Active compounds
775 compounds were ultimately reserved: 87 in Radix Astragali, 189 in Chuanxiong Rhizoma, 119 in Radix paeoniae Rubra, 125 in Angelicae Sinensis Radix, 66 in Persicae Semen, 189 in Carthami Flos, and 4 in Pheretima. After ADME screening, 78 potential compounds (OB≥30%, DL≥0.18, Caco-2≥0.4) of seven herbal medicines in BHD were identified, including 16 from Radix Astragali, 6 from Chuanxiong Rhizoma, 20 from Radix Paeoniae Rubra, 2 from Angelicae Sinensis Radix, 13 from Persicae Semen,21 from Carthami Flos, and 0 from Pheretima. The details of candidate ingredients are described in Table 1. Radix Atragali, Chuanxiong Rhizoma, Radix Paeoniae Rubra, Angelicae Sinensis Radix, Persicae Semen and Carthami Flos are simplified as RA, CR, RPR, ASR, PS, and CF respectively.
Targets of BHD
As 18 compounds of BHD had no targets in TCMSP, SymMap and TCM-MESH, Canonical SMILES of these compounds were found in Pubchem. Based on chemical structural similarity, we used databases like STITCH, SWISS and SEA, to predict their targets. These compounds were excluded because of the targets score less than 50% eventually. These compounds were isoflavanone, senkyunone, paeoniflorin_qt, Albiflorin_qt, Paeo niflorigenone, 9-ethyl-neo-paeoniaflorinA_qt, evofolinB, 1-o-beta-d-glucopyranosylpaeonisuffrone_qt, 4-ethyl-paeoniflorin_qt, 4-o–methyl-pa eoniflorin_qt, GA122, Popul oside_qt, Flavoxanthin, lupeol-palmitate, Phytoene, phytofluene, 6-Hydroxynaring enin, and 1-o-beta-d-glucopyranosyl-8-o–benzoylpaeo nisuffrone_qt. In brief, 235 targets were adopted in this research.
By means of the six available resources, namely, DisGeNET, OMIM, TTD, DrugBank, PharmGKB and MalaCards databases, we obtained 460 IS-related targets.
Based on targets of the candidate ingredients and IS,intersection targets were got by R software. 62 intersection genes were found eventually, shown as Fig. 3A. The details of intersection targets are described in Table 2.
Gene Ontology and Pathway Enrichment Analysis
Gene Ontology
GO analysis of 62 candidate targets for BHD against IS was performed using the DAVID database to understand the relationship between functional units and their underlying significance in the biological system networks. The result was divided into three parts, biological processes, cellular component, and molecular function, as shown in Fig. 2 A, B, and C.
We found that biological processes were related to inflammatory response, negative regulation of apoptotic process, response to estradiol, lipopolysaccharide-mediated signaling pathway, positive regulation of nitric oxide biosynthetic process, positive regulation of protein phosphorylation, positive regulation of NF-kappa B transcription factor activity, response to hypoxia, immune response and MAPK cascade. The cellular component was related to extracellular space, extracellular region, plasma membrane, cell surface, extracellular exosome, cytosol, nucleus, nucleoplasm, cytoplasm and integral component of plasma membrane. Finally, molecular function was related to cytokine activity, enzyme binding, protein binding, heme-binding, steroid hormone receptor activity, peroxidase activity, transcription factor binding, calcium ion binding, metallopeptidase activity and kinase activity.
Pathway Enrichment
Through comprehensive analysis, we obtained an integrated IS pathway based on our current knowledge of IS pathogenesis to illuminate the integral role of BHD in treating IS. TOP 10 KEGG signaling pathways of BHD were obtained and constructed based on P-Value.
Based on this systems-level picture, we picked and constructed ten therapeutic pathways of TNF signaling pathway, PI3K-Akt signaling pathway, NF-kappa B signaling pathway, MAPK signaling pathway, Complement and coagulation cascades, T cell receptor signaling pathway, Toll-like receptor signaling pathway, HIF-1 signaling pathway, Estrogen signaling pathway and VEGF signaling pathway, as shown in Fig. 2D.
BHD-IS PPI network analysis
BHD-IS PPI network
62 intersection targets were imported into the String database, and TSV text showing the interaction relationship was obtained, as shown in Fig. 3B. Then, the network topology analysis was applied by the software of Cytoscape 3.6.0. By integrating IS network and compound-compound target network, we could get BHD-IS network, as shown in Fig. 3C. This network contained 59 nodes and 664 edges. In this network, the rose red nodes had higher degrees. The number of those nodes’ edges was 36 in IL6, 31 in TNF, 29 in VEGFA, 28 in AKT1, 27 in MMP9, 26 in IL1B, 23 in MAPK1, 22 in ICAM1, 22 in PTGS2 and 20 in IL10 respectively. This suggested that these genes might be the key or central genes in IS development. Bar graph of all protein nodes degree related to the targets is shown in Fig. 3D. The target proteins in the PPI network were modularized and analyzed by using the plug-in of cluster maker of the software Cytoscape 3.6.0. The results showed that 62 targets were divided into four modules, including 24 in module one, 18 in module two, 11 in module three and 9 in module four, as shown in Fig. 3E.
Compound-compound target- IS network analysis
This network was composed of 284 nodes (253 compound target nodes and 49 compound nodes) and 1192 edges. In this network we could find that one target could be hit by several compounds (central nodes, such as IL6, MMP9, TNF, AKT1, ICAM1, IL1B, PTGS2, IL-10, VEGFA, and MAPK1), but some were modulated by only one compound in this network. Furthermore, one potential active compound could correspond to multiple targets. Top ten compounds with high degree were Quercetin, beta-Sitosterol, Kaempferol, Stigmasterol, Baicalein, Luteolin, Hederagenin, 7-O-methylis omicron ulatol, Formononetin, Isorhamnetin, and Myricanone. It could be seen that the neuroprotective mechanism of BHD had the characteristics of multi-component, multi-target and multi mechanism. The compound-compound target-IS network is shown in Fig. 4.
Pathway of drug-compound target-IS network
By importing all targets into DAVID, we could get 20 IS-related pathways. Radix Astragali and Carthami Flos had the highest degree, which means that the two herbs might be the main herbs in treating IS. Meanwhile, TNF signaling pathway showed the highest degree, followed by PI3K-Akt signaling pathway, MAPK signaling pathway, NF-kappa B signaling pathway, Toll-like receptor signaling pathway, and T cell receptor signaling pathway respectively. The pathway of drug-compound target-IS network is shown in Fig. 5.
Molecular docking verification
Compound-target interactions with binding energy less than -5.0 kcal/mol are shown in Fig. 6, including IL6 with Luteolin(A), MMP9 with Luteolin(B), TNF with Kaempferol(C), AKT1 with Kaempferol(D), ICAM1 with Kaempferol(E), IL1B with Quercetin(F), PTGS2 with Quercetin(G), IL-10 with Quercetin(H), VEGFA with Baicalein(I), and MAPK1 with 7-O-methylisomucronulatol(J).
Target path analysis
The pathway map of BHD in treating IS was obtained from KEGG PATHWAY Database, as shown in Fig. 7. The related pathways were marked in red, and the targets of BHD in treating IS were marked in rose red. The results showed that the main pathways of BHD in treating IS included TNF signaling pathway, MAPK signaling pathway, NF-κB signaling pathway and PI3K/AKT signalling pathway.