Does ginsenoside Rg1 promote intervertebral disc repair? An experimental study insights into ferroptosis mechanism

Ginsenoside Rg1 (Selleck, S3923, Purity: 99.18%), CCK8 kit (SAB, CP002), Reactive oxygen detection kit (Beyotime, S0033), DCFH-DA (Beyotime, China), FerroOrange (Dojindo, Japan), Malondialdehyde (MDA), Superoxide Dismutase (SOD), Glutathione Peroxidase (GSH-Px) assay kit, (A003-1-2, A001-3-2, A006-2-1, Nanjing Jiancheng Technology Co., Ltd, China). RIPA lysis buffer (Thermo Scientific™, USA), The antibodies mentioned in the experiments included: Collagen II (abcam, ab34712), MMP-3 (abcam, ab52915), Nuclear factor erythroid 2-related factor 2 (NRF2, CST, 20733), Glutathione peroxidase 4 (GPX4, abcam, ab125066), Ferritin light chain 1 (FTL1, abcam, ab69090), Solute carrier family 7 member 11 (SLC7A11, abcam, ab175186), GAPDH (Proteintech, 60004-1-Ig), goat anti-rabbit IgG H&L (Beijing Zhong Shan-Golden Bridge Biological Technology Co.,Ltd, ZB-2301), goat anti-rat IgG H&L (Zhong Shan-Golden Bridge Biological Technology CO.,LTD, ZB-2305) ECL chemiluminescence reagent kit (Millipore, USA), NRF2 inhibitor ML385 (Selleck, S8790).

This study was conducted in accordance with ethical guidelines and received approval from the Medical Ethics Committee of the Qingpu Branch of Zhongshan Hospital affiliated with Fudan University (approval code: qingyi2024-62). All patients provided informed consent. Nucleus pulposus specimens were collected from 9 patients who underwent surgical procedures between 2023 and 2024. The cohort included 5 females and 4 males, aged between 50 and 80 years. Prior to surgery, all patients underwent T2-weighted MRI scans to assess the degree of disc degeneration using the Pfirrmann grading system. Table 1 provides detailed information about the specimens utilized in this study.

Forty-eight male Sprague-Dawley rats (8 week) were applied for in vivo evaluation of ginsenoside Rg1. All animal protocols were approved by the Department of Animal Welfare and Ethics, Experimental Animal Science, Fudan University (approval code: 20240124 S). The rats were randomly divided into four groups: the control group, the IVDD group (receiving saline solution daily), the IVDD + ginsenoside Rg1-L group (receiving 20 mg/kg/day of Rg1), the IVDD + ginsenoside Rg1-H group (receiving 40 mg/kg/day of Rg1), with 12 rats in each group. The 36 SD rats were anesthetized by intraperitoneal injection of phenobarbital sodium (5 mg/100 g). After disinfection with iodinated polyvinyl pyrrolidone, a dorsal longitudinal skin incision was made. The Co7/8 disc was identified as the target segment. After fixation and disinfection, a longitudinal skin incision was made, and a syringe needle was inserted vertically (4–5 mm depth), rotated 360°, held for 30s, then removed before wound closure. Control rats received no puncture. This model effectively mimics human IVDD progression, inducing sustained severe degeneration within an appropriate timeframe while replicating the human degenerative cascade [23].

Intervertebral disc tissues were collected and stored in 10% neutral buffered formalin before being embedded in paraffin. Using a microtome, the paraffin-embedded specimens were sectioned into 5 μm slices. These sections were processed according to standard protocols and stained with hematoxylin and eosin (H&E) for histopathological analysis and safranin O solid green for the assessment of intervertebral disc morphology.

Total RNA was extracted using the Solarbio RNA extraction kit, followed by cDNA synthesis with the TransGen Biotech reverse transcription kit. Gene mRNA levels were quantified via qRT-PCR (TOLOBIO SYBR Master Mix) using primers listed in Table S1 (BLAST-verified). Data were normalized to GAPDH and analyzed by the 2^(−ΔΔCT) method.

Total proteins were extracted from intervertebral disc NP tissues using RIPA buffer, separated by SDS-PAGE, and transferred to PVDF membranes. After blocking with 5% skimmed milk, membranes were incubated overnight at 4 °C with primary antibody, followed by 1 h secondary antibody incubation at room temperature. Protein bands were detected using ECL and analyzed with Image J.

Immunofluorescence analysis was performed to investigate the localization and expression of target proteins in both cells (n = 3 per group) and IVDD tissue samples (n = 3 per group). NP tissue and cell sections were immunostained with Collagen II, MMP3, and NRF2 antibodies, incubated overnight at 4 °C, then treated with fluorescent secondary antibody (1 h, dark). After DAPI counterstaining, images were captured using a Nikon ECLIPSE Ni microscope. Images were captured and analyzed using image analysis software to quantify the fluorescence intensity. The relative expression levels and subcellular localization of the target proteins were compared between IVDD and control groups. Statistical analysis was performed to determine significant differences in protein expression and localization, with appropriate statistical tests employed to ensure rigor and reproducibility.

Serum was collected from rat abdominal aorta blood, centrifuged (4 °C, 10 min), and analyzed for MDA, SOD, and GSH-Px levels using ELISA kits according to the manufacturer's protocol.

Under anesthesia, rat tail X-rays were acquired (RADspeed M, Japan; 250 mA, 50 kV, 20ms) in the sagittal plane. Intervertebral disc height was measured using ImageJ and normalized to adjacent vertebrae to calculate the disc height index (DHI). MRI (GE 1.5T) with T2-weighted imaging (TR 3500ms, TE 120ms) assessed NP hydration. NP contours were delineated based on high signal intensity and analyzed using ImageJ. Four blinded researchers classified the disc images into five grades using the Pfirrmann classification system.

First, the target of Rg1 was found in PubChem database (https://pubchem.ncbi.nlm.nih.gov/↗) and the 2D structure was downloaded. The corresponding targets were retrieved from Swiss target prediction database (http://www.swisstargetprediction.ch/↗). GeneCards (https://www.genecards.org/↗) was selected as databases for disease target retrieval. Ferroptosis-related targets were screened from the FerrDB database (https://www.zhounan.org/ferrdb/↗). The retrieval results were combined and deduplicated. The overlapping genes of herbs and diseases were then used for the construction of the protein-protein interaction (PPI) network. Finally, import the PPI network into Cytoscape 3.7.2 software for visualization, and use CytoNCA plugin to analyze the preliminary network topology parameters to screen out core targets.

Transcriptomic analysis was conducted by Wuhan Saiweier Biotechnology Co., Ltd. NP tissues were randomly selected from rats in the Control, IVDD and Rg1-H groups (n = 3). The process involved enriching mRNA with polyA structure in total RNA using Oligo (DT) magnetic beads. Subsequently, the first strand of cDNA was synthesized in the M-MuLV reverse transcriptase system. The double-stranded cDNA was purified, subjected to end repair, add A-tail addition, and sequencing adapter ligation. Following further purification, PCR amplification was performed. Differential gene expression analysis was conducted using DESeq, with differentially expressed genes identified based on the criteria of expression fold change (|log2FoldChange| >1) and statistical significance (p value < 0.05).

The 2D structures of active compounds were obtained from PubChem (https://pubchem.ncbi.nlm.nih.gov/↗), while core targets (Homo sapiens) were retrieved from the PDB database (https://www.rcsb.org/↗), selecting high-resolution crystal structures with original ligands. Molecular docking was performed using Autodock, with results visualized and analyzed in PyMOL to generate interaction models.

NP tissues were isolated from caudal IVDs (Co1-Co6) of euthanized 8-week-old SD rats (pentobarbital sodium, 50 mg/kg). After collagenase II digestion (2 mg/ml, 37 °C, 2 h), cells were centrifuged, resuspended, and cultured in DMEM with 15% FBS.

Log-phase NP cells were trypsinized, counted (5 × 10⁴ cells/ml), and seeded in 96-well plates (5 × 10³ cells/well, triplicates). The plates were cultured at 37℃ overnight. Hydrogen peroxide (H₂O₂) can induce an IVDD model in human NP cells. H₂O₂ rapidly promoted ROS production and DNA damage in NP cells of human [24]. The screening drug concentrations for H₂O₂ were 10µM, 50µM, 100µM, 200µM, and 400µM. The screening drug concentrations for ginsenoside Rg1 were 10µM, 50µM, 100µM, 200µM, 400µM and 800µM. The plates were cultured for 48 h, after which a fresh culture medium containing a 10% Cell Counting Kit-8 (CCK-8) solution was added to the cells and incubated at 37℃ for 1 h. The optical density 450 nm was detected using a multimode microplate reader (Beijing Stawo Technology Co., Ltd, China).

To assess ferroptosis, log-phase NP cells (5 × 10⁵ cells/ml) were seeded in 6-well plates. After 24 h adhesion, cells were stained with 10 µM DCFH-DA or 5 µM C11 BODIPY (20 min, dark), then analyzed by flow cytometry (480/525 nm excitation/emission). ROS-positive cells showed FITC-channel fluorescence.

NP cells were immunostained for Collagen II, MMP3, and GPX4. After PBS washing and 4% formaldehyde fixation (30 min), cells were permeabilized (0.5% Triton X-100, 10 min), incubated with primary antibodies (4 °C, overnight) and fluorescent secondary antibodies (1 h, dark), then counterstained with DAPI. Images were captured using a Nikon ECLIPSE Ni microscope.

To observe the morphology of mitochondria, the cells were rapidly fixed with an electron microscope fixative solution at 4 °C for 2 to 4 h. After washing with 0.1 M PBS, the cells were fixed with a 2.5% glutaraldehyde fixed solution at room temperature for 2 h. Following a gradient dehydration process, the cells were embedded, sectioned, and then stained with 3% uranyl acetate and lead citrate for 15 min. Transmission electron microscopy (Hitachi SU8100) was utilized for image acquisition and analysis.

ChIP assay was undertaken using EZ ChIP Chromatin Immunoprecipitation Kit (Millipore, Billerica, MD, USA) following the manufacturer's protocols. Briefly, 2 × 10⁷ cells were used for each individual reaction. The rat NP cells were fixed by 1% formaldehyde and lysed. The genome was sonicated 200 ~ 1000 bp DNA fragments. Ten microgram antibodies specific for NRF2 (CST, #20733, USA) were added into the cell lysate for incubation at 4 ℃ overnight, than incubated with Protein A Agarose/Salmon Sperm DNA beads (50% Slurry). The input DNA and immunoprecipitated DNA was purified and detected by qPCR using the primer sequences (Supplementary materials 1).

All data were processed with Paragraph Prism 9.0.0 (GraphPad Software, La Jolla, CA) software. All data were represented as mean ± standard deviation (SD) and p < 0.05 was considered statistically different. Statistical analysis was performed by one-way analysis of variance (ANOVA) tests. Each experiment was repeated at least three times independently to confirm the consistency of the findings. The level of statistical significance was set at p < 0.05 for all analyses. Significance levels in the figures were denoted as follows: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****). Non-significant results were denoted as "ns".

To investigate ferroptosis in IVDD, NP tissues from 9 patients across Pfirrmann grades were analyzed. MRI (Fig. 1A-B) and histopathology (Fig. 1C) revealed degenerative changes. qRT-PCR and WB showed progressive downregulation of GPX4, FTL1, and SLC7A11 with increasing degeneration (Fig. 1D-F), demonstrating ferroptosis' critical role in human IVDD progression.

A detailed flowchart of the in vivo assays was shown in Fig. 2A. Ginsenoside Rg1 treatment significantly reduced MDA levels (Fig. 2B) and restored SOD/GSH activity (Fig. 2C, D) in IVDD rats. Calculation of the disc height index (DHI) showed that the degenerated intervertebral disc height of IVDD group decreased over 70%. Compared with the control group, and the ginsenoside Rg1 group had less disc height loss, restored the height to about 50% of the control group (Fig. 2E, G). The MRI results showed that the nucleus pulposus and endplates exhibited a black color in the IVDD group, while the Pfirrmann grade was elevated. In contrast to these findings, the Pfirrmann grade decreased in the ginsenoside Rg1-H and ginsenoside Rg1-L groups, the endplate gradually became white, and the signal gradually recovered (Fig. 2F, H). The results of HE staining and safranin O solid green staining revealed a significant reduction in the presence of NP tissue within the IVDD group, accompanied by a pronounced severity of fibrosis. However, the intervertebral disc nucleus pulposus in the ginsenoside Rg1 intervention group was relatively full (Fig. 2I, J). Immunohistochemistry demonstrated Rg1 upregulated Collagen II while downregulating MMP3 (Fig. 2K-M), confirming its protective effect against IVDD progression. Thus, ginsenoside Rg1 can inhibit IVDD progression in the rat tail puncture model.

Through network pharmacology analysis, 107 potential Rg1 targets were identified (Fig. 3A), with 24 overlapping targets between IVDD, ferroptosis, and Rg1 (Fig. 3B). PPI network analysis (median Degree = 19, BC = 4.11, CC = 0.85) revealed 14 core targets including NRF2 (Fig. 3C). Transcriptome sequencing confirmed differential expression of ferroptosis-related genes (NRF2, GPX4, SLC7A11, FTL1) (Fig. 3D-F). GSEA showed ferroptosis genes were predominantly expressed in the IVDD group (Fig. 3G).

Integrated network pharmacology and transcriptomics identified NRF2 as the key target through which ginsenoside Rg1 may ameliorate IVDD by regulating ferroptosis. Molecular docking revealed strong binding affinities between Rg1 and ferroptosis-related proteins: NRF2 (− 7.4 kcal/mol), GPX4 (− 5.6 kcal/mol), SLC7A11 (− 9.1 kcal/mol), and FTL1 (− 8.5 kcal/mol) (Fig. 3H) (Table S2). PCR analysis demonstrated that Rg1 significantly upregulated NRF2 expression in IVDD rats (Fig. 3I) while reversing the IVDD-induced downregulation of GPX4, SLC7A11, and FTL1, suggesting its therapeutic potential through modulation of these critical ferroptosis pathway components.

NRF2 is essential for the defense against ferroptosis. WB and immunohistochemistry confirmed ginsenoside Rg1 upregulated NRF2 expression in IVDD rats (Fig. 4A, B,F). As NRF2's downstream ferroptosis targets, GPX4, FTL1 and SLC7A11 protein levels were significantly restored by Rg1 treatment (Fig. 4A-E). Immunofluorescence demonstrated enhanced GPX4 expression (Fig. 4G), while FerroOrange assays showed Rg1 reduced Fe²⁺ accumulation (Fig. 4H, I), collectively indicating Rg1's anti-ferroptotic effect through NRF2 pathway activation.

Integrated analyses identified ferroptosis as crucial in IVDD (Fig. 3). Treatment with 200 µM Rg1 effectively normalized Lipid-ROS levels (Fig. 5A), reduced MDA (Fig. 5B), and restored SOD/GSH-Px activity (Fig. 5C, D). WB and immunofluorescence demonstrated Rg1 reversed H₂O₂-induced ECM dysregulation by increasing Collagen II and decreasing MMP3 expression (Fig. 5E-K), confirming its protective role in NP cell homeostasis.

NRF2 and GPX4 are key oxidative stress sensors in NP cells and play pivotal roles in ferroptosis. Network pharmacology and transcriptome analysis identified FTL1 and SLC7A11 as additional ferroptosis-related targets. NRF2, a critical ferroptosis suppressor, mitigates ferroptosis by upregulating downstream targets. WB results revealed elevated NRF2 levels in IVDD NP cells, further increased by ginsenoside Rg1 treatment (Fig. 6A, B). GPX4, negatively regulating ferroptosis, prevents lipid peroxide accumulation; its decline in H₂O₂-induced NP cells was reversed by Rg1. Reduced FTL1 (causing Fe²⁺ overload) and decreased SLC7A11 (a ferroptosis inhibitor) in H₂O₂-treated cells were also restored by Rg1 (Fig. 6A-E). Fe²⁺ probe assays confirmed that 50/100/200µM Rg1 significantly reduced iron overload in H₂O₂-stimulated NP cells after 48 h (Fig. 6F, G).

The NRF2 inhibitor ML385 reduced NP cell viability (Fig. 7A) and increased Lipid-ROS accumulation (Fig. 7B). Co-treatment with ginsenoside Rg1 + ML385 elevated MDA levels while suppressing SOD and GSH-Px activity (Fig. 7C-E). Although Rg1 alone upregulated Collagen II and downregulated MMP3 in H₂O₂-induced NP cells, ML385 reversed these effects (Fig. 7F-H). Immunofluorescence confirmed that ML385 counteracted Rg1's benefits, reducing Collagen II (Fig. 7I, J) and increasing MMP3 (Fig. 7K, L). These results demonstrate that ginsenoside Rg1's protective effects against H₂O₂-induced damage in NP cells are NRF2-dependent.

NRF2 inhibition by ML385 significantly decreased NRF2 protein levels (Fig. 8A, B) and downregulated ferroptosis-related proteins GPX4, FTL1, and SLC7A11 (Fig. 8C-F). Fe²⁺ probe analysis revealed elevated iron levels in the Rg1 + ML385 group compared to Rg1 alone, indicating ferroptosis reactivation (Fig. 8G, H). ChIP assays demonstrated that Rg1 enhanced NRF2's binding to GPX4 and FTL1 promoters, promoting their transcription (Fig. 8I-J). TEM analysis showed that Rg1 rescued H₂O₂-induced mitochondrial damage (shrinkage, cristae condensation, and membrane rupture), but ML385 abolished this protective effect (Fig. 8K), confirming NRF2's essential role in Rg1-mediated ferroptosis suppression and mitochondrial preservation.