SYR, HSYA, p-CA, SCU, p-hydroxybenzaldehyde (p-HBA), rutin (RU), quercetin, trans-cinnamic acid, kaempferol, adenosine, uridine, guanosine, cytidine, and riboflavin [internal standard (IS)] with a purity of >98% were purchased from Shanghai Ronghe Medicine Technology Development Co., Ltd. SIs (batch Nos. 19050511, 18113011, and 18101511) were manufactured by Langzhi Group Wanrong Pharmaceutical (Wanrong, Shanxi Province, China). Methanol and acetonitrile were HPLC grade (Sigma, United States). Ultra-pure water was purified by the Mini D system (Kertone, Changsha, China). Acetic acid was purchased from Aladdin (Shanghai, China). Fetal bovine serum (FBS) and trypsin were obtained from Gibco (NY, United States). Penicillin (P), streptomycin (S), and Dulbecco’s modified Eagle’s medium (DMEM, high glucose) were obtained from HyClone (United States). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and Dimethylsulfoxide (DMSO) were provided by Aladin (Shanghai, China). 30% H₂O₂ solution was obtained from Sinopharm Chemical Reagent Co., Ltd., (Shanghai, China). Vitamin C (Vc) (purity: 98%) was obtained from Sigma-Aldrich Co., (St. Louis, MO, United States).

The animal experiments were approved by the Animal Care and Use Committee of the College of Animal Science (College of Bee Science), Fujian Agriculture and Forestry University (Approval Number: PZCASFAFU21012). 250 ± 20 g of male Sprague-Dawley (SD) rats were provided by Lab Animal Center in Fujian Medical University (Fuzhou, China). The rats were acclimated to 22°C ± 2°C, relative humidity at 40%–70%, and a 12 h light and 12 h dark cycle, and the animal could access to feed and water ad libitum. Before administration, all rats were fasted for 12 h, but free access to water.

HPLC analysis was carried out on an Agilent 1290 Infinity LC instrument (Agilent, Waldbronn, Germany) consisting of a binary pump, a diode-array detector, an auto-sampler and a column compartment. The samples were separated on a RD-C₁₈ column (4.6 mm × 50 mm, 3.5 μm) (Zhongpu Science Inc., Fuzhou, China). The mobile phase was a stepwise gradient of water (containing 0.5% acetic acid, v/v) and acetonitrile (0 min, 98: 2; 20 min, 85: 15; 30 min, 70: 30; 50–52 min, 5: 95; 52.1–60 min, 98: 2). The column temperature was 30°C, the flow rate was 0.5 ml/min and the injection volume was 5 μl. The HPLC system was connected to an Agilent 6520 Q-TOF mass spectrometer (Santa Clara, CA, United States) equipped with an electrospray ionization (ESI) interface. Mass spectra were recorded at m/z 100–1000, and all masses were corrected by the internal standards provided by the Agilent Technologies (Agilent Part Number: G1969-85001) with m/z at 112.98559 and 1033.98811 in (–)ESI mode. The data were processed with Agilent MassHunter Workstation Software version B.06.00 (Agilent Technologies).

For qualitative analysis of the components in SI, the mixture methanol solution of the 13 standard references, including SYR, HSYA, p-CA, SCU, p-HBA, RU, quercetin, trans-cinnamic acid, kaempferol, adenosine, uridine, guanosine, and cytidine (about 10 μg/ml for each ingredient), were used in the HPLC-MS identification.

To predict the targets of the multiple ingredients of SI, the PharmMapper on-line Server (http://www.lilab-ecust.cn/pharmmapper/) was used according to previous reports (Shi et al., 2020). 300 targets were obtained from the on-line target fishing. Further, the targets with z'-score >0 were selected to take part in the following Disease Ontology Semantic and Enrichment (DOSE) analysis and molecular docking according to our previous report (Shi et al., 2020; Xie et al., 2020; Xie et al., 2021).

H9c2 cardiomyocytes obtained from Cell Bank of the Chinese Academy of Sciences (Shanghai, China) were cultured in DMEM (high glucose) supplemented with 1% P/S and 15% FBS in 5% CO₂ at 37°C. The cells were collected with 0.25% trypsin, followed by re-seeding in 96-well multiplates with a density of 5×10³ cells per well. The cell experiments were grouped as follows: normal group, model group (H₂O₂), administration groups and positive control groups (N-Acetyl-L-cysteine, NAC). After incubation for 24 h, the model groups were injured with 250 μM H₂O₂ for 2 h, while before exposed to 250 μM H₂O₂ for 2 h cardiomyocytes in administration groups and positive control groups were pretreated with HSYA (1, 5, and 10 μM), SYR (1, 5, and 10 μM), and NAC (1 mM) for 24 h, respectively. After that, cells were incubated with 0.5 mg/ml of MTT for 4 h under standard condition, followed by addition of 150 µl DMSO into every well. The cell supernatants were finally removed and the optical density (OD) was recorded at 570 nm using a microplate reader (Thermo Fisher Scientific, United States). The protection rate of the compounds on H9c2 cardiomyocytes was calculated as (OD value_{administration group (or NAC group)}-OD value_{model group})/(OD value_{normal group}-OD value_{model group}) ×100%.

All data were expressed as mean ± SD and analyzed using the SPSS 18.0. Statistical comparisons between groups were performed via one-way ANOVA. Differences were considered significant when p < 0.05.

For quantitative proteomics, the 195 μM HSYA and 168 μM SYR were respectively used to treat with the cardiomyocytes injured by H₂O₂ to obtain apparent protection. The procedures for cell culture, group and administration are similar to those in “2.7. Cell culture and protection of HSYA and SYR on cardiomyocytes injured by H₂O₂” with little modification. Briefly, the difference is that H9c2 cells were seeded and cultured in a culture flask with a bottom area of 25 cm² with a density 1×10⁵ cells/ml. After treatment, about 3×10⁷ cells per group were collected for the next quantitative proteomics investigation, which was done by Jingjie PTM Biolabs (Hangzhou, China). All the procedures for quantitative proteomics can be found in the section of “5. Quantitative proteomics” of the Supplementary Material.

The UV chromatogram at 265 nm and total ion chromatogram in negative mode of SI are shown in Supplementary Figure S1. Thirteen ingredients, including HSYA, SYR, p-CA, SCU, p-HBA, RU, quercetin, trans-cinnamic acid, kaempferol, adenosine, uridine, guanosine, and cytidine, identified via HPLC-Q-TOF-MS with the aid of their previous reported MS information (Fan et al., 2019) and HPLC-MS-MS with the standard references, are listed in Supplementary Tables S1, S2.

Contents of guanosine and kaempferol were very low (below 0.5 ng/ml) after the preliminary experiment. The remaining 11 ingredients, including SYR, HSYA, p-CA, SCU, p-HBA, RU, quercetin, trans-cinnamic acid, adenosine, uridine, and cytidine, were then considered quantified. As shown in Supplementary Figure S2 and Supplementary Tables S3–S5, the developed HPLC-MS/MS method was suitable and applied to determine the content of the 11 ingredients for three batches of SI. The heatmap in Figure 1 demonstrated that HSYA presents the highest content (62% via normalization method), followed by the contents of SYR (11.0%), p-CA (9.0%), and uridine (13.5%) in all the three batches of investigated samples. The seven other ingredients can be clustered into one group because of their low contents via normalization method. These findings suggested that the main chemical substances of SI consist of HSYA, SYR, p-CA, and uridine, which account for about 95.5% of the total amount of the 11 ingredients according to the normalization method.

Only six ingredients, including HSYA, p-CA, SCU, p-HBA, SYR, and RU, can be detected 1 h after administration (2 ml/kg of SI) in a preliminary test. Concentration levels of the seven other ingredients (quercetin, trans-cinnamic acid, kaempferol, adenosine, uridine, guanosine, and cytidine) were below their LODs (about 0.5 ng/ml) and cannot be determined at the PK time point of 15 min after administration. Accordingly, we focused on the PK behavior of six ingredients in SI, namely, HSYA, SYR, p-CA, SCU, p-HBA, and RU, due to the easily detectable property in rat plasma.

Methodological validation, such as selectivity, calibration curves, precision, accuracy, extraction recovery, matrix effect, and stability, was performed on the six compounds in plasma for bioanalysis. Typical SIM or MRM chromatograms for the six compounds and IS in blank, spiked, and drug plasma after administration for 5 min are presented in Supplementary Figure S3. The absence of interference indicated the acceptable selectivity of the method.

All calibration curves are summarized in Supplementary Table S6. Linearity (r ² > 0.995) was satisfactory and within the tested range. As shown in Supplementary Table S7, RSD values were 0.23%–6.90% and 1.14%–4.70% while RE values were −5.00%–7.42% and −1.56%–5.58% for the respective intra- and interday precisions of all six analytes. This finding suggested that the proposed method is accurate and reliable. Extraction recoveries for the six analytes were in the range of 84.12%–103.45%, with RSD values of 2.62%–10.99%, and matrix effects were in the range of 99.77%–121.41%, with RSD values of 1.09%–12.20% (Supplementary Table S8). These results demonstrated that the proposed method is reliable for bioanalysis. The results of prepreparation, postpreparation, freezing, thawing, and long-term stability experiments are listed in Supplementary Table S9. These values indicated that analytes are stable and can be detected with acceptable accuracy (RE within ±15%).

The plasma drug concentration–time curves for the six ingredients (i.v., 1, 2, and 4 ml/kg of SI) are shown in Figure 2. The calculated PK parameters (Table 1) showed that HSYA, SYR, p-CA, p-HBA, and RU are eliminated rapidly (0.08 h ≤ t _1/2 ≤ 0.64 h), while SCU was eliminated slowly with a t _1/2 ≥ 2.58 h in rats. Meanwhile, the order of AUC values is HSYA > SYR > p-CA > SCU > p-HBA > RU. Notably, the systematic exposure extent of HSYA was 6-10 times higher than that of SYR (See the AUC values in Table 1). Meanwhile, SCU, p-HBA, and RU can’t be detected behind 15 min after i.v., low dose administration (1 ml/kg of SI is equivalent to a mean dosage of normal adult).

Meanwhile, the multiple integrated PK of HSYA, SYR, p-CA, SCU, p-HBA, and RU was used with integrated methods (Yao et al., 2017; Shi et al., 2018; Yao et al., 2018; Shi et al., 2020), namely, “plasma drug concentration sum” and “AUC weighting integrated” methods. As shown in Figure 2 and Table 1, differences between the individual HSYA PK and the integrated PK obtained from the two integrated methods were insignificant for nearly all the PK parameters. The results suggested that PK characteristics of HSYA can reflect the integrated in vivo process of multiple ingredients of SI because its PK parameters are similar to those of the multiple-ingredient integrated PK in rat plasma.

System pharmacology is an excellent tool for determining the effects and mechanisms of multiple-component systems, such as TCM (Cheng et al., 2022). PK studies showed that ingredients, such as HSYA, p-CA, SCU, p-HBA, and SYR, (except for RU) demonstrate favorable PK properties (easily detected at PK time points and high system exposure). These six ingredients were utilized in the systematic pharmacology study. Potential targets with a z'-score > 0 for ingredients HSYA, p-CA, SCU, p-HBA, SYR, and RU were obtained through target fishing. These ingredients were subjected to DOSE analysis using R-package. The remaining five ingredients, except for p-HBA, can collect a series of diseases using their relative potential targets. As shown in Figure 3, HSYA and SYR can show “rank 1” relevance to arteriosclerotic cardiovascular disease, while p-CA, SCU, and RU can be highly correlated to lung disease, tauopathy, and nutrition disease. Meanwhile, p-CA, SCU, and RU can also be correlated with arteriosclerotic cardiovascular disease. However, their contents in SI and exposure extents in systematic circulation are minimal. In addition, a DOSE graph for p-HBA is unavailable due to the low relevance of its potential target proteins to diseases. The results suggested that HSYA and SYR may be the main active components that exert the effects of anti-CVDs for SI.

Molecular docking was performed on the five ingredients of HSYA, SYR, p-CA, SCU, and RU using their potential target proteins, which are correlated with arteriosclerotic cardiovascular disease in the DOSE analysis, to support this conjecture on active components of SI against CVDs further. As shown in Figure 4 and Supplementary Table S10, “rank 1” affinity targets of HSYA, SYR, p-CA, SCU, and RU (docking total score >7.0 and binding free energy ΔG ≤ −4 kcal/mol) are leukotriene A4 hydrolase (LTA4H), sorbitol dehydrogenase (SDH), deoxycytidine kinase, cell division protein kinase 2, and cathepsin B (CTSB), respectively, while p-HBA fails to find a target protein with a docking total score of >7.0 and binding free energy of ΔG ≤ −4 kcal/mol. Particularly, LTA4H and SDH are key targets for cardiovascular disease treatment (Sandanayaka et al., 2010; Tang et al., 2010). Therefore, the molecular docking results further supported the conjecture of roles of HSYA and SYR on anti-CVDs. CTSB can be used as an indicator for myocardial ischemia/reperfusion (I/R) injury (Qiao et al., 2021), thereby suggesting the possible positive effect of RU on CVDs. The consistency between molecular docking and DOSE analysis results confirmed that HSYA and SYR may be the main active components that exert effects of anti-CVDs in SI.

The protection of HSYA, SYR, p-CA, SCU, p-HBA, and RU was evaluated on cardiomyocytes injured by H₂O₂ to provide the direct evidence on the anti-CVD activity of ingredients. The PK test revealed that the detected mean maximum plasma concentrations for HSYA, SYR, p-CA, SCU, p-HBA, and RU are 14.07 (23 μM), 0.93 (2.5 μM), 1.25 (8.5 μM), 0.327 (0.7 μM), 0.074 (0.61 μM), and 0.039 (0.06 μM) μg/mL, respectively, after the administration of 4 ml/kg of SI, at which dosage the plasma concentration Cmax value of HSYA in rats is very close to that in human beings when administrating a single dose of 140 mg of HSYA injection to an adult (Yang et al., 2009). These ingredients with their maximum detected plasma concentration levels in protection tests were evaluated preliminarily. The results presented that 23 μM of HSYA and 2.5 μM of SYR achieve protection rates of about 26% and 10%, respectively, while the positive drug (1 mM of NAC) shows a protection rate of 6.45% to injured cardiomyocytes. However, 8.5 μM of p-CA, 0.7 μM of SCU, 0.61 μM of p-HBA, and 0.06 μM of RU demonstrated the absence of protection. Furthermore, HSYA and SYR were compared at the three concentration levels of 1, 5, and 10 μM on the basis of cardiomyocytes injured by H₂O₂. As shown in Figure 5A, 5- and 10-μM SYR and 10-μM HSYA groups show higher cell viability than the model group (p < 0.05 or p < 0.01). Although no statistical difference exists between the same concentration of HSYA and SYR on protection rates, both compounds show a concentration-dependent anti-H₂O₂ injury effect on cardiomyocytes (Figure 5B). The results directly confirmed that HSYA and SYR can be the main active substances exerting cardiomyocyte protection, especially when the normal dosage of SI is administered to human beings.

Twelve samples, including three control (C), three model (M), three HSYA-treated (HSYA), and three SYR-treated (SYR) groups, were monitored via SDS-PAGE analysis before the TMT quantitative proteomics study. As shown in Figure 6A, protein bands were clear and uniform, the parallelism of each lane in the group was satisfactory, and the electrophoretic behavior between groups was different when observed with the naked eye. This finding preliminary confirmed that significant differences exist in the protein expression between various treatment groups as well as provided support to the succeeding quantitative proteomics study.

The quantitative proteomics study was carried out according to our previous report (Li et al., 2022). Relative quantitative values of proteins in two samples were compared via t-test. The protein expression was significantly upregulated when the p-value and fold change of the differential expression were <0.05 and ≥1.5, respectively. The protein expression was significantly downregulated when the p-value and the fold change of the differential expression were <0.05 and ≤1/1.5, respectively. A total of 7,380 proteins were identified and 6,595 proteins were quantified in the twelve samples for the four groups. The identified difference proteins between groups were subjected further to principal component analysis (PCA). The results showed that the cluster degree of samples in each group is high and samples are significantly separated between different treatment groups (Figure 6B). Notably, both HSYA and SYR groups showed recovery in the direction of PC2 toward the C group, thereby suggesting their possible interference effect against H₂O₂ injury to cardiomyocytes. As shown in Figure 6C and Supplementary Figure S4, 245 proteins were upregulated (such as Atpefg, Ndufa10, and Atp6l) and 754 proteins were downregulated (such as Tpm3, Parv, and Msn) in the M group compared with those in the C group; 148 proteins significantly upregulated (such as Glg1, Esl1, and Tpm3) and 205 proteins significantly downregulated (such as Grb2, Mapkapk2, and Cdh3) in the HSYA group compared with those in the M group, while 131 proteins (such as Rrs1, Ndufa8, and Lmnb2) upregulated and 231 proteins (such as Mcm4, Map2k1, and Dab2ip) downregulated in the SYR group compared with those in the M group. The subcellular structure location was analyzed using these proteins as objectives. As shown in Figures 6D–F, all treatments, including H₂O₂ injury, H₂O₂ injury plus HSYA interference, and H₂O₂ injury plus SYR interference, mainly changed the protein expression located at the cytoplasm by 33.83% in the M group vs. C group, 41.64% in the HSYA group vs. M group, and 42.54% in the SYR group vs. M group as well as the nucleus by 30.73%, 27.48%, and 30.39%, respectively.

Furthermore, Gene Ontology (GO) classification was performed to understand the functional classification of all differentially expressed proteins (DEPs) between groups. In the annotation of biological process (BP), the majority of DEPs involved in the cellular process under all comparisons of M versus C (Figure 6G), HSYA versus M (Figure 6H), and SYR versus M (Figure 6I) groups. The majority of DEPs in the annotation of cellular component (CC) came from cells and organelles under all comparisons. Molecular function (MF) analysis demonstrated that the majority of DEPs are concerned with binding and catalytic activities. The GO analysis indicated that DEPs play predominant roles in the structural composition of ribosome, structural molecular activity, and peptide biosynthetic process for the comparisons between HSYA and M groups as well as SYR and M groups (Supplementary Figure S5). Notably, the results of GO enrichment for both of HSYA versus M and SYR versus M are very similar with consideration for BP, CC, and MF enrichments. This finding suggested that the interference effect of HSYA and SYR is also similar to the metabolism pathways of injured cardiomyocytes via H₂O₂.

Finally, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment of DEPs was carried out between groups. As shown in Figure 6J, H₂O₂ treatment mainly downregulated the activity of the spliceosome pathway and upregulated the valine, leucine, and isoleucine degradation pathways due to the stress alternation of cardiomyocytes against injury. However, the interference of both HSYA and SYR mainly downregulated the ribosome pathway and upregulated the lysosome pathway in H₂O₂-injured cardiomyocytes (Figures 6K,L). The significant downregulation of the ribosome pathway in long-lived individuals (LLIs) (Xiao et al., 2022) suggested that downregulating the ribosome pathway can delay cell aging and death. The upregulation of the lysosome function can promote autophagy of damaged cells and then play a role in myocardial protection (Gu et al., 2020). These results demonstrated that both HSYA and SYR can protect cardiomyocytes from death and the effect mechanism can be primarily related to the downregulation of the ribosome pathway and the upregulation of the lysosome pathway by the two ingredients. In addition, two pathways related to energy metabolism, including glycolysis/gluconeogenesis and oxidative phosphorylation, were downregulated and upregulated, respectively, and contributed to the effect and partly accounted for the mechanism of both HSYA and SYR against cardiomyocyte injury.

Quantitative proteomics further confirmed the cardiomyocyte protection of the two ingredients and revealed their effect mechanisms related to multiple pathways, including ribosome, lysosome, glycolysis/gluconeogenesis, and oxidative phosphorylation.

As shown in Figure 7A, chemical substance basis analysis indicated that the SI mainly consists of HSYA, SYR, p-CA, and uridine, and the component with absolute predominance in content is HSYA. The multiple-component PK study further confirmed that five ingredients, including HSYA, SYR, p-CA, SCU, and p-HBA, present favorable PK properties while only HSYA shows very similar PK parameters (no significant difference) to multiple ingredient-integrated PK. This finding is consistent with the PK representative potential of the whole SI preparation in vivo. System pharmacology predicted that HSYA and SYR may be the main anti-CVD active ingredients, which are validated by the experiments of protective effect study on cardiomyocytes injured by H₂O₂. Ultimately, the proteomics study further confirmed the anti-CVD effect and mechanisms of the two ingredients HSYA and SYR. According to the results, HSYA can be considered a PK marker to represent the whole in vivo process of multiple ingredients of SI not only for its favorable PK properties but also satisfactory cardiomyocyte protection when 4 ml/kg of SI is administered. This dosage is clinically equivalent to that of pure HSYA injection for adults. Although the systematic exposure of SYR is over 10 times lower than that of HSYA when administered, its plasma drug concentration within the first 15 min after administration is about 1–1.25 μM, which achieves cardiomyocyte protection rates of 6%–10%. Therefore, SYR should be considered an important ingredient due to its possible contribution to the therapeutic effect of SI and noted in the whole characterization preparation PK behavior. On this basis, the effect weighting-integrated PK was explored with our previous suggestion (Shi et al., 2018). Briefly, the area under the protection rate-drug concentration curve (AUC _effect ) was applied to calculate the coefficient of effect weighting and the integrated plasma drug concentrations as follows: ω i = A U C i e f f e c t ∑ 1 n A U C e f f e c t , (1) ∑ 1 n A U C e f f e c t = A U C 1 e f f e c t + A U C 2 e f f e c t + … + A U C i e f f e c t + … + A U C n e f f e c t , (2) C Z = ω 1 C 1 + ω 2 C 2 + ω 3 C 3 + … + ω n C n , (3) where ω is the effect weighting coefficient, i is a certain ingredient, n is the number of investigated ingredients, C ₁ –C _n are the concentrations of each ingredient at one PK time point, and C _Z is the integrated plasma drug concentration. At present, only the two ingredients (HSYA and SYR) are involved in the integrated PK study due to their positive anticardiomyocyte injury effect and favorable PK properties. The PK profiles indicated that the plasma drug concentration range of HSYA and SYR is 0.02–23 and 0.02–2.5 μM and the calculated average AUC _effect within their plasma drug concentration ranges are 473.19% and 12.18% µM, respectively. According to Eqs 1, 2, the calculated effect weighting coefficients of HSYA and SYR are 97.5 and 2.5, respectively. As results, with the PK profiles for HSYA and SYR (Figure 2) and above Eq. 3, the calculated integrated concentrations, PK curves, and parameters are presented in Figure 7B and Table 1.

Notably, although the resulting effect weighting coefficient for HSYA is 39 times higher than that for SYR under the condition of adult administration dosage, the therapeutic role of SYR should not be neglected due to its good protection on cardiomyocytes injured in the preparation of SI. Table 1 presents that differences in PK parameters between HSYA and the effect weighting integrated curves are insignificant. This finding supports the PK representative role of HSYA for the overall in vivo process characterization of SI, and of course, HSYA and SYR can be recommended as quality markers (Lu et al., 2022) for quality control of safflower preparations.