Shenlian Capsule C-T network analysis
Twenty-six of the 143 potential active compounds obtained by the ADME model were excluded due to they did not predict a target for interaction with humans. The C-T network of the Shenlian Capsule consists of 491 nodes and 3390 edges. The network node consists of 117 compounds and 264 targets (Fig. 1E). It can be known from the C-T network that different nodes play different roles in the network. For example, some target nodes have large degree value in the network, indicating that there are more compounds involved in regulating the target which means that the target is more strongly regulated by Shenlian Capsule. For instance, the average degree value of all regulated targets is 10.50, but the degree values of the top five regulated target nodes are: AR: 130, ESR1: 124, PTGS2: 107, PIM1: 103, CDK2: 99.
Shenlian Capsule C-D-T network analysis
The network analysis plug-in of Cytoscape software were used to intersect the C-T network with the D-T network (Fig. 2B) to obtain a total of 47 targets (Fig. 2C). The C-D-T network was constructed by combining intersection targets with compounds (Fig. 2D). The C-D-T network consists of 141 nodes and 541 edges. The nodes are composed of 47 NSCLC related targets and 94 potential compound nodes. By analyzing the C-T-D network, 11 disease-related targets were be found with degree value greater than 10,which PTGS2, CDK2, JUN, TNF, BCL2, IL6, BAX, AKT1, PPARG, HMOX1 and MMP1 target proteins were sequentially 107, 97, 19, 17, 15, 14, 13, 12, 12, 11, 11 regulatory numbers of potential active compounds. In addition, from the C-D-T network, potential active compounds in Shenlian Capsule can regulate multiple protein targets. For example, Quercetin can regulate 39 disease-related targets, luteolin regulates 24 disease-related targets and kaempferol regulate 15 Disease-related targets. Through analysis C-D-T network find that Shenlian Capsule can treat NSCLC through multi-component and multi-target.
Analysis of target protein enrichment based on C-D-T network
By enriching the biologiacal process of target proteins in the C-D-T network, Shenlian Capsule can regulate angiogenesis, nitric oxide synthesis, cell proliferation, cell apoptosis, and cell response to DNA damage stimuli, ERK1 and ERK2 Positive cascade regulation, protein phosphorylation, ERBB2 signaling pathway and other processes to fight NSCLC. Through molecular function analysis of Go enrichment analysis, it was found that the target nodes in the C-D-T network are mainly related to enzyme binding, transcription factor binding, protein phosphatase binding, receptor signaling protein tyrosine kinase activity and other functions. The cellular component analysis of the Go enrichment analysis found that the target nodes in the C-D-T network mainly acted on the nucleus and cytosol of the cells (Fig. 2E).
According to pathway enrichment analysis of 47 disease-related proteins, it can be seen that Shenlian Capsule are enriched in multiple pathways. Such as Pathways in cancer, T cell receptor signaling pathway, B cell receptor signaling pathway, TNF signaling pathway, NF-kappa B signaling pathway, PPAR signaling pathway, p53 signaling pathway, Wnt signaling pathway, VEGF signaling pathway, Thyroid hormone signaling pathway, Sphingolipid signaling pathway, MAPK signaling pathway, Ras signaling pathway, PI3K-Akt signaling pathway and other pathways(Fig. 2F).
Cluster Analysis of Core Targets for PPI Network
Targets in the C-D network were placed in Bisogenet to construct a PPI network for the target of Shenlian Capsule. The network consists of 8442 target nodes and 186422 interactions (Fig. 3A-a). The NSCLC related targets PPI network constructed as described above, the target network and a total of 6,554 targets and 160,234 edges (Fig. 3A-b).
The drug PPI network is further intersected with the disease PPI network to obtain the C-D-T PPI network. The network consists of 5,574 nodes and 143,784 interactions (Fig. 3B-a). In order to obtain the core targets of drug-acting diseases, the targets were first screened with a median value greater than 2 times the degree, and a network of 1367 nodes and 60209 edges was obtained (Fig. 3B-b). The core targets network was further screened by topology method, and the core targets of drug affected diseases were screened by used degree > 136, BC > 555.53, EC > 0.0160, LAC > 15.205 and NC > 127.70. The network consists of only 102 nodes and 2442 edges (Fig. 3B-c).
Table 1
Clustering Analysis Results of Core Targets in C-D-T PPI Network
| Cluster | Score | Nodes | Edges | Node name |
| 1 | 19.91 | 22 | 209 | RPS7, RPS8, NPM1, HNRNPA1, RPS6, RPS3, RPL11, RPS2, RPS4X, RPL13, RPS3A, RPLP0, RPSA, RPL10, RPL5, HNRNPU, RPS14, NTRK1, RPS16, VCAM1, RPL23, UBL4A |
| 2 | 16.91 | 36 | 296 | RACK1, CCDC8, TARDBP, CUL2, CUL3, CUL1, MCM2, PABPC1, NOP56, U2AF2, DHX9, CUL5,PAN2, HIST1H3H, FN1, CAND1, HIST1H3J, HIST1H3F, ILF3, ILF2, HIST1H3B, HIST1H3E, HIST1H3C, HIST1H3G, HIST1H3I, HIST1H3D, HIST1H3A, STAU1, NCL, RPL6, YWHAG, HDAC1, FUS, EP300, HSP90AA1, PARP1 |
| 3 | 12.62 | 30 | 183 | ITGA4, HNRNPD, YWHAQ, CDK2, SNW1, SUZ12, TUBG1, ESR1, NEDD8, VCP, CUL4B,EEF1A1, EGFR, UBC, CUL7, CDC5L, SYNCRIP, OBSL1, SMARCA4, YWHAZ, HDAC2, RPA1, HUWE1, XPO1, HSPA8, HNRNPM, EIF4A3, MCM5, HSP90AB1, CREBBP, |
| 4 | 3.56 | 10 | 16 | HNRNPK, SRRM2, GRB2, RNF2, YWHAE, TUBB, RPS27A, DDX5, HDAC5, HSPA5 |
Four clusters were obtained by clustering the core targets of the C-D-T PPI network by used MCODE plug-in. Cluster1 is consist of 22 nodes and 229 edges, cluster2 is consist of 36 nodes and 296 edges, cluster3 is consist of 30 nodes and 183 edges, cluster4 is consist of 10 nodes and 16 edges (Fig. 3C and Table 1).
Cluster target GO enrichment and pathway analysis
To illustrate the biological functions of the 102 major targets, through classified the core PPI network of drug diseases into 4 clusters (Fig. 4A), performed GO (Fig. 4B) and KEGG pathway (Fig. 4C) enrichment analysis on each cluster.
Based on these GO terms enrichment analysis data found that (a) gene expression, RNA binding, rRNA binding, protein kinase binding, skinase binding;(b) negative regulation of cell proliferation, protein monoubiquitination, ubiquitin protein ligase activity, receptor tyrosine kinase binding; (c) DNA repair, DNA replication;(d) positive regulation of transcription from RNA polymerase II promoter, negative regulation of epidermal growth factor receptor signaling pathway, ERBB2 signaling pathway, negative regulation of apoptotic process, ubiquitin protein ligase binding, cadherin binding involved in cell-cell adhesion. Enrichment analysis by Go suggesting that Shenlian Capsule treat NSCLC from multiple angles.
By analyzing the KEGG pathway data of each pathway, we consider that the key signaling pathways of Ribosome, Cell cycle, Viral carcinogenesis, PI3K-Akt signaling pathway, Notch signaling pathway and FoxO signaling pathway, the top 6 KEGG pathways in four clusters, might be the most important pharmacological mechanism of Shenlian Capsule for NSCLC.
Effect of Shenlian Capsule on the role of LUSC survival-related gene
LUSC Differential gene
Through differential gene analysis, we obtained a heatmap of the differential genes of LUSC disease (Fig. 5A-a) and a volcano map of the expression levels of differential genes (Fig. 5A-b). A total of 4,664 differential genes which 2836 genes were up-regulated and 1,828 genes were down-regulated.
Survival related genes of LUSC treated by Shenlian Capsule
A survival analysis was performed for each of the above-mentioned differential genes and a total of 256 differences were acquired. The 256 differences were intersected with the C-D network to obtain CTSD and PLAU survival-related genes (Fig. 5A-c). CTSD is regulated by MOL000098 active compounds and PLAU is regulated by two active compounds.
Differential expression of normal tissue and cancer tissue of LUSC survival related genes treated by Shenlian Capsule
CTSD expression in normal tissue was 594.57\(\pm\)235.35, cancer tissue expression was 293.66\(\pm\)165.48, cancer tissue expression was significantly lower than normal tissue, p < 0.01(Fig. 5B-a). The expression of PLAU in normal tissue was 12.17\(\pm\)6.69 and cancer tissue was 87.05\(\pm\)92.70, the expression of PLAU cancer tissue was higher than normal tissue, p < 0.01 (Fig. 5B-c).
Survival curve of LUSC survival related genes treated by Shenlian Capsule
The CTSD gene survival curve was significantly different in the low expression group and the higher expression group, p < 0.01(Fig. 5B-b). The survival time of the low expression group of the PLAU survival curve is better than that of the high expression group, p < 0.01(Fig. 5B-d).
Molecular docking verify the effect of active compounds on survival-related targets
After docking the MOL00098 active compound with the CTSD protein, the selected conformational energy is -7.1Kcal/mol. The active compound has five hydrogen bonds to interact with the CTSD protein. From the two-dimensional structure after docking, it can be seen that the active compound has two hydrogen bonds with the Tyr residue at position 205 and the distance are 2.77 Å.In addition, the active compound has hydrogen bonding interactions with Ser residue at position 80, Ile residue at position 142 and Gly residue at position 233. The hydrogen bond distances are 2.90 Å, 2.81 Å and 2.82 Å (Fig. 5C-a). Figure 5C-b shows the three-dimensional structure diagram of the active compound MOL00098 after docking with CTSD, and the two-dimensional structure diagram after docking produces good mutual confirmation.
After docking the MOL000449 active compound with PLAU protein, the selected conformational energy is -7.5Kcal/mol. The active compound has 3 hydrogen bonds to interact with the PLAU protein. According to the two-dimensional structure after docking, it can be seen that the active compound has 2 hydrogen bonds with the Arg residue at position 239, the distances are 2.97 Å and 3.05 Å, respectively. In addition, the active compound has a weak hydrogen bond interaction with the Cys residue at position 13 with a hydrogen bond distance of 3.27 Å (Fig. 5C-c). Figure 5C-d is a three-dimensional structure diagram of the docking of active compound MOL00449 and PLAU, showing the most stable conformation of the docking of active compound with PLAU protein.
Effect of Shenlian Capsule on the role of LUAD survival-related targets
LUAD Differential gene
Through differential gene analysis obtained a heat map (Fig. 6A-a) of the differential genes of LUAD and a volcano map (Fig. 6A-b) of the expression levels of differential genes. A total of 6776 differential genes which 5176 genes were up- regulated and 1600 genes were down-regulated.
Survival related genes of LUAD treated by Shenlian Capsule
A survival analysis was performed for each of the above-mentioned differential genes, and a total of 490 difference genes were obtained. Intersecting 490 difference genes with the C-D network yielded eight survival-related genes for ADRB2, BIRC5, CCNA2, CCNB1, CDK1, CHEK1, F2 and MMP3 (Figs. 6A-c). The above eight survival-related genes are regulated by 46, 74, 3, 3, 58, 72 and 1 active compounds, respectively.
Differential expression of normal tissue and cancer tissue of LUAD survival related genes treated by Shenlian Capsule
The ADRB2 expression in normal tissue was 12.80 ± 5.09, cancer tissue expression was 1.71 ± 1.59 and cancer tissue expression was significantly lower than normal tissue, p < 0.01 (Fig. 6C-a). The expression of BIRC5 in normal tissue was 0.81 ± 0.60 and cancer tissue was 11.46 ± 10.41, cancer tissue is higher than normal tissues, p < 0.01 (Fig. 6C-b).The expression of CCNA2 in normal tissue was 0.97 ± 0.56 and cancer tissue is 7.97 ± 7.23, the expression of cancer tissue is higher than normal tissues, p < 0.01 (Fig. 6C-c).The expression of CCNB1 in normal tissue was 2.01 ± 0.91 and cancer tissue is 20.97 ± 14.07, the expression of cancer tissue is higher than normal tissues, p < 0.01 (Fig. 6C-d).The expression of CDK1 in normal tissue was 1.56 ± 0.57 and cancer tissue was 9.51 ± 8.46, the expression of cancer tissue is higher than normal tissues, p < 0.01 (Fig. 6C-e).The expression of CHEK1 in normal tissue was 0.54 ± 0.13 and cancer tissue was 2.82 ± 2.30, the expression of cancer tissue is higher than normal tissues, p < 0.01 (Fig. 6C-f).The expression of F2 in normal tissue was 0.0045 ± 0.0074 and cancer tissue was 0.80 ± 7.13, the expression of cancer tissue is higher than normal tissues, p < 0.01 (Fig. 6C-g).The expression of MMP3 in normal tissue was 0.074 ± 0.11 and cancer tissue was 1.26 ± 2.87, the expression of cancer tissue is higher than normal tissues, p < 0.01 (Fig. 6C-h).
Survival curve of LUAD survival related genes treated by Shenlian Capsule
The ADRB2 survival curve was significantly different in the low expression group and the higher expression group, p < 0.01 (Fig. 6D-a). The greater the number of ADRB2 gene expressions, the longer the patient's survival time. The survival time of the low expression group of the BIRC5, CCNA2, CCNB1, CDK1, CHEK1, F2 and MMP3 gene survival curve is better than the high expression group, p < 0.01 (Fig. 6D(b-h)). The lower the amount of tumor tissue expression of the above seven genes, the longer the patient's survival time.
Molecular docking used to verify the effect of active compounds on survival-related genes
Forty-six active compounds were docked with the ADRB2 protein, the lower energy conformation was selected for display. The result was the conformation with the lower energy after the MOL07111 active compound was docked with the ADRB2 protein, the selected conformation energy is -7.8Kcal/mol. The active compound has 3 hydrogen bonds to interact with the protein. From the two-dimensional and three-dimensional structure after docking, it can be seen that the active compound has two hydrogen bonds at the amino terminal of the Tyr residue at the O1 position and the amino terminal of the Gly residue at the 25 position. The two hydrogen bonding distances are 3.07 Å and 3.24 Å. In addition, the O2 terminal of the active compound has a hydrogen bonding interaction with the amino terminal of the Arg residue at position 80 and distance is 3.00 Å (Fig. 6E-a, Fig. 6E-b).
Three active compounds were docked with the BIRC5 protein, the conformation with the lower energy for docking MOL00098 active compounds with the BIRC5 protein was been selected. Its conformational energy was − 7.1Kcal/mol. The active compound has 3 hydrogen bonds to interact with the BIRC5 protein. From the two-dimensional and three-dimensional structures after docking, it can be seen that the active compound has two hydrogen bonds at the amino terminal of the 16 position Asp residue at the O4 position and the amino terminal of the Lys residue at the 15 position. The distances between the two hydrogen bonds are 2.71 Å and 3.21 Å. In addition, the O6 terminal of the active compound has a hydrogen bonding interaction with the amino terminal of the Val residue at position 89 and distance is 2.74 Å (Fig. 6E-b, Fig. 6F-b).
Seventy-eight active compounds were docked with the CCNA2 protein, the conformation with the lower energy for docking MOL07150 active compounds with the CCNA2 protein was been selected. Its conformational energy was − 10.0 Kcal/mol. The active compound has 4 hydrogen bonds to interact with the protein. From the two-dimensional and three-dimensional structures after docking, the O4 terminal of the MOL07150 compound has two hydrogen bonds with the hydroxyl position of the Gln residue at position 81 and the hydroxyl position of the Tyr residue at position 113. The hydrogen bond distances are 3.15 Å and 3.28 Å, respectively. The O5 terminal of the MOL07150 compound has two hydrogen bonds with the hydroxyl position of the Lys residue at position 78 and the hydroxyl position of the Tyr residue at position 113, the hydrogen bond distances are 2.77 Å and 2.99 Å (Fig. 6E-c, Fig. 6F-c).
Three active compounds were docked with the CCNB1 protein, the conformation with the lower energy for docking MOL002714 active compounds with the CCNB1 protein was been selected. Its conformational energy was − 7.0Kcal/mol. From the two-dimensional and three-dimensional structures after docking, the O2 terminal of the MOL002714 compound has a hydrogen bond interaction with the amino terminal of the Val residue at position 156 and distance is 3.12 Å(Fig. 6E-d, Fig. 6F-d).
Three active compounds were docked with the CDK1 protein, and the conformation with the lower energy for docking MOL002714 active compounds with the CDK1 protein was been selected. Its conformational energy was − 9.6 Kcal/mol. The active compound has 4 hydrogen bonds to interact with the CDK1protein. It can be seen from the two-dimensional and three-dimensional structures after docking that the active compound has two hydrogen bonds at the O4 position with the amino terminal and the hydroxyl terminal of the 12 position Glu residue. The distance between the two hydrogen bonds are 2.88 Å and 3.21 Å. In addition, the O5 terminal of the active compound produces a hydrogen bond interaction with the hydroxyl end of the Ile residue at position 10 and distance is 3.18 Å. The O7 terminal of the active compound produces a hydrogen bond interaction with the amino terminus of the Asp residue at position 146 and distance is 2.93 Å(Fig. 6E-e, Fig. 6F-e).
Fifty-eight active compounds were docked with the CHEK1 protein, the conformation with the lower energy for docking MOL002651 active compounds with the CHEK1 protein was been selected. Its conformational energy was − 9.7 Kcal/mol. It can be seen from the two-dimensional and three-dimensional structures after docking that the oxygen at the 3 position of the active compound has a weak hydrogen bond interaction with the amino terminus of the Cys residue at the 87 position and distance is 3.25 Å. In addition, the active compound has significant hydrophobic interactions with the surrounding 11 amino acids (Fig. 6E-f, Fig. 6F-f).
Seventy-two active compounds were docked with the F2 protein, the conformation with the lower energy for docking MOL007151 active compounds with the F2 protein was been selected. Its conformational energy was − 7.1 Kcal/mol. It can be seen from the two-dimensional and three-dimensional structures after docking that O4 and O5 terminal of the active compound with Glu residue at position 362 of the F2 protein have two hydrogen bonding interactions, the distance are 2.71 Å and 2.91 Å.In addition, the active compound has significant hydrophobic interactions with the surrounding 7 amino acid residues.
After docking the active compound MOL00098 with MMP3, the lower energy conformation was been selected, the energy after docking was − 7.3 Kcal/mol. It can be seen from the two-dimensional and three-dimensional structures after docking that O4 terminal of the active compound has hydrogen bonding interactions with the Tyr residue at position 220 and the Leu residue at position 222 of the MMP3 protein, the distances are 2.70 Å and 3.23 Å. In addition, the active compound has significant hydrophobic interactions with the surrounding 6 amino acids.
