After all crude botanical drugs, including Radix rehmanniae praeparata (30 g), P. sibiricum F. Delaroche (30 g), L. chinense Miller (30 g), P. notoginseng (Burkill) F. H. Chen ex C. H. Chow (10 g), L. chuanxiong Hort. (20 g), S. miltiorrhiza Bunge. (30 g), T. kirilowii Maxim. (60 g), and C. pinnatifida Bge. (30 g), were identified by the director Xu Yanfeng of the pharmacy department of our hospital, they were soaked in 2000 mL of water for 2 h. After that, they were boiled with water and simmered at low heat for 30 min. Then, they were filtered to collect the decoction. Meanwhile, after the raw residue was boiled with 2000 mL of water again, the second decoction was collected. After mixing the two decoctions (approximately 2,500 mL) and concentration, they were freeze-dried using a lyophilizer (LAB1-50, Biocool, Beijing, China). In short, the decoction was pre-frozen at −50 °C for 5 h before the vacuum drying. Then, the dryer was heated and finally fixed at −40 °C for 72 h to prepare lyophilized powder. Finally, a total of approximately 560 mg lyophilized powder of BSHXP was obtained and used for subsequent studies. All voucher specimens were deposited in our department.

The standards of Tanshinone IIA (purity: 99.78%, Lot., HY-N0135), salvianolic acid B (purity: 99.92%, Lot., HY-N1362), ferulic acid (purity: 99.97%, Lot., HY-N0060), Ginsenoside Rg1 (purity: 99.91%, Lot., HY-N0045), Ginsenoside Rb1 (purity: 99.28%, Lot., HY-N0039), Rehmannia glutinosa D (purity: 99.56%, Lot., HY-N0025), betaine (purity: 99.84%, Lot., HY-N0710), and chlorgenic acid (purity: 99.53%, Lot., HY-N0055) were weighed each 1-2 mg, and dissolved in methanol (HPLC-grade, Lot., A456-4, Fisher Scientific, Pennsylvania, USA) to obtain the stock solutions of each reference substance. All standards were bought from MedChemExpress (New Jersey, USA). Take each reference substance stock solution and configure it into a mixed reference substance solution with a concentration of 1 μg/mL. All samples were stored at 4 °C until analysis.

After accurately weighing 100 mg lyophilized powder of BSHXP, they were dissolved in accurately measured 1 mL ultrapure water, and centrifuged at 12,000 rpm at 4 °C for 15 min. After accurately absorbing 10 μL supernatant, the ultrapure water was added to dilute to 100 times. Then, the 80 μL of diluent was accurately absorbed and added to the liquid phase vial for sample loading analysis.

The chromatographic separation was achieved on an ACQUITY UPLC® HSS T3 column (2.1 mm × 100 mm, 1.8 μm, Waters Corporation, Milford, MA, United States) tandem with a guard column using a Thermo Scientific™ Q Exactive™ Quadrupole-Orbitrap Mass Spectrometer system (Thermo Fisher Scientific Inc., Grand Island, NY, United States). The UPLC system consisted of a Thermo Scientific Dionex Ultimate 3000 Series RS pump coupled with a Thermo Scientific Dionex Ultimate 3000 Series TCC-3000 RS column compartments and a Thermo Fisher Scientific Ultimate 3000 Series WPS-3000 autosampler controlled by Chromeleon 7.2 Software. The cooling autosampler was set at 10 °C and protected from light, and the column heater was set at 45 °C. A gradient elution programmer was used for chromatographic separation with mobile phase C (0.1% formic acid (HPLC-grade, purity >99%, CAEQ-4-000,313-0050, CNW Technologies, Bayern, Germany) in water) and mobile phase D [acetonitrile (HPLC-grade, Lot., A996-4, Fisher Scientific, Pennsylvania, United States)] mixed as follows: 0–12.0 min, 5% D - 95% D; 12.01–14.0 min, 95% D; 14.01–16 min, 5% D. The flow rate of the mobile phase was set at 0.3 mL/min, and the injection volume was 3 μL.

A Thermo Scientific™ Q-Exactive™ Quadrupole-Orbitrap Mass Spectrometer system (Thermo Fisher Scientific Inc., Grand Island, USA) connected to the UPLC system via a heat electrospray ionization interface and controlled by Xcalibur 4.1 software was used for data capture and analysis. MS/MS was operated in positive and negative mode using the following operating parameters: capillary temperature, 320 °C; Spray voltage, +3.5 KV/-2.8 KV; aux gas volume flow rate, 13 L/min; sheath gas flow rate, 80 L/min; aux gas heating temperature, 350 °C; resolution, 70,000; scan range, 80∼1,200 m/z.

Based on the limiting conditions of OB ≥ 30% and DL ≥ 0.18, the Traditional Chinese Medicine Systems Pharmacology and Analysis Platform (TCMSP) database (http://tcmspw.com/tcmsp.php) was applied to screen the active metabolites of various botanical drugs in BSHXP. After retrieving the SMILES formula of active metabolites in the PubChem database (https://pubchem.ncbi.nlm.nih.gov/), the potential targets of active metabolites were predicted by the Swiss Target Prediction (http://www.swisstargetprediction.ch). Finally, after transforming target protein names into corresponding gene names using the Uniprot database (https://www.uniprot.org/), an active ingredients-targets network was established using the Cytoscape 3.8.0 software.

After inputting the keyword “Atherosclerosis,” the disease targets of AS were screened through the GeneCards database (https://www.genecards.org/). After removing duplicate targets, the intersection of drug and disease targets was visualized in a Venn diagram. An active ingredients-targets network was established again to select the core metabolites with the top 12 Degree value.

Based on the STRING database (https://string-db.org/), the PPI networks were established by selecting the “Multiple Proteins” tool, defining the Organism as “Homo Sapiens,” setting high confidence (0.900), and hiding the free nodes. By using the Cytoscape 3.8.0 software, the PPI networks were adopted for further topological analysis to select hub targets with the top 10 Degree values.

The GO and KEGG enrichment analysis were accomplished using the DAVID database (https://david.ncifcrf.gov/). After setting the threshold to P < 0.05 and sorting by the enrichment count, the enrichment results were visualized in a bubble chart.

The HUVEC cell line was bought from Immocell (Xiamen, China). The HUVEC cells were cultured in the DMEM culture medium (Lot. 11965118, Gibco, Grand Island, United States) supplemented with 10% fetal bovine serum (Lot. FSD500, Excell Bio, Shanghai, China), and 1% penicillin-streptomycin (Lot. C0222, Beyotime, Shanghai, China) with surroundings of 37 °C, 5% CO₂, and 95% humidity environment. Based on the previous study, this study used ox-LDL (100 μg/mL, Lot. 20605ES05, Yesen, Shanghai, China) to irritate HUVEC cells for 24 h to imitate AS in vitro (Men et al., 2024).

To determine the effects of BSHXP on lipid metabolism, HUVEC cells were stochastically separated into five groups, namely, the control, ox-LDL, BSHXP-L, BSHXP-M, and BSHXP-H groups. Apart from the control group, HUVEC cells in other groups were treated with ox-LDL (100 μg/mL) for 24 h (Gong et al., 2019). In addition, HUVEC cells in the last three groups were treated with 12.5, 25, and 50 μg/mL BSHXP lyophilized powder dissolved in DMEM culture medium for 24 h, respectively. To determine the effects of regulating FABP1 on the ferroptosis of HUVEC cells, they were stochastically divided into four groups, namely, the control, ox-LDL, sh-NC, and sh-FABP1 groups. HUVEC cells were normally cultured in the control group and were treated with ox-LDL in the other groups. In addition, HUVEC cells in the last two groups were respectively transfected with sh-NC (5′-CAC​CGT​TCT​CCG​AAC​GTG​TCA​CGT​TTC​AAG​AGA​ACG​TGA​CAC​GTT​CGG​AGA​ATT TTT​TG-3′) and sh-FABP1 (5′-AGC​AAA​ACC​TAG​TGA​AAC​CTG​TTC​AAG​AGA​CAG​GTT​TCA​CTA​GGT​TTT​GCT-3′) for 72 h using Lipofectamine 3000 (Lot. L3000015, Thermo Fisher Scientific, Waltham, USA). To investigate whether BSHXP alleviated HUVEC cell injuries by regulating FABP1, they were stochastically divided into five groups, namely, the control, ox-LDL, BSHXP, BSHXP + oe-NC, and BSHXP + oe-FABP1 groups. HUVEC cells were normally cultured in the control group and were treated with ox-LDL in the other groups. Before being treated with 50 μg/mL BSHXP, HUVEC cells in the last two groups were respectively transfected with oe-NC (5′-GAT​CCA​AAA​AAT​TCT​CCG​AAC​GTG​TCA​CGT​TCT​CTT​GAA​ACG​TGA​CAC​GTT​CGG AGA​AC-3′) and oe-FABP1 (5′-AAG​CAA​AAC​CTA​GTG​AAA​CCT​TTC​AAG​AGA​AGG​TTT​CAC​TAG​GTT​TTG​CTT-3′) for 72 h. To investigate whether BSHXP inhibited ferroptosis by regulating FABP1, HUVEC cells were stochastically divided into five groups, namely, the control, ox-LDL, BSHXP, BSHXP + oe-FABP1, and BSHXP + erastin groups. The treatments of HUVEC cells in the first four groups were the same as mentioned above. Apart from being treated with ox-LDL and BSHXP for 24 h, HUVEC cells in the last group were also treated with erastin (10 μM, ferroptosis inducer, Lot. HY-15763, MedChemExpress, Princeton, USA) for 24 h (Lv et al., 2023). All transfectants were synthesized by Sangon (Shanghai, China).

The MTT assay was adopted in this study to determine the safe concentrations of BSHXP on HUVEC cells. In short, after HUVEC cells were respectively treated with BSHXP lyophilized powder dissolved in DMEM culture medium (0, 6.25, 12.5, 25, 50, and 100 μg/mL) for 24 h, 10 μL MTT agent (0.5 mg/mL, Lot. C0009S, Beyotime, Shanghai, China) was supplemented and incubated for 4 h. Then, after supplementing with 100 μL dimethyl sulfoxide, the optical density (OD) values of HUVEC cells at 570 nm were detected by a Multiskan FC microplate reader 357-714018 (Thermo Fisher Scientific, Waltham, USA). The cell viability of HUVEC cells was calculated based on their OD values.

After finishing corresponding treatments, the HUVEC cells were fixed with 4% paraformaldehyde (Lot. P0099, Beyotime, Shanghai, China). Then, the HUVEC cells were successively stained with 1 mL staining solution composed of Nile red reagent (Lot. C2051S, Beyotime, Shanghai, China) and the Hoechst 33,342 solution (Lot. C1029, Beyotime, Shanghai, China) for 20 min in the dark. The stained results were observed using a BZ-H4XD fluorescence microscope (Keyence, Osaka, Japan), among which lipid droplets exhibited red fluorescence and cell nuclei displayed blue fluorescence. The Image-Pro Plus 6.0 software (Media Cybernetics Inc., Rockville, USA) was employed to quantify the fluorescence intensity to evaluate the lipid droplet content of HUVEC cells.

After finishing corresponding treatments, the HUVEC cells were fixed with oil red O fixed liquid. Then, they were washed with 60% isopropanol for 30 s. Next, after HUVEC cells were stained by the oil red O dye solution (Lot. C0157S, Beyotime, Shanghai, China) for 30 min, they were counterstained by Mayer’s hematoxylin solution (Lot. PR30004, Proteintech, Wuhan, China) for 15 min. Finally, the stained results were recorded by an Axio Imager M1 optical microscope (Zeiss, Oberkochen, Germany).

The colorimetric method was applied to detect the intracellular Fe²⁺ level. In short, after finishing the corresponding treatments, the HUVEC cells were collected, resuspended in 250 μL lysis solution, and cracked at 4 °C for 30 min. According to the instructions of the cell ferrous iron colorimetric assay kit (Lot. E-BC-K881-M, Elabscience, Wuhan, China), the solutions in standard wells, measuring wells, and control wells were separately prepared. The intracellular Fe²⁺ level was determined based on the OD values at 593 nm detected by the microplate reader. The operation processes for detecting intracellular glutathione (GSH) level were similar to those for intracellular Fe²⁺ level. The differences were that the reduced GSH detection kit (Lot. ZC-S0329, Zcibio, Nanjing, China) was employed, the intracellular GSH level was determined based on the calibration curve method, and the OD values were detected at 412 nm.

After finishing the corresponding treatments, the HUVEC cell supernatant was gathered. The interleukin-1β (IL-1β), interleukin-6 (IL-6), monocyte chemoattractant protein-1 (MCP-1), and vascular cell adhesion molecule-1 (VCAM-1) levels in the supernatant were respectively determined using corresponding human ELISA kits based on the producer’s instructions, among which the human IL-1β ELISA kit (Lot. PI305), human IL-6 ELISA kit (Lot. PI330), human CCL2/MCP-1 ELISA kit (Lot. PC130), and human VCAM-1/CD106 ELISA kit (Lot. PV955) were all bought from Beyotime (Shanghai, China).

To determine the apoptosis rate of HUVEC cells, after collecting HUVEC cells with corresponding treatments, they were successively stained by 5 μL Annexin V-FITC and PI (Lot. C1062L, Beyotime, Shanghai, China) in the dark for 20 min. To determine the ROS level of HUVEC cells, after finishing corresponding treatments, HUVEC cells were immersed in 1 mL DCFH-DA (10 µM) diluted by serum-free DMEM for 30 min in the dark. After finishing the above staining, the HUVEC cells were collected to determine the apoptosis rate and ROS level of HUVEC cells using an Attune NxT flow cytometer (Invitrogen, Carlsbad, United States).

After extracting total RNAs in HUVEC cells using the MiPure Cell/Tissue miRNA Kit (Lot. RC201, Vazyme, Nanjing, China), the total RNAs were reversely transcribed into cDNA and cDNA was amplified with the help of the miRNA 1st Strand cDNA Synthesis Kit (Lot. MR101, Vazyme, Nanjing, China) and miRNA Universal SYBR qPCR Master Mix (Lot. MQ101, Vazyme, Nanjing, China), respectively. The reverse transcription and cDNA amplification were performed in a CFX96 Touch 1855195 qRT-PCR instrument (Bio-Rad, Hercules, United States). The 2^−ΔΔCT approach was employed for the quantification of FABP1 expression level. The primer sequences of FABP1 and GAPDH are listed in Table 1, in which GAPDH is the internal reference.

After the total proteins in HUVEC cells were extracted and denatured in boiling water, they were loaded on the SDS-PAGE gel and transferred to the PVDF membranes. Then, after being blocked by skim milk, the membranes were immersed in the primary antibodies at 4 °C for one night, including anti-FABP1 (1: 1,000, Lot. #13368), anti-sterol regulatory element-binding protein 2 (SREBP2, 1: 200, Lot. PA5-88943), anti-solute carrier family 7 member 11 (SLC711A, 1: 1,000, Lot. ab307601), anti-glutathione peroxidase 4 (GPX4, 1: 500, Lot. ab125066), anti-acyl-coA synthetase long chain family member 4 (ACSL4, 1: 2000, Lot. ab155282), anti-4-Hydroxy-2-nonenal (4-HNE, 1: 3500, Lot. ab46545), and anti-GAPDH (1: 2000, Lot. bsm-33033M). Apart from the anti-FABP1, anti-SREBP2, and anti-GAPDH were respectively bought from Cell Signaling Technology (Danvers, United States), Thermo Fisher Scientific (Waltham, USA), and Bioss (Beijing, China), other antibodies were purchased from Abcam (Cambridge, UK). Finally, after incubating with the goat anti-mouse IgG H&L/HRP (1: 20,000, Lot. bs-0296G-HRP, Bioss, Beijing, China) for 30 min and being stained by ECL chemiluminescent substrate, the grayscale value of blots was quantified using the Image-Pro Plus 6.0 software.

All data produced from independently triplicate experiments were exhibited as mean ± standard deviation. Statistical significance was determined using Student’s t-test or one-way ANOVA, followed by Tukey’s post hoc test, with the help of GraphPad Prism 8.0.2 (La Jolla, California, USA). A p-value of less than 0.05 was considered statistically significant.

The main parameters for standards are shown in Table 2. Representative single ion monitoring (SIM) chromatograms and total ion flow chromatograms of a mixed standard solution and sample are shown in Figures 1A–F. The external standard one-point method was used to analyze BSHXP lyophilized powder samples. The analysis results of the BSHXP lyophilized powder are listed in Table 3. A total of 8 metabolites were detected in the BSHXP lyophilized powder (Tanshinone IIA, salvianolic acid B, ferulic acid, Ginsenoside Rg1, Ginsenoside Rb1, Rehmannia glutinosa D, betaine, and chlorgenic acid). Among them, the content of salvianolic acid B, Ginsenoside Rg1, Rehmannia glutinosa D, and betaine is higher.

Based on the TCMSP and Swiss Target Prediction database, 102 active metabolites of BSHXP and 786 of their targets were predicted. The active ingredients-targets network was presented in Figure 2A. Based on 786 drug targets mentioned above and 4043 disease targets predicted by the GeneCards database, a Venn diagram was constructed and presented in Figure 2B, in which there were 540 common targets. According to the 540 targets mentioned above, the active ingredients-targets network was constructed again, among which the darker color represented the higher Degree value. As illustrated in Figure 2C, the core metabolites with top 12 Degree value were as follows, MOL006331 (4′,5-dihydroxyflavone), MOL002714 (baicalein), MOL001792 (liquiritigenin), MOL004941 ((2R)-7-hydroxy-2-(4-hydroxyphenyl) chroman-4-one), MOL000006 (luteolin), MOL000098 (quercetin), MOL007077 (sclareol), MOL007079 (tanshinaldehyde), MOL002135 (myricanone), MOL002151 (senkyunone), MOL001494 (mandenol), and MOL001495 (ethyl linolenate). After constructing the PPI network, hub targets with top 10 Degree values were screened and listed as follows: PIK3R1, PTPN11, EGFR, PIK3CA, PIK3CB, JAK2, SRC, PTK2, ERBB2, and PDGFRB (Figure 2D).

The GO enrichment results showed that, in the category of biological process, these targets were dominantly enriched in signal transduction, protein phosphorylation, positive regulation of transcription from RNA polymerase II promoter, positive regulation of cell proliferation, inflammatory response, negative regulation of apoptotic process, G-protein coupled receptor signaling pathway, response to xenobiotic stimulus, negative regulation of transcription from RNA polymerase II promoter, and positive regulation of gene expression (Figure 3A). In the category of cellular component, these targets were dominantly enriched in plasma membrane, cytosol, cytoplasm, nucleus, integral component of membrane, nucleoplasm, membrane, integral component of plasma membrane, extracellular exosome, and extracellular region (Figure 3B). In the category of molecular function, these targets were dominantly enriched in protein binding, ATP binding, identical protein binding, protein serine/threonine/tyrosine kinase activity, metal ion binding, protein kinase activity, enzyme binding, zinc ion binding, protein serine/threonine kinase activity, and protein homodimerization activity (Figure 3C). In addition, the KEGG enrichment results presented that these targets were primarily enriched in metabolic pathways, pathways in cancer, PI3K-Akt signaling pathway, neuroactive ligand-receptor interaction, lipid and atherosclerosis, MAPK signaling pathway, calcium signaling pathway, proteoglycans in cancer, microRNAs in cancer, and chemical carcinogenesis-receptor activation (Figure 3D).

First, we adopted the MTT approach to exclude the toxic dosage of BSHXP for HUVEC cells. The cell viability of HUVEC cells was prominently reduced when treated with 100 μg/mL BSHXP (Figure 4A). Therefore, the 100 μg/mL BSHXP was excluded, and 12.5, 25, and 50 μg/mL BSHXP were respectively considered the BSHXP-L, BSHXP-M, and BSHXP-H in subsequent studies. As the above results predicted that the PPAR signaling pathway was one enriched pathway of BSHXP, we subsequently investigated whether BSHXP regulated FABP1-mediated lipid metabolism (one regulatory molecule of the PPAR pathway) in HUVEC cells (Yan et al., 2023). As illustrated in Figure 4B, BSHXP-L, BSHXP-M, and BSHXP-H effectively reversed ox-LDL-induced elevation in the FABP1 protein expression level of HUVEC cells. In addition, BSHXP-L, BSHXP-M, and BSHXP-H prominently decreased the lipid droplet content (Figures 4C,D) and SREBP2 protein expression level (Figure 4E) of HUVEC cells in a dose-dependent way. The above results suggested that BSHXP regulated lipid metabolism in HUVEC cells, which might be related to mediating FABP1.

First, FABP1 mRNA expression level was determined to select the optimal transfectant. As presented in Figure 5A, the FABP1 mRNA expression level of HUVEC cells was dramatically downregulated after being transfected with sh-FABP1-1 and sh-FABP1-2, in which the FABP1 mRNA expression level was lower when being transfected with sh-FABP1-2. Therefore, the sh-FABP1-2 was used to knock down FABP1 expression in subsequent studies. As the above results indicated that FABP1 might be involved in the effects of BSHXP on lipid metabolism, we subsequently investigated the effects of knocking down FABP1 on the lipid metabolism of HUVEC cells. Knocking down FABP1 effectively reversed ox-LDL-induced elevation in the lipid droplet content (Figures 5B–D) and SREBP2 protein expression level (Figure 5E) of HUVEC cells. More interesting, as shown in Figures 5F–I, we discovered that knocking down FABP1 also substantially reversed ox-LDL-induced elevation in the Fe²⁺ and ROS levels as well as the ACSL4 and 4-HNE protein expression levels and reduction in the GSH level and GPX4 and SLC711A protein expression levels. The above results indicated that knocking down FABP1 mitigated ferroptosis by regulating the lipid metabolism of HUVEC cells.

As the above studies have confirmed that knocking down FABP1 reduced ferroptosis by regulating lipid metabolism, we next investigated whether BSHXP improved HUVEC cell injuries by regulating FABP1. Since the previous studies have demonstrated that AS belonged to the pathological process of chronic inflammation, the inflammation-related biomarkers were first detected (Wolf and Ley, 2019; Kong et al., 2022). BSHXP memorably downregulated IL-1β, IL-6, MCP-1, and VCAM-1 (Figures 6A–D) levels of HUVEC cells. Moreover, BSHXP also markedly reduced the apoptosis rate of HUVEC cells (Figure 6E) of HUVEC cells. In addition, overexpressing FABP1 effectively reversed the alterations of these indicators mentioned above. The above results revealed that BSHXP improved HUVEC cell injuries by suppressing FABP1.

As the above studies have affirmed that BSHXP improved HUVEC cell injuries by inhibiting FABP1, and knocking down FABP1 alleviated ferroptosis of HUVEC cells, we subsequently investigated whether BSHXP improved HUVEC cell injuries by regulating FABP1-mediated ferroptosis. Our research results displayed that BSHXP significantly downregulated the Fe²⁺ (Figure 7A), ROS (Figure 7C), IL-1β, IL-6, MCP-1, and VCAM-1 (Figures 7E–H) levels, apoptosis rate (Figure 7I), and ACSL4 and 4-HNE (Figure 7D) protein expression levels of HUVEC cells, and upregulated the GSH level (Figure 7B) as well as GPX4 and SLC7A11 (Figure 7D) protein expression levels of HUVEC cells. In addition, overexpressing FABP1 exerted the same effects as erastin, both of which effectively reversed the effects of BSHXP on the indicators mentioned above. The above results implied that BSHXP improved HUVEC cell injuries by suppressing FABP1-mediated ferroptosis.