Isolation and ADMET information collection of Syringin
By using TLC, HPLC, NMR, and LC-MS, Syringin was isolated and purified from AS. The extraction and separation process of it is shown in the appendix, and its purity of it was >98.0%.
The first step of bioinformatics analysis of Syringin is to collect its identity information and analyze the possibility of its development into a drug with ADMET. TCMSP(http://lsp.nwu.edu.cn/tcmsp.php) [13], SwissADME (http://www.swissadme.ch/index.php) [14], pkCSM (HTTP: //biosig.unimelb.edu.au/pkcsm/prediction) databases [15] and PubChem (https://pubchem.ncbi.nlm.nih.gov/) [16] were used to collect the above information.
Cell culture
MDA-MB-231 and MCF-7 cells were obtained from Hubei University of Chinese Medicine (Wuhan, China). The cells were cultured in Dulbecco's modified eagle medium basic (DMEM, 1X, Gibco, MD, USA) supplemented with 12% fetal bovine serum (FBS) (Gibco, MD, USA) and penicillin-streptomycin solution (The concentration of penicillin in cell culture medium was 100 µ/ml, and that of streptomycin was 0.1 mg/ml). And they were incubated at 37 °C under 5% CO2 in a carbon dioxide incubator (HHCP-01, Shanghai BAIDIAN Instrument Equipment Co., Ltd, China).
Effect of syringin on cell proliferation
Firstly, MDA-MB-231 and MCF-7 cells were incubated in different 96-well plates at a density of 6.0×10^3 cells per well. After cell adhesion, each group was given different concentrations of Syringin (0, 20, 40, 80, 160, and 320 µg/mL) for 24 h and 48 h respectively. Then, 20 µL MTT (Sigma, MO, USA) was added to each well, and cells were cultured for another 4 h. After removing the supernatant, 150 µL dimethyl sulfoxide (DMSO) (Sigma) was added to each well and stirred on QB-9001 micropore rapid shaker (Kylin-Bell Lab Instruments Co. Ltd., Jiangsu, China) for 10 minutes to obtain crystal violet product. Finally, the absorbance of each well was read at 490 nm using a spark 10 m microplate reader (Tecan, Männedorf, Switzerland).
Candidate genes collection
To screen the targets of Syringin against these two BC cell lines, 6 databases were used in this study. Firstly, importing the structure of Syringin into SwissTarget Prediction (https://labworm.com/tool/swisstargetprediction) and PharmMapper databases (http://www.lilab-ecust.cn/pharmmapper/submitfile.html) to obtain all targets of Syringin. Then, the species was limited to humans, and the Gene Official Symbol format of targets was obtained via the Retrieve/ID mapping of Uniprot (https://www.uniprot.org/) database. The related genes of these two BC cell lines were obtained with the keyword “Breast cancer” by GeneCards (https://www.genecards.org/) and DrugBank (https://www.drugbank.ca/) databases at the same time. Finally, the Venny 2.1.0 database (https://bioinfogp.cnb.csic.es/tools/venny/) was used to find their intersection, which was the candidate targets.
Gene ontology and KEGG pathway enrichment
By using Database for Annotation, Visualization, and Integrated Discovery (DAVID) database (https://david.ncifcrf.gov/), the candidate targets related to the Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway were analyzed. Then, the top 25 KEGG pathways and GO entries (P < 0.05) were drawn as bubble plots via the Omicshare database (http://www.omicshare.com) [17].
Screening of hub genes
Using the plug-in of Cytoscape 3.8.1 software (https://cytoscape.org/download.html), a “C-T-P” network was constructed to analyze the association among them [18]. According to the topological parameters of CytoHubba, the hub genes were selected. Then, using the STRING database (https://string-db.org/) to construct a “PPI” network about the 12 candidate genes [19]. After a comprehensive analysis of the two results, the hub genes were obtained.
Molecular docking
Semi-flexible molecular docking technology is widely used to study the binding energy between drugs and genes, to verify the interaction sites and amino acid residues between them. AutoDockTools-1.5.6, PyMOL, Chem Draw, RSCBPDB (https://www.rcsb.org/), and PubChem (https://pubchem.ncbi.nlm.nih.gov/) are the most commonly used software or databases for molecular docking [20]. First, importing the hub targets into the RSCBPDB to download their PDB formats, and then download the 3D structure of Syringin through PubChem. After processing with PyMOL and Chem Draw respectively, they were imported into AutoDockTools-1.5.6 to dock. By analyzing the clusters in docking results, the optimal conformations of Syringin and hub genes were obtained according to the binding energy. It is commonly believed that the docking energy is inversely proportional to the interaction force between them, and the threshold is - 4 kcal/mol. At the same time, the maximum distance of hydrogen bond between component and protein was set to 2 nm to ensure the accuracy of molecular docking.
Then, the protein-ligand interaction profiler (https://plip.biotec.tu-dresden.de/plip-web/plip/index) and ZBH-Center for Bioinformatics (https://proteins.plus/) databases were used to describe the binding amino acid residues of the components and genes in the original PDB database [21, 22]. Finally, the differences and similarities between them were analyzed.
Plate cloning experiment
The 6-well plates were divided into three groups, and 2.0×10^3 cells (MDA-MB-231 and MCF-7) were added to each well. After adherence, added 2 mL of Syringin with the concentration of 0, 160, 320 µg/mL to each well respectively. 48 h later, replaced the medium in each well. On the 15th day, the medium of each well was discarded and the cells were fixed with 4% polyoxymethylene for 15 minutes. After washing with PBS 3 times, 1.5 mL crystal violet was added to each well. 20 minutes later, the 6-well plates were cleaned with flowing water until the background was clear and photos were taken.
Hoechst 33342/PI staining experiment
The 6-well plates were divided into three groups, and 5.0×10^5 cells (MDA-MB-231 and MCF-7) were added to each well. After adherence, added 2 mL of Syringin with the concentration of 0, 160, 320 µg/mL was added for 48 h, respectively. Then, the cells were fixed with 4% polyoxymethylene for 15 minutes. After washing with PBS 3 times, 1.0 mL Hoechst 33342 (Beyotime, 8 µg/ml) to each well for 15 minutes at 37℃. Then, added 1.0 mL PI (Biosharp, 20 µg/ml) to each well at 4℃ for 10 minutes, and observed the cells with a fluorescence microscope.
Cell migration experiment
The 6-well plates were divided into three groups, and 5.0×10^5 cells (MDA-MB-231 and MCF-7) were added to each well. After adherence, scratched the bottom of each well with the tip of a 1 mL sterilized pipette. Added 2 mL of Syringin with the concentration of 0, 160, 320 µg/mL to each well for 48 h. Then the migration of cells in each well was observed by microscope at different times.
Analysis of Western blot
MDA-MB-231 and MCF-7 cells were inoculated in 6-well plates at the density of 5.0×10^5 cells/well and treated with Syringin (0, 160, 320 µg/mL) for 48 h. Then, the RIPA lysis kit (BL504A, Biosharp, Anhui, CHINA) was mixed with PMSF (BL504A, Biosharp, Anhui, CHINA) and PI (BL615A, Biosharp, Anhui, CHINA) for cell lysis. After 12000 rpm centrifugation for 10 min, the total proteins of the whole cell lysate were quantified by a BCA protein assay kit (69107317, Biosharp, Anhui, CHINA). The same total proteins were separated on 10% SDS-PAGE gel electrophoresis (AS1012, ASPEN, CHINA) and then electrophoretically transferred to a nitrocellulose membrane by electrophoresis (DYY-6C, Beijing, CHINA). After sealing with 5% BSA at room temperature for 1.0 h, the membranes were washed 3 times and incubated with 13 primary antibodies respectively (including p-EGFR, PIK3CA, p-PIK3CA, etc.) at 4°C overnight. The next day, washed the membranes with tris buffered saline + Tween (TBST) 5 times for 30 minutes. Fresh ECL (170-5060, Bio-Rad, USA) mixed solution was added to the protein side of the membranes and detected by luminescence. The film was scanned and archived, and the optical density of the band was analyzed by the AlphaEaseFC (Alpha Innotech, USA).
Screening of genes significantly related to survival and prognosis of BC patients
Studying the relationship between hub genes and the survival rate of tumor patients is an important basis to judge whether they can be developed as biomarkers. The relationship between hub genes and stage of tumor also provides a valuable reference for the judgment of patients’ condition. Therefore, the Kaplan Meier plotter (http://kmplot.com/analysis/index.php) and GEPIA 2.0 (http://gepia.cancer-pku.cn/index.html) databases were used respectively to explore their relationship [23, 24]. The Kaplan Meier plotter includes data from GEO, METABRIC, and TCGA databases. Its purpose was to evaluate the effects of 54,000 different genes (including mRNA, miRNA, and protein) on the survival rate of 21 cancer types, to find and verify survival biomarkers. And using the “Single gene analysis” module in GEPIA 2.0, the expression of hub genes in different stages in BC was analyzed. Based on 9736 tumors and 8587 normal tissues from TCGA and GTEX projects, genes expression in various tumors was analyzed by the ANOVA method (setting |lg FC| < 1 and Q-value < 0.05).
Mutation of genes significantly associated with the survival and prognosis of BC patients
The mutation of genes plays an important role in the occurrence and development of tumors. It causes changes in gene structure and quantity and results in the loss of gene function. According to the types of mutated genes, targeted therapy is carried out, which provides a new idea for cancer therapy and prevention. Then using the “Mutation” of the cBioPortal database (http://www.cbiobortal.org) to analyze the 9536 samples of 18 studies about BC based on TCGA [25].
Real-time PCR and the Human Protein Atlas Analysis
Total RNA was extracted from each group of cells by Trizol Reagent (15596018, Invitrogen, California, USA) according to the manufacturer’s guidelines. The gene primers were designed and synthesized by PINUOFEI Biotechnology Co., Ltd (Wuhan, CHINA). Then, using FastStart Universal SYBR Green Master (04913914001, Roche, Basel, Switzerland) and SLANR Real-Time PCR (Shanghai Hongshi Medical Instrument Co., Ltd, CHINA) to determine the expression of mRNAs. The following thermal cycling conditions were adopted: pre-denaturation at 95℃ for 10 minutes, denaturation at 95℃ for 15 seconds, denaturation at 60℃ for 60 seconds (40 cycles). The following condition of the melting curve was used: 60℃→95℃, 1℃ was increased every 20 seconds. GAPDH was used as an internal reference and the 2-ΔΔCT method was used for relative quantification. At the same time, using the Human Protein Atlas database (https://www.proteinatlas.org/) to analyze the gene expression in clinical tumor pathology and normal tissues.
Construction of “mRNA-miRNA-lncRNA” network
LncRNA indirectly inhibits the negative regulation of miRNA on genes by competing with miRNA to bind the 3'-UTR of mRNA. Therefore, exploring the relationship between these three is conducive to a more comprehensive exposition of the mechanism of Syringin against these two BC cell lines. Using miRTarBase (http://mirtarbase.mbc.nctu.edu.tw) and miRWalk (http://mirwalk.umm.uni-heidelberg.de/) databases to screen miRNAs corresponding to biomarkers [26, 27]. After the miRNAs significantly associated with prognosis of patients with BC were obtained via Kaplan Meier plotter and GEPIA 2.0, TarBase v.7 (https://doi.org/10.1093/nar/gkx1141) and lnCAR (https://lncar.renlab.org/explorer) databases were used to screen lncRNA significantly associated with miRNA (score threshold was 0.9) [28, 29]. Then, the relationship between lncRNA and prognosis of BC patients was analyzed by “Survival Analysis” in GEPIA 2.0 databases. Finally, a “mRNA-miRNA-lncRNA” (ceRNA) network is constructed by Cytoscape 3.8.1 software.
Statistical analysis
In this study, the significant data of bioinformatics annotation were screened according to P < 0.05 of students’ t-test. Kaplan-Meier survival plot was used to calculate the hazard ratio of 95% confidence intervals and log-rank P-value. And using ANOVA to analyze the relationship between hub genes and the stages of BC in the GEPIA 2.0 database. The thresholds of |Log2fc| and Q-value were 1 and 0.05, respectively. CBioPortal database was used to obtain information about mutations in hub genes with Z-score. SPSS 22.0 software (IBM, USA) was used for in vitro experiments. All experiments were repeated three times, expressed as the mean ± SD, and P < 0.05 was considered significant.