The GSE dataset (GSE211949) was downloaded from NCBI GEO DataSets (https://www.ncbi.nlm.nih.gov/geo/). This dataset, generated using the GPL24247 platform, includes gene expression profiles from three normal mouse cardiomyocyte cells and 3 mouse cardiomyocyte cells treated with PM_2.5 at a concentration of 100 μg/mL. Differentially expressed genes (DEGs) were identified using the R package DESeq2 (version 1.40.2). A |log₂FoldChange|≥1.5 and adjusted p < 0.05 were set as the screening criteria. Pearson correlation coefficients were calculated using the cor. test function in R (version 4.0.3). Principal component analysis (PCA) was conducted using the FactoMineR package. The volcano plot and heatmap were generated using the “ggplot2” and “pheatmap” packages in R, respectively. The “clusterProfiler” package in R was used to perform Kyoto Encyclopaedia of Genes and Genomes (KEGG) enrichment analysis on the DEGs. The significance of enrichment analysis results was determined by a p-value below 0.05.

The candidate targets of Ses were predicted using Swiss Target Prediction (http://swisstargetprediction.ch/), Bioinformatics Analysis Tool for Molecular mechanism of Traditional Chinese Medicine (BATMAN-TCM, http://bionet.ncpsb.org.cn/batman-tcm/) (Liu et al., 2016), TargetNet (http://targetnet.scbdd.com/calcnet/index/) and Traditional Chinese Medicine Systems Pharmacology (TCMSP, https://www.tcmsp-e.com/#/home). Targets from the four aforementioned databases were synthesised to remove duplicate values and obtain the Ses-related targets. After organising the data, the VennDiagram package (version 1.7.3) in R software was used to find the intersection between the DEGs and the Ses targets. Subsequently, KEGG enrichment analysis was performed for the genes at the intersection. PPI analysis was conducted using the STRING database (STRING v11.5, https://cn.string-db.org) (Szklarczyk et al., 2019), and PPI networks were analysed using Cytoscape (version 3.9.1). The cytoHubba plugin was used to calculate clustering coefficient scores, which were ranked from high to low.

The 2D molecular structure of the ligand, Ses, was obtained from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/), and the AlphaFold structure of ACSL4 (receptor) was retrieved from the UniProt database (https://www.uniprot.org/). The 2D structure of Ses was converted into a 3D configuration using ChemBio3D Ultra (version 21.0.0.28), and the MM2 force field was used to optimise energy. Hydrogen atoms and charges were added to the ligand and receptor files using AutoDock Tools (version 1.5.7), and the files were subsequently exported in PDBQT format. After processing the structures of Ses and ACSL4, molecular docking was performed using AutoDockVina (version 1.1.2) (Trott and Olson, 2010) and visualised with PyMOL software (version 2.5).

H9C2 cell lines were supplied by Solarbio Science & Technology Co., Ltd. (Beijing, China). The PM_2.5 was collected using a membrane collection method on the Hebei Medical University campus. Ses (HPLC>98%) was purchased from Shanghai Aladdin Biochemical Technology Co., Ltd., (Shanghai, China). DMEM high-glucose medium, foetal bovine serum (FBS) and pancreatic digestive fluid (containing EDTA) were obtained from ZETA LIFE Inc. (CA, United States). Dimethyl sulfoxide (DMSO) and the ROS kit were purchased from Shanghai Biyotime Bio-technology (Shanghai, China). RIPA lysate, phosphate buffer (PBS), Phenylmethanesulfonyl fluoride protease inhibitor, skim milk powder, ECL Western blotting Substrate (ELC Plus), Tris-Base, Tween-20 and Fe²⁺ kit were purchased from Solarbio Science (Beijing, China). Anti-ACSL4 antibody, anti-β-actin polyclonal antibody, anti-GPX4 antibody and anti-SLC7A11 antibody were purchased from ABclonal Biotech Co., Ltd. (Wuhan, China). MDA, SOD, LDH, GSH, GSH-Px and BCA kits were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). The C11 BODIPY^581/591 and JC-1 fluorescence probes were purchased from MedChemExpress (Shanghai, China). The CCK-8 kit was obtained from ZETA LIFE Inc. (CA, United States).

H9C2 cells were cultured in DMEM high-glucose medium containing 10% FBS and 1% dual antibiotics (containing 100 U/mL penicillin and 100 mg/mL streptomycin) at 37°C in a 5% CO₂ incubator.

The collection and preparation of PM_2.5 sample were performed in accordance with our previous study (Zhang et al., 2021). Briefly, PM_2.5 samples were collected using an air sampler, and the collected PM_2.5 quartz fibre filters were sectioned into small fragments, suspended in PBS and sonicated for 30 min. After filtration through gauze, the PM_2.5 suspension was subjected to vacuum freeze-drying for 48 h to obtain PM_2.5 dry powder, which was then stored at −80°C. The detailed chemical components of PM_2.5 have also been reported in our previous study.

CCK-8 kits were used to assess cell viability in this study. The H9C2 cell line was plated in 96-well plates at a density of 5 × 10³ cells per well and cultured in 10% FBS medium for 12 h. Cells were then treated with different concentrations of Ses or PM_2.5 for 24 h. The CCK-8 solution was added to the culture medium at a ratio of 1:10 in each well. Cells were incubated at 37°C in a 5% CO₂ atmosphere for an additional 1 h. After the treatment, absorbance was measured using an enzyme-labelling instrument at 450 nm.

Multiple oxidative stress markers (MDA, SOD and GSH) and Fe²⁺ levels in H9C2 cells were measured using assay kits according to the manufacturer’s instructions.

The H9C2 cell line was plated in 6-well plates at a density of 3 × 10⁵ cells per well and cultured in 10% FBS medium for 12 h. Drug intervention was applied after 12 h of cultures. Lipid ROS levels and mitochondrial membrane potential were measured by flow cytometry to determine whether PM_2.5 induced ferroptosis in H9C2 cells.

The expression levels of ferroptosis-related proteins, including GPX4, SLC7A11, ACSL4 and LPCAT3, were measured by Western blot. Total protein from H9C2 cells was extracted using RIPA buffer containing protease inhibitors. The supernatant from the H9C2 cells was collected and subsequently denatured by boiling in 5× loading buffer for 5 min. After denaturation, equal amounts of proteins from each sample were electrophoresed by 10%–12% SDS–PAGE and transferred to a PVDF membrane. The membrane was then blocked in 5% skim milk for 2 h. Primary antibodies were incubated overnight at 4°C. The membrane was washed three times with TTBS, each wash lasting 10 min. Subsequently, the membrane was incubated with the secondary antibody for 40 min at room temperature. Finally, protein bands were detected using enhanced chemiluminescence.

The study analysis was performed using SPSS and GraphPad Prism. Data are expressed as the means ± SEM. Multiple group comparisons were conducted using one-way analysis of variance (one-way ANOVA). Differences were considered statistically significant when p-values were less than 0.05 (p < 0.05).

A correlation analysis was conducted to investigate the associations amongst different samples. A correlation coefficient close to one indicates a stronger positive correlation, whilst one close to −1 signifies a stronger negative correlation. Conversely, a correlation coefficient near 0 indicates a negligible or non-existent correlation between the two samples. The results indicate a strong positive correlation amongst samples within the same group and a strong negative correlation between samples from different groups (Figure 1A). The results of the unsupervised PCA demonstrated a distinct separation between the control group and the PM_2.5 exposure group samples. Principal component 1 (PC1) and principal component 2 (PC2) accounted for 62.59% and 23.66% of the total variance, respectively (Figure 1B). Differential expression analysis was performed using the ‘DESeq2’ R package to identify DEGs. The number of DEGs was illustrated using a volcano plot. A total of 3,212 DEGs were identified, meeting the threshold setting (|log₂FoldChange|≥1.5 and adjusted p < 0.05). Amongst these DEGs, 1729 genes were upregulated, and 1,483 genes were downregulated (Figure 1C). Subsequently, a heatmap of DEGs was presented in Figure 1D. Cluster analysis revealed the expression patterns of upregulated and downregulated genes across the samples. In the heatmap, red indicates upregulation, and blue indicates downregulation. KEGG enrichment analysis was used to determine the functions of DEGs. As illustrated in Figure 1E, the DEGs were primarily involved in ferroptosis, MAPK signalling pathway, hypertrophic cardiomyopathy, dilated cardiomyopathy, viral myocarditis, cardiac muscle contraction, chemical carcinogenesis-reactive oxygen species, oxidative phosphorylation, HIF-1 signalling pathway and glutathione metabolism. The detailed information of the DEGs involved in ferroptosis were shown in Supplementary Table S1.

H9C2 cells were treated with different concentrations of PM_2.5 for varying durations to determine the appropriate treatment time. Figure 2A shows a notable decrease in cell viability with increasing PM_2.5 concentration.

After exposure to PM_2.5, a decline in cell viability was observed, with the decrease being concentration-dependent. Notably, this decline was more pronounced after 24 h of exposure compared to 12 and 48 h of exposure. The results indicated that the half inhibition concentration (IC50) of PM_2.5 was close to 100 μg/mL (Figure 2A). The IC 50 value of erastin and the optimal Fer-1 dose in H9C2 cells were then measured to determine the dose for subsequent experiments (Figures 2B,C, respectively). The H9C2 cell viability in Figure 2B exhibited a notable concentration-dependent decrease in response to erastin. In addition, at concentrations up to 10 μΜ, Fer-1 did not demonstrate remarkable cytotoxicity (Figure 2C). Cell morphology, cell viability and LDH release assays confirmed the damaging effects of PM_2.5 exposure, and Fer-1 can attenuate these changes (Figures 2D–F). Subsequently, the oxidative stress indices, including SOD, GSH, GSH-Px and MDA, were examined (Figures 2G–J). As expected, exposure to PM_2.5 remarkably decreased the levels of SOD, GSH and GSH-Px whilst increasing the level of MDA. However, Fer-1 substantially reversed these changes and alleviated oxidative stress. As two important biomarker proteins of ferroptosis, GPX4 and SLC7A11 exhibited relatively decreased expressions due to PM_2.5 exposure. In contrast, Fer-1 pretreatment remarkably increased the expressions of GPX4 and SLC7A11 (Figures 2K, L).

A total of 16,360 targets of Ses were obtained from the Swiss Target Prediction, BATMAN-TCM, TargetNet and TCMSP databases, following the removal of partially overlapped targets. A total of 10 targets were obtained from Swiss Target Prediction, 16,359 targets from BATMAN-TCM, 4 targets from TargetNet and 17 targets from TCMSP (Figure 3A). According to Figure 3B, a total of 1859 targets were identified by integrating the Ses targets with the DEGs, namely, the targets of Ses involved in the treatment of PM_2.5-induced cardiomyocyte cells injury. Subsequently, KEGG enrichment analysis of the aforementioned intersection targets was conducted. The results showed that ferroptosis, chemical carcinogenesis-reactive oxygen species, glutathione metabolism, oxidative phosphorylation and cardiac muscle contraction were substantially enriched (Figure 3C). The PPI network was further constructed based on the genes enriched in ferroptosis and ranked by the clustering coefficient algorithm from high to low using the cytoHubba plugin (Figures 3D, E). The results revealed that ACSL4 had the highest score, followed by Slc39a14, Fth1, Cp, Hmox1, Tfrc, Prnp and Map1lc3a.

As shown in Figures 4A–D, after treatment with erastin, the cells shrank and lost their normal shape. However, pretreatment with Ses drastically inhibited this damage. H9C2 cells were treated with different concentrations of Ses (25–400 μM), and a CCK-8 assay was used to determine the viability of H9C2 cells. The results indicated no remarkable cytotoxicity of Ses at concentrations up to 100 μΜ (Figure 4E). Hence, concentrations of Ses, which include 25, 50 and 100 μΜ, were chosen as the intervention concentrations for subsequent experiments. An iCELLigence device was also used to monitor the cell index in real time (Figure 4F). The trend was consistent with the results of the CCK-8 assay. Based on the results of the CCK-8 assay and LDH release assays, Ses noticeably weakened the cellular damage caused by erastin (Figures 4G, H). Similarly, a decrease in the levels of GSH and an increase in the levels of MDA were observed after the administration of erastin. However, these alterations can be reversed by pretreatment with Ses (Figures 4I, J). The abovementioned results indicate that Ses is a potential candidate ferroptosis inhibitor and attenuates ferroptosis caused by erastin in H9C2 cells.

Figures 5A–C illustrate the protective effect of different concentrations of Ses on cell morphology, cell viability and LDH release. A dose-dependent phylactic effect of Ses pretreatment was also observed.

Concomitant with the progressive impairment of cardiomyocyte integrity, various markers of oxidative stress, such as SOD, GSH and GSH-Px, exhibited a notable decrease. However, with increased treatment levels of Ses, the levels of these crucial ROS scavengers were upregulated (Figures 6A, C, D). The MDA level, as the end product of lipid peroxidation, was downregulated after Ses administration (Figure 6B). The DCFH-DA staining (green) indicates an accumulation of ROS. The results revealed that Ses prevented ROS accumulation stimulated by PM_2.5 in H9C2 cells, and a dose-dependent therapeutic effect was notably observed (Figures 6E, F). In addition, the damage to mitochondrial membrane potential caused by PM_2.5 was remarkably reversed by Ses (Figures 6G, H). Collectively, these results indicate that Ses exerts strong antioxidant effects and inhibits mitochondrial damage in cardiomyocyte exposed to PM_2.5.

In ferroptosis, iron accumulation and lipid peroxidation are the two most important signs. Consequently, reducing the excess iron content is crucial for restraining ferroptosis. The present results indicate that PM_2.5 facilitates iron accumulation, whilst relatively high doses of Ses inhibit this progress (Figure 7A). The expression levels of ferroptosis-related proteins, including GPX4, SLC7A11, ACSL4 and LPCAT3, were measured to determine whether Ses inhibits cardiomyocyte ferroptosis. The results reveal that PM_2.5 downregulated the expression of GPX4 and SLC7A11 in H9C2 cells, whereas Ses restored the expression of both proteins (Figures 7B, C). Through molecular docking, the interaction between Ses and ACSL4 was confirmed (Figure 7D). The binding sites between Ses and ACSL4 were THR-159, ALA-182 and ARG-380. The binding energy was −9.1 kcal/mol (<−5 kcal/mol), indicating a strong binding affinity between the ligand and the protein. Similarly, Ses downregulated the elevated levels of ACSL4 and LPCAT3 caused by PM_2.5 (Figures 7E, F). C11-BODIPY^581/591 is a fluorescent probe developed to detect lipid peroxidation. As shown in Figures 7G, H, the fluorescence signal intensity of the oxidised BODIPY remarkably increased after PM_2.5 exposure, whilst it was reduced by Ses pretreatment.