Scutellaria baicalensis in the Treatment of Hepatocellular Carcinoma: Network Pharmacology Analysis and Experimental Validation

Objective The aim of the study was to use a network pharmacological method and experimental validation to examine the mechanism of Scutellaria baicalensis (SB) against hepatocellular carcinoma (HCC). Methods The traditional Chinese medicine systems pharmacology database and analysis platform (TCMSP) and GeneCards were used for screening of targets of SB for the treatment of HCC. Cytoscape (3.7.2) software was used to construct the “drug-compound-intersection target interaction” interaction network. The STING database was used to analyze the interactions of the previous intersecting targets. The results were visualized and processed by performing GO (Gene Ontology) enrichment analysis and KEGG (Kyoto Encyclopedia of Genes and Genomes) signaling pathway enrichment analysis at the target sites. The core targets were docked with the active components by AutoDockTools-1.5.6 software. We used cellular experiments to validate the bioinformatics predictions. Results A total of 92 chemical components and 3258 disease targets including 53 intersecting targets were discovered. The results showed that wogonin and baicalein, the main chemical components of SB, could inhibit the viability and proliferation of hepatocellular carcinoma cells, promote apoptosis through the mitochondrial apoptotic pathway, and effectively act on AKT1, RELA, and JUN targets. Conclusion SB has multiple components and targets in the treatment of HCC, providing possible potential targets for the treatment of HCC and providing a basis for further research.


Introduction
With roughly 906,000 new cases and 830,000 deaths from primary liver cancer worldwide in 2020, it is the sixth most prevalent cancer and the third main cause of cancer death globally, with typically three times as many men as women afected [1]. Te main risk factors for hepatocellular carcinoma (HCC) are chronic infection with hepatitis B virus (HBV) or hepatitis C virus (HCV), which account for 56% and 20% of liver cancer deaths worldwide, respectively. Additionally, liver cancer can occur as a result of smoking, drinking excessive amounts of alcohol, being overweight, having type 2 diabetes, and eating food contaminated with afatoxin. Since HCC starts insidiously and less than 30% of patients are suitable for radical treatment when frst diagnosed, drug therapy is very important in the treatment of HCC [2]. Currently, some patients with hepatic insufciency are not tolerant to frst-line therapeutic agents and the treatment methods are expensive, which causes physical and psychological stress to these patients. Terefore, fnding new efective drugs is an urgent matter to be addressed at present.
Many Chinese traditional herbs can have specifc efcacy in disease treatment, but the specifc mechanism of action is not clear. Network pharmacology has enabled some herbal components to be explored and studied, allowing more specifc components to be discovered and studied, and the combination with molecular docking technology is benefcial to discover new drug targets and active drug components [3]. For example, Tu discovered artemisinin in Artemisia annua and thus efectively acted on malaria patients [4]. Previously, it was thought that "one drug for one target," but modern research has demonstrated that the action of drugs in the body is far more complex than previously thought and that one drug may act on multiple targets and one target may be acted on by multiple drugs, and this idea of one drug for one target is no longer applicable to today's drug research [5]. Cyber pharmacology searches for key components of potential drug target combinations and facilitates drug research, especially anticancer drug research [6].
Scutellaria baicalensis (SB) has been used in China for thousands of years and was recorded in Shennong ben cao jing (Te Classic of Herbal Medicine) as early as 200 AD, where it was used to treat liver and lung problems and windcold fever [7,8]. Modern research has found that SB has antitumor efects on a variety of cancer cells, such as prostate cancer and head and neck tumors [9,10]. Te treatment of dividing cells with SB induces DNA damage and leads to cell death but does not cause cell mutation, which is a major side efect of conventional anticancer drugs [11]. Tis implies that SB has the potential to be a novel antitumor therapeutic agent.
Tis study determined the primary active ingredients in SB for the treatment of HCC based on network pharmacology and obtained its potential targets. Tis research established the molecular mechanism of SB and supported it with cytological tests, giving SB a rationale for use in the treatment of HCC. Te overall structure fowchart is shown in Figure 1.

Acquisition of Active Constituents and Targets of SB.
Te chemical components of SB were searched by the traditional Chinese medicine systems pharmacology database and analysis platform (TCMSP) (https://tcmspw.com/ tcmsp.php). Oral bioavailability (OB) and drug likeness (DL) were used as the screening criteria, and chemical components with OB ≥ 30% or DL ≥ 0.18 were selected.

Screening of the Action Targets of SB.
Te corresponding targets of the obtained compounds were collected in the TCMSP database, and the obtained targets were subjected to gene name transformation and specifcation in the UniProt database (https//www.uniprot.org/).

Acquisition of Targets for Hepatocellular Carcinoma.
We searched the GeneCards (https://www.genecards.org/) database for genes related to hepatocellular carcinoma using the keyword "Hepatocellular Carcinomas" and obtained the targets related to hepatocellular carcinoma after deduplication.
2.4. Screening of Targets of SB for the Treatment of Hepatocellular Carcinoma. Using Venny2.1 (https://www. liuxiaoyuyuan.cn/), the targets of SB component action and the targets related to hepatocellular carcinoma were intersected, and the overlapping targets were the targets related to the possible action of SB in hepatocellular carcinoma.

Constructing Drug-Compound-Intersection Target
Interaction Networks. Cytoscape (3.7.2) software was used to construct the "drug-compound-intersection target interaction" interaction network. Te nodes in the network diagram represent the disease, drug, active ingredients, and targets, respectively. Te interactions between active ingredient and target, target and disease, or target and drug were represented by the connection of edges. A network analysis, a plug-in of Cytoscape, was used to analyze the topological properties of the network, and the network was arranged according to the degree value to obtain a network graph that can characterize the relationships between drugs, diseases, ingredients, and targets.

Protein-Protein Interaction (PPI) Network Construction.
Te STING database (https://string-db.org/) was used to analyze the interactions of the previous intersecting targets, import the target information into the database, select multipleproteins, set the species to Homo sapiens, set the confdence level to 0.9, hide the scattered nodes, and obtain the protein-protein interaction of the intersecting target genes. Te protein-protein interaction (PPI) network of intersecting target genes was obtained. Cytoscape 3.7.0 software was applied to visualize the results, and topological analysis was performed with the aid of CytoNCA plug-in. Te core targets were calculated and visualized with the median values of the parameters of mediator centrality, proximity centrality, degree centrality, eigenvector centrality, local average connectivity, and network centrality as the fltering conditions, respectively. Tese six parameters represent the topological importance of each individual node in the whole network, and the larger their quantifed values, the more important the node is in that network.

GO Functional Annotation and KEGG Pathway Enrichment Analysis.
Te results were visualized and processed by performing GO (Gene Ontology) enrichment analysis and KEGG (Kyoto Encyclopedia of Genes and Genomes) signaling pathway enrichment analysis at the target sites, and the diferences were considered statistically signifcant at P < 0.05. We use Sangerbox online processing tool (https:// www.sangerbox.com/), a comprehensive and interactive friendly platform for clinical bioinformatics analysis. Tey used SpringCloud framework, java 11,10.2.30-MariaDB to build the website, website backend, and database. R 3.6.3, R 4.0.1, Perl, and JavaScript languages were applied to write and convert to arithmetic and analytical scripts.

Molecular Docking Study of Target Proteins and Active
Components of SB. We download the mol2 format fle of the 2D structure of the screened key compound from the Pub-Chem database (https://pubchem.ncbi.nlm.nih.gov/), import it into AutoDockTools-1.5.6 software, check its spatial structure for errors, add the atomic charges, and assign the atomic types; all fexible keys are changed to default rotatable and saved in the PDBQT format as docking ligands. We select the core target of the previous screening, look for the 3D concept of the active ingredient in the UniProt database (https://www.uniprot.org/), enter the query PDB id into the PDB database (https://www.rcsb.org/), download the 3D structure of its corresponding protein, and save it in the PDB format. We import this 3D structure into AutoDockTools-1.5.6 software, perform water molecule removal, hydrogen addition, charge addition, and various optimizations, then set the grid box (bird box) as the default value and save it, and fnally save the optimized protein in the PDBQT format as a pair of acceptors. According to the coordinates of the ligand in the target protein and the target protein activity pocket that had been set, molecular docking was performed using AutoDock Vina 1.1.2 to obtain the results, and the obtained results were imported into PyMOL software for visualization of the results. 2.10. CCK8 Assay. Cell viability was detected by employing the CCK8 assay (Meilunbio, CN). HCC cells in the logarithmic growth phase were seeded in 96-well plates, and 3 replicate wells were taken from each group. Cells were grown adherent and then replaced, and having taken 0.625∼10 μM as a concentration gradient, the cells were treated with wogonin and baicalein. After 24 hours, 10 μl CCK8 solution was added to each well. Te absorbance at a wavelength of 450 nm was measured after 1 hour using a microplate reader.

ROS Assay.
Te intracellular ROS levels were measured using a reactive oxygen species assay kit (Beyotime Biotechnology, China). Briefy, the cells were seeded in 6-well plates and exposed to wogonin and baicalein for 24 h. Following the treatment, the cells were incubated with DCFH-DA for 20 min at 37°C and then observed using Olympus IX71 fuorescence microscopy (Japan).

Data Analysis.
Statistical analysis was performed with GraphPad Prism 6.0. Image processing is conducted by image J and Sanger box 3.0. All data were presented as the mean ± SD. Te one-way analysis of variance was applied for comparisons between the groups, and P < 0.05 indicated that the diference was statistically signifcant.

Main Active Components and Teir Targets of SB.
A total of 105 compounds was founded in SB by TCMSP search, and 36 active ingredients were obtained after screening. Te basic information about the active ingredients of SB is shown in Table 1. Te gene names of the action targets were collected through the UniProt database, and the invalid and duplicate genes were deleted, and a total of 92 action targets of active ingredients of SB were obtained.

Target Screening of SB for the Treatment of HCC.
Te keyword "Hepatocellular Carcinomas" was used to collect 3,258 genes for related diseases in the GeneCards database. Te screened 92 baicalin targets were put into the online website Venny2.1 to take the intersection with 3,258 primary liver cancer targets, and a total of 53 common targets were collected ( Figure 2).

"Drug-Compound-Intersection Target" Analysis.
Te "drug-compound-intersection target" interaction network is shown in Figure 3. In the fgure, the central diamond node represents SB, the peripheral oval node represents the SB active ingredient, the number on the node is the corresponding TCMSP number of the compound, and the rectangle represents the "drug-disease" intersection target. "Te color shades of the active ingredient nodes represent the number of targets connected to them. As seen in Figure 3, the network includes 83 nodes and 230 edges. Among the active ingredients of SB, the top fve in the degree value are wogonin (30), baicalein (21), rivularin (14), betasitosterol (13), and oroxylin A (13), and the top fve protein targets were PTGS2, (27) PTGS1, (23) NCOA2 (17), AR (15), and NOS2 (14). We subsequently selected wogonin and baicalein for the corresponding cytological experiments.

Construction and Analysis of Target PPI Network.
Te 53 intersecting targets were imported into the STRING online data platform, the species genus was selected as Homo sapiens, the confdence level was set to 0.95, and the PPI graph was plotted (Figure 4(a)), where the nodes represent proteins and the edges represent protein-protein interactions. Bar graphs were made and analyzed to obtain the top 30 PPI core genes (Figure 4

Key Signaling Pathways of SB in the Treatment of HCC.
To analyze the cellular components, molecular functions, and biological processes enriched in the 53 intersecting targets, we performed GO analysis on the intersecting targets, and the results showed that the key targets of SB for hepatocellular carcinoma were enriched in a total of 1,199 GO entries, including 1,087 bioinformatic processes (BP), 25 cellular components (CC), and 87 molecular functions (MF).
Te BP results suggest that it may be related to response to inorganic substance, apoptotic signaling pathway, and response to bacterium. Te CC results suggest that it may be related to transcription regulator complex, cytoplasmic vesicle lumen, and vesicle lumen. Te MF results suggest that it may be related to protein domain specifc binding, transcription factor binding, and DNA-binding transcription factor binding. Figure 5(a) shows the results of the top 10 GO analysis for BP, CC, and MF. KEGG analysis of key targets of SB for hepatocellular carcinoma showed that the key targets were enriched in a total of 238 pathways, including those in various cancers, such as small cell lung cancer and colorectal cancer, involving hepatitis B, hepatitis C, Epstein-Barr virus, infuenza A and PI3K-Akt signaling pathway, AGE-RAGE signaling pathway in diabetic complications, IL-17 signaling pathway, TNF signaling pathway, and p53 signaling pathway. Te bubble plots are sorted by the gene ratio, the higher the number of genes in the top pathway, the larger the corresponding bubble in the plot, as shown in Figure B. JUN, RELA, and AKT1 were ranked in the top 5 of the core genes of the PPI network, which indicated that JUN, RELA, and AKT1 might be the key targets of SB for the treatment of hepatocellular carcinoma.
3.6. Molecular Docking. We used the molecular docking technique to investigate the interaction between the main active components of SB and JUN, RELA, and AKT1 proteins. Te stability of the binding between the receptor and the ligand depends on the binding energy. If the binding energy is less than −5 kJ/mol, it indicates that the target compound has some binding activity with the compound, and the lower the binding energy, the more stable the binding conformation of the receptor and the ligand. As 4 Evidence-Based Complementary and Alternative Medicine shown in Table 2, baicalein, oroxylin A, rivularin, wogonin, and β-sitosterol and JUN, RELA, and AKT1 are used as examples to analyze their possible modes of action with receptors ( Figure 6).

Wogonin and Baicalein Afects Cell Morphology and
Proliferation. To verify the efect of the active ingredient of SB on the growth of HepG2 cells and Huh7 cells, we compared the morphology of HepG2 cells and Huh7 cells in the blank group and the group treated with wogonin and baicalein. As shown in Figure 7(a), the cells in the blank group are in good growth condition, with the cells growing against the wall, clearly outlined, closely arranged, and proliferating at a faster rate. When treated with the drug for 24 h, a signifcant decrease in the number of cells was observed by microscopy, and the cells were crumpled and deformed and their proliferation was inhibited, as shown in Figure 7(a). Te cell scratch width of the drug-treated group was signifcantly larger than that of the blank control group after 24 h. We assumed that the drug inhibits the proliferation of HepG2 cells and Huh7 cells, as shown in Figures 7(b)-7(e).

Wogonin and Baicalein Exerts an Inhibitory Role in the Viability of HCC Cells.
To further validate the efect of wogonin and baicalein on HCC cells, diferent concentrations (0∼10 μM) of wogonin and baicalein were employed to act on

Wogonin and Baicalein May Act on HCC Cells through
Reactive Oxygen Species. To further investigate by what way wogonin and baicalein induce apoptosis, we used a reactive oxygen species assay kit to detect intracellular reactive oxygen species (ROS) levels, as shown in Figure 9, which were signifcantly higher than in the native control groups under the efect of the drug.

Wogonin and Baicalein May Act on JUN, RELA, and AKT1.
To verify the targets obtained from bioinformatic analysis, we used qPCR to detect the expression of JUN, RELA, and AKT1 mRNA after 24 h of drug treatment and found that JUN, RELA, and AKT1 mRNAs were signifcantly higher than those of the untreated group, as shown in Figure 10. JUN  TP53  RELA  AKT1  MAPK14  ESR1  TNF  IL6  FOS  CCND1  HIF1A  CDKN1A  CASP3  VEGFA  AR  CXCL8  CASP8  RXRA  BCL2  MCL1  NOS2  CCL2  CYCS  GSK3B  MMP1  ESR2  PTGS2 BAX NCOA2 MMP9

Discussion
HCC is a common cancer with extremely high incidence and mortality rates worldwide, and patients are often diagnosed at an advanced stage. Surgical treatment options are not applicable to these advanced HCC patients, making the development of new drugs critical for patients with advanced disease [12,13]. Chinese herb plays an important role in the treatment of HCC, and it can improve the therapeutic efect together with chemotherapy drugs. Chinese herbs can successfully treat clinical symptoms and enhance patients' quality of life, according to clinical practice [14][15][16]. SB, a Chinese herb, is commonly used in the treatment of HCC. But further research is needed to understand how it treats HCC. In this study, the main components and targets of the action of SB in HCC were explored using network pharmacology and cytology experiments. Network pharmacology is an emerging paradigm to systematically reveal the multicomponent and network regulation of prescriptions and has been widely applied to the screening of bioactive compounds and the study of efcacy mechanisms of various Chinese medicine prescriptions [17]. We determined that wogonin and baicalin are the primary active ingredients in SB in this investigation. We screened the SB potential therapeutic targets and liver cancer-related targets, and then we took the intersection to get 53 common targets that represent potential SB-HCC-therapy targets. We mapped PPI and examined key genes such as JUN, TP53, RELA, AKT1, MAPK14, ESR1, TNF, IL6, FOS, and CCND1 using STRING analysis of protein interactions. We enriched the common target genes in the GO pathway and KEGG pathway and found that AKT1, RELA, and JUN are likely to be potential therapeutic targets for hepatocellular carcinoma. AKT1 and AKT2 are widely distributed in the liver, and a recent study demonstrated that AKT1 and AKT2 mice are more susceptible to liver cancer [18]. Mice defcient in Akt1 and Akt2 exhibit impaired liver regeneration and increased mortality [19]. Xu et al. demonstrated that the mTORC2-Akt1 cascade pathway plays an important role in the development of HCC [20]. RELA, also known as p65, was shown by Murray to attenuate the proliferation of pancreatic ductal adenocarcinoma cells and prolong the survival time of mice [21]. JUN is the oncogenic transcription factor [22]. JNK/c-JUN is a recognized proapoptotic and proinfammatory pathway, and Lauricella et al. demonstrated that silencing of c-JUN reduced the susceptibility of HepG2 cells to apoptosis [23]. Te results of the present experiment showed that AKT1 mRNA, RELA mRNA, and JUN mRNAs were signifcantly upregulated in HepG2 cells in the presence of two active components of SB, which is consistent with the previous fndings. Our analysis of the GO pathway and KEGG pathway revealed that the main mechanisms of action of the active ingredients of SB were enriched in the PI3K-AKT signaling pathway, the MAPK signaling pathway, and the IL-17 signaling pathway. Several studies have demonstrated that the PI3K/AKT/mTOR pathway is a classical dysregulated pathway involved in the pathogenesis of HCC [24]. Moon and Ro demonstrated that the MAPK/ERK signaling pathway is often activated in HCC [25]. Ma et al. demonstrated that IL-17A is a tumor-promoting factor that regulates alcoholic fatty liver, liver fbrosis and HCC [26]. Te conclusions from this study by bioinformatics were consistent with the previous results, and the analysis showed activation of the apoptotic pathway after drug treatment. Te results of the cellular assay showed a decrease in cellular activity with increasing drug concentration. Further investigation of the cause of apoptosis revealed an increase in intracellular ROS levels after drug treatment, and it was hypothesized that the active ingredient of SB may cause apoptosis through the mitochondrial pathway. Previous studies have demonstrated that wogonin induces apoptosis in hepatocellular carcinoma cells through intracellular H 2 O 2 production and Ca 2+ release [27], Wang et al. demonstrated that baicalein induces apoptosis and autophagy in hepatocellular carcinoma cells through endoplasmic reticulum stress [28], and Huang et al. demonstrated that wogonin induces apoptosis by regulating PI3K-AKT signaling in breast cancer cells [29]. Tere is also a recent study demonstrating that wogonin inhibits the invasive efect of hepatocellular carcinoma cells through matrix metalloproteinase-9 [30]. Tis is consistent with the fndings of the present study. Although some preliminary fndings were obtained in the present study, further experimental validation is needed.

Conclusion
In this study, we investigated the drug-component-targetdisease interactions of SB in the treatment of HCC using network pharmacology. SB may induce apoptosis in HCC cells by modulating the PI3K-AKT signaling pathway through the mitochondrial pathway. Te fndings of this study shed light on SB's anticancer properties and probable biological pathways in HCC and lay the groundwork for future research into SB or its active constituents as a viable adjunctive therapy to SB treatment.

Data Availability
Te data used and analyzed during the current study are available from the corresponding author upon reasonable request.

Disclosure
Cai Xin and Peng Shi are regarded as the co-frst authors.