Journal of Clinical Trials
Open Access

Identification and Analysis of Molecular Pathways and Key Candidate Biomarkers in Glioblastoma Multiforme

Aditi Stuti Kumari^{* *}Correspondence: Aditi Stuti Kumari, Department of Nursing, Stevens Institute of Technology, Hoboken, USA, Email:

Author info »

Introduction

Glioblastoma Multiforme (GBM) is a grade IV malignant primary tumor present in the central nervous system. It affects glial cells in the central nervous system and may originate anywhere in the brain or the spine; however, it does not spread to other organs or parts of the body. Most people live for 15 months after being affected, but 90% of people die at 24 months [1]. GBM has an incidence rate of 3.21 per 100,000 people globally; however, despite being rare, it is the most common glioma in all age groups and has a poor prognosis as the median survival is 14 months-15 months after being diagnosed, making it a significant health issue. The low prognosis is due to the high molecular heterogeneity and high invasiveness. CT scan, MRI scan, or a brain PET scan, MR spectroscopy and a tissue biopsy are often used to diagnose a patient with GBM. Treatment options include chemotherapy, radiation therapy, targeted therapy, and surgery [2]. However even with large advances in treatments, it is still incurable.

Despite extensive studies done in GBM, it still remains to be poorly understood. Multiple factors cause the tumor, including mutations. In order to form a deeper understanding of GBM, it is crucial that its molecular mechanisms be further analyzed to open more treatment options as well as provide patients with better care. Additionally, finding molecular biomarkers along with a potential diagnosis would allow for more personalized treatment and improved prognosis.

Microarrays were utilized in the screening of Differentially Expressed Genes (DEGs), allowing for the identification of genes that are closely related to the occurrence of the tumor. However, due to the tissue heterogeneity, determining specific molecular mechanisms of the tumor is a challenge.

In this research, potential prognostic and therapeutic biomarkers were discovered using public databases. Identifying prognostic biomarkers would be advantageous as it would allow determining who is at a higher risk of getting the tumor and identify the life threatening tumor early on, allowing treating GBM earlier and increasing survival rates [3]. The goal of this study was to identify candidate biomarkers that can be used in the identification and treatment of glioblastoma multiforme. Additionally, the goal was to identify the functional pathways regulated in glioblastoma multiforme.

This study analyzed one microarray dataset consisting of 3 glioblastoma tissue and 3 normal, healthy tissue. DEGs were filtered using GEO2R, which were then analyzed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) functional annotation tool to access the gene ontology and the KEGG pathway analysis [4]. Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) was used to construct a Protein Protein Interaction (PPI) network, and Cytoscape allowed for further analysis of the network and identified the hub genes. Gene Expression Profiling Interactive Analysis (GEPIA) was utilized to validate the research findings. PROGgeneV2 allowed for survival analysis and prognostic data to be analyzed and further interpreted.

Materials and Methods

GBM tissue and normal tissue were accessed using the accession code GSE100675 on Gene Expression Omnibus (GEO) database, allowing access to database using next-generation sequencing along with microarray datasets. The gene expressions were further evaluated using GEO2R. DEGs were determined by comparing gene expression in GBM and normal tissue. In order to determine what genes were upregulated, the fold change value was to be greater than 1. If it was less than -1, it was to be downregulated. This was done to shorten the list of genes to analyze further. DEGs were identified by using a p<0.05 on the short listed genes.

The DAVID functional annotation tool was used to do the KEGG pathway analysis and access the gene ontology of the dataset with a p<0.05. The gene ontology categorizes the DEGs into three functional groups which are cellular component, molecular function, and biological processes.

To construct a PPI network, the STRING database was utilized by using the proteins the DEGs coded for. The network was then put into cytoscape, which was utilized in analyzing the interactions between the candidate proteins the DEGs transcribed for. Cytoscape also identified the degree of connectivity for each node along with the total number of connections that were in the PPI and the total number of nodes [5].

GEPIA database was used in the validation of the research and confirming its reliability. The database accessed data from The Cancer Genome Atlas (TCGA) and showed the expression of a gene in a cancer in relation to a healthy individual. Validation was done by comparing the regulation of genes collected from the dataset accessed in this study to that of a larger dataset from TCGA.

PROGgeneV2, a prognostic database, allowed for analysis of the prognosis and survival of a patient in relation to the hub genes. Data derived from TCGA was used to indicate survival rate. Kaplan-Meier analysis was done on the hub genes. Patients that had GBM were classified into either having a higher expression or a lower expression of the gene. The survival of each patient was then plotted on a graph, allowing a visual representation of survivability of the patient in relation to expression of the gene. A p<0.05 was used to determine which plotted dataset has a difference considered statistically significant [6].

Results

Identification of DEGs

Dataset was accessed from gene expression omnibus (GSE100675). It consists of 9 samples: 3 of glioblastoma tissues, 3 normal tissues, and 3 paired tissues.

Differentially expressed genes in the dataset were identified using microarray analysis. When glioblastoma tissue was compared with normal tissue, it was found that there were 693 significantly upregulated genes and 916 significantly downregulated genes, a total of 1,609 DEGs (Figure 1 and Table 1).

Figure 1: The boxplots above show that the dataset has been normalized by eliminating variation that can be accounted for between GBM and normal tissue.

Table 1: Table 1 below shows a list of the genes that were identified as DEGs from the comparison between glioblastoma tissue and normal tissue. The genes that have been bolded are those that were identified as the 10 hub genes.

Gene ontology: The DAVID functional annotation tool was used to analyze the gene ontology, comparing glioblastoma tissue with normal tissue. After the DEGs were sorted into the three categories, the top 10 most significant go terms were identified for each category. In biological processes, the upregulated DEGs were mostly involved in nucleosome assembly, negative regulation of gene expression, and chromatin silencing at DNA, while the downregulated DEGs were mostly involved in chemical synapse transmission and nervous system development.

In cellular components, upregulated genes were involved with extracellular exosome and membrane, while downregulated genes were mostly involved with plasma membrane and cell junction [7].

In molecular functions, upregulated DEGs were enriched mostly in protein binding and protein heterodimerization activity, while downregulated DEGs were enriched in calcium ion binding and calmodulin binding. These findings suggest that the upregulated DEGs are mostly involved with protein binding, extracellular exosome, and membrane, and the downregulated DEGs are mostly involved with plasma membrane, cell junction, and chemical synaptic transmission. It is also demonstrated that chemical synaptic transmission is the most enriched term in biological processes, plasma membrane in cellular components, and protein binding in molecular functions (Figures 2 and 3).

Figure 2: The graph above depicts the gene ontology when comparing significantly upregulated genes in normal tissue and GBM tissue and obtained using the DAVID functional annotation tool. Due to the vast number of genes listed, the top ten for each ontological category were looked at. These are functions that may be crucial in the development of the tumor.

Figure 3: The graph above depicts the gene ontology when comparing significantly downregulated genes in normal tissue and GBM tissue and obtained using the DAVID functional annotation tool. Due to the vast number of genes listed, the top ten for each ontological category were looked at similar to the figure above. These are also functions that may be crucial in the development of the tumor.

KEGG pathway analysis: The DAVID functional annotation tool was utilized to do KEGG pathway analysis as well, where pathways that have been upregulated and downregulated were identified. Online publications and research were used to determine which pathways are needed to look at based on its relevance to glioblastoma and whether it has been proved to be linked to its progression [8]. The most enriched upregulated and downregulated pathway are the pathways in cancer pathway and cAMP signaling pathway respectively (Tables 2 and 3).

Table 2: Table 2 was constructed after a KEGG pathway analysis was done using the DAVID functional annotation tool. Through the analysis, 4 pathways were found to be significantly upregulated. These pathways are significant in the development of the tumor due to the statistical significance and number of genes present in each pathway.

Table 3: Table 3 was constructed after a KEGG pathway analysis was done using the DAVID functional annotation tool. Through the analysis, 8 pathways were found to be significantly downregulated. These pathways are also significant in the development of the tumor due to the statistical significance and number of genes present in each pathway.

Construction of the PP1 network and identification of hub genes: The STRING database and cytoscape software were utilized in constructing a PPI network. A total of 148 DEGs were put into the software and database, resulting in a PP1 network that consisted of 130 nodes and 484 edges. From the 130 nodes, the 10 nodes with the highest degrees of connectivity were identified [9]. Cytoscape was used to analyze and identify the top 10 hub genes. The top ten hub genes with a degree of connectivity of 14 or higher were ITGB1, GNG3, GNG5, NRAS, CAMK2A, CAMK2B, ITGAV, ITGB3, ADCY5, and STAT3. ITGB1 has the highest degree of connectivity at 22 (Figure 4).

Figure 4: A protein-protein interaction network was constructed from STRING using genes from all of the significantly regulated pathways. It is composed of 130 nodes and 484 edges and was constructed at the highest confidence level (0.900).

Validation: GEPIA was used to validate the research. Boxplots were employed to compare the regulation of the genes in the sample accessed to a larger sample. It was determined if genes have been upregulated or downregulated in a larger sample size to ensure that it matches with the sample size. Analyzing if the gene is regulated similarly in both sample sizes validated the findings. The regulation of the genes in the smaller dataset composing of 3 GBM tissues and 3 normal tissues accessed through GEO was compared to a larger dataset consisting of 163 GBM tissues and 207 normal tissues (Figure 5).

Figure 5: The boxplots above were obtained from GEPIA when comparing the regulation of each gene to a larger sample to ensure that the gene regulation matched. The red boxplots represent GBM tissue and the gray boxplots represent normal tissue. The larger sample consisted of 163 tumor tissues and 207 normal tissues. The regulation of all ten genes in the sample from the dataset matched that of a larger dataset, validating the research.

Survival analysis: When using PROGgeneV2, line graphs were utilized to determine how the regulation of a specific gene may affect the survival rate of a person while also making sure that the results are statistically significant with a p>0.05 (Table 4). Based on PROGgeneV2, high expression of the STAT3 gene is negatively associated with patients with a lower survival (Figure 6).

Figure 6:The above figure was obtained using PROGgeneV2 from The Cancer Genome Atlas (TCGA). Data from the figures above show a lower survival rate of the individual when the gene is upregulated. The low P-value conveys that this difference in survival rate is significant and should be taken into consideration.

Table 4: The below table was obtained using PROGgeneV2 from The Cancer Genome Atlas (TCGA). Data from the figures above show a lower survival rate of the individual when the gene is upregulated. The low P-value conveys that this difference in survival rate is significant and should be taken into consideration.

Discussion

In this study, a total of 1,609 Differentially Expressed Genes (DEGs) (915 downregulated and 694 upregulated) were identified. The gene ontology revealed that the upregulated DEGs were involved mostly with cell proliferation and cell cycle, while the downregulated DEGs were involved mostly in synapses and nervous system development. The progression of GBM may be connected to multiple cellular components, biological processes, and molecular processes, as signified by the gene ontology. Studies show that disruption in cell proliferation is involved with the progression and development of the tumor. Functions from gene ontology and pathways from the KEGG pathway analysis may be crucial in the development of the tumor, requiring further study to be done.

The DEGs were put into STRING, creating a Protein- Protein Interaction network (PPI) composed of 130 nodes and 484 edges (connections to other genes). Using the DEGs and the PPI, 10 hub genes were identified through Cytoscape. A Kaplan-Meier analysis of the hub genes showed that overexpression of STAT3 is associated negatively with the patients, decreasing the survival rate significantly. Hence, STAT3 may be a key gene in GBM, whose functions may be integral in the tumor progression [10]. Signal Transducer and Activator of Transcription 3 (STAT3) genes is a cytoplasmic transcription factor regulated by cytokine and growth factor receptor, with a location of 17q21.2. The gene promotes transcription of multiple genes contributing to the cancer, including angiogenesis, anti-apoptosis, invasion, and cell cycle progression. It is proven to be found at a high regulation in many tumors, including GBM, and has been identified to be crucial in tumor progression.

Conclusion

The results of this study suggest that the Pathways and other processes identified through gene ontology and KEGG pathway analysis may be in close relation with the occurrence of GBM and its progression. STAT3 is found to be a key gene that can be associated with the prognosis of the tumor.

Due to its crucial role in the tumor progression and the poor survival rate when the gene is excessively expressed, STAT3 gene can be used as a prognostic biomarker with an upregulation being a sign. Additionally, identification of the upregulation may be helpful in identifying GBM cells. This gene can be used for therapeutic purposes and can be used in targeted therapy for the tumor by attacking the cells that have a high expression of the STAT3 gene. Further research is needed to evaluate STAT3 as a GBM biomarker and potential applications of pathways in the treatment of the tumor.

References

1. Bossy-Wetzel E, Schwarzenbacher R, Lipton SA. Molecular pathways to neurodegeneration. Nat Med. 2004;10(7):S2-S9.

   [Crossref] [Googlescholar] [Indexed]

2. Pieczenik SR, Neustadt J. Mitochondrial dysfunction and molecular pathways of disease. Exp Mol Pathol. 2007;83(1):84-92.

   [Crossref] [Googlescholar] [Indexed]

3. Al‐Sohaily S, Biankin A, Leong R, Kohonen‐Corish M, Warusavitarne J. Molecular pathways in colorectal cancer. J Gastroenterol Hepatol. 2012;27(9):1423-1431.

   [Crossref] [Googlescholar] [Indexed]

4. Labunskyy VM, Hatfield DL, Gladyshev VN. Selenoproteins: Molecular pathways and physiological roles. Physiol Rev. 2014;94(3):739-777.

   [Crossref] [Googlescholar] [Indexed]

5. Weis SM, Cheresh DA. Tumor angiogenesis: Molecular pathways and therapeutic targets. Nat Med. 2011;17(11):1359-1370.

   [Crossref] [Googlescholar] [Indexed]

6. Urbanska K, Sokolowska J, Szmidt M, Sysa P. Glioblastoma multiforme–an overview. Contemp Oncol (Pozn). 2014;18(5):307-312.

   [Crossref] [Googlescholar] [Indexed]

7. Krex D, Klink B, Hartmann C, Von Deimling A, Pietsch T, Simon M, et al. Long-term survival with glioblastoma multiforme. Brain. 2007;130(10):2596-2606.

   [Crossref] [Googlescholar] [Indexed]

8. Alifieris C, Trafalis DT. Glioblastoma multiforme: Pathogenesis and treatment. Pharmacol Ther. 2015;152:63-82.

   [Crossref] [Googlescholar] [Indexed]

9. Gan HK, Kaye AH, Luwor RB. The EGFRvIII variant in glioblastoma multiforme. J Clin Neurosci. 2009;16(6):748-754.

   [Crossref] [Googlescholar] [Indexed]

10. Roth JG, Elvidge AR. Glioblastoma multiforme: A clinical survey. J Neurosurg. 1960;17(4):736-750.

    [Crossref] [Googlescholar] [Indexed]