Characterization of Prognostic Apoptosis-related Gene Signature to Evaluate Glioma Immune Microenvironment and experimental verification

doi:10.21203/rs.3.rs-3330482/v1

Download PDF

Research Article

Characterization of Prognostic Apoptosis-related Gene Signature to Evaluate Glioma Immune Microenvironment and experimental verification

https://doi.org/10.21203/rs.3.rs-3330482/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 01 Jan, 2024

Read the published version in Genetic Testing and Molecular Biomarkers →

Version 1

posted

You are reading this latest preprint version

Purpose

Apoptosis-related gene has been demonstrated to modulate the cancer immune in recent studies. Nonetheless, comprehensive roles of apoptosis-related genes in glioma immune microenvironment (GIME) remain unknow. This article was intended to explore the prognostic of apoptosis-related genes in glioma.

Methods

Doxorubicin was used to induce glioma cells apoptosis, and four differentially expressed apoptosis-related genes were identified, namely CREM, TNFSF12, PEA15 and PRKCD. Kaplan-Meier analyses, receiver operating characteristic curve (ROC) analyses and nomogram were established to construe the relationship between risk markers and the prognosis of glioma patients.

Results

Risk biomarkers were dramatically associated with overall survival in glioma patients. The high and low risk groups were in relation to immune cell infiltration and immune checkpoints. Somatic mutation and anti-PD-1/L1 immunotherapy demonstrated a worse prognosis in the high-risk group on anti-PD1/PDL1 therapy. In addition, the expression of these four apoptosis-related genes was verified by qPCR and immunohistochemistry, and the relationship between these four genes and apoptosis was examined by flow cytometry.

Conclusions

This research proclaimed the apoptosis related genes played a critical function in shaping of GIME. Assessing the apoptotic patterns of individual tumors will enhance our understanding of the infiltration feature of GIME and lead better strategies for immunotherapy.

glioma

apoptosis

immune microenvironment

risk signature

Glioma is a primary brain tumor derived from neuroectodermal lobe and is the most frequent intracranial tumor [1]. The common types of gliomas are astrocytoma, glioblastoma, oligodendroglioma, ependymoma, etc [2]. Glioma can be separate into grade I-IV according to its malignancy, and its grade is inversely proportional to prognosis. Glioblastoma multiforme (GBM) is the type with the highest degree of malignancy, dominate about 60% of gliomas, and has a bad prognosis with a median survival of only 15 months [3]. Maximum safety surgical excision and chemoradiotherapy are the standard treatments for GBM, but relapse occurs in almost all patients and the prognosis for patients remains poor [4]. Recent years, emerging glioma treatment ways have emerged, such as targeted therapy, immunotherapy, electric field therapy, etc [5]. Immunotherapy has achieved good results in some hematologic tumors, but little breakthrough has been made in glioblastoma and other solid tumors [6]. These methods are a new treatment option for glioma, but these treatments have failed to significantly improve the survival of glioma patients [7]. Consequently, it is urgent to explore new therapeutic approach for glioma [8].

Apoptosis is a kind of procedural death of cells autonomously controlled by genes [9]. Different from necrosis, apoptosis is of great significance to the human body. It can remove unwanted and harmful cells and maintain the dynamic balance of cell number and function [10]. The main characteristic of apoptosis is: the cell becomes round, breaks away from neighboring cells, condenses the nucleus and cytoplasm, degrades the DNA nucleosome junction area, breaks the nuclear membrane, expands the endoplasmic reticulum, and invaginates the cell contents into some vesicles, which are then engulfed by surrounding cells [11]. Apoptosis plays a critical role in tumors, there is a loss of balance between cell multiplication and cell death, and this problem can occur at any step in the apoptosis process [12]. The down-regulation of p53 facilitates the decrease of apoptosis and the enhancement of tumor growth [13]. Up-regulated or down-regulated expression of some genes has been found to promote carcinogenesis by reducing apoptosis of tumor cells [14]. Research suggested that under certain conditions, in the absence of caspase, apoptosis can trigger an adaptive anti-tumor immune response through activating the NF-κB signaling pathway [15]. In the tumor microenvironment (TME), apoptosis induced by cytotoxic immune cells is the main way to kill cancer cells in TME [16]. In addition, the acidic, hypoxic and other immunosuppressive environment of TME also promoted the failure and apoptosis of cytotoxic immune cells, in which Treg, M2 macrophages and myeloid-derived suppressor cells (MDSC) converged [17]. Cytotoxic T cells depletion reduced the apoptosis of tumor cells, and thus weaken the anti-tumor effect [18]. This is the reason for the poor effect of glioma and solid tumor immunotherapy [19]. Consequently, it is prominent to clarify the apoptosis of tumor cells in TME.

Currently, apoptosis-related genes have been studied in gliomas; however, due to technical limitations, most of these studies have been conducted on apoptosis-related genes in individual gliomas, rather than pan-analysis. The characteristics of glioma immune microenvironment is the way a number of factors in a highly coordinated interaction. Accordingly, a holistic understanding of apoptosis-related gene-mediated immune cell infiltration characteristics will be conducive to improve our notion of glioma immune regulation. In this article, we induced apoptosis of glioma cell lines by doxorubicin, sequenced 1626 differentially related genes, and analyzed the expression of these 1626 apoptosis-related genes in 667 glioma patients using a complete transcriptome data from the Cancer Genome Atlas (TCGA) database. We acquired four apoptosis-related genes prominently relevant to immune infiltration and established risk profiles for four apoptosis-related genes with prognostic value.

2.1 Cell line and cell culture

Human U251 glioma cell lines were acquired from American Type Culture Collection (ATCC). Using DMEM (HyClone, USA) with 10% fetal bovine serum (Gibco, USA) and 1% penicillin/streptomycin (Biosharp, China) for the cell culture at 37℃ in 5% CO₂.

2.2 Induction and validation of apoptosis

The density of U251 cells was 1 × 10⁵ cells per well (6-well plate). For apoptosis induction, cells were treated with 0, 100, 200,500 nM Doxorubicin (Selleck, USA) for 4 days after tests of several concentration gradients. Effects of apoptosis were evaluated using a Cell apoptosis detection kit (Beyotime, China). Apoptosis was induced in U251 cells successfully for three times with DMSO (Beyotime, China) as a vehicle control, and the cell samples were stored in Trizol reagent (Invitrogen, USA) for stepwise transcriptome RNA sequencing (RNA-seq).

2.3 Online data collection

The FPKM transcriptome and clinical data of patients with low-grade glioma (LGG) and glioblastoma (GBM) (n = 667) were excerpted from The Cancer Genome Atlas Database (TCGA, https://cancergenome.nih.gov/). Validation FPKM transcriptome data and clinical features of patients (n = 970) from the Chinese Glioma Genome Atlas (CGGA) were obtained on the official website (http://www.cgga.org.cn/). The flow diagram for the design of this study was shown in Supplementary Figure S1.

2.4 Correlation analysis

Spearman correlation analyses and CIBERSORT deconvolution algorithm was employed to figure out the correlation index between expression level of senescence-related genes and T-cell immune infiltration data using R software. Normal distributions were tested using Kolmogorov-Smirnova test and Shapiro-Wilk test.

2.5 Screening of apoptosis and T-cell infiltration-related genes

Using transcriptome RNA sequencing (RNA-seq), apoptosis-related genes were sifted through the criteria (p value < .05, |log₂FC| >1). Using online data collected from TCGA datasets and CIBERSORT deconvolution algorithm, T-cell infiltration-related genes were screened out with the criteria (|cor| >0.4, p < .05). Taken the intersection of the two gene sets, apoptosis and T-cell infiltration-related genes were finally selected.

2.6 Establishment of a risk signature based on apoptosis-related prognostic genes

Prognostic analysis was applied to analyze the relationship between each apoptosis-related genes and overall survival (OS) of patient by a univariate Cox model. Least absolute shrinkage and selection operator (LASSO) regression was utilized to stepwise screen out the congenial apoptosis-related prognostic genes combination and to construct the risk signature. On the basis of LASSO analysis, a prognostic risk score equation was established, and we used a combination weighted by expression level and coefficient of regression as follows:

Risk score = Gene 1 expression × β1 + Gene 2 expression × β2 +......+ Gene n expression × βn. There were 667 patients with glioma were classified into high and low-risk subgroups using the median risk score as the cutoff value. Kaplan-Meier (KM) curves and time-dependent receiver operating characteristic (ROC) curves were utilized to validate the prognostic significance of the risk signature. The CGGA database containing 970 patients was utilized to further verify the effectiveness of the risk signature.

2.7 Evaluating the effectiveness of nomogram in predicting 1-, 3- and 5-year survival rate

To apply the risk signature to the clinical feature, we performed univariate and multivariate Cox regression analyses to confirm the connection between these characters (gender, age, grade, isocitrate dehydrogenase (IDH) status, 1p/19q codeletion, history of radiotherapy or chemotherapy and risk score) and patient overall survival. A nomogram was established and validated by using the TCGA database (training cohort) and CGGA dataset (validation cohort). The calibration plots and C-index were applied for both datasets. The prognostic capability of the nomogram and other predictors for 1-, 3-, 5-year OS was assessed by ROC curves.

2.8 Evaluation of overall survival (OS) and clinical variables in different risk subgroups

Kaplan-Meier curve was generated using R package “survminer” on R software to explore the OS of the risk subgroups. Meanwhile, a heatmap showing the clinical variables of different subgroups was generated using R package “pheatmap”.

2.9 Functional enrichment analysis

The differentially expressed genes (DEGs) of the high-/low-risk subgroups were screened out according to the criteria of |log2FC| > 1 and p-value < 0.05 using the R package “edgeR”. To analyze the functional enrichment of DEGs, Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were performed. The results were exhibited with R package “ggpolt2”.

2.10 Immune landscape and ICI therapy in different risk subgroups

The ESTIMATE algorithm was implemented to compute the ESTIMATE score, stromal score and tumor purity using R package “estimate” on R software. Infiltration of 22 species of immune cells of per sample were quantized using CIBERSORT (https://cibersort.stanford.edu/) algorithm. The response to ICI therapy of per sample was forecasted utilizing tumor immune dysfunction and exclusion (TIDE; http://tide.dfci.harvard.edu) algorithm.

2.10 Tumor mutation burden (TMB) in the risk subgroups

The TMB data of each patient were acquired from the TCGA database. The somatic variant information was saved in mutation annotation format (MAF). Furthermore, the acquired mutation files were processed and quantized using R package “Maftools”.

2.11 Tissue samples and immunohistochemistry (IHC)

All glioma and adjacent normal tissues were acquired from patients with glioma in Wuhan Union Hospital. Informed consents were acquired from included patients and ethical approval was acquired from Medical Ethics Committee. Immunohistochemistry (IHC) was implemented to verify protein level of finally-screened genes. Briefly, formalin-fixed, paraffin-embedded samples were sectioned in 4µm slices, dewaxed, hydrated and incubated in sodium citrate (15min, 95°C). Then, slices were washed with PBS buffer, added primary antibodies (4°C, overnight), washed again and incubated with secondary antibody (HRP polymer, 40min, Room temperature). Finally, color reaction was performed with diaminobenzidine and hematoxylin to manifest appropriate brown color.

2.12 RNA purification and quantitative real-time polymerase chain reaction (qRT-PCR)

Total RNA of both U251 cell and tissue involved were extracted by RNAiso Plus (Takara 9109). Then, followed by reverse transcription performed, expression level of selected genes was evaluated on LightCycler 480 Real-Time PCR system with TB Green® Premix Ex Taq™ II (Takara RR820A) and compared using the 2 − ΔΔCt method after normalized by GAPDH. The primers sequences were listed in Supplementary Table S4.

2.13 Flow cytometry

Annexin V-PE/ 7-AAD apoptosis detection kit was utilized to detect apoptosis level. After washing the control cells and knockout cell lines, discarded the supernatant and the cells were resuspended with the diluted 1×Annexin V Binding Buffer. Annexin V-PE and 7-AAD staining solution were put into the cell suspension. The cells were incubated at 25℃ for 15-20min away from light. Then flow cytometry was performed.

2.14 Statistical analysis

Statistical analyses and figures were accomplished by R software (version 4.0.3). Statistics were evaluated as significant different if the calculated two-tailed p value was less than 0.05. A P-value < 0.05 was considered statistically significant, “***”, p-value < 0.001, “**”, p-value < 0.01, “*”, p-value < 0.05.

3.1 Identification of apoptosis-related genes in glioma patients

We treated glioma cell line U251 with different concentrations of doxorubicin (0, 100, 200, 500 nM) for four days to induce apoptosis in Supplementary Figure S2. After sequencing, 1626 differentially expressed apoptosis-related genes were acquired. We obtained 667 glioma samples from the TCGA database. The expression data of these genes were obtained from TCGA database. Through univariate regression analysis, we obtained 218 apoptosis-related genes associated with prognosis. By LASSO regression analysis, there were 27 apoptosis-related genes (Fig. 1A and 1B). In order to study apoptosis and immunity, we finally screened out four genes related to T lymphocytes.

3.2 Accessing prognostic merit of the risk signature

After screening, we obtained the most suitable 4 apoptosis-related and T-cell-related genes with prognostic value, which are CREM, TNFSF12, PEA15 and PRKCD (Fig. 1C). The survival analysis was displayed in Supplementary Figure S3. The risk characteristics of 667 glioma patients were constructed, and then computed the risk scores of the patients. On the basis of median score, 667 patients were classified into the high-risk and low-risk groups (Supplementary Table S1 and S2). In the high-risk group, CREM, TNFSF12 and PRKCD were upregulated, while PEA15 was downregulated. Kaplan-Meier (KM) curves demonstrated that the OS rate of the high-risk group was dramatically lower than the overall survival rate of the low-risk group (p < 0.001) (Fig. 1D and 1G). The risk curve and scatterplot revealed that the death rate of the high-risk group was higher than the death rate of the low-risk group (Fig. 1E and 1H). Time-varying receiver operating characteristic (ROC) curves were used to verify the predictive significance of the risk signatures, and the AUC values of the 1-, 3- and 5-year survival rates in the TCGA database were 0.881, 0.929, and 0.905, respectively (Fig. 1F). Furthermore, we validated the efficacy of this risk signature in the CGGA database, and the same results were achieved, with AUC of 1-, 3- and 5-year survival rates of 0.733, 0.722, and 0.737, respectively (Fig. 1I).

3.3 Exploration of the relationship between risk signature and clinical feature

A heat map was drawn to study the relationship between the risk signature and glioma patient feature, and the results showed the relationship between risk groups and the clinical parameters of the patients such as gender, age, pathological grade, IDH mutation status and 1p/19q deletion (Fig. 2A). There seemed to be no prominent difference in gender between the high- and low-risk groups, but there were more elderly patients in the high-risk group. Patients ≤ 60 years old had lower risk scores than those > 60 years old (p < 0.001). When comparing the WHO classification in the high- and low-risk groups, the high WHO grade was more likely to move towards the direction of the high-risk group, while the low-risk group was more likely to have more low WHO grade patients (p < 0.001). As for IDH mutation status, the risk scores of IDH wild type patients were higher than the risk scores of IDH mutation patients; that is, the high-risk group was mostly IDH wild type (p < 0.001). Then, we analyzed the status of the 1p/19q deletion, and our results showed that those in the high-risk group were more possibly to be the 1p/19q non-codeletion type group(p < 0.001) (Fig. 2B-F). The clinical feature of glioma patients in CGGA database were displayed in Supplementary Figure S4.

3.4 Establishment of a nomogram for predicting the 1-, 3- and 5-year survival rates of patients with glioma

We conducted univariate and multivariate Cox analyses, and our results indicated that this risk signature can be recognized as an independent prognosis factor for glioma patients (Fig. 3A and 3B). For better glioma patient prediction depending on the analysis results above, we utilized the TCGA database to construct a nomogram of risk score, age, grade and IDH mutation and validated it with the CGGA database (Fig. 3E and Supplementary Figure S5). The calibration plots showed an ideal match between the actual and nomogram-predicted probability of 1-, 3- and 5-year OS in the TCGA and the CGGA cohort (Fig. 3C). Furthermore, a time-dependent ROC curve was performed to evaluate the predictive capability of the nomogram and predictors (risk score, age, grade and IDH status (Fig. 3D). The results showed that the nomogram could well predict 1-year, 3-year and 5-year OS with AUC values of 0.628, 0.580, and 0.565 in TCGA database and of 0.603, 0.698, and 0.738 in CGGA database. The nomogram is more accurate in predicting the 1, 3 and 5-year survival rates of patients with glioma. These results demonstrate that the nomogram is more credible and precise prognostic predictor.

3.5 Relevance between the risk signature and immune landscape of the glioma microenvironment

Through functional enrichment analyses, we obtained the potential biological processes and relevant signaling pathways in which the apoptosis related genes risk signature was involved in glioma. As shown in Fig. 4A-D, the biological processes neuroactive ligand-receptor interaction, cell adhesion molecules (CAMs), ECM-receptor interaction and many other processes were markedly relevant to the signature. Regarding the KEGG pathway, immunity-related pathways (human T- cell leukemia virus 1 infection, complement and coagulation cascades, peptide ligand-binding receptors, regulation of IGF transport and uptake by IGFBPS, neurotransmitter receptors and postsynaptic signal transmission, GPCR ligand binding) and apoptosis-related pathways (P53 signaling pathway, cell cycle, cellular senescence, degradation of the extracellular matix) were significantly associated with the signature. The immune infiltrating cells of the high- and low-risk groups was indicated in Fig. 4E. We observed that Tregs, activated NK cells, monocytes, M0 macrophages, M1 macrophages, activated CD4 memory T cells, mast cells, follicular helper T cells and neutrophils were all markedly different between the two risk groups (p < 0.05). However, CD8 + T cells were not significantly different in the two groups. Due to the dual importance of CD8 + T cell number and function, we further calculated the relevance between risk scores and immune checkpoint expression. In Fig. 5E, we found that the number of immune checkpoints such as CD274, CD276, CTLA4, CD70, CD86, LILRB2 and TIMGD2, was positively relevant to the risk score, which fully demonstrated an immunosuppressive microenvironment in the high-risk group of patients with glioma.

3.6 Risk signature with tumor mutation burden and anti-PD-1/L1 immunotherapy

We analyzed the distribution of tumor mutation burden between low- and high-risk groups in TCGA cohort using “maftools” package. As indicated in Fig. 5A and 5B, low-risk group displayed more extensive tumor mutation burden than the high-risk group, with the rate of the 20th most noticeable mutated gene 99.37% versus 89.71%. More precisely, IDH1 and ATRX mutation happen more frequently in low-risk group than high one (92% vs 50%, 37% vs 31%), which predicts a better prognosis and is consistent with the results above. While several oncogenic mutations in high-risk group such as EGFR and PTEN happen a few times more than in low-risk group (14% vs 2%, 10% vs 2%). These all proved that high-risk group are more aggressive. Further, immunotherapies of PD-1 and PD-L1 blockade has indubitably arose a momentous breakthrough in tumor therapy. We inquired whether the apoptosis-related gene signature could predict patients’ response to immune checkpoint blockade therapy. The result from TIDE demonstrated that 58% patients have a response for anti-PD-1/L1 immunotherapy in low-risk group, while 69% patients have a response for anti-PD-1/L1 immunotherapy in high-risk group (Fig. 5D). The result from ImmuCellAI demonstrated that 39% patients have a response for anti-PD-1/L1 immunotherapy in low-risk group, while 54% patients have a response for anti-PD-1/L1 immunotherapy in high-risk group (Fig. 5C).

3.7 Expression level of the four apoptosis-related genes

We validated the level of the 4 apoptosis-related genes at mRNA and protein levels in the tissue sections of 28 gliomas and 15 paracancer patients. Quantitative PCR and immunohistochemistry showed that CREM, TNFSF12 and PRKCD were highly expressed in gliomas, while PEA15 was low expressed in gliomas (Fig. 6A-E).

3.8 Verified the relationship between these four genes and apoptosis by flow cytometry analysis

We constructed CREM, TNFSF12, PRKCD and PEA15 knockdown U251 glioma cell lines. As indicated in Fig. 6F, compared to the control group, the cell vitality of CREM, TNFSF12, PRKCD knockdown group decreased significantly, while that of PEA15 knockdown group increased significantly. Flow cytometry analysis demonstrated an increase in apoptotic cells in CREM, TNFSF12, PRKCD knockdown cells, and a decrease in apoptotic cells in PEA15 knockdown cells compared with the control group. We can conclude that CREM, TNFSF12 and PRKCD knockdown can promote glioma cells apoptosis, and PEA15 knockdown can restrain apoptosis of glioma cells (Fig. 6G-J).

Glioma is a refractory brain tumor. According to the latest studies, the tumor immune microenvironment significantly contributes to glioma therapeutic resistance and has attracted wide attention from both researchers and clinicians [20]. Therefore, immunotherapy has gradually become a mainstream treatment strategy for glioma. Immunotherapy is a novel and potential therapy that can mobilize immune cells to enhance their antitumor immune function. Significant therapeutic effects have been achieved in lymphoma, leukemia and other hematological diseases [21]. However, the effects of immunotherapy in glioma are not as optimistic as the effects in nonsolid tumors, which may be attributed to the inhibitory immune microenvironment [22]. In addition, the mutational rate of glioma is relatively low when compared with other malignancies, which causes the “cold” characteristic, and thus, there are fewer intra-tumoral inflammatory cells [23]. Therefore, researchers are devoted to finding an immune-related checkpoint that can switch “cold tumor” to “hot tumor” [24]. Many basic studies and clinical trials have been performed to remodel the glioma immune microenvironment, but less progress has been made due to the low therapeutic response rate [25].

Apoptosis plays a momentous function in embryogenesis, organ development and tumor occurrence [26]. Apoptosis is different from other death modes because it has a series of special morphological and biochemical characteristics. Studies have shown that apoptosis is an important process to prevent the emergence of tumors, and tumor cells can inhibit apoptosis [27]. In tumor cells, the activity of Caspase is inhibited. Caspases are a family of evolutionarily conserved cysteine-dependent endonuclease enzymes. Caspase is a key protein in initiating and executing cell apoptosis. Reduced expression level or impaired function of caspase is closely related to cancer progression [28]. In tumor cells, the death receptor signaling pathway is impaired. The anti-apoptotic gene Bcl2 is activated, which promotes the development of lymphoma. Tumor suppressor gene p53 is a key regulator of apoptosis. Down-regulation of p53 enhances the occurrence of many tumors. McL-1 is high expressed in a variety of tumors and is relevant to apoptosis resistance in tumor cells [29]. Endothelial cells promote angiogenesis, which is significantly related to apoptosis resistance [30]. Adipocytes play a momentous part of the tumor microenvironment and facilitate tumor progression by secreting growth factors. Adipocytes can make leukemia cells resistant to chemotherapy by promoting anti-apoptotic proteins Pim-2 and Bcl-2 expression [31].

As a classical regulator of almost all tumor pathologic processes, apoptosis regulates the formation and development of tumor immune microenvironment [32]. Thus, we purposely developed a risk signature containing four apoptosis-related genes (CREM, TNFSF12, PRKCD and PEA15) that was relevant to the survival of glioma patients. Among these genes, cyclic AMP response element (CRE) modulator (CREM) is highly expressed in gastric adenocarcinoma and is closely related to poor overall survival [33]. CREM is also relevant to T cell depletion and M2 polarization. CREM may regulate the expression of Serpin family F member 1 (SERPINF1) in gliomas [34]. Tumor necrosis factor (ligand) superfamily member 12, also called TWEAK, is negative correlation with patient survival [35]. TNFSF12-fibroblast growth factor-induced 14 (Fn14) signaling system regulates glioma cell survival through NF-κB pathway activation and BCL-XL/BCL-W expression [36]. In addition, TNFSF12 facilitates glioma cell invasion by inducing NF-κB-induced kinase (NIK) conduction [37]. Protein kinase C (PKC) is involved in the proliferation and apoptosis of glial tumors and plays a considerable function in regulating glioma growth [38]. Protein kinase C delta (PRKCD) is low expressed in glioma. In human U87 glioma cell line, PRKCD overexpression inhibited cell proliferation and enhanced glial fibrillary acidic protein (GFAP) expression [39]. PKCdelta also keeps glioma cells from apoptosis induced by tumor necrosis factor-associated apoptosis-inducing ligand (TRAIL) [40]. The level of proliferation and apoptosis of adaptor protein 15 (PEA15) were decreased in glioblastoma, which affected the stability of p53, and was essential for the inhibition of p53-induced apoptosis [41]. The high level of PEA-15 was negatively relevant to overall survival (OS) of high-grade glioma patients [42].

Aiming at validating the feasibility of this model in clinical application, Kaplan-Meier analyses, ROC analyses and nomogram construction demonstrated great potential in predicting glioma patient survival. Meanwhile, functional analysis revealed that this risk signature was relevant to cancer immunity and cancer progression, and immune cell infiltration analysis verified that the risk signature could reflect the immune microenvironment by determining the proportion of multiple lymphocytes. Furthermore, immune checkpoint expression was proven to be connected to risk scores. Therefore, it is essential to have several detectable indicators that can reflect the tumor immune microenvironment, somatic mutation maps and predict the prognosis of patients quickly and effectively. Also, we found the high-risk group patients were more sensitive to anti-PD-1/L1, which may result from the high expressions of immune checkpoints in them and more infiltrated T cells in glioma. The four apoptosis-related genes in our study meet the two conditions mentioned above and can be of great potential in accurately diagnosing and treating glioma.

Our studies identified four apoptosis related genes as prognostic signatures for glioma. This signature was associated with the tumor immune microenvironment and can be a potential indicator to predict the effect of immunotherapy in glioma. More clinical validation studies are needed to verify the effectiveness of the study.

Funding: This research was funded by NATIONAL NATURAL SCIENCE FOUNDATION OF CHINA, grant number 81974390.

Ethics Statement: The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of TONGJI MEDICAL COLLEGE, HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY (protocol code: [2019]IEC(S742). date of approval: MAR 4^th 2019).

Informed Consent Statement: Informed consent was obtained from all subjects involved in the study, and all patients signed written informed consent for the use of their samples and data.

Data Availability Statement:

All RNA-seq files are available from the TCGA database (https://cancergenome.nih.gov/) and CGGA (http://www.cgga.org.cn/) database within the article.

Acknowledgments:

We would like to thank all the people who helped us in the current study.

Conflicts of Interest:

The authors declare no conflict of interest.

Author contributions

Conceptualization, Xiaobing Jiang; Data curation, Hao Yu; Formal analysis, Hao Yu; Investigation, Minjie Wang; Methodology, Jiapeng Yu; Project administration, Xiaobing Jiang; Software, Jiapeng Yu; Supervision, Xiaobing Jiang; Validation, Minjie Wang; Visualization, Hao Yu; Writing - original draft, Hao Yu; and Writing - review & editing, Xiaobing Jiang.

Lapointe S., A. Perry, N. A. Butowski. Primary brain tumours in adults. Lancet. 2018;392(10145):432-46. https://doi.org/10.1016/s0140-6736(18)30990-5.
Schaff L. R., I. K. Mellinghoff. Glioblastoma and Other Primary Brain Malignancies in Adults: A Review. Jama. 2023;329(7):574-87. https://doi.org/10.1001/jama.2023.0023.
Ostrom Q. T., G. Cioffi, K. Waite, et al. CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2014-2018. Neuro Oncol. 2021;23(12 Suppl 2):iii1-iii105. https://doi.org/10.1093/neuonc/noab200.
Ricard D., A. Idbaih, F. Ducray, et al. Primary brain tumours in adults. Lancet. 2012;379(9830):1984-96. https://doi.org/10.1016/s0140-6736(11)61346-9.
Xu S., L. Tang, X. Li, et al. Immunotherapy for glioma: Current management and future application. Cancer Lett. 2020;476:1-12. https://doi.org/10.1016/j.canlet.2020.02.002.
Xun Y., H. Yang, B. Kaminska, et al. Toll-like receptors and toll-like receptor-targeted immunotherapy against glioma. J Hematol Oncol. 2021;14(1):176. https://doi.org/10.1186/s13045-021-01191-2.
Bunse L., T. Bunse, C. Krämer, et al. Clinical and Translational Advances in Glioma Immunotherapy. Neurotherapeutics. 2022;19(6):1799-817. https://doi.org/10.1007/s13311-022-01313-9.
Hua T., H. Shi, M. Zhu, et al. Glioma‑neuronal interactions in tumor progression: Mechanism, therapeutic strategies and perspectives (Review). Int J Oncol. 2022;61(3). https://doi.org/10.3892/ijo.2022.5394.
D'arcy M. S. Cell death: a review of the major forms of apoptosis, necrosis and autophagy. Cell Biol Int. 2019;43(6):582-92. https://doi.org/10.1002/cbin.11137.
Xu X., Y. Lai, Z. C. Hua. Apoptosis and apoptotic body: disease message and therapeutic target potentials. Biosci Rep. 2019;39(1). https://doi.org/10.1042/bsr20180992.
Lin X., S. Ouyang, C. Zhi, et al. Focus on ferroptosis, pyroptosis, apoptosis and autophagy of vascular endothelial cells to the strategic targets for the treatment of atherosclerosis. Arch Biochem Biophys. 2022;715:109098. https://doi.org/10.1016/j.abb.2021.109098.
Wong R. S. Apoptosis in cancer: from pathogenesis to treatment. J Exp Clin Cancer Res. 2011;30(1):87. https://doi.org/10.1186/1756-9966-30-87.
Wang X., E. R. Simpson, K. A. Brown. p53: Protection against Tumor Growth beyond Effects on Cell Cycle and Apoptosis. Cancer Res. 2015;75(23):5001-7. https://doi.org/10.1158/0008-5472.Can-15-0563.
Kumar S., M. S. Tomar, A. Acharya. HSF1-mediated regulation of tumor cell apoptosis: a novel target for cancer therapeutics. Future Oncol. 2013;9(10):1573-86. https://doi.org/10.2217/fon.13.106.
Hou J., J. M. Hsu, M. C. Hung. Molecular mechanisms and functions of pyroptosis in inflammation and antitumor immunity. Mol Cell. 2021;81(22):4579-90. https://doi.org/10.1016/j.molcel.2021.09.003.
Pfeifer E., J. M. Burchell, F. Dazzi, et al. Apoptosis in the Pancreatic Cancer Tumor Microenvironment-The Double-Edged Sword of Cancer-Associated Fibroblasts. Cells. 2021;10(7). https://doi.org/10.3390/cells10071653.
Dong F., X. Qin, B. Wang, et al. ALKBH5 Facilitates Hypoxia-Induced Paraspeckle Assembly and IL8 Secretion to Generate an Immunosuppressive Tumor Microenvironment. Cancer Res. 2021;81(23):5876-88. https://doi.org/10.1158/0008-5472.Can-21-1456.
Hu Y., D. Chen, M. Hong, et al. Apoptosis, Pyroptosis, and Ferroptosis Conspiringly Induce Immunosuppressive Hepatocellular Carcinoma Microenvironment and γδ T-Cell Imbalance. Front Immunol. 2022;13:845974. https://doi.org/10.3389/fimmu.2022.845974.
Wang H., T. Xu, Q. Huang, et al. Immunotherapy for Malignant Glioma: Current Status and Future Directions. Trends Pharmacol Sci. 2020;41(2):123-38. https://doi.org/10.1016/j.tips.2019.12.003.
Lieberman N. A. P., K. Degolier, H. M. Kovar, et al. Characterization of the immune microenvironment of diffuse intrinsic pontine glioma: implications for development of immunotherapy. Neuro Oncol. 2019;21(1):83-94. https://doi.org/10.1093/neuonc/noy145.
Ansell S. M. Immunotherapy in Hodgkin Lymphoma: The Road Ahead. Trends Immunol. 2019;40(5):380-6. https://doi.org/10.1016/j.it.2019.03.003.
Sampson J. H., M. V. Maus, C. H. June. Immunotherapy for Brain Tumors. J Clin Oncol. 2017;35(21):2450-6. https://doi.org/10.1200/jco.2017.72.8089.
Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 2008;455(7216):1061-8. https://doi.org/10.1038/nature07385.
Khasraw M., D. A. Reardon, M. Weller, et al. PD-1 Inhibitors: Do they have a Future in the Treatment of Glioblastoma? Clin Cancer Res. 2020;26(20):5287-96. https://doi.org/10.1158/1078-0432.Ccr-20-1135.
Feng M., W. Jiang, B. Y. S. Kim, et al. Phagocytosis checkpoints as new targets for cancer immunotherapy. Nat Rev Cancer. 2019;19(10):568-86. https://doi.org/10.1038/s41568-019-0183-z.
Lyons S. K., A. R. Clarke. Apoptosis and carcinogenesis. Br Med Bull. 1997;53(3):554-69. https://doi.org/10.1093/oxfordjournals.bmb.a011630.
Ritter V., F. Krautter, D. Klein, et al. Bcl-2/Bcl-xL inhibitor ABT-263 overcomes hypoxia-driven radioresistence and improves radiotherapy. Cell Death Dis. 2021;12(7):694. https://doi.org/10.1038/s41419-021-03971-7.
Liu J., F. Wu, M. Wang, et al. Caspase-3-Responsive Fluorescent/Photoacoustic Imaging of Tumor Apoptosis. Anal Chem. 2023;95(25):9404-8. https://doi.org/10.1021/acs.analchem.3c01387.
Liu T., V. Lam, E. Thieme, et al. Pharmacologic Targeting of Mcl-1 Induces Mitochondrial Dysfunction and Apoptosis in B-Cell Lymphoma Cells in a TP53- and BAX-Dependent Manner. Clin Cancer Res. 2021;27(17):4910-22. https://doi.org/10.1158/1078-0432.Ccr-21-0464.
Cheng H. W., Y. F. Chen, J. M. Wong, et al. Cancer cells increase endothelial cell tube formation and survival by activating the PI3K/Akt signalling pathway. J Exp Clin Cancer Res. 2017;36(1):27. https://doi.org/10.1186/s13046-017-0495-3.
Behan J. W., J. P. Yun, M. P. Proektor, et al. Adipocytes impair leukemia treatment in mice. Cancer Res. 2009;69(19):7867-74. https://doi.org/10.1158/0008-5472.Can-09-0800.
Deepak K. G. K., R. Vempati, G. P. Nagaraju, et al. Tumor microenvironment: Challenges and opportunities in targeting metastasis of triple negative breast cancer. Pharmacol Res. 2020;153:104683. https://doi.org/10.1016/j.phrs.2020.104683.
Yu K., L. Kuang, T. Fu, et al. CREM Is Correlated With Immune-Suppressive Microenvironment and Predicts Poor Prognosis in Gastric Adenocarcinoma. Front Cell Dev Biol. 2021;9:697748. https://doi.org/10.3389/fcell.2021.697748.
Song L., X. Huo, X. Li, et al. SERPINF1 Mediates Tumor Progression and Stemness in Glioma. Genes (Basel). 2023;14(3). https://doi.org/10.3390/genes14030580.
Pelekanou V., G. Notas, M. Kampa, et al. BAFF, APRIL, TWEAK, BCMA, TACI and Fn14 proteins are related to human glioma tumor grade: immunohistochemistry and public microarray data meta-analysis. PLoS One. 2013;8(12):e83250. https://doi.org/10.1371/journal.pone.0083250.
Tran N. L., W. S. Mcdonough, B. A. Savitch, et al. The tumor necrosis factor-like weak inducer of apoptosis (TWEAK)-fibroblast growth factor-inducible 14 (Fn14) signaling system regulates glioma cell survival via NFkappaB pathway activation and BCL-XL/BCL-W expression. J Biol Chem. 2005;280(5):3483-92. https://doi.org/10.1074/jbc.M409906200.
Cherry E. M., D. W. Lee, J. U. Jung, et al. Tumor necrosis factor-like weak inducer of apoptosis (TWEAK) promotes glioma cell invasion through induction of NF-κB-inducing kinase (NIK) and noncanonical NF-κB signaling. Mol Cancer. 2015;14(1):9. https://doi.org/10.1186/s12943-014-0273-1.
Bredel M., I. F. Pollack. The role of protein kinase C (PKC) in the evolution and proliferation of malignant gliomas, and the application of PKC inhibition as a novel approach to anti-glioma therapy. Acta Neurochir (Wien). 1997;139(11):1000-13. https://doi.org/10.1007/bf01411552.
Mandil R., E. Ashkenazi, M. Blass, et al. Protein kinase Calpha and protein kinase Cdelta play opposite roles in the proliferation and apoptosis of glioma cells. Cancer Res. 2001;61(11):4612-9.
Okhrimenko H., W. Lu, C. Xiang, et al. Protein kinase C-epsilon regulates the apoptosis and survival of glioma cells. Cancer Res. 2005;65(16):7301-9. https://doi.org/10.1158/0008-5472.Can-05-1064.
Renner G., H. Janouskova, F. Noulet, et al. Integrin α5β1 and p53 convergent pathways in the control of anti-apoptotic proteins PEA-15 and survivin in high-grade glioma. Cell Death Differ. 2016;23(4):640-53. https://doi.org/10.1038/cdd.2015.131.
Watanabe Y., F. Yamasaki, Y. Kajiwara, et al. Expression of phosphoprotein enriched in astrocytes 15 kDa (PEA-15) in astrocytic tumors: a novel approach of correlating malignancy grade and prognosis. J Neurooncol. 2010;100(3):449-57. https://doi.org/10.1007/s11060-010-0201-1.

No competing interests reported.

SupplementaryFigureS1.tif
Supplementary Figure S1. Flow chart of the study design for this article.
SupplementaryFigureS2.tif
Supplementary Figure S2. Doxorubicin (Doxo) induced apoptosis of U251 glioma cell line. (A) DMSO, 100 nM Doxo, 200 nM Doxo and 500 nM Doxo were added in U251 cells for 4 days. 1, DMSO; 2, 100 nM Doxo; 3, 200 nM Doxo; 4, 500 nM Doxo. (B) The relative number of apoptotic glioma cells was verified by qPCR. Data shown represent mean ± SD from three independent experiments. Comparisons were made using the Student’s t-test. ***p < 0.001; ****p < 0.0001; ns, not significant. (C) Differentially expressed gene after doxorubicin induction in U251 cells.
SupplementaryFigureS3.tif
Supplementary Figure S3. The Kaplan Meier (KM) curve showed that the overall survival rate of the 4 apoptosis-related genes. (A) CREM (B) TNFSF12 (C) PRKCD (D) PEA15.
SupplementaryFigureS4.tif
Supplementary Figure S4. Relationship between risk signature and clinical characteristics of glioma in CGGA database. (A) The heatmap showed the relationship between high and low risk groups and clinical features of glioma patients. Risk scores of high and low groups of (B) Age, (C) Gender, (D) IDH mutation status, (E) Radiotherapy, (F) Chemotherapy, (G) 1p19q codeletion and (H) Grade.
SupplementaryFigureS5.tif
Supplementary Figure S5. Calibration plots were used to validate the efficacy in the CGGA cohort (C-E). ROC curves were used to evaluate the predictive ability of the nomogram and other predictors (F-H). Univariate Cox regression analysis (A) and Multivariate Cox regression analysis (B) in CGGA database.
supplementaryfiles.docx

Download PDF

Journal Publication

published 01 Jan, 2024

Read the published version in Genetic Testing and Molecular Biomarkers →

Version 1

posted

You are reading this latest preprint version

Characterization of Prognostic Apoptosis-related Gene Signature to Evaluate Glioma Immune Microenvironment and experimental verification

Status:

Journal Publication

Version 1

Abstract

Figures

1. INTRODUCTION

2. METHODS AND MATERIALS

2.1 Cell line and cell culture

2.2 Induction and validation of apoptosis

2.3 Online data collection

2.4 Correlation analysis

2.5 Screening of apoptosis and T-cell infiltration-related genes

2.6 Establishment of a risk signature based on apoptosis-related prognostic genes

2.7 Evaluating the effectiveness of nomogram in predicting 1-, 3- and 5-year survival rate

2.8 Evaluation of overall survival (OS) and clinical variables in different risk subgroups

2.9 Functional enrichment analysis

2.10 Immune landscape and ICI therapy in different risk subgroups

2.10 Tumor mutation burden (TMB) in the risk subgroups

2.11 Tissue samples and immunohistochemistry (IHC)

2.12 RNA purification and quantitative real-time polymerase chain reaction (qRT-PCR)

2.13 Flow cytometry

2.14 Statistical analysis

3. RESULTS

3.1 Identification of apoptosis-related genes in glioma patients

3.2 Accessing prognostic merit of the risk signature

3.3 Exploration of the relationship between risk signature and clinical feature

3.5 Relevance between the risk signature and immune landscape of the glioma microenvironment

3.6 Risk signature with tumor mutation burden and anti-PD-1/L1 immunotherapy

3.7 Expression level of the four apoptosis-related genes

3.8 Verified the relationship between these four genes and apoptosis by flow cytometry analysis

4. DISCUSSION

5. CONCLUSIONS

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1