The genes significantly associated with an improved prognosis and long-term survival of glioblastoma

Hong Gyu Yoon; Jin Hwan Cheong; Je Il Ryu; Yu Deok Won; Kyueng-Whan Min; Myung-Hoon Han

doi:10.1371/journal.pone.0295061

Abstract

Background and purpose

Glioblastoma multiforme (GBM) is the most devastating brain tumor with less than 5% of patients surviving 5 years following diagnosis. Many studies have focused on the genetics of GBM with the aim of improving the prognosis of GBM patients. We investigated specific genes whose expressions are significantly related to both the length of the overall survival and the progression-free survival in patients with GBM.

Methods

We obtained data for 12,042 gene mRNA expressions in 525 GBM tissues from the Cancer Genome Atlas (TCGA) database. Among those genes, we identified independent genes significantly associated with the prognosis of GBM. Receiver operating characteristic (ROC) curve analysis was performed to determine the genes significant for predicting the long-term survival of patients with GBM. Bioinformatics analysis was also performed for the significant genes.

Results

We identified 33 independent genes whose expressions were significantly associated with the prognosis of 525 patients with GBM. Among them, the expressions of five genes were independently associated with an improved prognosis of GBM, and the expressions of 28 genes were independently related to a poorer prognosis of GBM. The expressions of the ADAM22, ATP5C1, RAC3, SHANK1, AEBP1, C1RL, CHL1, CHST2, EFEMP2, and PGCP genes were either positively or negatively related to the long-term survival of GBM patients.

Conclusions

Using a large-scale and open database, we found genes significantly associated with both the prognosis and long-term survival of patients with GBM. We believe that our findings may contribute to improving the understanding of the mechanisms underlying GBM.

Citation: Yoon HG, Cheong JH, Ryu JI, Won YD, Min K-W, Han M-H (2023) The genes significantly associated with an improved prognosis and long-term survival of glioblastoma. PLoS ONE 18(11): e0295061. https://doi.org/10.1371/journal.pone.0295061

Editor: Syed M. Faisal, University of Michigan Medical School, UNITED STATES

Received: September 30, 2023; Accepted: November 14, 2023; Published: November 29, 2023

Copyright: © 2023 Yoon et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: This study was funded by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF-2022R1F1A1063739).

Competing interests: The authors have declared that no competing interests exist.

Introduction

Glioblastoma multiforme (GBM) is the most common and devastating primary brain tumor, which is characterized by infiltrative growth and resistance to treatment and leads to an extremely poor prognosis. Despite aggressive treatment strategies against GBM, including chemotherapy, radiotherapy, immunotherapy, and surgical resection, only a few patients survive 2.5 years, and less than 5% of patients survive 5 years following their diagnosis [1].

Extensive studies have focused on the genetics of GBM to improve the understanding of the underlying mechanisms of GBM and to contribute to an improved prognosis of patients with GBM [2]. We also previously identified a DKK3 gene from the Wnt/β-catenin pathway and 12 genes from 10 oncogenic signaling pathways associated with GBM prognosis using The Cancer Genome Atlas (TCGA) database [3, 4]. It is well known that TCGA is the world’s largest publicly accessible genomic database. It includes information on digital pathologic slides, mRNA expression data, clinicopathological information, and DNA methylation and mutation data. However, there has not been a study aiming to identify the genes significantly related to the prognosis of GBM by assessing the direct association between the gene expression levels in GBM tissue and both the lengths of the overall survival (OS) and the progression-free survival (PFS) in patients with GBM, using large gene expression datasets of GBM. In addition, we hypothesize that if genes related to long-term survival in patients with GBM are found, it may help predict the future prognosis or treatment of patients with GBM.

Therefore, this study aimed to investigate specific genes, using the TCGA database, whose expressions are significantly related to both the lengths of OS and PFS in patients with GBM. Next, we aimed to classify the identified genes significantly associated with the prognosis of GBM, according to the Gene Ontology (GO) terms using bioinformatics. Finally, this study aimed to identify which genes, among the identified genes significantly related to the GBM prognosis, were significantly associated with the long-term survival of patients with GBM. A schematic flow chart depicting the steps involved in this research is presented in Fig 1.

Download:

Fig 1. Schematic diagram detailing the process of selecting the independent genes significantly associated with the prognosis of GBM for our study.

https://doi.org/10.1371/journal.pone.0295061.g001

Methods

Study patients

We obtained 1,149 glioma cases, consisting of 619 GBM cases and 530 low-grade glioma cases with mRNA gene expression data from the TCGA database (https://gdc.cancer.gov/about-data/publications/pancanatlas and https://www.cbioportal.org/) [5]. We initially selected 594 GBM cases with virtual histopathological slides and clinical data out of 619 GBM cases. We excluded 594 GBM cases with significantly incomplete mRNA gene expression information and clinical data. Therefore, the 525 GBM cases with complete virtual histopathological slides, mRNA expression data, and clinical information were finally included in the study as described elsewhere [3, 4]. Log 2 (x + 1) transformation normalized all mRNA gene expression values before analysis [6].

Informed consent was not required because the data were obtained from the publicly available TCGA database.

Study design

In Fig 1, the study design is shown as follows: (1) we initially observed a dataset from the TCGA database containing mRNA expression information for 12,042 genes from 525 GBM tissues; (2) then excluded 11,187 genes whose expressions showed no significant association with the lengths of the OS or PFS in the study’s patients, according to Pearson correlation analysis (p ≥ 0.01); (3) excluded 819 genes with a low strength of correlation: Genes showing a Pearson coefficient absolute value of less than 0.2, according to a previous study [7]; (4) after adjusting for clinical variables, three genes whose expressions were not significantly associated with the lengths of the OS or PFS were further excluded (Table 1);

Download:

Table 1. Multivariable linear regression analysis of the lengths of the OS and PFS according to the 36 significant genes in patients with GBM.

https://doi.org/10.1371/journal.pone.0295061.t001

(5) A total of 33 genes whose expressions showed significant independent associations with both the lengths of the OS and PFS in patients with GBM were finally enrolled for the study. We also present the results of the univariate linear regression analysis of the lengths of the OS and PFS according to the 36 significant gene expressions in patients with GBM in the S2 Table. The raw data related to the study design can be found in the S1 Data.

In silico flow cytometry

As previously reported, we analyzed tumor-infiltrating lymphocytes in GBM tissues using CIBERSORT (https://cibersort.stanford.edu), a versatile computational method for quantifying the immune cell-type fractions. This method relies on a validated leukocyte gene signature matrix containing 547 genes and 22 human immune cell subpopulations [3, 4]. The gene expression profiles of the GBM tissues from the TCGA were entered into CIBERSORT for analysis, and the algorithm was run using the LM22 signature matrix at 100 permutations.

CD8+ T-cells are major drivers of antitumor immunity, and elevated CD8+ T-cell counts in the tumor microenvironment are related to a good prognosis in cancer [8]. In addition, as we have previously described, CD4+ T-cells, CD8+ T-cells, regulatory T-cells (Tregs), B-cells, and antigen-presenting cells are reported to play an important role in the immune microenvironment of GBM [3]. Therefore, we included the following eight representative immune cells for the study to evaluate the relationships between the status of the GBM immune microenvironment and specific gene expressions: CD8+ T-cells, regulatory T-cells, naive CD4+ T-cells, resting and activated memory CD4+ T-cells, memory B-cells, plasma B-cells, and activated dendritic cells [3].

Bioinformatics analysis

We performed bioinformatics analysis using Cytoscape (version 3.9.1) software (https://cytoscape.org/). We used ClueGo and CluePedia plugins that enabled functional Gene Ontology and pathway network analyses in Cytoscape to interpret the biological roles and interactions of the 33 selected significant genes in GBM [9]. We analyzed the biological function annotated pathways based on 33 significant genes related to the prognosis of GBM. We also activated the cerebral view function in the ClueGO application of the Cytoscape to estimate the approximate location of any significant proteins in the cell. We also conducted pathway-based network analysis using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database version 11.5 (http://www.string-db.org/) to further investigate the inter-relationship between these 33 significant gene expressions. The STRING provides known and predicted protein-protein association data from a large database based on co-expression analysis, signals across genomes, and automatic text-mining of the biomedical literature. All interaction sources, text-mining, experiments, databases, co-expression, neighborhood, gene fusion, and co-occurrence were activated in the STRING setting.

Statistical analysis

Heatmap analyses of 33 significant gene expressions and immune cell infiltrations in 525 GBM tissues were performed using R software’s “pheatmap” package (version 4.1.2).

Pearson correlation coefficients and significance levels were calculated to evaluate the associations between the 33 significant gene expressions and the lengths of the OS and PFS in patients with GBM and the immune cell infiltrations in GBM tissues. We used the “corrplot” package of R software with the clustering technique (R code: corrplot, M, order = “hclust”, sig. level = 0.01, method = “square”) to visualize the correlations. A scatterplot with a linear regression line was used to visualize the relationship between several significant gene expressions and the lengths of the OS and PFS in patients with GBM. The OS and PFS months were transformed to the natural log scale to normalize the distributions for the analysis. We calculated the OS and PFS rates using Kaplan–Meier analysis based on the gene expression quartiles in patients with GBM.

Receiver operating characteristic (ROC) curve analysis was performed to determine the genes significant for predicting the 2.5-year and 5-year survivals in patients with GBM, defined as showing the shortest distance from the upper left corner (where sensitivity = 1 and specificity = 1).

A p-value < 0.05 was considered statistically significant. All statistical analyses were performed using R software version 4.1.2 and SPSS for Windows version 24.0 (IBM, Chicago, IL).

Results

Characteristics of the study patients

A total of 525 patients with GBM from the TCGA database were included in this study. The mean patient age at the diagnosis of GBM was 57.7 years, and 39.0% of patients were female. A total of 435 (82.9%) patients underwent radiation treatment, and further detailed information, including immune cell fractions in GBM tissues, is shown in the S1 Table.

Identification of significant genes associated with the prognosis of GBM

Through the process shown in Fig 1, among the 12,042 observed genes, we identified 33 independent genes whose mRNA expressions were significantly associated with both the lengths of the OS and PFS in patients with GBM. The identified 33 independent and significant genes are: ADAM22, AEBP1, ATP5C1, C13orf18, C1RL, CBR1, CCL2, CHI3L1, CHL1, CHST2, CLEC5A, DHRS2, DYNLT3, EFEMP2, EMP3, F3, FBXO17, FLJ11286, MSN, NSUN5, PDPN, PGCP, PPCS, RAC3, SERPING1, SHANK1, SLC25A20, SLC2A10, STEAP3, SWAP70, TIMP1, TMEM22, and TRIP6. Among these 33 genes, there were 5 genes (ADAM22, ATP5C1, DHRS2, RAC3, and SHANK1) whose expressions were positively correlated with the lengths of the OS and PFS. The expressions of the remaining 28 genes exhibited negative correlations with the lengths of the OS and PFS in patients with GBM.

Expression patterns of the 33 significant genes and immune cells in GBM

The heat map shows different mRNA expression patterns between the 33 significant genes in 525 GBM tissues (Fig 2A).

Download:

Fig 2. Gene expression patterns of the 33 independent and significant genes with comparisons of the immune cell fractions in GBM.

The correlations between the 33 significant genes, the OS and PFS lengths, and fractions of representative immune cells in GBM. (A) A hierarchically clustered heatmap showing the expression patterns of the 33 significant genes related to the prognosis of GBM. Gene expression levels were log2 transformed, and a color density indicating levels of log2 fold changes is presented. Red and blue represent up- and downregulated expression, respectively, in GBM; (B) a bar plot indicating average expression levels of the 33 significant genes in GBM tissue; (C) a hierarchically clustered heatmap showing the expression patterns of eight representative immune cells in GBM; (D) boxplots showing the differences in eight representative immune cell fractions in GBM; (E) Pearson correlation coefficients and significance levels were calculated between the expressions of the 33 significant genes and lengths of the OS and PFS in patients with GBM; (F) Pearson correlation coefficients and significance levels were calculated between the expressions of the 33 significant genes and fractions of representative eight immune cells in GBM. The color-coordinated legend indicates the value and sign of the Pearson correlation coefficient. The number in the box indicates the Pearson correlation coefficient. The ‘x’ in the box indicates a p-value ≥ 0.01.

https://doi.org/10.1371/journal.pone.0295061.g002

GBM: glioblastoma multiforme; OS: overall survival; PFS: progression-free survival

There were three genes whose mRNA expression levels were noticeably increased in GBM, and those genes were ATP5C1, CHI3L1, and TIMP1 (Fig 2B). Among the five genes associated with a good prognosis of GBM, the expressions of DHRS2, ADAM22, RAC3, and SHANK1 were relatively reduced in GBM tissues. The heatmap also showed differences in eight immune cell fractions between the 525 GBM tissues (Fig 2C). Heterogenous infiltrations were observed in CD8+ T-cells, resting CD4+ T-cells, naive CD4+ T-cells, and memory B-cells between the 525 GBM tissues. Boxplots show overall fractional differences between eight representative immune cells in the GBM tissues (Fig 2D).

Correlations between the expressions of the 33 genes, the lengths of the OS and PFS, and the immune cells in GBM

We visualized the correlations between the mRNA expressions of the 33 significant genes and the lengths of the OS and PFS in patients with GBM (Fig 2E). The expressions of 5 genes (ADAM22, ATP5C1, DHRS2, RAC3, and SHANK1) showed positive correlations with the lengths of the OS and PFS by providing Pearson coefficients greater than 0.2. The remaining 28 genes showed negative correlations with the lengths of the OS and PFS, providing Pearson coefficients less than –0.2. When we estimated correlations between the expressions of the 33 significant genes and the infiltrations of the eight immune cells from the 525 GBM tissues, there were significant correlations (p < 0.01) between the expressions of the 32 genes and the CD8+ T-cell infiltrations, except for the ATP5C1 gene (an x in the box indicates a p-value ≥ 0.01) (Fig 2F). We also found that the expressions of C13orf18, CHI3L1, CHL1, and CHST2 showed significant correlations with all eight immune cell fractions in GBM.

Associations between the expressions of the selected genes and the lengths of the OS and PFS in patients with GBM

We observed significant positive linear associations between the expression of ADAM22, ATP5C1, RAC3, and SHANK1 and the lengths of the OS and PFS in patients with GBM (Fig 3A).

Download:

Fig 3. Scatter plot with linear regression line between several significant gene expressions and log2-transformed lengths of the OS and PFS in patients with GBM.

Kaplan–Meier analysis showing the OS and PFS rates based on several significant gene expression quartiles in patients with GBM. The ROC curves to identify significant genes associated with 2.5-year and 5-year survivals in patients with GBM. (A) Linear regression lines showing the associations between ADAM22, ATP5C1, RAC3, and SHANK1 expressions and the lengths of the OS and PFS in patients with GBM; (B) Kaplan–Meier curves showing the OS and PFS rates according to ADAM22, ATP5C1, RAC3, and SHANK1 expression quartiles in patients with GBM; (C) linear regression lines showing the associations between C13orf18, CHI3L1, CHL1, and CHST2 expressions and the lengths of the OS and PFS in patients with GBM; (D) Kaplan–Meier curves showing the OS and PFS rates according to C13orf18, CHI3L1, CHL1, and CHST2 expression quartiles in patients with GBM; (E) ROC curves showing the significant genes both positively and negatively associated with a 2.5-year survival in patients with GBM; (F) ROC curves showing the significant genes both positively and negatively associated with 5-year survival in patients with GBM.

https://doi.org/10.1371/journal.pone.0295061.g003

GBM: glioblastoma multiforme; OS: overall survival; PFS: progression-free survival; ROC: receiver operating characteristic.

Using the Kaplan–Meier survival analysis, the fourth quartiles of ADAM22, ATP5C1, RAC3, and SHANK1 expressions showed significantly greater OS and PFS rates than those in the first, second, and third quartiles, except for the fourth quartile analysis of RAC3 for PFS (p = 0.1) (Fig 3B). Among the 28 genes associated with poor prognosis of GBM, we observed that the expressions of C13orf18, CHI3L1, CHL1, and CHST2, which were associated with all eight immune cell fractions, showed significant negative linear associations with the lengths of the OS and PFS in patients with GBM (Fig 3C). The first quartiles of C13orf18, CHI3L1, CHL1, and CHST2 expressions were significantly associated with greater OS and PFS rates compared to other quartile groups (Fig 3D). We also analyzed the OS and PFS in patients with GBM according to the quartile groups of the remaining 25 gene expressions, which are not included in the main figures (S1 and S2 Figs). We observed that both OS and PFS were statistically significant in all the remaining genes except for DHRS2 and SWAP70.

Identification of genes whose expressions predict long-term survival of patients with GBM

According to the ROC analysis of our study, when only the top four genes with the highest area under the curve (AUC) were included, higher expressions of ATP5C1 (AUC = 0.682; p < 0.001), RAC3 (AUC = 0.677; p < 0.001), ADAM22 (AUC = 0.643; p < 0.001), and SHANK1 (AUC = 0.605; p = 0.007), and lower expressions of C1RL (AUC = 0.711; p < 0.001), CHL1 (AUC = 0.701; p < 0.001), EFEMP2 (AUC = 0.699; p < 0.001), and PGCP (AUC = 0.695; p < 0.001) in GBM tissues were associated with the long-term survival (more than 2.5 years) in patients with GBM (Fig 3E). When predicting the long-term survival of more than 5 years in patients with GBM, the identification of higher expressions of ATP5C1 (AUC = 0.814; p < 0.001), SHANK1 (AUC = 0.746; p < 0.001), RAC3 (AUC = 0.733; p = 0.001), and ADAM22 (AUC = 0.731; p = 0.001), alongside lower expressions of EFEMP2 (AUC = 0.810; p < 0.001), CHST2 (AUC = 0.765; p < 0.001), AEBP1 (AUC = 0.756; p < 0.001), and CHL1 (AUC = 0.736; p < 0.001), in GBM tissue, provided significant associations with a long-term survival (more than 5 years) in patients with GBM (Fig 3F).

Functional gene ontology and pathway network analyses

The ClueGO and the CluePedia plugins of Cytoscape were performed to identify the enriched pathways to investigate the functionally grouped networks of the 33 significant proteins in GBM. We found three significant GO terms, which are ‘neuromuscular process controlling balance’, ‘mitochondrial proton-transporting ATP synthase complex, catalytic sector F(1)’, ‘carbonyl reductase (NADPH) activity’ among the five significant proteins (ADAM22, ATP5C1, DHRS2, RAC3, and SHANK1) associated with an improved prognosis of GBM (Fig 4A and 4B).

Download:

Fig 4. Bioinformatic analysis of the significant genes associated with the prognosis of GBM using Cytoscape with ClueGo and CluePedia plugins and STRING database.

(A) Grouping of the networks of the significant genes associated with an improved prognosis of GBM based on functionally enriched GO terms and pathways using the ClueGo and CluePedia plugins of Cytoscape; (B) functionally grouped networks based on the GO terms of the genes significantly associated with an improved prognosis of GBM, showing three significant GO terms. The cerebral view shows the approximate location of those significant proteins in the cell; (C) a protein-protein interaction network was constructed among the genes associated with an improved prognosis of GBM; (D) grouping of the networks of the genes significantly associated with a poorer prognosis of GBM, based on functionally enriched GO terms and pathways using the ClueGo and CluePedia plugins of Cytoscape; (E) functionally grouped networks based on the GO terms of the genes significantly associated with a poorer prognosis of GBM, showing 14 significant GO terms. The cerebral view shows the approximate location of the significant proteins in the cell; (F) a protein-protein interaction network was constructed among the genes associated with a poorer prognosis of GBM, showing that they were roughly divided into two clusters.

https://doi.org/10.1371/journal.pone.0295061.g004

GBM: glioblastoma multiforme; GO: gene ontology; STRING: Search Tool for the Retrieval of Interacting Genes/Proteins.

When protein-protein interaction was analyzed using STRING, only RAC3 and SHANK1 demonstrated a significant interaction (Fig 4C). There were 14 significant GO terms for the genes associated with poor prognosis in patients with GBM (Fig 4D and 4E). Among the 14 GO terms, the top four significant GO terms were ‘negative regulation of myeloid cell apoptotic process’, ‘formation of fibrin clot (clotting cascade)’, ‘regulation of extracellular matrix organization’, and ‘T-cell aggregation’ (Fig 4D and 4E). Following further analysis of the protein-protein interactions between the 28 genes associated with poor prognosis in patients with GBM, we found that the genes were roughly divided into two clusters (Fig 4F). These findings and possible mechanisms for the 33 significant genes affecting the OS and PFS in patients with GBM based on previous studies are summarized (Table 2).

Download:

Table 2. Classification of the 33 significant genes according to their GO terms alongside the possible mechanisms of the 33 significant proteins affecting the OS and PFS in GBM patients.

https://doi.org/10.1371/journal.pone.0295061.t002

Discussion

In this study, we identified 33 independent genes, among 12,042 genes from the TCGA database, whose expressions were significantly associated with the prognosis of 525 patients with GBM. Among them, the expressions of five genes were independently associated with an improved prognosis of GBM, while the expressions of the other 28 genes were independently related to a worse prognosis of GBM. Moreover, the genes associated with long-term survival were identified in GBM patients. Among the five genes associated with an improved prognosis of GBM, the genes whose expressions were significantly associated with long-term survival of GBM patients were ADAM22, ATP5C1, RAC3, and SHANK1. Alternatively, among the 28 genes that were associated with a worse prognosis in GBM patients, the expressions of AEBP1, C1RL, CHL1, CHST2, EFEMP2, and PGCP were negatively related to the long-term survival of GBM patients. When bioinformatics analysis was performed, there were three significant GO terms among the genes associated with an improved prognosis of GBM, whereas, 14 significant GO terms were among genes associated with a worse prognosis of GBM.

We classified the 33 significant genes according to their GO terms and the possible roles of those proteins on the prognosis of GBM based on the GeneCards database (www.genecards.org) and previous studies (Table 2) [10–44]. GeneCards is known as a comprehensive, authoritative compendium of annotative information about human genes, which are automatically mined and integrated from over 80 digital sources, resulting in a web-based deep-linked card for each of > 73 000 human gene entries [45].

Consequently, we found that the expression of the genes involved in the GBM immune microenvironment most commonly influences the GBM prognosis. To support this, our study showed significant correlations between the expressions of all 32 significant genes (except ATP5C1) and CD8+ T-cell infiltrations in the 525 GBM tissues. A recent study also reported that GBM cases with high-risk scores were involved in immune and inflammatory processes or pathways [46]. Based on our investigation, among the 33 significant genes, there were 12 significant genes that appeared to be related to the GBM immune microenvironment and may affect the prognosis of GBM: C1RL, CCL2, CHI3L1, CLEC5A, EMP3, FBXO17, MSN, SERPING1, STEAP3, SWAP70, TIMP1, and TMEM22. According to our findings, these 12 genes were associated with a worse prognosis for GBM; therefore, we hypothesized that they might be involved in the immunosuppression of the GBM microenvironment. Our findings support this hypothesis since all of these 12 genes were negatively correlated with CD8+ T-cell infiltrations in the GBM tissues. Moreover, we observed that these 12 genes are almost identical to the genes belonging to the red cluster in Fig 4F. The immune microenvironment of GBM is highly immunosuppressive due to the lack of a number of tumor-infiltrating lymphocytes and other immune effector cells in the GBM microenvironment [21]. This immunosuppressive GBM microenvironment results in resistance to immunotherapy and promotes a poor prognosis in GBM patients. Among the 12 significant genes involved in the immunosuppression of GBM, CCL2 recruits Tregs and myeloid-derived suppressor cells, which play a critical role in the immunosuppressive glioma microenvironment [20]. High levels of CHI3L1 are positively related to the infiltration of Tregs, neutrophils, and resting NK cells, which induces limitations in the effective anti-tumor immune response to GBM [21]. In addition, EMP3 is an important immunosuppressive factor for recruiting tumor-associated macrophages in GBM, which induces suppression of T-cell infiltration and leads to tumor progression [27]. Furthermore, C1RL may play an immunosuppressive role in the pathogenesis of glioma by triggering the activation of haptoglobin and complement component 1 [18].

The second most common possible mechanism related to the effect these 33 significant genes could produce on the prognosis of GBM was through cell adhesion or structural and extracellular matrix. According to our findings, 10 genes including ADAM22, AEBP1, CHL1, EFEMP2, PDPN, PGCP, RAC3, SHANK1, SWAP70, and TRIP6 appeared to influence the prognosis of GBM through mechanisms involving cell adhesion or structural and extracellular matrix. Among the genes associated with a good prognosis in GBM patients, ADAM22, RAC3, and SHANK1 are thought to inhibit GBM progression in an integrin-dependent manner [10, 13–16]. Meanwhile, based on our investigation, AEBP1, EFEMP2, and PGCP, which were negatively related to long-term survival in GBM patients are thought to affect the prognosis of GBM through matrix metalloproteinases (MMPs)-related mechanisms [17, 26, 34]. Low expressions of MMP9 in GBM tissues are associated with a good response to temozolomide and longer survival of patients with GBM [47]. In addition, CHL1, which is also negatively associated with long-term survival in GBM patients, promotes the survival of glioma cells by inhibiting the apoptosis of glioma cells via the phosphatidylinositol 3-kinase (PI3K)/AKT signaling pathway [22].

Meanwhile, among the 33 independent and significant genes, CHST2, PPCS, and FBXO17 were considered to influence the prognosis of GBM in relation to metabolism [23, 29, 35, 36]. The role of CHST2 in GBM is largely unknown, however, it is thought to have a negative influence on long-term survival in GBM patients in our study. Moreover, it has been previously reported that the CHST family may cause GBM cell proliferation through the WNT/β-catenin pathway [23]. Furthermore, according to our study, the genes related to the blood coagulation cascade, such as F3 and SERPING1, may affect the prognosis for GBM. F3 encodes coagulation factor III, which promotes hypercoagulation status. The hypercoagulation status increases the risk of thromboembolic events and promotes the growth and progression of brain tumors by stimulating intracellular signaling pathways [28]. In addition, according to our study, an increased expression of ATP5C1, which is involved in mitochondrial ATP synthesis, was significantly associated with the long-term survival of GBM patients. A metabolic switch from respiration (in the mitochondria) to glycolysis (in the cytosol) is a common feature in tumor cells. However, increased expression of ATP5C1 may also be related to maintaining the activities of ATP synthase and cellular respiration, which leads to the inhibition of tumor progression [11].

In summary, the overexpression of C1RL, CCL2, CHI3L1, CLEC5A, EMP3, FBXO17, MSN, SERPING1, STEAP3, SWAP70, TIMP1, and TMEM22 genes appears to influence the prognosis of patients with GBM by causing an immune-suppressive GBM microenvironment. Immunotherapy holds tremendous promise for revolutionizing cancer therapies, but the significant immunosuppression seen in patients with GBM inhibits the effectiveness of immunotherapy. Therefore, reversing this GBM-mediated immune suppression is critical to increase the effectiveness of immunotherapy for GBM [48]. Consequently, we believe it is meaningful to validate whether blocking the above 12 genes, which are associated with immunosuppression in GBM, affects the prognosis of GBM in this study. Secondly, ADAM22, AEBP1, CHL1, EFEMP2, PDPN, PGCP, RAC3, SHANK1, SWAP70, and TRIP6 genes may impact the prognosis of GBM through mechanisms involving cell adhesion or structural and extracellular matrix. ADAM22, RAC3, and SHANK1 were associated with a favorable prognosis in patients with GBM, and the expression of the remaining genes was associated with a poor prognosis. Focal adhesion is at the center of signaling pathways crucial for tumor development and may mediate radioresistance, chemotherapy, and resistance to targeted therapy in glioma [49]. Consequently, we believe that the above cell adhesion-related genes associated with the GBM prognosis identified in this study may have clinical implications for the future treatment of GBM. Finally, our results demonstrate that CHST2, PPCS, and FBXO17 may influence the prognosis of GBM through metabolism pathways. CHST2 could impact the WNT/β-catenin pathway, F3 and SERPING1 through blood coagulation cascade, and ATP5C1 through mitochondrial ATP synthesis. Therefore, based on our findings, we are planning future in vitro and/or in vivo experiments to validate the relationship between the identified genes and GBM prognosis. We expect that future experimental studies may contribute to improving the treatment of GBM.

This study has several limitations: Firstly, we obtained all clinical and mRNA expression data from the TCGA database, which is retrospective. Thus, further planned studies are required to verify these results. However, since public TCGA data was used and all the raw data is presented as Supplementary Data 1, our results can be evaluated and validated by other researchers. Secondly, the fraction of immune cells in GBM was estimated using in silico flow cytometry-based analysis, although this may not accurately reflect the actual number of immune cells. Thirdly, the current findings were not verified through experimental analyses; therefore, further in vitro and/or in vivo studies are required. Fourth, there are missing clinical and mRNA expression data that were unavailable in the TCGA dataset, potentially influencing the results of the statistical analyses in the study. Lastly, this study is subject to potential bias because it only used data from a single TCGA database. Therefore, verifying the results in future studies using different databases is necessary.

Conclusion

Overall, we investigated significant genes related to both length of OS and PFS in patients with GBM using a large-scale, open database. According to our findings, there were 33 independent genes among 12,042 human genes whose expressions were significantly associated with the prognosis of GBM. Among these 33 significant genes, the expressions of five genes were associated with an improved prognosis of GBM, while numerous other genes were related to a worse prognosis in patients with GBM. In addition, expressions of ADAM22, ATP5C1, RAC3, SHANK1, AEBP1, C1RL, CHL1, CHST2, EFEMP2, and PGCP genes were either positively or negatively related to the long-term survival of GBM patients. Although our findings are required to be validated in the future, we believe that they may contribute to improving the understanding of the mechanisms underlying the pathophysiology of GBM.

Supporting information

S1 Data. The clinical information and mRNA expression data from the TCGA database of 525 GBM cases.

https://doi.org/10.1371/journal.pone.0295061.s001

(XLSX)

S1 Fig. Kaplan–Meier curves showing overall survival (OS) and progression-free survival (PFS) rates according to DHRS2, AEBP1, C1RL, CBR1, CCL2, CLEC5A, DYNLT3, EFEMP2, EMP3, F3, FBXO17, FLJ11286, MSN, NSUN5, PDPN, and PGCP expression quartiles.

https://doi.org/10.1371/journal.pone.0295061.s002

(TIF)

S2 Fig. Kaplan–Meier curves showing overall survival (OS) and progression-free survival (PFS) rates according to PPCS, SERPING1, SLC25A20, SLC2A10, STEAP3, SWAP70, TIMP1, TMEM22, and TRIP6 expression quartiles.

https://doi.org/10.1371/journal.pone.0295061.s003

(TIF)

S1 Table. Clinical and immune cell characteristics in patients with GBM.

https://doi.org/10.1371/journal.pone.0295061.s004

(DOCX)

S2 Table. Univariable linear regression analysis of the lengths of the OS and PFS according to the 36 significant gene expressions in patients with GBM.

https://doi.org/10.1371/journal.pone.0295061.s005

(DOCX)

References

1. Tamimi AF, Juweid M. Epidemiology and Outcome of Glioblastoma. In: De Vleeschouwer S, editor. Glioblastoma. Brisbane (AU): Codon Publications; 2017. Available: http://www.ncbi.nlm.nih.gov/books/NBK470003/
2. Bikfalvi A, da Costa CA, Avril T, Barnier J-V, Bauchet L, Brisson L, et al. Challenges in glioblastoma research: focus on the tumor microenvironment. Trends in Cancer. 2023;9: 9–27.
- View Article
- Google Scholar
3. Han M-H, Min K-W, Noh Y-K, Kim JM, Cheong JH, Ryu JI, et al. High DKK3 expression related to immunosuppression was associated with poor prognosis in glioblastoma: machine learning approach. Cancer Immunol Immunother. 2022;71: 3013–3027. pmid:35599254
- View Article
- PubMed/NCBI
- Google Scholar
4. Han M-H, Min K-W, Noh Y-K, Kim JM, Cheong JH, Ryu JI, et al. Identification of genes from ten oncogenic pathways associated with mortality and disease progression in glioblastoma. Frontiers in Oncology. 2022;12. Available: https://www.frontiersin.org/articles/10.3389/fonc.2022.965638
- View Article
- Google Scholar
5. Weinstein JN, Collisson EA, Mills GB, Shaw KRM, Ozenberger BA, Ellrott K, et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nat Genet. 2013;45: 1113–1120. pmid:24071849
- View Article
- PubMed/NCBI
- Google Scholar
6. Sun Q, Li M, Wang X. The Cancer Omics Atlas: an integrative resource for cancer omics annotations. BMC Medical Genomics. 2018;11: 63. pmid:30089500
- View Article
- PubMed/NCBI
- Google Scholar
7. Zou KH, Tuncali K, Silverman SG. Correlation and Simple Linear Regression. Radiology. 2003;227: 617–628.
- View Article
- Google Scholar
8. van der Leun AM, Thommen DS, Schumacher TN. CD8+ T cell states in human cancer: insights from single-cell analysis. Nat Rev Cancer. 2020;20: 218–232. pmid:32024970
- View Article
- PubMed/NCBI
- Google Scholar
9. Bindea G, Mlecnik B, Hackl H, Charoentong P, Tosolini M, Kirilovsky A, et al. ClueGO: a Cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks. Bioinformatics. 2009;25: 1091–1093. pmid:19237447
- View Article
- PubMed/NCBI
- Google Scholar
10. D’Abaco GM, Ng K, Paradiso L, Godde NJ, Kaye A, Novak U. ADAM22, expressed in normal brain but not in high-grade gliomas, inhibits cellular proliferation via the disintegrin domain. Neurosurgery. 2006;58: 179–186; discussion 179–186.
- View Article
- Google Scholar
11. Loo LWM, Cheng I, Tiirikainen M, Lum-Jones A, Seifried A, Dunklee LM, et al. cis-Expression QTL analysis of established colorectal cancer risk variants in colon tumors and adjacent normal tissue. PLoS One. 2012;7: e30477. pmid:22363440
- View Article
- PubMed/NCBI
- Google Scholar
12. Zhou Y, Wang L, Ban X, Zeng T, Zhu Y, Li M, et al. DHRS2 inhibits cell growth and motility in esophageal squamous cell carcinoma. Oncogene. 2018;37: 1086–1094. pmid:29106393
- View Article
- PubMed/NCBI
- Google Scholar
13. Wang M, Shen S, Hou F, Yan Y. Pathophysiological roles of integrins in gliomas from the perspective of glioma stem cells. Frontiers in Cell and Developmental Biology. 2022;10. Available: https://www.frontiersin.org/articles/10.3389/fcell.2022.962481
- View Article
- Google Scholar
14. Song Y, Ma R. Multiple Omics Analysis of the Rac3 Roles in Different Types of Human Cancer. IEEE Access. 2022;10: 92633–92650.
- View Article
- Google Scholar
15. Haataja L, Kaartinen V, Groffen J, Heisterkamp N. The Small GTPase Rac3 Interacts with the Integrin-binding Protein CIB and Promotes Integrin αIIbβ3-mediated Adhesion and Spreading*. Journal of Biological Chemistry. 2002;277: 8321–8328.
- View Article
- Google Scholar
16. Lilja J, Zacharchenko T, Georgiadou M, Jacquemet G, Franceschi ND, Peuhu E, et al. SHANK proteins limit integrin activation by directly interacting with Rap1 and R-Ras. Nat Cell Biol. 2017;19: 292–305. pmid:28263956
- View Article
- PubMed/NCBI
- Google Scholar
17. Guo K, Song L, Chang J, Cao P, Liu Q. AEBP1 Promotes Glioblastoma Progression and Activates the Classical NF-κB Pathway. Behavioural Neurology. 2020;2020: e8890452. pmid:33224311
- View Article
- PubMed/NCBI
- Google Scholar
18. Wang J, Tong L, Lin G, Wang H, Zhang L, Yang X. Immunological and clinicopathological characteristics of C1RL in 2120 glioma patients. BMC Cancer. 2020;20: 931. pmid:32993564
- View Article
- PubMed/NCBI
- Google Scholar
19. Yun M, Choi AJ, Lee YC, Kong M, Sung J-Y, Kim SS, et al. Carbonyl reductase 1 is a new target to improve the effect of radiotherapy on head and neck squamous cell carcinoma. Journal of Experimental & Clinical Cancer Research. 2018;37: 264. pmid:30376862
- View Article
- PubMed/NCBI
- Google Scholar
20. Chang AL, Miska J, Wainwright DA, Dey M, Rivetta CV, Yu D, et al. CCL2 Produced by the Glioma Microenvironment Is Essential for the Recruitment of Regulatory T Cells and Myeloid-Derived Suppressor Cells. Cancer Research. 2016;76: 5671–5682. pmid:27530322
- View Article
- PubMed/NCBI
- Google Scholar
21. Li F, Qi B, Yang L, Wang B, Gao L, Zhao M, et al. CHI3L1 predicted in malignant entities is associated with glioblastoma immune microenvironment. Clinical Immunology. 2022;245: 109158.
- View Article
- Google Scholar
22. Lin W-W, Ou G-Y, Lin J-Z, Yi S-J, Yao W-C, Pan H-C, et al. Neuregulin 1 enhances cell adhesion molecule L1 like expression levels and promotes malignancy in human glioma. Oncology Letters. 2020;20: 326–336. pmid:32565959
- View Article
- PubMed/NCBI
- Google Scholar
23. Wang J, Xia X, Tao X, Zhao P, Deng C. Knockdown of carbohydrate sulfotransferase 12 decreases the proliferation and mobility of glioblastoma cells via the WNT/β-catenin pathway. Bioengineered. 12: 3934–3946. pmid:34288811
- View Article
- PubMed/NCBI
- Google Scholar
24. Tong L, Li J, Choi J, Pant A, Xia Y, Jackson C, et al. CLEC5A expressed on myeloid cells as a M2 biomarker relates to immunosuppression and decreased survival in patients with glioma. Cancer Gene Ther. 2020;27: 669–679.
- View Article
- Google Scholar
25. Hu Y, Wang J, Chen Z. BIOM-33. LOW EXPRESSION OF DYNLT3 PREDICTS BETTER PROGNOSIS FOR FEMALE GLIOBLASTOMA PATIENTS. Neuro Oncol. 2020;22: ii8.
- View Article
- Google Scholar
26. Wang L, Chen Q, Chen Z, Tian D, Xu H, Cai Q, et al. EFEMP2 is upregulated in gliomas and promotes glioma cell proliferation and invasion. Int J Clin Exp Pathol. 2015;8: 10385–10393.
- View Article
- Google Scholar
27. Chen Q, Jin J, Huang X, Wu F, Huang H, Zhan R. EMP3 mediates glioblastoma‐associated macrophage infiltration to drive T cell exclusion. Journal of Experimental & Clinical Cancer Research. 2021;40: 160. pmid:33964937
- View Article
- PubMed/NCBI
- Google Scholar
28. Mandoj C, Tomao L, Conti L. Coagulation in Brain Tumors: Biological Basis and Clinical Implications. Frontiers in Neurology. 2019;10. Available: https://www.frontiersin.org/articles/10.3389/fneur.2019.00181
- View Article
- Google Scholar
29. Wang N, Song Q, Yu H, Bao G. Overexpression of FBXO17 Promotes the Proliferation, Migration and Invasion of Glioma Cells Through the Akt/GSK-3β/Snail Pathway. Cell Transplant. 2021;30: 09636897211007395. pmid:33853342
- View Article
- PubMed/NCBI
- Google Scholar
30. Wang Q, Lu X, Zhao S, Pang M, Wu X, Wu H, et al. Moesin Expression Is Associated with Glioblastoma Cell Proliferation and Invasion. Anticancer Res. 2017;37: 2211–2218.
- View Article
- Google Scholar
31. Ansa-Addo EA, Zhang Y, Yang Y, Hussey GS, Howley BV, Salem M, et al. Membrane-organizing protein moesin controls Treg differentiation and antitumor immunity via TGF-β signaling. J Clin Invest. 2017;127: 1321–1337. pmid:28287407
- View Article
- PubMed/NCBI
- Google Scholar
32. Janin M, Ortiz-Barahona V, de Moura MC, Martínez-Cardús A, Llinàs-Arias P, Soler M, et al. Epigenetic loss of RNA-methyltransferase NSUN5 in glioma targets ribosomes to drive a stress adaptive translational program. Acta Neuropathol. 2019;138: 1053–1074. pmid:31428936
- View Article
- PubMed/NCBI
- Google Scholar
33. Modrek AS, Eskilsson E, Ezhilarasan R, Wang Q, Goodman LD, Ding Y, et al. PDPN marks a subset of aggressive and radiation-resistant glioblastoma cells. Frontiers in Oncology. 2022;12. Available: https://www.frontiersin.org/articles/10.3389/fonc.2022.941657
- View Article
- Google Scholar
34. Expression of CPQ in glioma—The Human Protein Atlas. [cited 18 Jan 2023]. Available: https://www.proteinatlas.org/ENSG00000104324-CPQ/pathology/glioma
- View Article
- Google Scholar
35. Lee JV, Berry CT, Kim K, Sen P, Kim T, Carrer A, et al. Acetyl-CoA promotes glioblastoma cell adhesion and migration through Ca2+–NFAT signaling. Genes Dev. 2018;32: 497–511. pmid:29674394
- View Article
- PubMed/NCBI
- Google Scholar
36. Iuso A, Wiersma M, Schüller H-J, Pode-Shakked B, Marek-Yagel D, Grigat M, et al. Mutations in PPCS, Encoding Phosphopantothenoylcysteine Synthetase, Cause Autosomal-Recessive Dilated Cardiomyopathy. Am J Hum Genet. 2018;102: 1018–1030. pmid:29754768
- View Article
- PubMed/NCBI
- Google Scholar
37. Xiao K, Tan J, Yuan J, Peng G, Long W, Su J, et al. Prognostic value and immune cell infiltration of hypoxic phenotype-related gene signatures in glioblastoma microenvironment. Journal of Cellular and Molecular Medicine. 2020;24: 13235–13247. pmid:33009892
- View Article
- PubMed/NCBI
- Google Scholar
38. Davis AE, Mejia P, Lu F. BIOLOGICAL ACTIVITIES OF C1 INHIBITOR. Mol Immunol. 2008;45: 4057–4063. pmid:18674818
- View Article
- PubMed/NCBI
- Google Scholar
39. Jiang L, Yang J, Xu Q, Lv K, Cao Y. Machine learning for the micropeptide encoded by LINC02381 regulates ferroptosis through the glucose transporter SLC2A10 in glioblastoma. BMC Cancer. 2022;22: 882. pmid:35962317
- View Article
- PubMed/NCBI
- Google Scholar
40. Expression of SLC25A20 in glioma—The Human Protein Atlas. [cited 18 Jan 2023]. Available: https://www.proteinatlas.org/ENSG00000178537-SLC25A20/pathology/glioma
- View Article
- Google Scholar
41. Han M, Xu R, Wang S, Yang N, Ni S, Zhang Q, et al. Six-Transmembrane Epithelial Antigen of Prostate 3 Predicts Poor Prognosis and Promotes Glioblastoma Growth and Invasion. Neoplasia. 2018;20: 543–554. pmid:29730475
- View Article
- PubMed/NCBI
- Google Scholar
42. Shi L, Liu H, Wang Y, Chong Y, Wang J, Liu G, et al. SWAP-70 promotes glioblastoma cellular migration and invasion by regulating the expression of CD44s. Cancer Cell International. 2019;19: 305. pmid:31832018
- View Article
- PubMed/NCBI
- Google Scholar
43. Liu L, Yang S, Lin K, Yu X, Meng J, Ma C, et al. Sp1 induced gene TIMP1 is related to immune cell infiltration in glioblastoma. Sci Rep. 2022;12: 11181. pmid:35778451
- View Article
- PubMed/NCBI
- Google Scholar
44. Lin VTG, Lin VY, Lai Y-J, Chen C-S, Liu K, Lin W-C, et al. TRIP6 regulates p27 KIP1 to promote tumorigenesis. Mol Cell Biol. 2013;33: 1394–1409. pmid:23339869
- View Article
- PubMed/NCBI
- Google Scholar
45. Safran M, Dalah I, Alexander J, Rosen N, Iny Stein T, Shmoish M, et al. GeneCards Version 3: the human gene integrator. Database. 2010;2010: baq020. pmid:20689021
- View Article
- PubMed/NCBI
- Google Scholar
46. Li Z-H, Guan Y-L, Zhang G-B. Genomic Analysis of Glioblastoma Multiforme Reveals a Key Transcription Factor Signature Relevant to Prognosis and the Immune Processes. Frontiers in Oncology. 2021;11. Available: https://www.frontiersin.org/articles/10.3389/fonc.2021.657531
- View Article
- Google Scholar
47. Li Q, Chen B, Cai J, Sun Y, Wang G, Li Y, et al. Comparative Analysis of Matrix Metalloproteinase Family Members Reveals That MMP9 Predicts Survival and Response to Temozolomide in Patients with Primary Glioblastoma. PLoS One. 2016;11: e0151815. pmid:27022952
- View Article
- PubMed/NCBI
- Google Scholar
48. Himes BT, Geiger PA, Ayasoufi K, Bhargav AG, Brown DA, Parney IF. Immunosuppression in Glioblastoma: Current Understanding and Therapeutic Implications. Frontiers in Oncology. 2021;11. pmid:34778089
- View Article
- PubMed/NCBI
- Google Scholar
49. Li H, Wang G, Wang W, Pan J, Zhou H, Han X, et al. A Focal Adhesion-Related Gene Signature Predicts Prognosis in Glioma and Correlates With Radiation Response and Immune Microenvironment. Frontiers in Oncology. 2021;11. Available: https://www.frontiersin.org/articles/10.3389/fonc.2021.698278
- View Article
- Google Scholar

[ref1] 1. Tamimi AF, Juweid M. Epidemiology and Outcome of Glioblastoma. In: De Vleeschouwer S, editor. Glioblastoma. Brisbane (AU): Codon Publications; 2017. Available: http://www.ncbi.nlm.nih.gov/books/NBK470003/

[ref2] 2. Bikfalvi A, da Costa CA, Avril T, Barnier J-V, Bauchet L, Brisson L, et al. Challenges in glioblastoma research: focus on the tumor microenvironment. Trends in Cancer. 2023;9: 9–27.
View Article
Google Scholar

[3] View Article

[4] Google Scholar

[ref3] 3. Han M-H, Min K-W, Noh Y-K, Kim JM, Cheong JH, Ryu JI, et al. High DKK3 expression related to immunosuppression was associated with poor prognosis in glioblastoma: machine learning approach. Cancer Immunol Immunother. 2022;71: 3013–3027. pmid:35599254
View Article
PubMed/NCBI
Google Scholar

[6] View Article

[7] PubMed/NCBI

[8] Google Scholar

[ref4] 4. Han M-H, Min K-W, Noh Y-K, Kim JM, Cheong JH, Ryu JI, et al. Identification of genes from ten oncogenic pathways associated with mortality and disease progression in glioblastoma. Frontiers in Oncology. 2022;12. Available: https://www.frontiersin.org/articles/10.3389/fonc.2022.965638
View Article
Google Scholar

[10] View Article

[11] Google Scholar

[ref5] 5. Weinstein JN, Collisson EA, Mills GB, Shaw KRM, Ozenberger BA, Ellrott K, et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nat Genet. 2013;45: 1113–1120. pmid:24071849
View Article
PubMed/NCBI
Google Scholar

[13] View Article

[14] PubMed/NCBI

[15] Google Scholar

[ref6] 6. Sun Q, Li M, Wang X. The Cancer Omics Atlas: an integrative resource for cancer omics annotations. BMC Medical Genomics. 2018;11: 63. pmid:30089500
View Article
PubMed/NCBI
Google Scholar

[17] View Article

[18] PubMed/NCBI

[19] Google Scholar

[ref7] 7. Zou KH, Tuncali K, Silverman SG. Correlation and Simple Linear Regression. Radiology. 2003;227: 617–628.
View Article
Google Scholar

[21] View Article

[22] Google Scholar

[ref8] 8. van der Leun AM, Thommen DS, Schumacher TN. CD8+ T cell states in human cancer: insights from single-cell analysis. Nat Rev Cancer. 2020;20: 218–232. pmid:32024970
View Article
PubMed/NCBI
Google Scholar

[24] View Article

[25] PubMed/NCBI

[26] Google Scholar

[ref9] 9. Bindea G, Mlecnik B, Hackl H, Charoentong P, Tosolini M, Kirilovsky A, et al. ClueGO: a Cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks. Bioinformatics. 2009;25: 1091–1093. pmid:19237447
View Article
PubMed/NCBI
Google Scholar

[28] View Article

[29] PubMed/NCBI

[30] Google Scholar

[ref10] 10. D’Abaco GM, Ng K, Paradiso L, Godde NJ, Kaye A, Novak U. ADAM22, expressed in normal brain but not in high-grade gliomas, inhibits cellular proliferation via the disintegrin domain. Neurosurgery. 2006;58: 179–186; discussion 179–186.
View Article
Google Scholar

[32] View Article

[33] Google Scholar

[ref11] 11. Loo LWM, Cheng I, Tiirikainen M, Lum-Jones A, Seifried A, Dunklee LM, et al. cis-Expression QTL analysis of established colorectal cancer risk variants in colon tumors and adjacent normal tissue. PLoS One. 2012;7: e30477. pmid:22363440
View Article
PubMed/NCBI
Google Scholar

[35] View Article

[36] PubMed/NCBI

[37] Google Scholar

[ref12] 12. Zhou Y, Wang L, Ban X, Zeng T, Zhu Y, Li M, et al. DHRS2 inhibits cell growth and motility in esophageal squamous cell carcinoma. Oncogene. 2018;37: 1086–1094. pmid:29106393
View Article
PubMed/NCBI
Google Scholar

[39] View Article

[40] PubMed/NCBI

[41] Google Scholar

[ref13] 13. Wang M, Shen S, Hou F, Yan Y. Pathophysiological roles of integrins in gliomas from the perspective of glioma stem cells. Frontiers in Cell and Developmental Biology. 2022;10. Available: https://www.frontiersin.org/articles/10.3389/fcell.2022.962481
View Article
Google Scholar

[43] View Article

[44] Google Scholar

[ref14] 14. Song Y, Ma R. Multiple Omics Analysis of the Rac3 Roles in Different Types of Human Cancer. IEEE Access. 2022;10: 92633–92650.
View Article
Google Scholar

[46] View Article

[47] Google Scholar

[ref15] 15. Haataja L, Kaartinen V, Groffen J, Heisterkamp N. The Small GTPase Rac3 Interacts with the Integrin-binding Protein CIB and Promotes Integrin αIIbβ3-mediated Adhesion and Spreading*. Journal of Biological Chemistry. 2002;277: 8321–8328.
View Article
Google Scholar

[49] View Article

[50] Google Scholar

[ref16] 16. Lilja J, Zacharchenko T, Georgiadou M, Jacquemet G, Franceschi ND, Peuhu E, et al. SHANK proteins limit integrin activation by directly interacting with Rap1 and R-Ras. Nat Cell Biol. 2017;19: 292–305. pmid:28263956
View Article
PubMed/NCBI
Google Scholar

[52] View Article

[53] PubMed/NCBI

[54] Google Scholar

[ref17] 17. Guo K, Song L, Chang J, Cao P, Liu Q. AEBP1 Promotes Glioblastoma Progression and Activates the Classical NF-κB Pathway. Behavioural Neurology. 2020;2020: e8890452. pmid:33224311
View Article
PubMed/NCBI
Google Scholar

[56] View Article

[57] PubMed/NCBI

[58] Google Scholar

[ref18] 18. Wang J, Tong L, Lin G, Wang H, Zhang L, Yang X. Immunological and clinicopathological characteristics of C1RL in 2120 glioma patients. BMC Cancer. 2020;20: 931. pmid:32993564
View Article
PubMed/NCBI
Google Scholar

[60] View Article

[61] PubMed/NCBI

[62] Google Scholar

[ref19] 19. Yun M, Choi AJ, Lee YC, Kong M, Sung J-Y, Kim SS, et al. Carbonyl reductase 1 is a new target to improve the effect of radiotherapy on head and neck squamous cell carcinoma. Journal of Experimental & Clinical Cancer Research. 2018;37: 264. pmid:30376862
View Article
PubMed/NCBI
Google Scholar

[64] View Article

[65] PubMed/NCBI

[66] Google Scholar

[ref20] 20. Chang AL, Miska J, Wainwright DA, Dey M, Rivetta CV, Yu D, et al. CCL2 Produced by the Glioma Microenvironment Is Essential for the Recruitment of Regulatory T Cells and Myeloid-Derived Suppressor Cells. Cancer Research. 2016;76: 5671–5682. pmid:27530322
View Article
PubMed/NCBI
Google Scholar

[68] View Article

[69] PubMed/NCBI

[70] Google Scholar

[ref21] 21. Li F, Qi B, Yang L, Wang B, Gao L, Zhao M, et al. CHI3L1 predicted in malignant entities is associated with glioblastoma immune microenvironment. Clinical Immunology. 2022;245: 109158.
View Article
Google Scholar

[72] View Article

[73] Google Scholar

[ref22] 22. Lin W-W, Ou G-Y, Lin J-Z, Yi S-J, Yao W-C, Pan H-C, et al. Neuregulin 1 enhances cell adhesion molecule L1 like expression levels and promotes malignancy in human glioma. Oncology Letters. 2020;20: 326–336. pmid:32565959
View Article
PubMed/NCBI
Google Scholar

[75] View Article

[76] PubMed/NCBI

[77] Google Scholar

[ref23] 23. Wang J, Xia X, Tao X, Zhao P, Deng C. Knockdown of carbohydrate sulfotransferase 12 decreases the proliferation and mobility of glioblastoma cells via the WNT/β-catenin pathway. Bioengineered. 12: 3934–3946. pmid:34288811
View Article
PubMed/NCBI
Google Scholar

[79] View Article

[80] PubMed/NCBI

[81] Google Scholar

[ref24] 24. Tong L, Li J, Choi J, Pant A, Xia Y, Jackson C, et al. CLEC5A expressed on myeloid cells as a M2 biomarker relates to immunosuppression and decreased survival in patients with glioma. Cancer Gene Ther. 2020;27: 669–679.
View Article
Google Scholar

[83] View Article

[84] Google Scholar

[ref25] 25. Hu Y, Wang J, Chen Z. BIOM-33. LOW EXPRESSION OF DYNLT3 PREDICTS BETTER PROGNOSIS FOR FEMALE GLIOBLASTOMA PATIENTS. Neuro Oncol. 2020;22: ii8.
View Article
Google Scholar

[86] View Article

[87] Google Scholar

[ref26] 26. Wang L, Chen Q, Chen Z, Tian D, Xu H, Cai Q, et al. EFEMP2 is upregulated in gliomas and promotes glioma cell proliferation and invasion. Int J Clin Exp Pathol. 2015;8: 10385–10393.
View Article
Google Scholar

[89] View Article

[90] Google Scholar

[ref27] 27. Chen Q, Jin J, Huang X, Wu F, Huang H, Zhan R. EMP3 mediates glioblastoma‐associated macrophage infiltration to drive T cell exclusion. Journal of Experimental & Clinical Cancer Research. 2021;40: 160. pmid:33964937
View Article
PubMed/NCBI
Google Scholar

[92] View Article

[93] PubMed/NCBI

[94] Google Scholar

[ref28] 28. Mandoj C, Tomao L, Conti L. Coagulation in Brain Tumors: Biological Basis and Clinical Implications. Frontiers in Neurology. 2019;10. Available: https://www.frontiersin.org/articles/10.3389/fneur.2019.00181
View Article
Google Scholar

[96] View Article

[97] Google Scholar

[ref29] 29. Wang N, Song Q, Yu H, Bao G. Overexpression of FBXO17 Promotes the Proliferation, Migration and Invasion of Glioma Cells Through the Akt/GSK-3β/Snail Pathway. Cell Transplant. 2021;30: 09636897211007395. pmid:33853342
View Article
PubMed/NCBI
Google Scholar

[99] View Article

[100] PubMed/NCBI

[101] Google Scholar

[ref30] 30. Wang Q, Lu X, Zhao S, Pang M, Wu X, Wu H, et al. Moesin Expression Is Associated with Glioblastoma Cell Proliferation and Invasion. Anticancer Res. 2017;37: 2211–2218.
View Article
Google Scholar

[103] View Article

[104] Google Scholar

[ref31] 31. Ansa-Addo EA, Zhang Y, Yang Y, Hussey GS, Howley BV, Salem M, et al. Membrane-organizing protein moesin controls Treg differentiation and antitumor immunity via TGF-β signaling. J Clin Invest. 2017;127: 1321–1337. pmid:28287407
View Article
PubMed/NCBI
Google Scholar

[106] View Article

[107] PubMed/NCBI

[108] Google Scholar

[ref32] 32. Janin M, Ortiz-Barahona V, de Moura MC, Martínez-Cardús A, Llinàs-Arias P, Soler M, et al. Epigenetic loss of RNA-methyltransferase NSUN5 in glioma targets ribosomes to drive a stress adaptive translational program. Acta Neuropathol. 2019;138: 1053–1074. pmid:31428936
View Article
PubMed/NCBI
Google Scholar

[110] View Article

[111] PubMed/NCBI

[112] Google Scholar

[ref33] 33. Modrek AS, Eskilsson E, Ezhilarasan R, Wang Q, Goodman LD, Ding Y, et al. PDPN marks a subset of aggressive and radiation-resistant glioblastoma cells. Frontiers in Oncology. 2022;12. Available: https://www.frontiersin.org/articles/10.3389/fonc.2022.941657
View Article
Google Scholar

[114] View Article

[115] Google Scholar

[ref34] 34. Expression of CPQ in glioma—The Human Protein Atlas. [cited 18 Jan 2023]. Available: https://www.proteinatlas.org/ENSG00000104324-CPQ/pathology/glioma
View Article
Google Scholar

[117] View Article

[118] Google Scholar

[ref35] 35. Lee JV, Berry CT, Kim K, Sen P, Kim T, Carrer A, et al. Acetyl-CoA promotes glioblastoma cell adhesion and migration through Ca2+–NFAT signaling. Genes Dev. 2018;32: 497–511. pmid:29674394
View Article
PubMed/NCBI
Google Scholar

[120] View Article

[121] PubMed/NCBI

[122] Google Scholar

[ref36] 36. Iuso A, Wiersma M, Schüller H-J, Pode-Shakked B, Marek-Yagel D, Grigat M, et al. Mutations in PPCS, Encoding Phosphopantothenoylcysteine Synthetase, Cause Autosomal-Recessive Dilated Cardiomyopathy. Am J Hum Genet. 2018;102: 1018–1030. pmid:29754768
View Article
PubMed/NCBI
Google Scholar

[124] View Article

[125] PubMed/NCBI

[126] Google Scholar

[ref37] 37. Xiao K, Tan J, Yuan J, Peng G, Long W, Su J, et al. Prognostic value and immune cell infiltration of hypoxic phenotype-related gene signatures in glioblastoma microenvironment. Journal of Cellular and Molecular Medicine. 2020;24: 13235–13247. pmid:33009892
View Article
PubMed/NCBI
Google Scholar

[128] View Article

[129] PubMed/NCBI

[130] Google Scholar

[ref38] 38. Davis AE, Mejia P, Lu F. BIOLOGICAL ACTIVITIES OF C1 INHIBITOR. Mol Immunol. 2008;45: 4057–4063. pmid:18674818
View Article
PubMed/NCBI
Google Scholar

[132] View Article

[133] PubMed/NCBI

[134] Google Scholar

[ref39] 39. Jiang L, Yang J, Xu Q, Lv K, Cao Y. Machine learning for the micropeptide encoded by LINC02381 regulates ferroptosis through the glucose transporter SLC2A10 in glioblastoma. BMC Cancer. 2022;22: 882. pmid:35962317
View Article
PubMed/NCBI
Google Scholar

[136] View Article

[137] PubMed/NCBI

[138] Google Scholar

[ref40] 40. Expression of SLC25A20 in glioma—The Human Protein Atlas. [cited 18 Jan 2023]. Available: https://www.proteinatlas.org/ENSG00000178537-SLC25A20/pathology/glioma
View Article
Google Scholar

[140] View Article

[141] Google Scholar

[ref41] 41. Han M, Xu R, Wang S, Yang N, Ni S, Zhang Q, et al. Six-Transmembrane Epithelial Antigen of Prostate 3 Predicts Poor Prognosis and Promotes Glioblastoma Growth and Invasion. Neoplasia. 2018;20: 543–554. pmid:29730475
View Article
PubMed/NCBI
Google Scholar

[143] View Article

[144] PubMed/NCBI

[145] Google Scholar

[ref42] 42. Shi L, Liu H, Wang Y, Chong Y, Wang J, Liu G, et al. SWAP-70 promotes glioblastoma cellular migration and invasion by regulating the expression of CD44s. Cancer Cell International. 2019;19: 305. pmid:31832018
View Article
PubMed/NCBI
Google Scholar

[147] View Article

[148] PubMed/NCBI

[149] Google Scholar

[ref43] 43. Liu L, Yang S, Lin K, Yu X, Meng J, Ma C, et al. Sp1 induced gene TIMP1 is related to immune cell infiltration in glioblastoma. Sci Rep. 2022;12: 11181. pmid:35778451
View Article
PubMed/NCBI
Google Scholar

[151] View Article

[152] PubMed/NCBI

[153] Google Scholar

[ref44] 44. Lin VTG, Lin VY, Lai Y-J, Chen C-S, Liu K, Lin W-C, et al. TRIP6 regulates p27 KIP1 to promote tumorigenesis. Mol Cell Biol. 2013;33: 1394–1409. pmid:23339869
View Article
PubMed/NCBI
Google Scholar

[155] View Article

[156] PubMed/NCBI

[157] Google Scholar

[ref45] 45. Safran M, Dalah I, Alexander J, Rosen N, Iny Stein T, Shmoish M, et al. GeneCards Version 3: the human gene integrator. Database. 2010;2010: baq020. pmid:20689021
View Article
PubMed/NCBI
Google Scholar

[159] View Article

[160] PubMed/NCBI

[161] Google Scholar

[ref46] 46. Li Z-H, Guan Y-L, Zhang G-B. Genomic Analysis of Glioblastoma Multiforme Reveals a Key Transcription Factor Signature Relevant to Prognosis and the Immune Processes. Frontiers in Oncology. 2021;11. Available: https://www.frontiersin.org/articles/10.3389/fonc.2021.657531
View Article
Google Scholar

[163] View Article

[164] Google Scholar

[ref47] 47. Li Q, Chen B, Cai J, Sun Y, Wang G, Li Y, et al. Comparative Analysis of Matrix Metalloproteinase Family Members Reveals That MMP9 Predicts Survival and Response to Temozolomide in Patients with Primary Glioblastoma. PLoS One. 2016;11: e0151815. pmid:27022952
View Article
PubMed/NCBI
Google Scholar

[166] View Article

[167] PubMed/NCBI

[168] Google Scholar

[ref48] 48. Himes BT, Geiger PA, Ayasoufi K, Bhargav AG, Brown DA, Parney IF. Immunosuppression in Glioblastoma: Current Understanding and Therapeutic Implications. Frontiers in Oncology. 2021;11. pmid:34778089
View Article
PubMed/NCBI
Google Scholar

[170] View Article

[171] PubMed/NCBI

[172] Google Scholar

[ref49] 49. Li H, Wang G, Wang W, Pan J, Zhou H, Han X, et al. A Focal Adhesion-Related Gene Signature Predicts Prognosis in Glioma and Correlates With Radiation Response and Immune Microenvironment. Frontiers in Oncology. 2021;11. Available: https://www.frontiersin.org/articles/10.3389/fonc.2021.698278
View Article
Google Scholar

[174] View Article

[175] Google Scholar

Figures

Abstract

Background and purpose

Methods

Results

Conclusions

Introduction

Methods

Study patients

Study design

In silico flow cytometry

Bioinformatics analysis

Statistical analysis

Results

Characteristics of the study patients

Identification of significant genes associated with the prognosis of GBM

Expression patterns of the 33 significant genes and immune cells in GBM

Correlations between the expressions of the 33 genes, the lengths of the OS and PFS, and the immune cells in GBM

Associations between the expressions of the selected genes and the lengths of the OS and PFS in patients with GBM

Identification of genes whose expressions predict long-term survival of patients with GBM

Functional gene ontology and pathway network analyses

Discussion

Conclusion

Supporting information

S1 Data. The clinical information and mRNA expression data from the TCGA database of 525 GBM cases.

S1 Fig. Kaplan–Meier curves showing overall survival (OS) and progression-free survival (PFS) rates according to DHRS2, AEBP1, C1RL, CBR1, CCL2, CLEC5A, DYNLT3, EFEMP2, EMP3, F3, FBXO17, FLJ11286, MSN, NSUN5, PDPN, and PGCP expression quartiles.

S2 Fig. Kaplan–Meier curves showing overall survival (OS) and progression-free survival (PFS) rates according to PPCS, SERPING1, SLC25A20, SLC2A10, STEAP3, SWAP70, TIMP1, TMEM22, and TRIP6 expression quartiles.

S1 Table. Clinical and immune cell characteristics in patients with GBM.

S2 Table. Univariable linear regression analysis of the lengths of the OS and PFS according to the 36 significant gene expressions in patients with GBM.

References