Molecular pathways in glioblastoma-derived stem cells to identify effective drug agents: A bioinformatics study.

Background and Aim: Glioblastoma multiform (GBM) is considered as one of the malignant brain tumors that affect a wide range of people every year. Cancer stem cells, as essential factors, are resistant to chemotherapy drugs and complicate treatments. Therefore, finding critical molecular pathways in GBM-derived stem cells, and selecting the appropriate drug agents can prove more effective treatment approaches for GBM. Method: In this study, using RNA-Seq data, we performed continuous bioinformatics analyses and examined the up-and down-regulated genes from GBM-derived stem cells samples. Afterward, we separated the signaling pathways using the KEGG database and measured the protein interactions with the STRING database. Then, using the Drug matrix database, we nominated drugs that could affect these genes.
Results: The first 20 pathways on tumorigenesis and 41 up-regulated and 73 down-regulated genes were selected. These genes were most active in the pathways involved in cell division, metabolism, cytoskeleton, cell adhesion molecules, and extracellular space. We then examined the candidate genes and the approach of the drugs that target these genes. Chlorambucil, cyclosporine A, doxorubicin, and etoposide were selected as the drug agents.
Conclusion: Using integrated bioinformatics analyses, it was found that prominent genes in the cell cycle and cytoskeletal pathways are more expressed in cancer stem cells and that Chlorambucil, cyclosporine A, doxorubicin, and etoposide can be effective compounds to attenuate these cells. Copyright:

Introduction

Glioblastoma multiform (GBM) is a malignant brain tumor that affects many people every year. Despite the new treatment options, its mortality is high. Moreover, due to the late diagnosis of asymptomatic patients in the early phase, cancer advances to an advanced stage, and as a result, treatment is difficult or even ineffective.[12] Another challenge is the blood-brain barrier (BBB) that makes it difficult for the medication to pass through. In addition, tumor cells have shown resistance to drug compounds during several chemotherapy sessions, and one of the main factors for this is the stem cells derived from cancerous tissues.[3]

Cancer stem cells constitute 1% of the total population of cancer cells. Despite chemotherapy, these cells can survive due to their small size and form new tumor cells, which due to their differentiation and automotive properties, are more resistant to drugs.[45] In one study, it was found that temozolomide was ineffective against GBM-derived stem cells.[2] Therefore, new strategies in the study of signaling pathways and molecules involved in cancer stem cells may help in identifying more suitable drug compounds to overcome cancer stem cells.

In the past few years, bioinformatics has proven to be very effective in identifying essential elements in cell physiology and their pathways that play a beneficial role in predicting molecular functions and the nature of genes and protein products.[678] Therefore, this study aimed to investigate the pathways of cancer stem cells derived from GBM patients. Finally, our attempt was to select the required genes in the development and progression of GBM and to determine the effective drug agents in this regard.

Methods and Materials

Selection of databases

In this study, we selected the RNA-Seq dataset (GSE92459) from the SRA database. This dataset contained 21 samples from stem cells derived from GBM patients and several cells as a control sample.

Classifying the data and performing bioinformatics analyses

In this step, we uploaded the GSE92459 to the Biojupies database and then separated the signaling pathways from the wiki pathway database and the gene ontology from the Enrichr database. In this section, all routes were classified according to the P value <0.05.

Investigation of protein association

In this section, we isolated the genes that were closely related to tumorigenesis and the progression of GBM, loaded them into the STRING database; and finally, isolated the protein network of the upregulated genes.

Choosing effective medications

Following bioinformatics analysis, we entered the critical genes, both in terms of ontology genes and the relationship between their proteins, in the Drug matrix database and then listed the drugs associated with cancer pathways along with their dosage.

Examining and plotting essential genes in the pathway of various cancers

In this section, essential genes associated with cancer were isolated and uploaded to the GEPIA database for testing with other common cancers and GBM. Also, a Kaplan Meyer box plot diagram was drawn to show the survival rate.

Results

Tumor-dependent neuronal migration, axon guidance, and cell cycle phase-dependent pathways were significantly expressed.

After bioinformatics analysis, mitotic prometaphase, resolution of sister chromatid cohesion, cell cycle, mitotic, gastric cancer network 1, cell cycle, M phase, and mitotic anaphase pathways showed up-regulation. The neuronal system, GABA synthesis, release, reuptake and degradation, axon guidance, neurotransmitter release cycle, and extracellular matrix organization showed down-regulation when GBM-derived stem cells were compared to normal cells [Table 1, Figure 1].

Table 1

Top 10 up/downregulated signaling pathways

Pathways	P	Genes
Upregulated pathways
Mitotic Prometaphase	1.60E-10	ZWILCH, PLK1, CDCA8, SMC4, CENPA, NCAPH, SKA1, CDC20, CCNB2, CENPE, CCNB1, KIF18A, CENPF, BIRC5, KIF2C, CENPN, BUB1, MAD2L1
Resolution of Sister Chromatid Cohesion	3.07E-09	ZWILCH, PLK1, CDCA8, CENPA, SKA1, CDC20, CCNB2, CENPE, CCNB1, KIF18A, CENPF, BIRC5, KIF2C, CENPN, BUB1, MAD2L1
Cell Cycle, Mitotic	2.21E-08	TOP2A, ZWILCH, CDCA8, FOXM1, SMC4, CENPA, NCAPH, SKA1, AURKA, CDC20, CCNB2, CCNB1, PTTG1, NEK2, MYBL2, BUB1, NEK9, CDKN2C, BORA, UBE2C, PLK1, KIF23, CDC25C, CCNA2, CENPE, TPX2, CENPF, KIF18A, DBF4, BIRC5, CENPN, KIF2C, KIF20A, MAD2L1
Gastric Cancer Network 1	2.28E-08	TOP2A, TPX2, CENPF, UBE2C, MYBL2, ECT2, E2F7, AURKA, KIF15
Polo-like kinase mediated events	5.51E-08	CCNB2, CENPF, CCNB1, PLK1, MYBL2, FOXM1, CDC25C
Cell Cycle	1.12E-07	TOP2A, ZWILCH, CDCA8, FOXM1, SMC4, CENPA, NCAPH, SKA1, AURKA, CDC20, CCNB2, CCNB1, TERT, PTTG1, MYBL2, NEK2, NBN, BUB1, NEK9, CDKN2C, BORA, UBE2C, PLK1, KIF23, CDC25C, CCNA2, CENPE, TPX2, CENPF, KIF18A, DBF4, DKC1, BIRC5, CENPN, KIF2C, KIF20A, MAD2L1
RHO GTPases Activate Formins	1.63E-07	ZWILCH, PLK1, CDCA8, CENPA, SKA1, CDC20, CENPE, KIF18A, CENPF, DIAPH 3, BIRC5, KIF2C, CENPN, BUB1, MAD2L1
M Phase	3.05E-07	NEK9, ZWILCH, UBE2C, PLK1, CDCA8, KIF23, SMC4, CENPA, NCAPH, SKA1, CDC20, CENPE, CCNB2, CENPF, CCNB1, KIF18A, PTTG1, BIRC5, KIF2C, CENPN, KIF20A, BUB1, MAD2L1
MHC class II antigen presentation	1.76E-06	SEC23A, SEC24A, KIF23, KIF15, CENPE, KIF18A, RACGAP1, KIF4A, KIF2C, KIF20A, OSBPL1A, SEC24D, SEC31A
Activation of HOX genes during differentiation	2.16E-06	EGR2, HOXA3, HOXB4, HOXA2, HOXB3, HOXC4, HOXA1, PAX6, HOXB2, HOXD4, HOXD3, HOXA4
Downregulated pathways
Arrhythmogenic right ventricular cardiomyopathy (ARVC)	2.81E-07	CACNB2, GJA1, CACNB3, TCF7L1, CACNB4, ITGB5, CDH2, JUP, CACNA2D1, CACNA1C, SLC8A1, ITGA9
Transmission across Chemical Synapses	3.75E-07	GABRB3, GRIA1, BCHE, SNAP25, SLC32A1, KCNJ12, CACNA2D1, GABRA4, GAD1, CHRNA7, CACFD1, RASGRF2, GAD2, SLC6A1, SYN2, GRIP2, GRIN2D, CACNB2, CACNB3, CACNB4
Neuronal System	2.29E-06	GABRB3, GRIA1, BCHE, SNAP25, KCND1, KCNH5, SLC32A1, KCNJ12, CACNA2D1, GABRA4, GAD1, CHRNA7, CACFD1, RASGRF2, GAD2, KCNAB2, SLC6A1, SYN2, GRIP2, GRIN2D, CACNB2, CACNB3, CACNB4
Phase 1 - inactivation of fast Na+channels	3.37E-06	CACNB2, CACNB3, KCND1, CACNB4, CACNA2D1, KCNIP4, CACNA1C
Depolarization of the Presynaptic Terminal Triggers the Opening of Calcium Channels	1.04E-05	CACNB2, CACNB3, CACNB4, CACFD1, CACNA2D1
NCAM1 interactions	3.14E-05	CACNA1I, CACNB2, CACNB3, CACNB4, ST8SIA2, GFRA1, CACNA1C
Nicotine addiction	5.34E-05	GABRB3, GRIA1, SLC32A1, CHRNA7, GABRA4,
GABA synthesis, release, reuptake and degradation	8.33E-05	SNAP25, SLC32A1, GAD1, GAD2, SLC6A1
Mecp2 and Associated Rett Syndrome	0.000177379	GRIA1, RBFOX1, DLX5, GAD1, OPRK1, FGF3, CDON
Phase 2 - plateau phase	0.000407637	CACNB2, CACNB3, CACNB4, CACNA2D1, CACNA1C

Figure 1

Volcano Plot. The figure contains an interactive scatter plot that displays the log2-fold changes and statistical significance of each gene calculated by performing a differential gene expression analysis. Every point in the plot represents a gene. Red points indicate significantly up-regulated genes; blue points indicate down-regulated genes

Investigation of gene ontology

We evaluated three general approaches to molecular functions, biological processes, and cellular components after examining the gene expression profile. Accordingly, in terms of molecular procedures and biological processes, transcriptional activator activity, RNA polymerase II transcription regulatory region sequence-specific binding (GO: 0001228), spindle (GO: 0005819), RNA polymerase II regulatory region sequence-specific DNA binding (GO: 0000977), mitotic sister chromatid segregation (GO: 0000070), antigen processing and presentation of exogenous peptide antigen via MHC class II (GO: 0019886), microtubule cytoskeleton (GO: 0015630), and chromosome, centromeric region (GO: 0000775) showed up-regulation, while nervous system development (GO: 0007399), chemical synaptic transmission (GO: 0007268), calcium ion transport into the cytosol (GO: 0060402), anterograde trans-synaptic signaling (GO: 0098916), neuron projection morphogenesis (GO: 0048812), amyloid-beta binding (GO: 0001540), dendrite (GO: 0030425), and Wnt-activated receptor activity (GO: 0042813) demonstrated down-regulation. The presence of protein products of genes in different cell locations can also be seen in [Figure 2].

Figure 2

Gene ontology enrichment analysis results. The figure contains interactive bar charts displaying the results of the gene ontology enrichment analysis generated using Enrichr. The x-axis indicates the -log10 (P-value) for each term. Significant terms are highlighted in bold. Additional information about enrichment results is available by hovering over each bar. a: Biological processes; b: Molecular functions; c: Cellular component

Correlation between protein networks

We examined 41 up-regulated genes that showed a more critical role and formed 54 nodes and 73 edges in the protein network. This network showed acceptable communication in cell adhesion molecules, TNF, and P53 signaling pathways [Figure 3].

Figure 3

Correlations between protein products of up-regulated genes. Cell cycle (red), cellular senescence (green), and P53 (blue) signaling pathway

Evaluation of drugs that can be effective in treatments of GBM

We isolated important genes that play a significant role in the main processes of tumorigenesis and GBM progression. Azathiprine (20 mg/kg), doxorubicin (3 mg/kg), cyclosporin A (350 mg/kg), chlorambucil (0.6 mg/kg), etoposide (100 mg/kg), leflunomide (30 mg/kg), thioguanine (12 mg/kg), cyclophosphamide (25 mg/kg), daunorubicin (3.25 mg/kg), and clobetasol propionate (17 mg/kg) were the top 10 drugs with high significance P value [Table 2].

Table 2

Candidate drugs associated with up-regulated genes

Drugs and dosage	P	Genes
Azathioprine (20 mg/kg)	1.01E-06	CCNA2;CENPE; CDKN2C; CDCA8;MYBL2;FOXM1;KIF15
Doxorubicin (3 mg/kg)	2.17E-05	CCNA2;CDKN2C; DKC1;CDCA8;MYBL2;KIF15
Cyclosporin A (350 mg/kg)	2.35E-05	CCNA2;NEK9;CDKN2C; CDCA8;MYBL2;KIF15
Chlorambucil (0.6 mg/kg)	2.40E-05	CCNA2;CENPE; CDKN2C; CDCA8;MYBL2;KIF15
Etoposide (100 mg/kg)	2.85E-05	CCNA2;CENPE; CDKN2C; CDCA8;MYBL2;KIF15
Leflunomide (30 mg/kg)	3.37E-05	CCNA2;CENPE; CDCA8;MYBL2;FOXM1;KIF15
Thioguanine (12 mg/kg)	4.11E-05	CCNA2;CENPE; CDKN2C; CDCA8;MYBL2;KIF15
Cyclophosphamide (25 mg/kg)	4.34E-05	CCNA2;CDKN2C; DKC1;CDCA8;MYBL2;KIF15
Daunorubicin (3.25 mg/kg)	1.96E-04	CCNA2;CDKN2C; CDCA8;MYBL2;KIF15
Clobetasol propionate (17 mg/kg)	2.00E-04	CCNA2;CDKN2C; CDCA8;MYBL2;KIF15

Discussion

In this study, the main purpose was to find essential signaling pathways in GBM-derived stem cells to select effective drugs by finding important genes and molecules in these pathways.

For the past decade, researchers have been interested in studying cancer stem cells that are necessary for the treatment of a wide range of cancers. Studies on brain cancers have shown that cancer stem cells have a high potential for tumorigenesis under hypoxia.[9] Safari et al.[10] demonstrated that GBM-derived stem cells have a high expression of O6-methylguanine methyltransferase, which plays a vital role in chemotherapy resistance. In addition, Warrier et al.[11] showed that GBM-derived stem cells have a significant expression for ABC family genes highly resistant to chemotherapy. Limited drugs have been used to address the GBM-derived stem cells that have not been very successful, including temozolomide, carboplatin, and 1,3-bis (2-chloroethyl)-1-nitroso-urea BCNU.[12]

In the current study, we isolated the most relevant and significant genes with high differential expression to find more accurate pathways of GBM-derived stem cells by bioinformatics analysis on RNA-Seq data.

CCNA2 is an essential gene in cell division that plays a significant role in the transition of G1/S to G2/M. A microarray data study showed that TAF7/CCNA2 could play a high pathogenic role in GBM and there is a close relationship between them.[13] Another survey found that miR-219 could control the CCNA2 gene and play a significant role in inhibiting cancer cells’ growth [Figure 4a].[14]

Figure 4

The expression of candidate genes in GBM has been compared by the GEPIA database (right) and the survival plot of these genes is also shown on the left. a: CCNA2; b: CDKN2C; c: DKC1; d: KIF15; e: MYBL2

CDKN2C is another candidate gene in this study, which was part of the INK4 family and is associated with CDK4/6, and acts by inhibiting CDK activity on cell division in the G1 phase. Various studies were performed on this gene, the high expression of which has been noticed and proven in the sample of GBM compared to normal brain tissue.[15] In the GBM xenograft model, CDKN2C expression was significantly higher than in the control group, playing a pathogenic role in GBM.[16] Previous studies on the GBM cell line using microarray analysis showed that 25 genes could be sensitive to chemotherapy drugs, including CDKN2C.[1718] Evidence demonstrated that CDKN2C is not directly involved in the development of tumor cells but may be included in the event of cancer by affecting cyclin D.[19] Mutations in CDKN2C can also be involved in the development of cancer cells [Figure 4b].[20]

DKC1 is one of the most important genes involved in telomerase, which plays a significant role in apoptosis and cellular aging. These genes have very high expression in GBM, the product of which can protect telomerase and cause cancer cell death. In a study using viral vectors on GBM, it was found that DKC1 -4.5 LogFC decreased expression and positively affected the apoptosis of cancer cells.[21] Miao et al.[22] showed increased abnormal expression, increased angiogenesis, division, and migration of cadherin N, HIF-1, and MMP2-mediated cancer cells. DKC1 expression has been reported in several other cancers. High expression of this gene is hazardous in lung cancer cells.[23] Also, by affecting the HIF-1 gene, it played a role in increasing the angiogenesis of colorectal cancer and its invasion to other tissues.[24] In addition, studies have been developed on the progression of prostate cancer [Figure 4c].[25]

The molecular dynamics of the cytoskeleton’ also play an essential role in stimulating the activity of cancer cells to grow and proliferate, as well as in their invasion. KIF15 was one of the critical items in this study that was identified. Wang et al.[26] showed that the product of this gene in GBM cells could increase cell division in the G1 phase, in which inhibition or extinction can also reduce tumorigenesis. PBK is a member of the MAPK family and is a mitogen activator. In GBM, PBK can increase cell proliferation by interacting with KIF15.[27] Also, Terribas et al.[28] indicated that KIF11/15/25 has a high expression in neural sheath tumors and plays a vital role in the survival of these cells. Moreover, KIF15 plays a crucial role in increasing cell division and tumorigenesis of cancer cells by acting on the MEK/ERK pathway [Figure 4d].[29]

Another significant gene identified in this study was MYBL2, which is effective as a nuclear transcription factor in cell cycle regulation. An in-silico study of GBM data showed that several genes, including MYBL2, play a significant role in cancer survival and division.[30] Zhang et al.[31] have found that by increasing MYBL2 expression the miR- 30e increases and the growth and invasion of glioblastoma cells are decreased. A clinical trial on GBM patients showed that MYBL2 is downstream of the AKT/FOXM1 pathway genes involved in cell division and apoptosis inhibition. When AKT inhibitors are active, and FOXM1 is silent, MYBL2 expression decreases and eventually results in cell cycle inhibition and apoptosis induction, indicating that the AKT, FOXM1, and MYBL2 are related to each other.[32] Another in-silico study by gene network analysis showed that FOXM1 and MYBL2 play a vital role in cancer cell growth and proliferation.[33] In gastric adenocarcinoma, it was found that high expression of MYBL2 causes cancer cells to differentiate and invade lymph nodes, the inhibition of which can be a binding antitumor effect [Figure 4e].[34]

The next step is to identify drug agents that are able to inhibit or reduce the expression of mentioned genes that have an influential role in attenuating cancer stem cells. Chlorambucil is used as a chemotherapy drug, which is mainly used for leukemia cancers. In a study of 297 patients, the effects of chlorambucil and almethosumab were evaluated, with 55% and 43% of cure rates for leukemia, respectively.[35] Hu et al.[36] studied chlorambucil with drug delivery and tissue engineering approach, which showed that when chlorambucil is combined with 1, 6-Hexanediamine hydrochloride (HDH) micelles, it has a high permeability into the cancerous tissue and physiological barriers and could have an acceptable therapeutic effect. Millard et al.[37] showed that chlorambucil could specifically affect the energy production pathways in mitochondria and increase the death of pancreatic and breast cancer cells by more than 80% by acting on mtDNA. Luo et al.[38] revealed that chlorambucil could increase the path of oxidative stress and be used as a viable treatment option for breast cancer.

Cyclosporine A is considered an immunosuppressive drug, and also plays a significant role in various cancers. A study of the C6 GBM cell line found that cyclophilin A can develop drug resistance in tumor cells. Cyclosporine A with sanglifehrin combined with cisplatin can reduce the expression of cyclophilin A and increase the apoptosis and the reactive oxygen species pathway.[39] A study on the T98G GBM cell line found that cyclosporine A with an effect on the morphine tolerance pathway could affect the NO/ERK pathway and ultimately play an inhibitory role in GBM cell division.[40] Sliwa et al.[41] demonstrated a new route of GBM with the use of cyclosporine A. Microglia in brain tissue could contribute to higher cell proliferation and tumorigenesis by activating the PI3K/AKT pathway in conditions where the tumor tends to invade the brain. Cyclosporine A with effect on this pathway and inhibiting microglia activity prevents the invasion and progression of the disease. Cyclosporine A induces apoptosis in gastric cancer by inhibiting the NF-KB pathway by Dastaxel.[42] In breast cancer, cyclosporine A reduces the drug resistance in this cancer by decreasing the expression of ABCG2.[43]

Doxorubicin is another chemotherapy drug that can induce apoptosis with high potency. Most studies of these drugs are used in tissue engineering and pass through the BBB in the form of micelles or nanoparticles of this drug, which can have an acceptable effect in reducing tumorigenesis.[44454647] It has also been used in other cancers, such as lung cancer.[48]

Etoposide is a very successful drug in chemotherapy. In a study of a mouse model, researchers showed that low doses of etoposide effectively induced apoptosis.[49] For this drug to have a better effect on brain tumors, the approach of drug delivery with lipid particles and nanoparticles has been a good option so far.[505152]

Conclusion

In conclusion, it can be argued that the use of appropriate drug regimens for GBM can be more effective in destroying cancer stem cells, especially in the GBM, and that etoposide, doxorubicin, cyclosporine A, and chlorambucil can be used and have good synergistic effectiveness.

Notes on Contributors

T.M. A.J., A.B., S.A.S.E., conception and design, acquisition of data, or analysis and interpretation of data; S.S., V.K., A.Z., drafting the article or revising it critically for important intellectual content. All authors revised final approval of the version to be submitted for publication

Financial support and sponsorship

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of interest

There are no conflicts of interest.