Therapeutic targets and signaling pathways of active components of QiLing decoction against castration-resistant prostate cancer based on network pharmacology

Cell culture and reagents

PC3 cells were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA), and cultured in a Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS), 100 units/mL penicillin, and 100 µg/mL streptomycin at 37 °C in a humidified 5% CO₂ atmosphere. Formononetin (cat. no. 20100311) was purchased from Jubang Biotechnology Co., Ltd. (Chengdu, China), and Calycosin (20575-57-9) was purchased from Sigma-Aldrich (St. Louis, MO, USA), both of which were dissolved with dimethyl sulfoxide (DMSO). The cells in the experimental groups were treated with formononetin (10 µM) or calycosin (20 µM), and the control cells were treated with DMSO.

Screening of anti-CRPC targets of QLD

The active components of QLD in each herb were identified by the Traditional Chinese Medicine Systems Pharmacology (TCMSP) (http://tcmspw.com/tcmsp.php) and the Traditional Chinese Medicine Integrated Database (TCMID). Oral bioavailability (OB) and drug likeness (DL) were used for the screening of active components, as previously reported. DL ≥ 0.18 and OB ≥ 30% were considered to be the threshold for candidate compounds.

To further identify the target proteins on the active components, we applied the Online Mendelian Inheritance in Man (OMIM; https://omim.org/), DisGeNET (http://www.disgenet.org/), and GeneCards (https://www.genecards.org/) to screen the pathological targets of CRPC. The active components of QLD and potential pathological targets evaluated were predicted using Venn diagrams. Cytoscape 3.8.2 (https://cytoscape.org/) was used to reveal the active components and potential targets in the form of a network diagram.

Screening of the most appropriate targets of QLD against CRPC and construction of an interrelated network

The mapped targets of QLD against CRPC were re-analyzed using the STRING database to establish a target-to-target function-related protein network. Cytoscape software (ver. 3.7.1) was used to visualize the results by analyzing the topological parameters (Su et al., 2014). The maximum degree-value in the topological data was set as the upper limit and twice the median of degree was taken as the lower limit (Li et al., 2019). The top five genes with the highest degree of connectivity were used for screening.

Gene ontology (GO) and kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analyses

RStudio 3.6.1 software (http://www.interactivenn.net) was used for the GO and KEGG pathway enrichment analyses to determine the role of target proteins that interact with the active components of QLD in gene function and signaling pathway. The data were acquired using a cut-off value adjusted to P < 0.05.

Establishment of CRPC xenograft models and immunohistochemistry (IHC) in vivo

To establish CRPC xenograft models, BALB/c nude mice (six week-old males) were provided by LONGHUA Hospital, which is affiliated with the Shanghai University of Traditional Chinese Medicine (Shanghai, China). The mice were injected with 100 µL phosphate-buffered saline (PBS) containing PC-3 cells into the left axilla of the mice, as described previously (Fu et al., 2019). The mice were then randomly divided into four groups (n = 6) on day 7 after the injection. The groups included a control group (vehicle solution was administered intragastrically), QLD group (9 g/kg QLD (body weight) was given intragastrically), Doc group (5 mg/kg QLD (body weight) was given intragastrically), and QLD + Doc group (QLD (9 g/kg) and Doc (5 mg/kg) were given intragastrically). The mice with QLD were fed once daily and while the Doc group was fed once a week. The treatment was maintained for 28 days and then all mice were euthanized. The tumor weight and volume were monitored. The tumor volume (mm³) was calculated as follows: tumor volume = (L × W²)/2 (L: length, W: width). IHC was conducted to detect the expression level of PCNA in cancer cells using PCNA antibody (cat. no. ab29, 1/10000; Abcam, Cambridge, UK), in accordance with previous reports. All applicable international, national, and/or institutional guidelines for the care and use of animals were followed. The protocol was approved by the Ethics Committee of LONGHUA Hospital Shanghai University of Traditional Chinese Medicine.

Cell counting kit-8 (CCK-8) assay

The PC3 cells were cultured in 96-well plates and when the cells reached a density of 2  × 10³ cells/well, 10% CCK-8 (Sigma-Aldrich) was added into 100 ul fresh mixture, followed by incubation at 37 °C for 1 h. Finally, the optical density (OD) value was measured at a wavelength of 450 nm using a spectrophotometer.

Colony formation assay

The PC3 cells treated with the indicated reagents were seeded into 6-well culture plates at a density of 1  × 10³ cells/well. Two weeks later, the colonies were fixed with 4% paraformaldehyde for 15 min at room temperature and visualized by a camera.

Western blot assay

Total proteins were extracted from cells using RIPA lysis buffer (Beyotime Institute of Biotechnology, Shanghai, China). The proteins were quantified using a BioRad Bradford protein assay kit and electrophoresed into sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Then, the proteins were transferred onto polyvinylidene fluoride (PVDF) membranes, followed by incubation with anti-HIF-α (Abcam, Cambridge, UK; cat. no. ab179483, 1:2000), anti-vascular endothelial growth factor A (VEGFA) (Abcam, Cambridge, UK; cat. no. ab214424, 1:1000), anti-tumor necrosis factor alpha (TNF-α) (Abcam, Cambridge, UK; cat. no. ab208156, 1:2000), anti-interleukin 6 (IL6) (Abcam, Cambridge, UK; cat. no. ab233706, 1:2000), anti-P53 (Abcam, Cambridge, UK; cat. no. ab26, 1:2000), anti-cleaved-caspase-3 (cat. no. ab32042, 1:2000), anti-Bax (Abcam, Cambridge, UK; cat. no. ab32503, 1:2000), and anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Abcam, Cambridge, UK; cat. no. ab8245, 1:1000) at 4 °C overnight, and then washed with Tris Buffered Saline with Tween (TBST) three times. The horseradish peroxidase (HRP)-labeled secondary antibodies (ab205718/ab150117, 1:20000, Abcam) were used for the secondary incubation for 1 h. Finally, imaging was performed by the ChemiScope 6200 Touch Chemiluminescence imaging system (Clinx Science Instruments Co. Ltd., Shanghai, China), and the band intensity was quantified using ImageJ software (National Institutes of Health, MD, USA).

Statistical analysis

All of the data were analyzed by GraphPad Prism (GraphPad Software Inc., San Diego, CA, USA) and SPSS 22.0 (IBM, Armonk, NY, USA) software, and were presented as mean ± standard deviation (SD). The t-test or one-way analysis of variance (ANOVA) were used to compare differences between groups. P < 0.05 was considered to be statistically significant.

Components of QLD

QLD consisted of 10 herbs, of which we identified six herbs with 1,157 active components using the TCMSP database. Specifically, there was one component in Baimaoteng, 515 components in Huangqi, 41 components in Jianghuang, 236 components in Shegan, 34 components in Shudihuang, and 330 components in Yimucao (File S1). A total of 51 potential components were screened based on the adopted criteria, including OB ≥ 30% and DL ≥ 0.18 (Table 1).

Screening of effective anti-CRPC components of QLD

Determining the effectiveness of QLD in preventing and treating CRPC depends on the synergic interaction between active components and their targets; thus, we evaluated the relationships among the active components and their targets using the compound-target network. The active components of six herbs of QLD and their target proteins related to CRPC were detected using TCMSP and TCMID. The interaction between the active components and target proteins was analyzed using the Cytoscape 3.2.1 software. The results are shown in Figs. 2 and 3.

Construction of the protein–protein interaction (PPI) network and identification of core target genes

A total of 1,086 pharmacological target genes were identified using the TCMID database based on the 51 primary components found above. The overlapping genes were removed and 149 proteins were identified, which were converted into gene names within the limited species of “Homo sapiens” via the UniProtKB. Next, the OMIM, DisGeNET, and GeneCards databases were used to collect the CRPC-related target genes. Venn diagrams were drawn to display the intersected genes of these two target sets, and 66 intersected genes were found. The PPI network was then established to display the intersection of the target genes of QLD against CRPC and CRPC-related biotargets based on the STRING database (Fig. 4). The mapped proteins were imported into Cytoscape in order to filter the core target genes, and five core hub target genes were obtained, including tumor protein P53 (TP53), IL6, VEGFA, caspase-3 (CASP3), and estrogen receptor 1 (ESR1) (Fig. 5).

Core target genes in GO and KEGG pathway enrichment analyses

The potential pathways of QLD were analyzed to better understand the mechanisms involved in the development of CRPC. GO and KEGG pathway enrichment analyses were conducted using the “clusterprofiler” package in R for the amelioration of CRPC. The GO enrichment and the KEGG pathway analyses were conducted to explore the potential biological processes (BP), cellular components (CC), and molecular functions (MF). A histogram and bubble chart were plotted using the ggplot2 package (version 1.0.2) in the R software (Fig. 6). The top 10 terms for GO-BPs were more abundant in the negative mediation of miRNAs production involved in gene silencing by miRNA, gene silencing by miRNA, posttranscriptional gene silencing, gene silencing by RNA, and gene silencing. The regulation of miRNAs production involved in gene silencing was performed by miRNA, and the production of small RNA involved in gene silencing was carried out by dsRNA processing. The remaining BPs are listed in File S2. Additionally, GO-MFs related to the target genes mainly included RNA polymerase II general transcription initiation factor binding, cytokine receptor binding, general transcription initiation factor binding, basal transcription machinery binding, basal RNA polymerase II transcription machinery binding, protease binding, growth factor receptor binding, growth factor activity, cytokine activity, and TFIID-class transcription factor complex binding. The remaining GO-MFs are shown in File S2.

KEGG pathway enrichment analysis demonstrated that the six hub target-related KEGG pathways included the MAPK signaling pathway, resistance to epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKIs), microRNAs in cancer, the PI3K–Akt signaling pathway, AGE–RAGE signaling pathway in diabetic complications, HIF–1 signaling pathway, endometrial cancer, central carbon metabolism in cancer, IL–17 signaling pathway, rheumatoid arthritis, P53 signaling pathway, platinum drug resistance, etc (Fig. 7). The remaining pathways are listed in File S3.

Establishment of network diagram

Cytoscape software was used to determine the targets of QLD against CRPC and the hub target-associated pathways, and a network was constructed. All of the drug protein targets and disease-related proteins were screened. Cytoscape was used to detect the targets of QLD against CRPC, the hub target-associated pathways, and a network visualization of QLD against CRPC targets. An interaction diagram for core target-related pathways was plotted (Fig. 8).

QLD suppressed the growth of CRPC tumor in vivo and in vitro

We established CRPC xenograft models using PC-3 cells to investigate whether QL has therapeutic effects on CRPC tumor growth in vivo. Considering the therapeutic significance of Docetaxel (Doc) on CRPC (Lowrance et al., 2018), Doc was chosen as the reference. The xenograft tumor in the QLD group was found to be significantly smaller and lighter than that in the control group. The QLD + Doc group exhibited the lightest and smallest xenograft tumor, with an approximately 76% repression in tumor growth (Figs. 9A–9C). Furthermore, IHC assay demonstrated that QLD efficiently inhibited the expression level of PCNA in the xenograft tumor. The combination of QLD + Doc was found to decrease the level of positive cells versus the use of single drugs (Fig. 9D). The in vivo assay suggested that QLD could effectively exert repressive effects on CRPC tumor growth. The molecular targets that might be affected by active components of QLD in CRPC were also investigated. The QLD treatment was found to suppress the activity of HIF-a/VEGFA and TNF-a/IL6 signaling pathways and increase the P53 expression level, which is consistent with the results of KEGG pathway enrichment analysis (Fig. 9E). Furthermore, we included two active components, formononetin and calycosin, which have anti-cancer effects on diverse types of cancer (Yang et al., 2021; Yu et al., 2020). Western blot analysis revealed that formononetin suppressed the growth of CRPC cells by blocking the activity of HIF-a/VEGFA and TNF-a/IL6 signaling pathways (Figs. 9F–9H). Calycosin also inhibited CPRC cells by inducing apoptosis and activating the P53 pathway (Figs. 9I–9K). QLD was found to efficiently suppress CRPC tumor growth.