Exploring the mechanism of cordycepin combined with doxorubicin in treating glioblastoma based on network pharmacology and biological verification

Reagents, antibodies, and chemicals

Fetal bovine serum, RPMI 1640, DMEM, MEM culture medium and penicillin and streptomycin were purchased from Gibco (Grand Island, NE, USA). Cordycepin, doxorubicin and MTT were purchased from Sigma (Ronkonkoma, NY, USA). Trypsin was purchased from Beijing Solaibao Technology Co., Ltd.. Dimethyl sulfoxide was purchased from Sinopill Chemical Reagents Co., Ltd.. Primary antibodies against GAPDH, E-cadherin, N-cadherin, Zinc finger E-box binding homeobox 1 (Zeb1), Twist1, NFKB, MAPK8, MMP-9, and cleaved caspase 3 were purchased from Cell Signaling Technology.

The cordycepin CAS number is 73-03-0, and its HPLC is >98%. The doxorubicin CAS number is 25316-40-9, and its purity is 98.0%–102.0%.

Cell culture

Human glioblastoma cells (LN-229, U251, T98G) came from the American Type Culture Collection (Manassas, VA, USA). After resuscitation, LN-229, U251 and T98G cells were respectively cultured in an RPM-1640, DMEM and MEM medium containing 10% fetal bovine serum and 1% antibiotics (penicillin and streptomycin) in a constant-temperature incubator with a volume fraction of 5% CO₂ at 37 °C. The cells were grown to 90% on the wall and were digested and subcultured with 0.25% trypsin. The logarithmic-phase cells were taken for experiments.

In this study, we dissolved cordycepin and doxorubicin in complete medium.

Cell morphology

LN-229, U251 and T98G cells in the logarithmic growth phase were diluted to 1 × 10⁴ cells/mL in a cell suspension, and 1 mL was inoculated on a six-well culture plate and cultured in an incubator at 37 °C with 5% CO₂. After the cells were fully expanded, a control group and experimental groups were formed. The control group had 2 mL of complete medium. The experimental groups contained 80 μmol/L of cordycepin and 1 μmol/L of doxorubicin separately and the same concentrations of each drug in combination. The solutions for culture had a final volume of 2 mL each and were maintained at 37 °C and 5% CO₂ in an incubator for 48 h. Microscopic observation at ×200 magnification was used to record the treatment group effects on glioblastoma cells.

The survival rate analysis was calculated as follows: cell survival rate = [OD_{test group} − OD_{zero hole}] ÷ [OD_{control group} − OD_{zero hole}] × 100%, in which OD represents optical density.

Cell viability assay

The experimental method refers to the method of Chen, Wang & Li (2012), Wang & Li (2013) and Syam et al. (2014) and is modified. LN-229, U251 and T98G cells in the logarithmic growth phase were selected and counted under a microscope; then, the cell concentration was diluted to 1 × 10⁴ cells/mL and seeded into 96-well plates, with 100 μL per well. After the cell adherent was fully expanded, the drug was administered by treatment group: cordycepin alone, doxorubicin alone, or the combined treatment. At the same time, the control group without medicine was established. Each group was placed into six parallel wells for culture and was incubated overnight in 5% CO₂ at 37 °C for 48 h. After the culture, the culture medium was discarded, and 20 μL of MTT (5 mg/mL) was added to each well. The culture was continued for 4 h; then, the culture medium was discarded, and 150 μL of dimethyl sulfoxide was added to each well and mixed thoroughly until the purple brown precipitate was completely dissolved. The OD was measured for each well at a wavelength of 490 nm. Survival analysis was calculated as follows: cell survival rate = [OD_{experiment group} − OD_{zero pore}] ÷ [OD_{control group} − OD_{zero pore}] × 100%. We conducted at least three repeated biological experiments to ensure the accuracy of the results.

Colony formation assay

We took logarithmic LN-229, U251 and T98G cells and the cell suspension was diluted to 1 × 10³ cells/mL and inoculated onto a six-well culture plate. The control group was a complete medium (2 mL); the experimental groups contained cordycepin (80 μM), doxorubicin (1 μM), or 80 μM of cordycepin combined with 1 μM of doxorubicin in culture to a final volume of 2 mL. Each group was added to three wells, and the culture medium was replaced every 3 days. If the number of cells observed in the colony was greater than 50, the clone had formed, and the culture could be terminated. The drug-containing culture medium was absorbed, discarded, and washed with phosphate-buffered saline (PBS) three times. We fixed the colonies with glutaraldehyde (6.0%v/v) and a sufficient amount of crystalline violet (0.1%) was added and stained at room temperature for 30 min; then, the crystalline violet was absorbed, phosphate-buffered saline was added and the solution was rinsed three times until no obvious purple background was present. We counted using a stereomicroscope. Counting was determined as follows: cloning formation rate = [clone number of experimental group ÷ clone number of control group] × 100%. We conducted at least three repeated biological experiments to ensure the accuracy of the results.

Wound healing

The experimental method refers to the method of Franken et al. (2006) and is modified. Three horizontal lines were drawn on the back of the sterile six-well plate and passed through the holes. The distance between each horizontal line was 0.5–1 cm. LN-299 cells in the logarithmic growth period were taken and counted under the microscope. Each well was inoculated with 1 mL of diluted cells to 1 × 10⁴ cells/mL and placed in an incubator at 37 °C and 5% CO₂ for additional culture. When the cell length in the hole reached approximately 80–90%, perpendicular to the horizontal line behind the hole, the six-well plate after the scratch was placed under a microscope and photographed, and the scratch width of the original cell was recorded as 0 h. A 10% serum culture medium was added to the control and the experimental groups (cordycepin 80 μM, doxorubicin 1 μM and the same doses of cordycepin + doxorubicin together) for incubation after 48-h cultivation. Wells were photographed under a microscope at 0 h, and a photographic record of cell scratch healing was obtained, with records of the scratch width of cells for 48 h, according to the following formula: cell migration rate = [(original scratch width − current scratch width) ÷ original scratch width] × 100%. We conducted at least three repeated biological experiments to ensure the accuracy of the results.

Network pharmacological analysis

Statistical analysis

Data were expressed as the mean ± standard deviation, and the data were analyzed with SPSS version 25.0 software. One-way analysis of variance was used to compare different treatment groups, and p < 0.05 and p < 0.01 indicated significant and extremely significant differences, respectively.

Drug combination to inhibit the activity of LN-229 cells

Cordycepin combined with doxorubicin can inhibit the growth of glioblastoma LN-229 cells and induce their apoptosis. We first observed the morphological changes of LN-229 cells after 48 h of treatment with cordycepin, doxorubicin and cordycepin combined with doxorubicin. Figure 1A shows that, after 48 h, the growth rate of cells in the control group was faster and there were more adherent cells with a regular morphology and spindle shape. Compared with the number in the control group, the number of adherent cells treated with 80 μmol/L of cordycepin slightly decreased, and the cell morphology became slightly round. The number of adherent cells with doxorubicin decreased significantly, and there were some suspended cells. All the cells treated with the combination showed obvious morphological changes; the suspended cells increased and were broken and floating in fragments. Compared with the numbers with cordycepin or doxorubicin alone, the number of adherent cells was the least, the number of apoptotic and rounded cells was the most, and surrounding cells were more obvious in the combination group. These results indicate that cordycepin and doxorubicin alone were both toxic to LN-229 cells, cordycepin combined with doxorubicin on LN-229 cells was more effective obvious.

We then measured the cytotoxicity of the combination on LN-229 cells. MTT analysis (Fig. 1B) showed that cordycepin (80 μmol/L), doxorubicin (1 μmol/L), and the combined treatment gradually increased the cell inhibition rate; the cell survival rates for the treatments, respectively, were 58.97 ± 3.05%, 44.85 ± 5.75%, and 10.85 ± 3.04%. Overall, these results demonstrate that both cordycepin and doxorubicin could inhibit the viability of LN-229 cells and that cordycepin combined with doxorubicin was more inhibitory than that of the single treatment.

In order to better study the effect of cordycepin combination on glioma, we used two other cell lines (U251, T98G) for verification. The results of MTT analysis (Fig. 1D) showed that cordycepin combined with doxorubicin can significantly reduce the survival rate of U251 and T98G cells. The survival rates of U251 cells and T98G cells are respectively 36.33 ± 2.73% and 47.56 ± 3.51%. This shows that the combination of drugs can effectively inhibit the viability of U251 and T98G cells, and that cordycepin combined with doxorubicin has certain toxicity to cells of glioblastoma cell lines.

Figure 1C shows that compared with the control group, the number of clones formed by the cells in the cordycepin combined with doxorubicin group is extremely small. After calculation, Fig. 1G shows that the clone formation rate of U251 cells treated with the combination drug was 1.25 ± 0.02%, while the clone formation rate of T98G cells treated with the combination drug was 0.59 ± 0.01%, which was significantly different from that of the control group (p < 0.01). This shows that the test concentration is appropriate, and the combination of drugs can effectively inhibit the proliferation of U251 and T98G cells.

Combined therapy inhibits the proliferation of glioblastoma cells

To study the effect of cordycepin combined with doxorubicin on the proliferation of LN-229 cells, a colony formation assay was carried out. The cell clone formation assay can be used to test the proliferative ability, invasiveness, sensitivity to killing factors, and other aspects of Fig. 1C shows that, compared with the number in the control group, the numbers of cell clones formed in the cordycepin group and the doxorubicin group were significantly reduced, and the number of cell clones formed in the combination group was very low. The ability to form clones successively decreased from the control group to the cordycepin group, the doxorubicin group, and the combined group. Figure 1D shows that the calculated clone formation rates of cordycepin, doxorubicin, and the combined group were, respectively, 34.18 ± 3.68%, 11.72 ± 0.94%, and 1.55 ± 0.83%, and the clone formation rates of the cordycepin and doxorubicin groups were, respectively, 22.05 and 7.56 times higher than that of the combined group (p < 0.01). These results indicate that the clone ability of LN-229 cells was significantly inhibited by cordycepin, doxorubicin, and the combination therapy, and the inhibition of the proliferation ability of LN-229 cells was stronger with the combination therapy than with the single therapy.

Drug combination inhibits the invasion and migration of LN-229 cells by regulating epithelial mesenchymal transformation

Cell migration is one of the key steps in the invasion and metastasis of malignant tumors. The effect of the drug combination on the migration ability of LN-229 cells was detected by a cell scratch test, and the results are shown in Fig. 2. As seen in Fig. 2A, compared with 0 h, scratches healed after 48 h of cell culture in the control group and the experimental group. The healing abilities of different treatment groups were as follows: the control group more than the cordycepin group, doxorubicin group and combined group, indicating that cells migrated during the culture process. The healing ability of the control group was the strongest, indicating the greatest degree of cell migration, but the scratch healing of the combined group was the weakest. The combination of cordycepin and doxorubicin could effectively inhibit the movement and migration of LN-229 cells. Figure 2B shows that the calculated cell migration rates of the cordycepin group, doxorubicin group, and combined treatment group were, respectively, 87.95 ± 1.12%, 81.05 ± 5.52%, and 56.07 ± 5.85%, and extremely significant differences were documented between the combined group and the cordycepin group and the doxorubicin group (p < 0.01). Combination medication can effectively inhibit the migration ability of U251 and T98G cells (Fig. 2E). Figure 2F shows that the cell migration rate of cordycepin combined with doxorubicin on U251 and T98G cells was 23.97 ± 3.22%, 9.70 ± 2.04%, respectively.

To examine the molecular mechanisms underlying the epithelial mesenchymal transformation (EMT) process, we measured the expression of E-cadherin, N-cadherin, Zeb1, and Twist1 by western blot. The cordycepin group, doxorubicin group and combined group could regulate the EMT process by regulating EMT markers (e.g., N-cadherin, E-cadherin, ZEB1, Twist1) and could inhibit the expression of proteins related to tumor migration. EMT is a biological process in which epithelial cells transform into mesenchymal stem cells with characteristics driven by specific physiological or pathological factors (Yang et al., 2021)^. EMT plays an important role in the process of invasion and migration in glioblastoma. Figure 2C shows that, compared with the control group, the cordycepin group, doxorubicin group, and combined treatment group could upregulate the expression of E-cadherin protein and downregulate the expression of N-cadherin protein, ZEB1, and Twist1 protein. E-cadherin is a major mediator of cell adhesion and epithelial tissue integration, and it plays an important role in maintaining cell adhesion (Bremnes et al., 2002). The low expression or deletion of E-cadherin can inhibit intercellular adhesion, trigger tumor detachment from the primary site, and lead to invasion. The decreased expression level of E-cadherin protein was significantly correlated with the invasion and metastasis ability of tumor cells. N-cadherin is an important marker of EMT. ZEB1 is an important regulatory gene for EMT; it can inhibit the expression of E-cadherin and promote EMT and tumor invasion and metastasis (Zhao et al., 2018). The results of this study showed that, compared with expression in the control group, the expression of E-cadherin was upregulated in all the experimental groups. The expression of E-cadherin protein was the highest in the combined treatment group, whereas the expression of N-cadherin protein was low, thus reducing the invasion and metastasis of glioblastoma.

Candidate targets of cordycepin, doxorubicin and glioblastoma

After human gene name replacement, we gained 146 and 2,494 potential targets of cordycepin and doxorubicin, respectively. The PubChem CID, molecular formula, and number of targets of cordycepin and doxorubicin collected from the PubChem compound database in NCBI are shown in Table 1.

A total of 4,159 glioblastoma-related targets were obtained from the GeneCards database using “glioblastoma” as the keyword. The targets of cordycepin, doxorubicin, and glioblastoma were imported into a Venn diagram, and we obtained 71 intersection targets (Fig. 3).

Biological process enrichment and pathway analysis

The potential targets of cordycepin combined with doxorubicin to treat glioblastoma were imported into the DAVID database for KEGG pathway enrichment analysis; the top 20 pathways with the smallest p values were arranged according to the p value, from small to large, and drawn into a bubble diagram (Fig. 4B). The abscissa of the bubble graph is –log10 (p value) of the pathway, and the ordinate is the name of pathway. The color represents the p value, and a redder p value indicates a more significant enrichment. The bubble size indicates the number of enrichment targets in the pathway, and the larger the bubble reflects more enriched genes.

Figure 4B shows that combination therapy to treat glioblastoma involves the pathways mainly enriched in proteoglycans in cancer, the tumor necrosis factor (TNF) signaling pathway, microRNA in cancer and pathways in cancer, among which the proteoglycans in cancer pathway had the smallest p value (Table 2). After analysis, 35 targets were enriched in the first four pathways. We imported 35 common targets into the DAVID database for GO biological process analysis and obtained 64 GO terms with very significant differences (p < 0.01). Figure 4A shows GO biological process entries with p values <0.01; −log10 (p value) is on the horizontal axis, and the GO term serial number is on the vertical axis (see the attached table for detailed information of GO terms). The entries related to cell apoptosis, proliferation, and migration are marked out. The blue entries represent the biological processes related to cell apoptosis; the red ones, cell proliferation; the black ones, cell migration. The analysis of biological processes showed that cordycepin combined with doxorubicin could regulate cell apoptosis, cell proliferation, cell migration, nucleotide synthesis, and transcription by acting on multiple targets, thus leading to the apoptosis of glioblastoma and inhibition of the migration of glioblastoma.

Construction of a drug-disease-target-pathway network and PPI network

In order to further explore the mechanism of cordycepin combined with doxorubicin in the treatment of glioblastoma, we selected the first four pathways with the lowest enriched p value and their related targets for analysis. The red nodes in the figure represent glioblastoma, the green nodes represent pathways, the related genes are yellow targets, and the edges between nodes represent the interaction between them.

In order to further confirm the possible important targets in combination, we constructed PPI network for 35 targets enriched in the first four pathways. The 35 targets were imported into the STRING database, the species was set to Homo sapiens, the medium confidence was set to 0.4, and the remaining settings were kept at defaults; the data were saved as a TSV file. TSV files were imported into Cytoscape software to build the PPI network, as shown in Fig. 5B. Overall, 35 nodes interacted with each other through 226 edges. The circular nodes represent target genes, and the straight lines between nodes indicate the interaction between two linked proteins. Thicker lines represent stronger interaction relationships. The size and color depth of nodes in the figure are set by the degree value; that is, the larger the node and the darker the color, the greater the degree value of the corresponding target gene and vice versa. As seen in Fig. 5B, the nodes of targets such as MYC, CASP3, MAPK8, SRC, NFKB1, MMP9, TLR4, and CTNNB1 are relatively large, and the color is relatively dark, indicating a relatively large degree value. These targets may play an important role in the treatment of glioblastoma using combined cordycepin and doxorubicin.

Validation of key targets

We use network pharmacology to predict the possible target of cordycepin combined with doxorubicin in the treatment of glioblastoma. Then 71 intersection genes were imported into the DAVID database for GO and KEGG analysis. GO analysis can be divided into three parts: Biological Process, Cell Component and Molecular Function. The analysis found that a total of nine items (1. negative regulation of apoptotic process; 2. apoptotic process 3. apoptotic signaling pathway 4. regulation of apoptotic process 5. negative regulation of intrinsic apoptotic signaling pathway 6. negative regulation of extrinsic apoptotic signaling pathway via death domain receptors 7. positive regulation of apoptotic process 8. positive regulation of neuron apoptotic process 9. negative regulation of cysteine-type endopeptidase activity involved in apoptotic process are enriched in relation to cell apoptosis), and two items (1. regulation of cell proliferation 2. cell proliferation) are related to cell proliferation. Two biological processes related to cell migration. After KEGG analysis, it was found that a total of 74 KEGG pathways were enriched, and the first top 20 pathways with p values from small to large were visualized using bubble charts. We analyzed the first top 4 pathways that may play an important role in the treatment of glioblastoma with cordycepin combined with doxorubicin, and were involved 35 targets. We use Cytoscape 3.8.0 to perform PPI network analysis on 35 targets. The greater the degree of the target, the more important the gene may be in the treatment of glioblastoma with cordycepin combined with doxorubicin. We detected the first few targets with a higher Degree value in the PPI network map by western blot, and explored their expression in the drug combination. Generally, the potential targets that have higher degrees value are more pharmacologically important. To explore the molecular mechanism of cordycepin combined with doxorubicin in the treatment of glioblastoma, we detected the expression levels of NFKB1, MAPK8, MYC, MMP9, and cleaved caspase 3 in LN-229 cells treated with cordycepin and doxorubicin for 48 h by western blot analysis (Fig. 6), and these genes were the targets with the highest degree values in the PPI network. NFKB1 is an important intracellular nuclear transcription factor, which is involved in inflammation and immune regulation and mainly plays a role in the nuclear factor kappa B pathway in the TNF signaling pathway (Peluso et al., 2019). MAPK8, also known as c-Jun terminal kinase 1, is an important member of the MAPK family. MAPK8 can phosphorylate and activate transcription factor activator protein-1, thereby activating the expression of a series of downstream genes, and it plays an important regulatory role in cell proliferation, cell differentiation, cell apoptosis, inflammation, and other pathological processes (Kini & Garg, 2015; Xu et al. 2018). MMP-9 is a matrix metalloproteinase that regulates cell adhesion and is associated with the occurrence, progression and local invasion of gliomas (Wu & Shen, 2021). Caspase 3 is a member of the Bcl-2 family and is the most important terminal splicing enzyme in the process of cell apoptosis; it can inhibit programmed apoptosis of tumor cells and promote proliferation of tumor cells (Humphreys, Espona-Fiedler & Longley, 2018). MYC is an oncogene encoding nuclear protein, which can be divided into C-Myc, N-Myc, and I-Myc; it is related to cell proliferation, tumorigenesis, and outcomes. We analyzed the genes related to the first top four pathways and performed PPI network analysis on them, and verified the more important targets. EMT is a relatively complex biological process, among which CASP3, MAPK8, MMP9, and NFKB1 are also EMT-related genes in the Genecards database. Onishi et al. (2011) have reported that MMP-9 maintain the rigidity of the extracellular matrix in vitro, and promote the structural rigidity, movement and proliferation of established glioma cell lines. What we verified is not genes that have nothing to do with cell migration. Each gene may be involved in a variety of biological processes. Among them, CASP3 and MYC are also genes related to cell apoptosis, and genes such as MAPK8 are related to cell proliferation, which also echoes the analysis of GO biological processes. The results of this study showed that, compared with the control group, the combined treatment group significantly downregulated the expression of NFKB1, MAPK8, MYC, and MMP-9 proteins; upregulated cleaved caspase 3; promoted the apoptosis of LN-229 cells; and inhibited the proliferation and migration of tumors.