Screening and enrichment analysis of candidate active components and potential drug targets of CR
To investigate the pharmacological action mechanism of CR, we first performed virtual screening based on the two important parameters of ADME, i.e., OB and DL, to obtain the potential active components in CR. Seven monomers met the requirements of OB ≥ 30% and DL ≥ 0.18. Then ,we also identified seven other monomeric components that either showed good pharmacological activity or were typical main components of CR reported in the literature. They were included for subsequent analysis; however, for the components, either OB was less than 30% or DL was less than 0.18. Therefore, a total of 14 compounds were considered potential monomerically active components of CR (Figure 1B & Table 1).
Table 1
The specific information of candidate active ingredients in CR obtained from a virtual screening and literature mining.
MOL number | Candidate ingredients | OB/% | DL |
MOL000019 | D-Camphene | 34.98 | 0.04 |
MOL000073 | ent-Epicatechin | 48.96 | 0.24 |
MOL000358 | Beta-sitosterol | 36.91 | 0.75 |
MOL000359 | Sitosterol | 36.91 | 0.75 |
MOL000492 | (+)-Catechin | 54.83 | 0.24 |
MOL000608 | ()-Terpinen-4-ol | 81.41 | 0.03 |
MOL000708 | Benzaldehyde | 32.63 | 0.01 |
MOL000991 | Cinnamaldehyde | 31.99 | 0.02 |
MOL001736 | (-)-Taxifolin | 60.51 | 0.27 |
MOL002225 | Styrone | 38.35 | 0.02 |
MOL002295 | Cinnamic acid | 19.68 | 0.03 |
MOL003530 | O-Methoxycinnamaldehyde | 26.52 | 0.04 |
MOL004576 | Ttaxifolin | 57.84 | 0.27 |
MOL011169 | Peroxyergosterol | 44.39 | 0.82 |
Subsequently, we further used two online databases, i.e., TCMSP and SwissTargetPrediction, to explore the candidate drug targets of the above 14 compounds and obtained a total of 142 drug targets (Supplementary Table 1). Interestingly, there were many overlapping targets among the 14 different components, indicating that these compounds may play key roles in generating synergistic effects. Then, we further used Cytoscape 3.2.1 software to construct a drug-target network to visualize the interactions between this systems (Figure 1C).
Next, we performed enrichment analysis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) molecular signaling pathways and disease types for the 142 drug targets that were obtained using DAVID v 6.8 and KOBAS 3.0 software, respectively. The results indicated that the above targets were significantly associated with “Pathways in cancer” and “Cancers” (Figure 1D & 1E). Therefore, these prediction results not only help reveal the pharmacological mechanism of CR but also suggest that CR has great potential in cancer treatment.
Oral administration of CR can inhibit the growth of CC cells in C57BL/6 mice
To investigate the direct tumor suppressor effect of CR in vivo, we assessed its specific effects on xenograft tumors in C57BL/6 mice (subcutaneously injected with MC-38 cells) (Figure 2A). As shown in Figure 2B and 2C, compared with the control, the oral administration of CR significantly inhibited the growth of MC38 xenograft tumors in vivo. By day 14, the tumor volume in the CR treatment group was approximately 2.6 times that in the control group (p < 0.05) (Figure 2D). Furthermore, there was a significant difference in tumor weight between these two groups (p < 0.05); however, there was no significant difference in the body weight of mice between these two groups (Figure 2E & 2F). Next, through hematoxylin and eosin (H&E) tissue staining, we found that compared with the control group, tumor tissue in the CR treatment group showed obvious necrotic cells; the subsequent immunohistochemical analysis showed that the number of Ki-67-positive cells was significantly reduced in the CR treatment group, thus indicating that CR had an antiproliferative effect on colon tumor cells (Figure 2G). Therefore, the above results provide convincing evidence that CR has direct anti-CC activity.
CR can significantly inhibit the viability and motility of CC cells
Subsequently, we performed a series of in vitro CR pharmacodynamics experiments using CC cells. The results of the CCK-8 assay showed that CR had a dose- and time-dependent inhibitory effect on the growth of HT-29, HCT-15 and MC-38 cells (Figure 3A). The half maximal inhibitory concentration (IC50) analysis at 24 h showed that the IC50 values for CR in HT-29 and HCT-15 human colon carcinoma cells were 0.278±0.019 mg/mL and 0.123±0.014 mg/mL, respectively. The results of the subsequent cell growth experiments indicated that the viability of HT-29 and HCT-15 cells significantly decreased at 48 h compared with the control group (Figure 3B). Meanwhile, the cell colony formation assay showed that the colony formation ability of the cells decreased in a dose-dependent manner after 24 h of treatment with CR (Figure 3C). In addition, we also used a high-content microscope to track the motility of cells in real time. The average cumulative movement of the cells in the CR treatment group was less than the control group at 24 h, indicating that the cell viability in the CR treatment group was significantly inhibited (Figure 3D). Therefore, the above results indicate that CR has a direct inhibitory effect on the survival and motility of CC cells.
Construction of CR anti-CC PPI systematic network and core target enrichment analysis
With the help of network pharmacology and bioinformatics techniques, we used the Bisogenet plugin in Cytoscape 3.2.1 software to perform PPI analysis of the 142 potential drug therapeutic targets identified for CR. A total of 5,575 nodes and 103,469 edges were obtained. Then, through the GEO online database, the gene expression results of two microarrays of human CC and paired paracancerous tissues (GSE44076 and GSE13471) were downloaded and analyzed. After screening based on the conditions of p < 0.01 and fold change > 2, 114 relevant disease targets corresponding to CC were identified (Figure 4A-4C & Supplementary Table 2). Subsequently, the Bisogenet plugin was used to perform a PPI analysis of the above 114 disease targets. A total of 2027 nodes and 34,539 edges were obtained. Subsequently, to more accurately predict the potential core targets of CR for the treatment of CC, we used the CytoNCA plugin in Cytoscape software to systematically integrate the PPI analysis results for the drug targets and the disease targets. The conditions 'DC' > 48 and 'DC' > 80, 'BC' > 0.002, and 'CC' > 0.478 were used to perform core mining of the integrated network at the topology level, and 111 potential core targets were predicted (Figure 5A & Supplementary Table 3).
To further predict the potential biological processes and related molecular mechanisms involving the abovementioned 111 core targets, we used Metascape and DAVID v6.8 to perform an enrichment analysis of the biological functions (GO-Biological Process, GO-BP) and molecular mechanisms (KEGG). The results of the GO-BP analysis indicated that the above targets were closely associated with processes such as “pathways in cancer”, “apoptotic signaling pathways”, and “cell cycle” (Figure 5B), and the results of the KEGG analysis indicated that the above targets were closely associated with mechanisms such as 'pathways in cancer', 'viral carcinogenesis', the 'cell cycle', the 'PI3K-Akt signaling pathway', and the 'MAPK signaling pathway' (Figure 5C). The results of these “virtual studies” provide clues for further revealing and verifying the pharmacological mechanism of CR in the treatment of CC.
CR can induce apoptosis and G2/M phase arrest in CC cells
Next, we conducted a series of cell function experiments to verify the results of the above enrichment analysis. First, compared with the control group, the cells after CR treatment showed typical characteristics of apoptosis, such as shrinkage, rounding, and loss of contrast. Meanwhile, fluorescence observation after Hoechst 33342 staining revealed that nuclei after CR treatment exhibited dense staining or dense fragmented staining (Figure 6A). In addition, the flow cytometry results indicated that the apoptotic cell population stained with Annexin-V FITC significantly increased in a dose-dependent manner after CR treatment (Figure 6B-6D). The WB results also indicated that CR promoted the accumulation (dose-dependent) of the proapoptotic protein Bax and cleaved caspase-9 and PARP and the downregulation of antiapoptotic protein Bcl-2 (Figure 6H & Supplementary Figure 1).
In addition, we also performed cell cycle analysis via BrdU incorporation and flow cytometry to evaluate the effect of CR on the cell cycle of CC cells. The results showed that CR treatment arrested the cells at the G2/M phase, thereby significantly inhibiting their growth and proliferation (Figure 6E-6G). The WB results confirmed the accumulation of the G2/M phase-specific marker cyclin B1 and the downregulation of cyclin D1, CDK2 and cyclin A2 in CR-treated CC cells (Figure 6H & Supplementary Figure 1). In summary, the above results all indicated that CR inhibited the growth of CC cells by inducing apoptosis and cell cycle arrest.
CR can inhibit the growth of CC cells by blocking the Akt/ERK signaling pathways
To further explore the potential molecular mechanism by which CR inhibits the growth of CC, we next examined the key signaling pathways involved in cell proliferation and viability. Among them, we selected the PI3K-Akt and MAPK signaling pathways based on the results of the KEGG pathway enrichment analysis. The WB results indicated that after CR treatment, the expression levels of key protein factors, such as p-Akt (T308 & S473) and p-ERK (T202/Y204), in the above pathways in the two CC cells were significantly inhibited (Figure 7 & Supplementary Figure 2). This result indicates that the effects of CR, i.e., inducing apoptosis and cell cycle arrest and inhibiting the growth of CC cells, may be the result of the simultaneous inhibition of the Akt/ERK signaling pathway, thus demonstrating once again that CHM has “multi-component, multi-target, and multi-function” pharmacological characteristics.
Taxifolin is a potential key active component of CR in the treatment of CC
Next, we used ultraperformance liquid chromatography-tandem mass spectrometry (UPLC–MS/MS) to identify the potential active components in CR. A total of 5 key components were identified from the crude extract of CR, i.e., cinnamic acid (tR: 26.30 min), 2-hydroxycinnamic acid (tR: 11.70 min), protocatechuic acid (tR: 4.17 min), catechinic acid (tR: 6.93 min), and taxifolin (tR: 9.12 min) (Figure 8A & 8B & Table 2). Intriguingly, 3 of the 5 components detected above are potential active components that were previously screened based on ADME-related characteristics or literature searches, thus again confirming the reliability of the network pharmacology prediction method. To further explore the molecular mechanism by which CR affects CC at the protein level, we used AutoDock 4.2 software to perform molecular docking analysis between the above active ingredients and key targets. The key targets were obtained by combining the previously obtained 142 drug targets of CR with 111 core targets of CR for CC and identifying areas of overlap; ultimately, a total of 9 key targets were obtained (Figure 8C). Next, we predicted the docking efficiency between the above 5 active ingredients and the 9 key targets. The results indicated that the binding affinities between taxifolin and the targets HSP1A1, HSP90AB1 and PARP1 (docking energies of -9.25, -9.17 and -9.05 kcal/mol, respectively) were all higher than those between the remaining 4 components and other proteins (Figure 8D & Table 3). Therefore, taxifolin is very likely to be a potential key active component of CR in the treatment of CC; however, further experimental verification is needed.
Table 2
Identification of main ingredients in CR by UPLC-MS/MS data.
No. | tR (min) | Molecular formula | Selected ion | Theoretical | Experimental | MS/MS fragmentions | Compounds |
1 | 26.30 | C9H8O2 | [M-H]+ | 148.16 | 149 | 130.74, 102.73 | Cinnamic acid |
2 | 11.70 | C9H8O3 | [M-H]+ | 164.16 | 165 | 119.37, 90.92 | 2-Hydroxycinnamic acid |
3 | 4.17 | C7H6O4 | [M-H]− | 154.12 | 153.51 | 109.14, 92.29, 80.73 | Protocatechuic acid |
4 | 6.93 | C15H14O6 | [M-H]− | 290.27 | 289.03 | 244.85, 202.74, 109.09 | Catechinic acid |
5 | 9.12 | C15H12O7 | [M-H]− | 304.25 | 302.85 | 192.91, 134.74 | Taxifolin |
Table 3
The results of molecular docking studies of five main ingredients in the active sites of nine proteins performed using AutoDock.
Proteins | Ligands |
Cinnamic acid | 2-Hydroxycinnamic acid | Protocatechuic acid | Catechinic acid | Taxifolin |
PRKCA | -5.47 | -5.83 | -6.00 | -7.44 | -7.55 |
HSP90AB1 | -7.33 | -4.58 | -6.88 | -6.27 | -9.17 |
APP | -5.68 | -5.61 | -5.09 | -7.52 | -8.04 |
RELA | -7.09 | -6.72 | -6.52 | -7.10 | -7.05 |
JUN | -4.20 | -4.66 | -3.98 | -4.71 | -5.21 |
MDM2 | -5.49 | -5.53 | -5.13 | -6.71 | -6.83 |
PRKACA | -7.09 | -6.81 | -6.00 | -7.93 | -8.19 |
HSPA1A | -6.24 | -5.52 | -5.39 | -7.56 | -9.25 |
PARP1 | -6.79 | -6.62 | -6.41 | -8.62 | -9.05 |
(kcal/mol) |