaptors were ligated. Size selection (250–300 bp) was performed using AMPure XP beads (A63880, Beckman Coulter, Brea, CA, USA), followed by USER enzyme treatment (37 °C, 15 min) and PCR amplification with Phusion polymerase. Final libraries were quality-controlled on an Agilent 2100 Bioanalyzer (G2939A, Agilent Technologies, Santa Clara, CA, USA). Transcript quantification derived from fragments per kilobase per million mapped reads. Differential gene expression screening implemented the DESeq2 algorithm. Genes exhibiting adjusted p-values 0.05 alongside fold changes exceeding 2 underwent classification as differentially expressed, proceeding to subsequent enrichment analyses.

Western blot (WB) employed a total protein extraction kit (KGP2100, KeyGEN BioTECH, Nanjing, Jiangsu, China) to isolate proteins from murine spinal cords and microglial cells separately. Protein concentration quantification utilized an enhanced BCA assay kit (P0012, Beyotime, Shanghai, China). Prior to electrophoresis, gel formulation deployed a polyacrylamide gel electrophoresis (PAGE) gel rapid preparation kit (PG211, Epizyme, Shanghai, China). Protein transfer onto polyvinylidene fluoride (PVDF) membranes (IPVH00010, Millipore, Billerica, MA, USA) occurred via rapid transfer solution (35 min duration). Post-transfer, PVDF membranes underwent ddH₂O rinsing followed by immersion in freshly constituted 5% skimmed milk (1160GR500, BioFroxx, Guangzhou, Guangdong, China) for 1.5 hours at ambient temperature to effect blocking. Subsequently, membranes underwent primary antibody incubation overnight at 4 °C, succeeded by secondary antibody co-incubation at room temperature for 1 hour the ensuing day. Antibody specifications appear in Table 1. Protein signal acquisition employed an automated chemiluminescence imaging platform (Tanon 5200, Tanon Science & Technology, Shanghai, China), with quantification executed via ImageJ software (Version 1.54r, NIH, Bethesda, MD, USA).

For the spinal cord paraffin sections, the first step is to deparaffinize and hydrate, then use the sodium citrate fixative solution (P0081, Beyotime) for thermal antigen retrieval, followed by adding the endogenous peroxidase strong blocking solution (P0100B, Beyotime) and incubating at room temperature for 30 min. For the cell slide sections, the spinal cord paraffin sections and cell slide sections start to be consistent from the following experimental steps. For immunofluorescence, specimens received immunostaining blocking solution (P0102, Beyotime) incubated at ambient temperature for 1 hour. Primary antibody application preceded overnight incubation at 4 °C. The subsequent day witnessed incubation with corresponding fluorescent secondary antibodies at ambient temperature for 1 hour (under light-restricted conditions). Antibody particulars are cataloged in Table 1. 4^′,6-Diamidino-2-phenylindole (DAPI, G1012, Servicebio) solution application ensued until complete specimen coverage, followed by 10-minute ambient incubation. Conclusively, anti-fluorescence quenching mounting medium (G1401, Servicebio) was applied before coverslip-mediated sample sealing. The sections were observed with a microscope (DM2500, Leica). The specific parameters are as follows: Fluorescence imaging, 20$\times$, DAPI: excitation filter: 360/40 nm, emission filter: 460/50 nm, exposure time: 80 ms; FITC: excitation filter: 480/30 nm, emission filter: 460/50 nm, exposure time: 150 ms; Tetramethylrhodamine isothiocyanate (TRITC): excitation filter: 560/25 nm, emission filter: 630/60 nm, exposure time: 220 ms. Then use the measurement tools in ImageJ to quantify the fluorescence intensity, and ultimately compare the fluorescence levels among different experimental groups.

BV-2 microglial cells (CL-0493, Wuhan Pricella Life Science & Technology, Wuhan, Hubei, China) were sustained in Dulbecco’s Modified Eagle Medium (DMEM; KGM12800, KeyGEN, Nanjing, Jiangsu, China) enriched with 10% fetal bovine serum (FBS; 10099141, Gibco, Waltham, MA, USA) and 1% Penicillin-Streptomycin (15140122, Gibco, Grand Island, NY, USA) under 37 °C/5% CO₂ conditions. The microglial BV-2 cell line was authenticated by short tandem repeat (STR) profiling. Mycoplasmic contamination was absent in the authenticated BV2 line by using both PCR-based and colorimetric assay methods. To establish a neuroinflammatory paradigm, cultures received 20 µM POL (HY-N0033, MedChemExpress, Woodbridge, NJ, USA) pretreatment for 24 hours preceding 1 µg/mL lipopolysaccharide (LPS; L2880, Sigma-Aldrich) stimulation for 6 hours.

Cellular viability underwent quantification via Cell Counting Kit-8 (CCK-8; C6005, NCM Biotech, Suzhou, Jiangsu, China). BV-2 progeny were subjected to 5 µM, 10 µM, 20 µM, 40 µM, and 80 µM POL exposure for 6 hours, succeeded by CCK-8 incubation (10 µL, 2 hours). Absorbance measurements (450 nm) utilized a Biotek microplate reader (Synergy H1, Agilent BioTek, Winooski, VT, USA), with untreated specimens representing 100% viability controls.

The 3D structure of Poliumoside was obtained from PubChem in structured data file (SDF) format and converted to mol2 format using OpenBabel (Version 2.4.1, Open Babel development community). Protein targets (PI3K, AKT, and mTOR) were retrieved from the UniProt database, and their protein data bank (PDB) files were processed in PyMOL (Version 2.5.x, Schrödinger, LLC, New York, NY, USA) to remove water molecules and ligands. Molecular docking was performed using AutoDock Vina (v1.5.6, Olson Lab, La Jolla, CA, USA), where protein structures were hydrogenated, and the ligand (POL) was optimized by assigning charges and rotatable bonds. Molecular docking positioned the search box centrally within the active site. Optimal binding conformations were selected according to docking scores (kcal/mol) and visualized via PyMOL. We considered the target proteins (PI3K, AKT, and mTOR) to exhibit strong binding affinity with Poliumoside when the binding energy was –7 kcal/mol.

Statistical analyses deployed GraphPad Prism (Version 10.4.1, GraphPad Software, San Diego, CA, USA). Data represent mean $\pm$ SD from $\ge$3 autonomous replicates. Intergroup comparisons employed one-way ANOVA with Tukey’s post hoc analysis ($\ge$3 cohorts). Statistical significance threshold: *p 0.05.

Histopathological interrogation of principal organs (brain, spinal cord, heart, liver, spleen, lung, kidney) via H&E staining. Murine subjects received POL (15 or 30 mg/kg) across seven days. Treated cohorts manifested no significant pathological deviations relative to controls, indicating absent POL-induced organic toxicity at examined dosages (Fig. 1C).

To evaluate the therapeutic effects of POL treatment on motor function recovery in SCI mice, we used BMS scoring, hindlimb reflex assessment, and footprint analysis. POL-treated SCI mice demonstrated significantly higher BMS scores than did untreated SCI controls, with particularly notable improvements at 21 and 28 dpi (p 0.05) (Fig. 2A). At 28 dpi, these cohorts additionally manifested diminished hindlimb reflex scores (p 0.05) (Fig. 2B). Plantar imprint analysis disclosed SCI mice administered 15 mg/kg or 30 mg/kg POL exhibited greater anteroposterior stride dimensions and lateral base spans versus SCI counterparts (Fig. 2C), cumulatively denoting augmented locomotor restitution. Particularly, stride magnitude in SCI+POL (30 mg/kg) specimens substantially exceeded SCI group measurements (p = 0.0021) (Fig. 2D,E). To evaluate POL’s neuropreservative capacities, histological examinations (H&E, Nissl, LFB staining) were conducted at 28 dpi. H&E analysis demonstrated POL administration significantly contracted neural lesion areas dose-dependently (p = 0.005) (Fig. 2F,G). LFB staining confirmed that POL treatment attenuated demyelination at the lesion site, particularly at the 30 mg/kg dose (p = 0.0005) (Fig. 2H,I). Nissl staining revealed that while SCI caused extensive neuronal death in and around the injured area, the 30 mg/kg POL treatment preserved significantly more surviving neurons (p = 0.0016) (Fig. 2J,K). Based on these findings, we selected the dose of 30 mg/kg POL for subsequent experiments.

Fig. 2.

POL treatment alleviated pathological changes and improved motor function after SCI. (A) BMS metrics within 28 dpi across sham, SCI, POL-treated SCI cohorts. (B) Hindlimb reflex metrics at 28 dpi. (C) Exemplar plantar imprints at 28 dpi. Quantitative stride length (D) and base width (E) analyses. (F) H&E-stained transverse sections (28 dpi). Scale bars = 200 µm, 20 µm. (G) Lesion area quantification. (H) LFB-stained longitudinal sections depicting myelination (28 dpi). Scale bars = 200 µm, 20 µm. (I) Demyelinated area quantification. (J) Nissl-stained longitudinal sections (28 dpi). Scale bars = 200 µm, 20 µm. (K) Surviving neuron quantification. Data: mean $\pm$ SEM; N = 6/group (3 experiments). Statistical determinations deployed one-way ANOVA with Tukey’s post hoc analysis. *p 0.05 vs SCI, **p 0.01, ***p 0.001, ****p 0.0001, ns = non-significant.

To further elucidate POL’s mechanistic action in SCI intervention, transcriptomic interrogation was conducted. Sequencing outcomes delineated 28,940 differentially expressed genes distinguishing SCI cohorts from SCI+POL counterparts, comprising 1194 substantially upregulated and 1966 markedly downregulated genes (p $\le$ 0.05, $|$log₂FoldChange$|$ $\ge$0; Fig. 3A). The volcano plot analysis identified three significantly downregulated genes, with Scn10a (p = 9.68 $\times$ 10^-12) and Mrgprd (p = 4.83 $\times$ 10^-9) showing the most pronounced downregulation, both of which are functionally associated with neuropathic pain. Additionally, Igf1r (p = 0.000441), a target gene implicated in axonal regeneration, was also significantly downregulated. Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis prominently implicated inflammation-associated signaling cascades, notably the PI3K/AKT/mTOR axis, ras-proximate 1 (Rap1) signaling cascade, and Ras signaling pathway (Fig. 3B). Gene ontology (GO) enrichment evaluation postulated POL’s principal biological functionalities pertain to axonal reconstitution post-spinal cord injury (Fig. 3C). Reactome enrichment scrutiny further substantiated POL’s predominant involvement in neuronal system-relevant biological processes following SCI (Fig. 3D). Consequently, transcriptomic evidence directed subsequent investigation toward POL’s effects on neuroinflammation, axonal regeneration, and PI3K/AKT/mTOR signaling modulation.

Fig. 3.

POL treatment modulated gene expression and regulated axon regeneration-associated pathways after SCI. (A) Differential expression volcano map between SCI and SCI+POL cohorts. (B) KEGG pathway enrichment profiling. (C) Gene Ontology enrichment assessment. (D) Reactome pathway enrichment analysis. KEGG, Kyoto encyclopedia of genes and genomes. In panels B, C, and D, the red boxes highlight the KEGG pathway, GO term, and Reactome pathway, respectively, that were highly enriched and most directly related to the axon and neuroinflammation under investigation. In the three parentheses in the part of the figure references of the results, we have also added “red box” accordingly.

Following spinal cord injury, axonal reconstitution and myelin sheath remodeling constitute pivotal processes for neural function restoration. Our investigation documented significantly elevated neurofilament 200 (NF200) axonal density within lesion areas of POL-treated specimens versus SCI controls through immunofluorescence, indicating enhanced axonal regeneration (p 0.0001; Fig. 4A,B). Concurrently, myelin basic protein (MBP) myelin-positive regions expanded substantially in POL cohorts relative to SCI groups (p 0.0001; Fig. 4C,D). Growth-associated protein 43 (GAP43)—a neuron-specific phosphoprotein abundantly localized in axonal growth cones—mediates axonal development, guidance, and synaptic plasticity. WB analysis confirmed POL intervention restored GAP43 protein expression post-SCI, further evidencing axonal regeneration (p = 0.0427; Fig. 4E,F). For all original WB images in Fig. 4E see the Supplementary material. Post-traumatic glial scarring presents a primary physiological barrier to axonal regeneration [25]. We consequently evaluated glial scar formation via glial fibrillary acidic protein (GFAP) and ionized calcium-binding adapter molecule 1 (IBA1) immunofluorescence in spinal tissues at 7/28 dpi. SCI controls exhibited dense microglial and astrocytic accumulation within lesion epicenters. POL administration markedly attenuated microglial and astrocytic scar formation in injury cores, with superior efficacy at higher doses (p 0.0001; Fig. 4G–L). Collectively, these outcomes demonstrate POL treatment suppresses post-SCI glial cicatrization while facilitating axonal and myelin regeneration.

Fig. 4.

POL treatment enhanced axonal regeneration and myelin repair while reducing glial scar after SCI. (A) NF200 axonal distribution (red) in longitudinal sections (7 dpi). Scale bar = 50 µm. (B) NF200 fluorescence quantification. (C) MBP myelination (red; 7 dpi). Scale bar = 50 µm. (D) MBP fluorescence quantification. (E) Representative GAP43 WB (3 dpi). (F) GAP43 expression quantification. (G,J) Dual IBA1 microglia (green)/GFAP astrocytes (red) immunofluorescence at 28 dpi (H) and 7 dpi (K). Scale bar = 50 µm. (H,K) IBA1 fluorescence quantification. (I,L) GFAP fluorescence quantification. Blot data: mean $\pm$ SEM; N = 3/group (3 experiments). Immunofluorescence data: mean $\pm$ SEM; N = 6/group (3 experiments). Statistics: one-way ANOVA + Tukey’s post hoc. *p 0.05 vs. SCI group, **p 0.01, ****p 0.0001. NF200, Neurofilament 200; GAP43, Growth-associated protein 43.

Following spinal cord injury, activated microglia propagate inflammatory mediators iNOS and COX-2 while secreting cytokines TNF-$\alpha$, IL-1$\beta$, and IL-6. WB quantification revealed elevated protein expression of iNOS, COX-2, TNF-$\alpha$, IL-1$\beta$, and IL-6 post-SCI. Poliumoside intervention substantially attenuated all five inflammatory markers, demonstrating suppression of post-traumatic neuroinflammation (iNOS: p = 0.0003; COX-2: p = 0.0019; IL-1$\beta$: p 0.0001; IL-6: p 0.0001; TNF-$\alpha$: p = 0.0004; Fig. 5A–F). For all original WB images in Fig. 5A see the Supplementary material. Immunofluorescence corroborated POL’s capacity to mitigate microglia-mediated neuroinflammation. Co-staining of microglial marker IBA1 with iNOS demonstrated reduced iNOS fluorescence intensity in IBA1-positive cells following POL administration versus injury controls (p 0.0001; Fig. 5G,H). Additional co-localization studies of TNF-$\alpha$/IL-1$\beta$ and IL-6/COX-2 confirmed significantly diminished fluorescence intensities for all four markers in POL-treated cohorts relative to SCI groups (IL-1$\beta$: p 0.0001; TNF-$\alpha$: p 0.0001; IL-6: p 0.0001; COX-2: p 0.0001; Fig. 5I–N). These collective findings establish POL’s efficacy in alleviating microglial-driven neuroinflammatory responses post-SCI.

Fig. 5.

POL treatment attenuated microglial-mediated neuroinflammation after SCI. (A) Representative iNOS, COX-2, IL-1$\beta$, IL-6, TNF-$\alpha$ WB (3 dpi). (B) iNOS quantification. (C) COX-2 quantification. (D) IL-1$\beta$ quantification. (E) IL-6 quantification. (F) TNF-$\alpha$ quantification. (G) IBA1 (green)/iNOS (red) co-staining (7 dpi). Scale bar = 50 µm. (H) iNOS fluorescence quantification. (I) IL-1$\beta$ (green)/TNF-$\alpha$ (red) co-staining (7 dpi). Scale bar = 50 µm. (J) IL-1$\beta$ fluorescence quantification. (K) TNF-$\alpha$ fluorescence quantification. (L) IL-6 (green)/COX-2 (red) co-staining (7 dpi). Scale bar = 50 µm. (M) IL-6 fluorescence quantification. (N) COX-2 fluorescence quantification. Blot data: mean $\pm$ SEM; N = 3/group (3 experiments). IF data: mean $\pm$ SEM; N = 6/group (3 experiments). Statistics: one-way ANOVA + Tukey’s post hoc. **p 0.01 vs. SCI group, ***p 0.001, ****p 0.0001.

Compared to the injury cohort, POL administration post-SCI significantly mitigated levels of the oxidative stress-associated proteins NOX2 and NOX4 (NOX2: p = 0.0406; NOX4: p = 0.0128) (Fig. 6A–C). For all original WB images in Fig. 6A see the Supplementary material. The microglial biomarker IBA1 underwent concomitant staining with NOX4. Immunofluorescence analysis revealed POL substantially diminished NOX4 fluorescence intensity co-localized with IBA1 relative to injured controls (NOX4: p 0.0001) (Fig. 6D,E). These observations indicate POL treatment alleviated microglial oxidative stress following spinal cord injury.

Fig. 6.

POL treatment attenuated microglial-mediated oxidative stress after SCI. (A) Representative WB depict NOX2 and NOX4 levels within the spinal cord at 3 dpi. (B) Quantitative assessment of NOX2 expression. (C) Quantitative assessment of NOX4 expression. (D) Dual immunofluorescence labeling visualized microglia via IBA1 (green) and NOX4 (red) at 7 dpi. Scale bar = 50 µm. (E) Quantification of NOX4 fluorescence intensity. WB quantification reflects mean $\pm$ SEM derived from 3 independent experiments, each utilizing 3 mice per group (N = 3 per group). Immunofluorescence quantification presents mean $\pm$ SEM from 3 independent experiments, each employing 6 mice per group (N = 6 per group). Statistical significance was ascertained utilizing one-way ANOVA succeeded by Tukey’s post hoc test. *p 0.05 vs. SCI group, *p 0.05, **p 0.01, ****p 0.0001.

BV-2 microglial cellular viability under varying POL dosages underwent assessment via CCK-8 methodology. Incubation with POL exclusively (5, 10, 20, 40, 80 µM) for 24 hours manifested no viability alterations (Fig. 7A). This investigation consequently elected 20 µM POL for ensuing cellular assays. LPS activation of BV-2 microglia in vitro replicated post-spinal cord injury neuroinflammatory cascades. WB analyses revealed LPS markedly amplified iNOS and COX-2 protein quantities, which POL administration substantially attenuated (iNOS: p = 0.0388; COX-2: p = 0.0207; Fig. 7B–D). For all original Western blot images in Fig. 7B see the Supplementary material. Immunofluorescence imaging demonstrated inflammatory markers iNOS and COX-2 intensification 6 hours post-LPS provocation. Following 20 µM POL intervention, LPS-induced iNOS levels declined (Fig. 7E), with COX-2 exhibiting analogous diminishment (Fig. 7F). These outcomes validate POL’s neuroinflammation-inhibiting efficacy within microglia.

Fig. 7.

POL treatment suppressed LPS-induced neuroinflammation in BV-2 microglial cells. (A) CCK-8 assay quantifying cellular viability. (B) Representative WB depicting iNOS/COX-2 levels in BV-2 microglia after 24-hour POL exposure and 6-hour LPS stimulation. (C) iNOS expression quantitative evaluation. (D) COX-2 expression quantitative evaluation. (E) Representative immunofluorescence visualizing iNOS (green) in BV-2 microglia post-treatment. Scale bar = 50 µm. (F) Representative immunofluorescence visualizing COX-2 (green) under identical conditions. Scale bar = 50 µm. CCK-8 data: mean $\pm$ SEM from triplicate independent experiments, each encompassing six BV-2 microglial cells per cohort (N = 6/group). WB quantifications: mean $\pm$ SEM from triplicate experiments, each comprising three BV-2 microglial cells per cohort (N = 3/group). Statistical significances were ascertained via one-way ANOVA with Tukey’s post hoc analysis. *p 0.05 vs. SCI group, *p 0.05, **p 0.01. LPS, Lipopolysaccharide; CCK-8, Cell Counting Kit-8.

To delineate the precise molecular actions of POL against SCI, guided by RNA-Seq indications, molecular docking assessed POL’s binding capacity towards three pivotal targets: PI3K, AKT, and mTOR, followed by evaluation of the PI3K/AKT/mTOR cascade both in vivo and in vitro. Docking simulations revealed substantial binding affinities between POL and core PI3K/AKT/mTOR constituents, exhibiting energies of –7.5 kcal/mol (PI3K), –8.8 kcal/mol (AKT), and –9.1 kcal/mol (mTOR), all surpassing the established threshold for potent molecular binding (–7 kcal/mol) (Fig. 8A). Immunoblot analyses of spinal cord tissue demonstrated markedly diminished p-PI3K/PI3K, p-AKT/AKT, and p-mTOR/mTOR ratios at 3 dpi. Conversely, POL administration elevated these phosphorylated protein ratios significantly (p-PI3K/PI3K: p = 0.0021; p-AKT/AKT: p = 0.0013; p-mTOR/mTOR: p = 0.0005) (Fig. 8B–E). Analogously, within the microglial model, LPS challenge substantially reduced p-PI3K/PI3K, p-AKT/AKT, and p-mTOR/mTOR levels. POL treatment conversely augmented these ratios (p-PI3K/PI3K: p = 0.0017; p-AKT/AKT: p = 0.0002; p-mTOR/mTOR: p = 0.0355) (Fig. 8F–I). For all original WB images in Fig. 8B,F see the Supplementary material. These findings imply POL’s activation of the PI3K/AKT/mTOR signaling cascade post-SCI constitutes its principal therapeutic mechanism.

Fig. 8.

POL treatment activated the PI3K/AKT/mTOR signaling pathway after SCI. (A) Molecular docking models depicting POL’s binding interactions with PI3K, AKT, and mTOR. (B) Representative WB of phosphorylated and total PI3K/AKT/mTOR proteins in spinal cord tissue at 3 dpi. (C) Quantification of p-PI3K/PI3K expression levels. (D) Quantification of p-AKT/AKT expression levels. (E) Quantification of p-mTOR/mTOR expression levels. (F) Representative immunoblots depicting phosphorylation states and corresponding total PI3K, AKT, and mTOR proteins in BV-2 microglial cells following 24-hour POL administration and 6-hour LPS exposure. (G) Quantification of p-PI3K/PI3K expression levels. (H) Quantification of p-AKT/AKT expression levels. (I) Quantification of p-mTOR/mTOR expression levels; Findings are presented as mean $\pm$ SEM derived from three distinct experimental replicates, each encompassing three mice or BV-2 microglial cell samples per cohort (N = 3 per group). Statistical significance was assessed using one-way ANOVA with Tukey’s post hoc analysis. *p 0.05 vs. SCI group, *p 0.05, **p 0.01, ***p 0.001, ****p 0.0001. PI3K, phosphatidylinositol 3-kinase; AKT, protein kinase B; mTOR, mechanistic target of rapamycin.