2.1 Chemicals and reagents

Lipopolysaccharide (LPS) was supplied from Sigma-Aldrich Co. (Shanghai, China). Cell Counting Kit 8 (CCK-8) was purchased from Dojindo Laboratory (Kumamoto, Japan). Fetal bovine serum (FBS) and RPMI 1640 medium were provided by Biological Industries (Kibbutz Beit Haemek, Israel). Information about the antibodies used for immunoblotting is shown in Supplementary Table 1. Unless otherwise noted, all other chemicals were supplied by Sigma-Aldrich.

2.2 Establishment of the “ECG-ingredients-targets” interactions

To construct a network diagram of ECG-ingredients-targets interactions, two suggested drug screening criteria of TCMSP (https://tcmspw.com/tcmsp.php) database (Ru et al., 2014) such as oral bioavailability (OB) threshold ≥ 30 % and drug-likeness (DL) threshold ≥ 0.18, were used to screen the active ingredients and their related targets in ECG. Then, Cytoscape software (Version: 3.6.0) was used to establish a network diagram of “ECG-ingredients-targets” interactions.

2.3 Targets screening for ECG treatment of CB

To obtain CB-related targets, the Drugbank (https://www.drugbank.ca/) (Wishart et al., 2008), Genecards (https://www.genecards.org/) (Rebhan et al., 1997), OMIM (https://omim.org/) (Amberger et al., 2015), and DigSee (http://210.107.182.61/geneSearch/) (Kim et al., 2013) databases were used to screen via searching “Chronic bronchitis” as the keyword. Subsequently, overlapping targets were obtained by screening with ECG active ingredients-related targets from TCMSP, and recognized as potential targets for ECG treatment of CB.

2.4 Construction of a protein–protein interaction network

The screened overlapping proteins were subjected to analysis using the STRING (Version: 11.5) data resources (https://cn.string-db.org/) (Szklarczyk et al., 2017), and then the tsv. format was exported to Cytoscape for construction and visualization of the protein–protein interaction (PPI) network.

2.5 Biological functional enrichment analysis

To further reveal the biological function of the overlapping target, the Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed using DAVID bioinformatics resources (https://david.ncifcrf.gov/) (Huang et al., 2009), a comprehensive set of functional annotation tools to find out the biological function and molecular pathways of the selected targets.

2.6 Molecular docking

Molecular docking of ECG active ingredients to targets was analyzed as previously described (Luo et al., 2022; Liu et al., 2018; Xie et al., 2019). Briefly, the structures of ECG active ingredients were retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) (Wang et al., 2009) and optimized using molecular mechanics to obtain the most stable systems. The RCSB PDB database was then used to retrieve the 3D crystal structure of each target (https://www.rcsb.org/) (Burley et al., 2019) and processed for structural correction. Molecular docking of active ingredients and targets were analyzed using SYBYL-X 2.2.1 software. The docking mode was set as Suflex-Dock Geom (SFXC), and then the protein was subjected to Hydrogenation and Charging, as well as Protomol Generation for Muti-channel Surface to obtain SFXC files. Next, fully flexible docking was selected for optimal docking conformation. Finally, active ingredients-proteins binding pattern schematization and related binding amino acids were analyzed using Discover Studio Visualizer 2016. The box sizes for molecular docking targets are shown in Supplementary Table 2.

2.7 Molecular dynamics simulation

To further investigate the stability of ECG active ingredients with key targets of CB in the binding pocket. Molecular dynamics (MD) simulations of 100 ns were done using the GROMACS v.2020.4 software package as previously described (Luo et al., 2022). Briefly, the crystal structures of nuclear factor κB (NF-κB) p65 (PDB_ID: 5URN) and Caspase3 (PDB_ID: 5IAE) proteins were obtained from the PDB bank. The molecular structures of the small molecules were retrieved from the PubChem database (Wang et al., 2009). To eliminate close contact between atoms, energy minimization of proteins using the steepest descent method. Energy optimization of small molecule structures using Chem3D16 and molecular docking was done by SYBYL-X 2.2.1 software. MD simulations were run with constant temperature, pressure, and periodic boundary conditions. The amber99sb-ildn force field and the TIP3P water model were used. The acpype.py script in AmberTools generates the force field parameters for small molecules. During the MD simulation, the LINCS technique with a 2.0 fs integration time step was used to limit hydrogen bonding. The Particle Mesh Ewald (PME) technique was used to compute the electrostatic interactions with a truncation value of 1.2 nm and the non-bonding interactions truncation value of 10 Å and every ten steps updated. All the MD simulations were run using the Berendsen method and V-rescale temperature coupling at constant temperatures of 300 K and 1 bar, respectively. Then, 100 ps of NVT and NPT equilibrium simulations at 300 K, followed by 100 ns of molecular dynamics simulations. For the protein–ligand complexes, a 100 ns MD simulation was completed. The binding free energy of complexes was determined using g_mmpbsa (Kumari et al., 2014). The output of the MD simulation was visualized using the software-embedded programs of GROMACS and Discover Studio Visualizer 2016.

2.8 ECG preparation

ECG herbal medicine (Batch No.: Z46-20220901), a monarch medicine in “Zhike Juhong Granules” prescriptions produced by Honz Pharmaceutical Co., Ltd. In the current study, the herb of ECG (100.1 g) was refluxed for 1 h in an 8-fold mixture of 50 % ethanol and water (v/v). After filtering through gauze, labeled as filtrate 1. Then the residue was refluxed with a 6-fold mixture of 50 % ethanol and water (v/v) for a further hour and also filtered through gauze, labeled as filtrate 2. Next, filtrate 1 and filtrate 2 were combined and concentrated. The concentrate was lyophilized at the end with a yield of 41.61 %, and the lyophilized sample was further analyzed by UPLC-Q/TOF-MS/MS (Supplementary Fig. 1).

2.9 Culture of cell

RAW264.7 Cells were cultured in RPMI 1640 medium (with 100 U/mL of penicillin, 100 μg/mL of streptomycin, and 10 % FBS) and incubated in a wet incubator at 37 ℃ with 5 % CO₂. Cells should be seeded onto culture dishes or plates for 24 h before performing all experiments. Cells were pre-treated with different concentrations of ECG (25, 50, 100 μg/mL) for 1 h and then exposed to 0.5 µg/mL LPS. After 24 h, the cells or the cell culture medium were then collected for subsequent analyses.

2.10 Cell viability assay

To determine the effect of ECG on cell viability, 2 × 10⁴ cells per well were added to the 96-well plate. After 24 h, cells were treated with various concentrations of ECG for 24 h, and the blank group of cells was added with the same amount of complete culture medium. Finally, cell viability was assessed according to CCK-8 assay requirements.

2.11 Griess assay

The cell culture medium level of nitric oxide (NO) produced by cells was measured according to the instructions of the NO kit (Beyotime Biotechnology, China, S0021S).

2.12 Measurement of IL1β, IL6 and TNF-α

IL1β, IL6, and TNF-α levels in cell culture medium were measured according to the instructions of commercial kits (eBioscience, USA, 88-7013A-86, 88–7064-86, 88–7324-86).

2.13 RNA sequencing

Total RNA was extracted from cells with Trizol reagent and reversed transcribed into cDNA, followed by PCR amplification and purification to obtain PCR products for transcriptomics as previously described (Li et al., 2020), Quantification and detection were performed using a Qubit2.0 Fluorometer and Agilent 2100 bioanalyzer to detect libraries. The Illumina NovaSeq 6000 platform was used to sequence and generate 150 bp paired-end reads for subsequent high-quality analysis. Indexes of the reference genome were constructed using HISAT2 (v2.0.5) and compared to the reference genome. Read mapped to each gene was calculated by featureCounts (1.5.0-p3) for the FPKM of each gene. Differential expression genes (DGEs) between groups were identified using DESeq2 software (1.20.0) and the resulting p values (padj) were adjusted by Benjamini and Hochberg's method to control the false discovery rate. GO enrichment analysis and KEGG pathway analysis of DGEs were performed respectively using clusterProfiler (3.8.1) software.

2.14 Immunofluorescence

Cells were perforated with 0.5 % Triton X-100 at room temperature for 20 min after being fixed with 4 % paraformaldehyde for 15 min. Then, cells were blocked by 5 % BSA solution for 1 h, incubated by NF-κB p65 antibody (1: 100) overnight at 4 ℃. After being washed, cells were incubated by fluorescent secondary antibody (1: 1000) for 1 h at room temperature. Finally, cells were covered by an antifade mounting medium containing DAPI. Pictures were observed under a fluorescence microscope (Nikon, Shanghai, China).

2.15 Apoptosis assay

Cells were collected and stained with Annexin V-FITC and propidium iodide (PI) following the instructions of the Annexin V-FITC apoptosis detection kit (Beyotime Biotechnology, China, C1062L). Images were observed by fluorescence microscope (Leica, Wetzlar, Germany).

2.16 Western blot

Protein samples were extracted from cells after treatment, and their concentrations were measured by a BCA kit (Beyotime Biotechnology, China, P0010). Proteins were separated by 10 % SDS-polyacrylamide gel and transferred to PVDF membranes. The membranes were then incubated with different primary antibodies overnight at 4 ℃, and followed by a secondary antibody. ECL kit (Guangdong Prochem Biotechnology Co.,Ltd. China, ECL-0102) was used and images were obtained using Molecular Imager® (BIO-RAD, CA, United States).

2.17 Statistical analysis

The values were expressed as the mean ± SD and analyzed using GraphPad Prism 6.0 (San Diego, CA, USA). The significance of the differences was analyzed using a one-way ANOVA followed by Bonferroni’s multiple comparisons test. P < 0.05 was considered statistically significant.

3.1 Construction of a “ECG-ingredients-targets” network

To obtain the active ingredients of ECG, the TCMSP database was used for screening. In this study, 10 active ingredients were recognized (Supplementary Table 3), and 99 potential targets were predicted. As shown in Fig. 1, the ECG-active ingredients-related targets network was constructed.

Close

Fig. 1

Network pharmacology predicted the active ingredients-targets interactions for ECG. (Nodes with orange, white, and blue represent ECG, ingredients, and targets, respectively.). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Export to PPT

3.2 PPI network construction

The active ingredients-related targets of ECG may be closely associated with CB-related targets obtained from Drugbank, Genecards, OMIM, and DigSee databases. To further investigate the underlying pharmacological effects of ECG on CB, a Venn diagram was utilized to visualize the correlation. As shown in Fig. 2A, 46 overlapping targets were screened, and their interactions were analyzed through STRING data resource. Then, the PPI network was constructed through Cytoscape software and presented 46 nodes and 1204 edges. Among them, AKT1 (degree = 42), MAPK3 (degree = 41), PTGS2 (degree = 39), CASP3 (degree = 38), NOS3 (degree = 34), MAPK1 (degree = 34), PTGS1 (degree = 33), MAPK8 (degree = 32), PIK3CG (degree = 31) and RELA (degree = 30) with a higher degree were most likely involved in the inflammations and pathogenesis processes of CB (Fig. 2B).

Close

Fig. 2

Venn diagram of ECG and CB-related targets (A), PPI network of potential targets (B).

Export to PPT

3.3 GO and KEGG pathway enrichment analysis

In the present study, GO enrichment analysis based on DAVID bioinformatics resources revealed that overlapping targets were involved in response to the drug, adenylate cyclase-inhibiting G-protein coupled acetylcholine receptor, positive regulation of apoptotic process, response to lipopolysaccharide, cellular response to reactive oxygen species, negative regulation of the apoptotic process, G-protein coupled acetylcholine receptor in biological processes (Fig. 3A). Enzyme binding, protein homodimerization activity, G-protein coupled serotonin receptor activity, ubiquitin protein ligase binding, protease binding, NADP binding, transcription coactivator binding, histone deacetylase binding, protein serine/threonine kinase activity, MAP kinase activity, enzyme binding, protein homodimerization activity in molecular functions (Fig. 3B). Plasma membrane, synapse, macromolecular complex, cytoplasm, cytosol, Golgi apparatus, mitochondrion, nucleus, nucleoplasm, endoplasmic reticulum in cellular components (Fig. 3C), suggesting that ECG may play a pharmacological role in treating CB through these biological functions.

Close

Fig. 3

GO enrichment and KEGG pathways analysis of the anti-CB targets of ECG. (A) Biological Processes, (B) Cellular Components, (C) Molecular Function, (D) KEGG Pathway.

Export to PPT

Furthermore, KEGG pathway enrichment analysis revealed the TNF signaling pathway, the IL-17 signaling pathway, and the PI3K-Akt signaling pathway may be remarkably related to the pathogenesis of CB, indicating that ECG can act as a therapeutic CB by ameliorating the inflammatory response that is closely linked to these pathways (Fig. 3D and Supplementary Table 4).

3.4 Interactions between active ingredients with targets

The above network pharmacology analysis highlighted the important role of ECG active ingredients in anti-inflammatory effects. Chemical studies showed that naringin, naringenin, and neohesperidin were highly abundant in ECG herbal medicine (Liu et al., 2022; Kong et al., 2022; Jiang et al., 2014). In this study, the experimental ECG sample was also examined by UPLC-Q/TOF-MS/MS technology. As expected, naringin, naringenin, and neohesperidin were identified (Supplementary Fig. 1, Supplementary Fig. 2, and Supplementary Table 5). Based on the results above, the screened active ingredients with key targets of CB inflammatory signaling, such as myeloid differentiation factor 88 (MyD88) and NF-κB p65 (RelA), were investigated in the binding mode by molecular docking (Supplementary Table 6–9). An ingredient binding to the target with a total score > 7 and a C-score close to 5 indicates strong activity (Liu et al., 2018; Xie et al., 2019). The results showed that neohesperidin docking to MyD88 (PDB_ID: 4DOM) gave a total score and C-score of 9.2265 and 1, respectively (Supplementary Table 7). The binding modes were that neohesperidin formed hydrogen bonds with protein residues SER 209, SER 194, ILE 165, and ASP 195, formed carbon-hydrogen bonds with protein residues LYS 231, TYR 167, interacted with van der Waals forces with protein residues ILE 212, PHE 235, VAL 193, CYS 168, ILE 172, PHE 164, GLU 232, hydrophobic interactions in the form of π bond with protein residues PRO 169, VAL 198, and formed lone pair–π interaction with protein residue TYR 167 (Fig. 4A-C). Naringenin docking to MyD88 (PDB_ID: 7BER) gave a total score and C-score of 7.2349 and 0, respectively (Supplementary Table 7). The binding modes were that naringenin formed hydrogen bonds with protein residues SER 209, SER 194, ILE 165, and ASP 195, formed a carbon-hydrogen bond with protein residue ILE 253, interacted with van der Waals forces with protein residues THR 272, PRO 254, LEU 295, LEU 293, ARG 218, ARG 288, and hydrophobic interactions in the form of π bond with protein residues LYS 250, ILE 271 (Fig. 4D-F). Similarly, neohesperidin docking to NF-κB p65 (PDB_ID: 5URN) gave a total score and C-score of 10.9476 and 2, respectively (Supplementary Table 9). The binding modes were that neohesperidin formed hydrogen bonds with protein residues GLU 10, MET 88, SER 535, LYS 11, MET 58, formed carbon-hydrogen bonds with protein residues ILE 8, PHE 9, HIS 87, ARG 86, MET 88, LYS 11, interacted with van der Waals forces with protein residues GLU 532, ASP 531, PHE 534, ILE 537, ASP 541, PHE 89, and hydrophobic interactions in the form of π bond with protein residues MET 59, LYS 57, ALA 538, MET 88 (Fig. 4G-I). Naringenin docking to NF-κB p65 (PDB_ID: 5URN) gave a total score and C-score of 8.0717 and 2, respectively (Supplementary Table 9). The binding modes were that naringenin formed hydrogen bonds with protein residues MET 88, GLU 10, and ASP 531, formed a carbon-hydrogen bond with protein residue ASP 531, interacted with van der Waals forces with protein residues PHE 534, MET 59, SER 90, PHE 89, ILE 8, PHE 9, SER 535, GLU 532, GLY 530, hydrophobic interactions in the form of π bond with protein residue MET 88, and interacted with protein residue ASP 531 as a π-bond in a lone pair electron cloud (Fig. 4J-L).

Close

Fig. 4

Molecular docking analysis revealed the bond pattern of neohesperidin and naringenin with MyD88 and NF-κB p65, respectively. (A-C) Molecular docking of neohesperidin and MyD88, (D-F) Molecular docking of naringenin and MyD88, (G-I) Molecular docking of neohesperidin and NF-κB p65, (J-L) Molecular docking of naringenin and NF-κB p65. (A, D, G, J): Areas of the donor and acceptor of hydrogen bond (H-bond). (B, E, H, K): Areas of hydrophobicity. (C, F, I, L): Two-dimensional patterns of the bond.

Export to PPT

In response to cellular stress, nucleolar translocation of NF-κB p65 directly affects a nucleolar pathway/protein(s) to mediate apoptosis (Khandelwal et al., 2011). Caspases belonging to cysteinyl proteases, such as Caspase3, 8, and 9, were well-characterized roles in apoptosis and cytokine maturation. In the present study, neohesperidin binding with Caspase9 (PDB_ID: 1JXQ) provided a total score and C-score of 15.4065 and 2, respectively (Supplementary Table 11). The binding modes were that neohesperidin formed hydrogen bonds with protein residues PRO 271, SER 144, ASN 148, ASP 150, ASP 326, ARG 395, TYR 383, LEU 321, GLY 275, formed carbon-hydrogen bonds with protein residues GLY 274, PRO 271, SER 144, GLY 267, interacted with van der Waals forces with protein residues THR 268, ASN 266, PRO 322, THR 323, LYS 276, LEU 273, SER 272, GLY 147, LEU 145, ALA 149, SER 325, ILE 382, MET 386, SER 37, formed T-type interaction forces with protein residues TYR 383, GLY 267 at π-π bonds with amide bonds (Fig. 5A-C). Naringenin binding with Caspase9 (PDB_ID: 4RHW) provided a total score and C-score of 8.905 and 3, respectively (Supplementary Table 11). The binding modes were that naringenin formed hydrogen bonds with protein residues SER 23, LYS 21, ASP 53, PRO 47, GLN 57, formed carbon-hydrogen bonds with protein residues GLY 274, PRO 271, SER 144, GLY 267, interacted with van der Waals forces with protein residues THR 22, SER 48, ARG 52, hydrophobic interactions in the form of π bond with protein residues ARG 44, ALA 46, and forms an interaction force with protein residues GLY 47 via π-amide bond (Fig. 5D-F). Neohesperidin docking to Caspase3 (PDB_ID: 5IAE) provided a total score and C-score of 8.4479 and 4, respectively (Supplementary Table 13). The binding modes were that neohesperidin formed hydrogen bonds with protein residues TYR 197, LYS 137, and GLU 167, formed carbon-hydrogen bonds with protein residues GLY 125, TYR 195, GLU 124, interacted with van der Waals forces with protein residues PHE 266, GLY 202, ARG 164, TYR 197, LYS 137, LEU 136, TYR 195, THR 140, hydrophobic interactions in the form of π bond with protein residues TYR 105, PRO 201, PRO 201 (Fig. 5G-I). Naringenin docking to Caspase3 (PDB_ID: 5IAE) provided a total score and C-score of 7.9554 and 2, respectively (Supplementary Table 13). The binding modes were that naringenin formed hydrogen bonds with protein residues GLU 190, GLU 124, LYS 137 GLY 125, formed a carbon-hydrogen bond with protein residue LYS 137, interacted with van der Waals forces with protein residues ALA 200, MET 268, PHE 266, ARG 164, ASP 135, TYR 197, ILE 160, THR 140, hydrophobic interactions in the form of π bond with protein residues PRO 201, LEU 136, and formed T-type interactions with protein residue TYR 195 through π-π bonds (Fig. 5J-L). Detailed information on the docking studies for other active ingredients was provided in Supplementary Table 10–13. All these results suggested that the ECG active ingredients could bind directly with MyD88, NF-κB p65, Caspase3, and Capase9, thereby inhibiting the inflammatory response and apoptosis to improve CB.

Close

Fig. 5

Molecular docking analysis revealed the bond pattern of neohesperidin and naringenin with Caspase9 and Caspase3, respectively. (A-C) Molecular docking of neohesperidin and Caspase9, (D-F) Molecular docking of naringenin and Caspase9, (G-I) Molecular docking of neohesperidin and Caspase3, (J-L) Molecular docking of naringenin and Caspase3, (A, D, G, J): Areas of the donor and acceptor of hydrogen bond (H-bond). (B, E, H, K): Areas of hydrophobicity. (C, F, I, L): Two-dimensional patterns of the bond.

Export to PPT

3.5 The dynamics stability simulation of active ingredients with targets

Research evidence has convincingly shown that the first line of protection against invading pathogens is innate immunity (Cameron et al., 2016). NF κB is a pleiotropic transcription factor involved in inflammatory and immune responses (Baeuerle and Henkel, 1994; Baeuerle and Baltimore, 1996). In the docking studies above, ECG active ingredients showed strong binding activity with the critical inflammatory target (RELA/p65). Thus, an MD simulation was carried out to further understand the binding pocket's stability. Root mean square deviation (RMSD) represents the sum of all atomic deviations of the conformation at a given moment from the target conformation, which is an essential measure of the system’s stability. As shown in Fig. 6A, the neohesperidin-NF-κB p65 system was in an equilibrium state from 30 to 100 ns with a mean RMSD value of 0.1168 ± 0.020, and the RMSD values of the whole system converged, and the system reached equilibrium. In contrast, the naringenin-NF-κB p65 system dramatically rose at 32 ns and then converged to equilibrium in the last 5 ns of time with a mean RMSD value of 0.1693 ± 0.017. In addition, the variation of solvent-accessible surface area (SASA) with time showed that the more converged values, the more structural stability. The neohesperidin-NF-κB p65 system and naringenin-NF-κB p65 system gradually stabilized in the 75 to 100 ns period with mean values of 206.3 ± 2.970 nm² and 202.4 ± 2.122 nm² (Fig. 6B), respectively. The results of the binding free energy analysis showed that the van der Waal energy between neohesperidin and NF-κB p65 was –222.503 ± 3.619 kJ/mol, the electrostatic energy was −128.713 ± 3.148 kJ/mol, the polar solvation energy was 325.487 ± 3.635 kJ/mol, the SASA energy was −25.258 ± 0.310 kJ/mol, and the total binding energy was −50.958 ± 3.175 kJ (Fig. 6C). By contrast, the van der Waal energy of naringenin with NF-κB p65 was −111.626 ± 1.690 kJ/mol, the electrostatic energy was −25.214 ± 1.475 kJ/mol, the polar solvation energy was 86.682 ± 2.350 kJ/mol, the SASA energy was −12.275 ± 0.206 kJ/mol, and the total binding energy was −62.510 ± 2.533 kJ/mol (Fig. 6D). The main contribution of amino acid residues in the binding regions of neohesperidin and naringenin to the binding free energy of NF-κB p65 were shown in Supplementary Table 14 and Supplementary Table 15.

Close

Fig. 6

Molecular dynamics simulations confirmed the stability of neohesperidin and naringenin with NF-κB p65 and Caspase3 in the binding pocket, respectively. (A) RMSD analysis for NF-κB p65 with neohesperidin and naringenin complexes. (B) The plot of SASA of NF-κB p65 with neohesperidin and naringenin complexes. (C) Binding free energy decomposition of neohesperidin-NF-κB p65 complex. (D) Binding free energy decomposition of naringenin-NF-κB p65 complex. (E) RMSD analysis for Caspase3 with neohesperidin and naringenin complexes. (F) The plot of SASA of Caspase3 with neohesperidin and naringenin complexes. (G) Binding free energy decomposition of the neohesperidin-Caspase3 complex. (H) Binding free energy decomposition of the naringenin-Caspase3 complex.

Export to PPT

As a crucial component of apoptosis, caspase3 is involved in the proteolytic cleavage of numerous essential proteins. (Fernandes-Alnemri et al., 1994). In the current study, the mean RMSD of the neohesperidin-Caspase3 system at equilibrium was 0.1352 ± 0.010. Similarly, the mean RMSD of the naringenin-Caspase3 system at equilibrium was 0.1151 ± 0.016 (Fig. 6E). The RMSD value of the whole system converged, and the system reached equilibrium. Furthermore, the results of SASA for the neohesperidin-Caspase3 and naringenin-Caspase3 systems gradually stabilized in the 90–––100 ns period with mean values of 96.02 ± 1.597 nm² and 102.2 ± 2.109 nm² (Fig. 6F), respectively. The results of the binding free energy analysis showed that the van der Waal energy between neohesperidin and Caspase3 was −255.197 ± 2.682 kJ/mol, the electrostatic energy was −31.387 ± 3.454 kJ/mol, the polar solvation energy was 211.44 ± 4.284 kJ/mol, the SASA energy was −26.469 ± 0.169 kJ/mol, and the total binding energy was −91.684 ± 3.950 kJ (Fig. 6G). Equally, the van der Waal energy of naringenin with Caspase3 was −123.763 ± 2.859 kJ/mol, the electrostatic energy was −74.308 ± 2.933 kJ/mol, the polar solvation energy was 165.957 ± 2.742 kJ/mol, the SASA energy was −14.316 ± 0.126 kJ/mol, and the total binding energy was −46.575 ± 2.933 kJ/mol (Fig. 6H). The main contribution of amino acid residues in the binding regions of neohesperidin and naringenin to the binding free energy of Caspase3 was shown in Supplementary Table 16 and Supplementary Table 17.

All of these MD simulations revealed that the active ingredients of ECG were stable in the binding pocket with NF-κB p65 and Caspase3, indicating that neohesperidin and naringenin were capable of binding directly interacting with critical targets in the inflammatory response and apoptotic signaling to exert their anti-CB pharmacological effects. Notably, these findings were in line with the results of network pharmacological and molecular docking studies discussed above.

3.6 ECG inhibited LPS-stimulated NF-κB signaling activation

Toll-like receptor 4 (TLR4), a macrophage LPS signaling receptor, cooperates with LY96 and CD14 to mediate the innate immune response and acts via MyD88, leading to NF-κB activation, thereby triggering the release of various inflammatory mediators (Raetz and Whitfield, 2002), such as IL1β, IL6, and TNF-α (Beutler and Cerami, 1988; Dinarello, 1991; Qu et al., 2020). Due to limitations in the establishment of animal models of CB, an in vitro LPS-induced RAW 264.7 macrophage inflammation model was used to investigate the anti-inflammatory mechanisms of ECG on CB in this study. As shown in Fig. 7A, ECG was not toxic to RAW264.7 cells but promoted cell proliferation above 200 ug/mL according to CCK-8 results. Exposure to LPS resulted in significant production of IL1β, IL6, TNF-α, and NO, while ECG pre-treatment significantly reduced those cytokines secretion (Fig. 7B-E). The transcriptome results showed that a total of 670 DEGs were screened in the LPS and CTRL (Con) comparison groups, of which 379 were up-regulated and 291 were down-regulated (Supplementary Table 18). Meanwhile, a total of 337 DEGs were screened between ECG and LPS groups, and among them, 104 were up-regulated and 233 were down-regulated (Fig. 7F-H, and Supplementary Table 19). By further KEGG analysis of these DEGs, the NF-kappa B signaling pathway, the Toll-like receptor signaling pathway, the IL-17 signaling pathway, and the TNF signaling pathway were found to be closely related pathways, indicating that ECG may significantly inhibit the activation of inflammatory signaling in ameliorating the development of CB (Fig. 7I-J). Furthermore, immunofluorescence and immunoblotting results confirmed the inhibition of TRL4/MyD88/NF κB signaling by ECG and showed that ECG significantly inhibited NF-κB p65 entry into the nucleus (Fig. 7K), down-regulated the expression of TLR4, MyD88, NF-κB p65, NF-κB p-p65 (S536), p-IκBα (S36), COX2, ICAM1, and iNOS proteins and up-regulated the expression of IκBα protein in LPS-induced RAW264.7 macrophages (Fig. 7L), suggesting that ECG could inhibit NF-κB signaling transduction, and thus resist the onset and progression of inflammation closely related to CB.

Close

Fig. 7

ECG inhibited LPS-stimulated TLR4/MyD88/NF-κB p65 signaling in RAW 264.7 cells. (A) Cell viability. IL1β (B), IL6 (C), and TNF-α (D) levels in culture supernatants were measured using ELISA kits. (E) Production of NO. (F) The hierarchical clustering heat maps of DGEs (n = 3). (G) Volcano map of LPS vs. CTRL (Con). (H) Volcano map of ECG vs. LPS. (I) KEGG pathway analysis for DGEs of LPS vs. CTRL (Con). (J) KEGG pathway analysis for DGEs of ECG vs. LPS. (K) Representative images of immunofluorescence of NF-κB p65. (L) Representative Western blots showing the proteins of TLR4, MyD88, NF-κB p65, NF-κB p-p65 (S536), IκBα, p-IκBα (S36), COX2, ICAM1, and iNOS changes and normalized to β-actin. Similar results were obtained from three independent experiments. Data are shown as the mean ± SD. *p < 0.05, ^**p < 0.01, ^***p < 0.001, and ^****p < 0.0001 vs. the control group. ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001 vs. the LPS group.

Export to PPT

3.7 ECG inhibited LPS-stimulated apoptosis

It is well known that activation of NF-κB signaling has numerously been demonstrated to be associated with apoptosis (Khandelwal, 2011; Lawrence et al., 2001; Lo et al., 2011). The above studies demonstrated the significant inhibitory ability of ECG on NF-κB signaling. To further explore the pharmacological mechanisms of ECG, Annexin V-FITC/PI staining and immunoblotting were carried out to study the effect of ECG on LPS-stimulated apoptosis in RAW264.7 cells. As expected, LPS stimulation of RAW264.7 macrophages resulted in significant apoptosis. Interestingly, pre-treatment with ECG suppressed this in a dose-dependent manner (Fig. 8A). Additionally, ECG significantly up-regulated the expression of anti-apoptotic protein Bcl-2 and down-regulated the expression of pro-apoptotic proteins including Bax, Caspase3, Caspase8, and Caspase9 (Fig. 8B), indicating that ECG played an essential role in the inflammatory response by inhibiting apoptosis.

Close

Fig. 8

ECG inhibited LPS-stimulated apoptosis in RAW 264.7 cells. (A) Cells were stained with Annexin V-FITC, and PI was analyzed by fluorescence microscope. (B) Representative Western blots showing the proteins of Bcl-2, Bax, Caspase3, Caspase8, and Caspase9 changes and normalized to β-actin. Similar results were obtained from three independent experiments.

Export to PPT