The Traditional Chinese Medicine Systems Pharmacology Database (TCMSP, http://lsp.nwu.edu.cn/), the Encyclopedia of Traditional Chinese Medicine Database (ETCM, http://www.tcmip.cn/ETCM/index.php/Home/Index/), and the TCMID Database (http://119.3.41.228:8000/tcmid) were used to collect the chemical compounds of SGR, the candidate compounds with oral bioavailability (OB) ≥ 30% and drug-likeness (DL) ≥ 0.18 were identified as active compounds through the Symmap Database (https://www.symmap.org/) indicating the compounds had chemical stability and solubility.

SGR active compounds were screened after duplication, and they were imported into the Batman Database (http://bionet.ncpsb.org/batman-tcm/index.php/Home/Index/index) to search and predict the targets of each active compound.

The original HF disease targets were obtained from the Genecards Database (https://www.genecards.org/) with a relevance score no less than 5. Then the intersection targets between HF diseases targets and predicted targets of SGR compounds were screened by Venn analysis. The intersection targets were identified as targets of SGR compounds that had a potential pharmacological effect on HF.

The active compounds of SGR and their predicted targets were introduced into Cytoscape 3.1.1 software to construct a network diagram, revealing the distribution characteristics of SGR active compounds and their targets. The network topology analysis was performed to screen the key active compounds of SGR according to the connection degree.

Protein–Protein Interaction (PPI) network of the intersection targets was obtained from Search Tool for the Retrieval of Interacting Genes/Proteins (STRING, a free biological database) (http://string-db.org/; version 11.0), with the species limited to “Homo sapiens,” PPI network was then imported into Cytoscape 3.1.1 software for network analysis of the core subsystem. Key targets (Hub-Target) were screened according to the connection degree after network topology analysis.

Gene Ontology Annotation enrichment analysis (GO) of screened hub targets was performed via DAVID Database (https://david.ncifcrf.gov), including cellular component (CC), molecular function (MF), and biological process (BP). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was conducted via the “org.Hs.eg.db” package in R software (Version 3.6.3). Enrichment bubbles were constructed through the OMICShare tool online platform (http://www.omicshare.com/tools).

Reference standards of Istidina, resveratrol, shikimic acid, epicatechin, and dihydrokaempferol were accurately weighed and dissolved with 80% methanol to obtain standard solutions at a concentration of 0.153, 0.110, 0.068, 0.065, and 0.098 mg/ml, respectively. Also, the standard solutions were filtered with a 0.22 syringe filter before analysis. A UHPLC system coupled with LTQ Orbitrap XL high resolution mass spectrometric conditions (Thermo Fisher, Scientific, United States) was applied for LC-MS analysis. The chromatographic separation was performed on an ACQUITY UPLC HSS T3 Column (2.1 × 100 mm, 1.8 µm) at a flow rate of 0.2 ml/min. The mobile phases were composed of solution A (water) and solution B (methanol) and the gradient elution was set as follows: 10% B (0–8 min), 10–35% B (8–15 min), 35–90% B (15–40 min), and 90–10% D (46–50 min). The injection volume was 4 µl. MS analysis was performed on a mass spectrometer equipped with an ESI source and the parameters were as follows: spray voltage, 3.8 KV; capillary temperature, 350.05°C; capillary voltage, −18.00 units; tube lens voltage, −67.68 V; sweep gas, 0.24 units; sheath gas and auxiliary gas were set as 29.95 and 4 units, respectively. Orbitrap mass analyzer was set up with the scan range of m/z 90–1,500 and the resolution of 30,000. The samples used negative ion mode analysis.

The protein structures of screened targets were downloaded from the PDB Database (https://www.rcsb.org/) and were preprocessed by PyMol software. The structural files of the key active compounds of SGR were downloaded from the PubChem Database (https://pubchem.ncbi.nlm.nih.gov/) and were saved in *mol2 format. The abovementioned structures were added hydrogens, and the charges were calculated by autodocktools1.5.6 software, respectively; then, they were saved in a PDBQT format. Finally, molecules and proteins were docked through the Autodock Vina 1.1.2 software, the model with the smallest binding energy value was selected, and their structural visualizations were conducted by PyMOL tools.

Smilax glabra Roxb was purchased from Kangmei Industry (LOP: 160505781). H9c2 rat cardiomyocytes (ATCC) were purchased from American Type Culture Collection (Rockville, MD, United States). Dulbecco’s Modified Eagle’s Medium, fetal bovine serum, 100  U/ml penicillin, and 100 μg/ml streptomycin were purchased from Gibco (Grand Island, NY, United States). Reactive oxygen species assay kit, mitochondrial permeability transition pore assay kit, Lyso-Tracker Red, and antifade mounting medium with DAPI were purchased from Beyotime (Beyotime Biotechnology Co. Ltd., Shanghai, China). Cell Meter™ Mitochondrial Hydroxyl Radical Detection Kit was purchased from AAT Bioquest. TMRE Mitochondrial Membrane Potential Assay Kit was purchased from Solarbio (Solarbio, Beijing, China). 3-(4,5-dimethylthiazol-2-yl) -2,5-diphenyltetrazolium bromide (MTT) was purchased from Sigma-Aldrich (MTT, Sigma-Aldrich, United States), Seahorse XF Assay Kit was purchased from Agilent (Agilent, United States). Primers were ordered from Invitrogen (Thermo Fisher). Antibodies including β-actin (ab182651), Bcl-2 (ab32124), HRP-conjugated goat anti-rabbit (ab6721), and anti-mouse IgG (ab6789) were purchased from Abcam Co. P38 MAPK (CST, 9212), SAPK/JNK (CST, 9252), and p44/42 MAPK (CST, 9102) were purchased from Cell Signaling Technology. Goat Anti-Rabbit IgG AF 488 (abmart, M21012F) was purchased from Abmart (Abmart, Shanghai).

H9c2 cells were cultured in Dulbecco’s Modified Eagle’s Medium with 10% fetal bovine serum, 100  U/ml penicillin, and 100 μg/ml streptomycin in a 37°C, 5% CO₂ incubator. Cells in the H₂O₂ group were exposed to 600 μM H₂O₂ for 3 h. In the SGR group, cells were pretreated with 200, 400, and 800 μg/ml SGR for 24 h, then they were exposed to 600 μM H₂O₂ for another 3 h. Cells in the control group were cultivated with a normal medium.

Cell viability was detected by MTT assay. In brief, H9c2 cardiomyocytes (1  ×  10⁴/well) were seeded into 96-well plates. After treatment, 10 μl of 5 mg/ml MTT solution was added to each well and then cells were incubated in the dark for 4 h. After removing the cultured medium, the crystals were dissolved in 150 μl of DMSO. The absorbance was measured at 490 nm with a BioTek plate reader (BioTek).

Lactate dehydrogenase (LDH) production was measured by using LDH cytotoxicity Assay Kit for detection of cytotoxicity in H9c2 cells, according to the manufacturer’s protocol. Absorbance was measured at 490 nm using a BioTek plate reader (BioTek), and the values were directly proportional to the enzyme activity. In cellular morphology observation, cellular morphology was imaged by using a fully automatic inverted fluorescence microscopic analysis system (ECLIPSE Ti2-E, Nikon, Japan).

Superoxide Dismutase (SOD) activity was measured by using Superoxide Dismutase Assay Kit for the measurement of SOD inhibition rate in H9c2 cells, according to the manufacturer’s protocol. Absorbance was measured at 450 nm using a BioTek plate reader (BioTek), and the values were directly proportional to the SOD Inhibition rate.

The level of cellular ROS production was determined by using Reactive oxygen species Assay Kit, according to the manufacturer’s protocol. After treatment, the H9c2 cells were stained with DCFH-DA (1 μM) and incubated in the dark for 30 min, and then were washed three times with PBS. The cellular ROS fluorescence intensity was determined using Flow Cytometry Assay (Agilent ACEA, Quanteon, United States).

Cell Meter™ Mitochondrial Hydroxyl Radical was detected by using MitoROS™ OH580. After treatment, cells were stained with MitoROS™ OH580 probe for 1 h at 37°C and washed twice with pre-warmed PBS, and then 100 µl Assay Buffer was added into each group and 5 μl DAPI was added for 5 min in the incubator subsequently. The intensity of fluorescence was imaged by using a laser scanning confocal microscope (Zeiss LSM710, Germany). Cell Meter™ Mitochondrial Hydroxyl Radical was represented as red fluorescence.

Mitochondrial membrane potential (ΔΨm) was detected using TMRE. In brief, after the indicated treatments, cells were stained with a TMRE probe (0.2 µM), for 30 min at 37°C and washed once with pre-warmed DMEM without serum. The fluorescence intensity of cells was detected by using a laser scanning confocal microscope (Zeiss LSM710, Germany). The degree of TMRE fluorescence was considered as an indication of the change of mitochondrial membrane potential.

A mitochondrial permeability transition pore assay kit was used to detect the opening of mPTP. After treatment, the culture medium was aspirated and cells were washed twice with PBS, 1 ml Calcein AM staining solution [containing 1 µl Calcein AM (×1,000), 10 µl Solvent promoting (×100), and 10 µl CoCl₂ (×100)] were added to each sample, and cells were incubated for 30 min in the dark. After incubation, Calcein AM staining solution was aspirated, cells were washed twice with PBS, and then 500 µl detection buffer was added to the sample and the opening of mPTP was analyzed by using a BD FACSCalibur cytometer (Becton Dickinson, San Jose, CA, United States).

The level of lysosome was determined using Lyso-Tracker Red according to the manufacturer’s protocol. After treatment, the H9c2 cells were stained with Lyso-Tracker Red (1 μM) solution and incubated in the dark for 10 min, then the Lyso-Tracker Red (1 μM) solution was replaced with fresh DMEM. The lysosome fluorescence intensity was determined using Flow Cytometry Assay (Agilent ACEA, Quanteon, United States).

Cellular oxygen consumption rate (OCR) indicating the mitochondrial respiration was measured by using the Seahorse XF24 Extracellular Flux analyzer and software (Yu et al., 2019). Firstly, H9c2 cells were cultured in XF24 cell culture plates (1 × 10⁴ cells/well). Before the assay, 1 mmol/L pyruvate, 2 mmol/L glutamine, and 10 mmol/L glucose were added into Seahorse XF DMEM for preparation. After treatment, the cell medium was replaced with 500 μl prepared Seahorse XF DMEM, and then cells were incubated at 37°C without CO₂ for 1 h. Thereafter, cells were exposed to oligomycin (15 Mm, complex V inhibitor), FCCP (10 μM, respiratory uncoupler), and rotenone plus antimycin A (5 μM, inhibitors of complex I and complex III) by using the XF Cell Mito Stress Test kit (Seahorse Bioscience) to measure the parameters of OCR. Finally, the OCR value was normalized by CCK8 analysis. Respiration rates were calculated as the mean of all determinations at a certain state.

H9c2 cardiomyocytes were seeded in a laser confocal dish at a density of 1 × 10⁵ cells/ml (200 μl per dish). After treatment, the cell culture medium was removed, cells were washed three times with DPBS and fixed with 4% paraformaldehyde for 15 min. Then, paraformaldehyde was removed, cells were washed three times with PBS and blocked with BSA for 15 min, subsequently, cells were incubated with primary antibody Bcl-2 overnight at 4°C. Thereafter, cells were incubated with fluorescent secondary antibody (1:3,000) for 1 h after washing three times with PBS. 5 μl DAPI was added for 5 min in the incubator before imaging by using a fully automatic inverted fluorescence microscopic analysis system (ECLIPSE Ti2-E, Nikon, Japan).

Real-time quantitative polymerase chain reaction (qRT-PCR) was used to evaluate the mRNA expression level of pathway genes. Total RNA was isolated from H9c2 cells using Trizol reagent, according to the standard protocol, and RNA concentration was determined by using a NanoDrop spectrophotometer (Thermo Scientific, Rockford, IL, United States). 3 μg of total RNA was employed for reverse-transcription to yield cDNA. Subsequently, the synthesized cDNA was amplified using a 10 μl reaction system by the SYBR master Mixture (TAKARA, Japan). Finally, the threshold cycle (CT) value and the relative mRNA expression of the genes were calculated. The sequences of primers were shown as follows Table 1:

After the intervention, H9c2 cells were harvested and the total protein was extracted by using RIPA Lysis Buffer (Beyotime Biotechnology Co., Ltd., Shanghai, China) with pierce protease inhibitor tablet (Thermo Scientific, Rockford, IL, United States) and pierce phosphatase inhibitor tablet (Thermo Scientific, Rockford, IL, United States). Then total protein was quantified and mixed with 5X SDS-PAGE protein buffer (Beyotime Biotechnology Co., Ltd., Shanghai, China). After electrophoresis, the separated proteins were transferred to a polyvinylidene difluoride (PVDF) membrane. Later, the membranes were incubated with 5% non-fat milk in TBST (Tris-buffered saline with 0.1% Tween-20) for 2 h at room temperature, washed three times with TBST, and then the membrane with target proteins was incubated overnight at 4°C with primary antibody (1:1,000) against Bax, Caspase-3, p38 MAPK, JNK, ERK, β-actin, and Tubulin overnight at 4°C. Thereafter, the membrane was incubated with a respective HRP conjugated secondary antibody at 1:3,000 for 1 h after washing three times with TBST. BioRad imaging system (BioRad, Hercules, CA, United States) was applied to detect the fluorescent signal and the signal strip was quantified using Image Lab Software (BioRad, Hercules, CA, United States).

The experimental data, obtained from at least three independent experiments, were presented as the mean ± SEM. Statistical analysis was performed by one-way ANOVA analysis followed by the Bonferroni test for multiple comparisons by using GraphPad Prism software, version 8.0 for Windows (GraphPad Software Inc., United States). A probability of less than 0.05 was defined as statistically significant.

According to the TCMSP, ETCM, and TCMID Database, 77 chemical ingredients in SGR were collected, then 55 active ingredients of SGR were screened through the Symmap Database with the condition of OB ≥ 30%.Then, 32 active compounds were selected after removing ethylene ingredients, volatile oil ingredients, and duplicate ingredients (Details in Supplementary Table S1). Finally,a total of 640 predicted targets of 32 active components of SGR were collected (Details in Supplementary Table S2). Active components–predicted targets network was constructed to screen the hub active components of SGR (Details in Supplementary Table S3). As shown in Figure 1A, this network consisted of 689 nodes and 902 edges. The average degree node of the common components was 28.19 . In total, 5 hub components were identified, whose node degrees were far greater than the average node degree in this network. The nodes interacted with others via numerous edges (226 in Stearic Acid, 183 in Oleic Acid, 128 in Dihydroresveratrol, 62 in trans-resveratrol, and 42 in Istidina). The results of this network suggested that these five hub components may be the potential compounds of SGR in the treatment of HF.

A total of 2,978 HF diseases targets with relevance scores no less than 5 were selected after duplication (Details in Supplementary Table S4), then 266 intersection targets were screened via Venn analysis between SGR candidate targets and HF targets (as Figure 1B). The intersection targets were considered as the targets acting on HF from SGR candidate targets (Details in Supplementary Table S5).

A PPI network of 266 intersection targets was constructed to illustrate the interactions between targets, as shown in Figure 1C, this network consisted of 232 nodes and 680 edges. The average degree node of the common targets was 5.86. In total, 25 hub targets were identified (Figure 1D), whose node degrees were two-fold greater than the average node degree in this network (38 in SRC, 34 in JUN, 31 in CTNNB1, 29 in MAPK1, 26 in IL6, 26 in TNF, 20 in NFKB1, 18 in IL1B, 18 in INS, 17 in ESR1, 17 in PPARG, 16 in PRKCA, 15 in PPARA, 14 in IGF1, 14 in SIRT1, 14 in PRKCB, 14 in SDHB, 13 in IL10, 13 in IFNG, 13 in SP1, and 13 in SCN5A) (Detailed in Supplementary Table S6).

GO enrichment analysis was conducted to investigate the enrichment of the associated targets in biological processes (BP), cellular component (CC), and molecular function (MF) using the DAVID Database (detailed in Supplementary Table S7). The associated targets enriched in CC included cytoplasm, cytosol, mitochondrion, cell surface, endoplasmic reticulum, mitochondrial inner membrane, and endoplasmic reticulum membrane (Figure 2B). MF enrichment mainly included protein binding, ATP binding, enzyme binding, and receptor binding (Figure 2C). As shown in Figure 2A, the main functional modules in BP included regulation of transcription, cell proliferation, oxidation-reduction process, apoptotic process, ERK1 and ERK2 cascade, response to hypoxia, MAPK cascade, and inflammatory response. Previous studies have validated that MAPK participates in cell proliferation and apoptosis, mitochondrial regulation, and heart function, but there is no literature exploring the relationship between SGR and MAPK signaling pathway, and the pharmacodynamics of SGR.

KEGG enrichment analysis was used to explore the molecular mechanism of SGR involving heart failure treatment (Detailed in Supplementary Table S8). The mechanisms mainly included MAPK signaling pathway, Alzheimer’s disease, calcium signaling pathway, cardiac muscle contraction, type II diabetes mellitus, dilated cardiomyopathy, hypertrophic cardiomyopathy (HCM), PPAR signaling pathway, GnRH signaling pathway, toll-like receptor signaling pathway, and NOD-like receptor signaling pathway. (Figure 2D). When we conducted a gene-concept network (Figure 2E) and enrichment map analysis (Figure 2F), it all indicated that the MAPK signaling pathway was the potential mechanism of the SGR. The intersection targets of SGR enriched in pathways correlated with HF were integrated into an “HF-pathway” network, as Figure 3 showed, the MAPK signaling pathway and aforementioned pathways were involved in the regulation of a variety of biological processes including cell proliferation, apoptosis, inflammation, and mitochondrial metabolism, all of which were closely associated with HF progression. Thus, we further verified whether the MAPK signaling pathway is involved in the pharmacologic effect of GSR.

As Figure 4B showed, the cell viability of the H₂O₂ group was significantly reduced, but cellular viability increased as the concentration increased in the SGR pretreatment group. MTT results indicated the concentration of SGR range 0–800 μg/ml had no obvious cytotoxicity (Figure 4A). Thus, 200, 400, and 800 μg/ml (with the best protective effect) of SGR were selected for later experiments. As Figure 4C showed, 200, 400, and 800 μg/ml SGR could restore the cellular morphology damage induced by H₂O₂. Cells in the control group arranged regularly and tightly, but they became sparsely arranged and round-shaped, and necrotic cells presented as white bright aperture in the H₂O₂ group. In the 200, 400, and 800 μg/ml SGR group, white bright aperture gradually decreased, the fusiform of the cell morphology was gradually restored, and the arrangement began to be regular. It was also verified by the result of LDH (Figure 4D), compared with H₂O₂ group, 200, 400, and 800 μg/ml SGR all inhibited the secretion of LDH, indicating that the cellular damage induced by H₂O₂ was reversed by SGR.

Overexpression of ROS damages mitochondrial function, which leads to further oxidative toxicity (Takano et al., 2003). Oxidative injury and mitochondrial dysfunction induce the significant release of intracellular ROS. As shown in Figures 5A–C, H₂O₂ exposure increased cellular ROS production reflected by that the fluorescence peak shifted to the right and average fluorescence increased in the H₂O₂ group. However, with the pretreatment of SGR, the fluorescence peak began to shift to the left gradually, they all showed significant differences compared with the H₂O₂ group. It was verified that SGR increased the expression of SOD, revealing SGR did inhibit the oxidative stress in the H₂O₂-induced model (Figure 5D). In addition, SGR treatment also significantly prevented H₂O₂-induced mitochondrial ROS production measured by MitoROS staining. Mitochondrial ROS (MitoROS) is the most important source of intracellular ROS. The results to explore the effect of SGR on the MitoROS production showed that H₂O₂ induced the elevation of MitoROS production, while SGR pretreatment suppressed this elevation (Figure 5E).

A lysosome is an important indicator of oxidative stress. Lysosomal dysfunction is associated with autophagy impairment, senescence and Lysosomal disruption is a hallmark of oxidative induced apoptosis (Rajapakse et al., 2017; Tai et al., 2017). In this study, we found that H₂O₂ inhibited Lysosome content compared with the control group, but with pretreatment of SGR, Lysosomal content was improved (Figures 6A–C), this result also suggested that SGR protected Lysosomal from H₂O₂-induced damage.

To determine whether SGR protected mitochondria damage in response to H₂O₂ injury, we examined H₂O₂-induced alterations in mitochondrial membrane potential (ΔΨm) presented by Mito-tracker live-cell tag probe and the opening of mitochondrial permeability transition pore (mPTP). mPTP opening assay results showed that mPTP was opening because of the damage of mitochondria induced by H₂O₂, but the excessive opening of mPTP was inhibited in the SGR groups, as Figures 7A,C. In addition, we also found that pretreatment with SGR reduced the ΔΨm depolarization, which was indicated by the increased red fluorescence (Figures 7B,D).

Figures 8A–D showed that compared with the H₂O₂ group, the OCR values of basal respiration, maximum respiration and ATP production of the SGR treatment group increased significantly. It revealed that SGR was capable of protecting mitochondrial respiratory and mitochondrial energy metabolism in rat H9c2 cardiomyocytes against H₂O₂-induced mitochondrial dysfunction.

To investigate the possible mechanism of SGR on H₂O₂-induced cellular damage and mitochondrial dysfunction, aiming to verify the pathway predicted by network pharmacological analysis, we explored the SGR modulation on the p38MAPK pathway. It indicated that compared with the H₂O₂ group, SGR decreased the mRNA levels of CDC42, PP2A, RAC1, Akt, NFkB (Figures 9N,K,L,H,I), and increased the mRNA levels of IKK, Sirt1, and Jun (Figures 9M,J,K). WB results revealed that H₂O₂ increased protein expression of ERK1, ERK2, JNKs, Bax, Caspase3 (Figures 9A,C–F) and decreased p38MAPK and Bcl-2 expression (Figures 9B,O,P), but these were all reversed by SGR pretreatment. All suggested that SGR significantly activated p38MAPK pathways that was inhibited by H₂O₂.

Figures 10A–H showed, extracting masses in negative ionization mode for compounds 5 suggested a molecular formula of C₆H₉N₃O₂, C₁₄H₁₂O₃, C₇H₁₀O₅, C₁₅H₁₄O₆, C₁₅H₁₂O₆, with the Molecular Weight of 155.15, 228.24, 174.15, 290.27, 288.25 Da, respectively. By referring to the retention time and secondary fragment ions, Istidina, resveratrol, shikimic acid, epicatechin, and dihydrokaempferol in SGR were accurately identified with the same retention time of 1.20, 24.53, 1.52, 17.47, 23.54 min, the molecular weight of 154.06, 227.07, 173.05, 289.07, 287.06 Da and the same fragment ions.

According to network topology analysis, Istidina and trans-resveratrol were identified as the key active components in SGR. Shikimic acid (6 in degree), epicatechin (7 in degree), and dihydrokaempferol (5 in degree) were also identified by the UHPLC-LTQ-Orbitrap-MSn method. SRC, JUN, CTNNB1, MAPK1, IL6, TNF, NFKB1, IL1B, INS, ESR1, PPARγ, CCL21, PPARA, IGF1, SIRT1, FGF2, IL10, PRKCA, SP1, and SCN5A were the key targets of SGR, and they were docked with the identified component of SGR to clarify the interaction relationships between the key targets and SGR active compound. The results showed that Istidina, Trans-resveratrol, Shikimic acid, Epicatechin, and Dihydrokaempferol had higher docking scores with SRC, JUN, TNF, PPARα, and Sirt1, respectively (Figure 11L), indicating these 5 drugs may have anti-heart failure effect by targeting SRC, JUN, TNF, PPARα, and Sirt1, respectively. Figures 11A–K showed the 2D and 3D docking structure of 5 components docked with their predicted targets.