The present study was granted approval by the Specialty Committee on Ethics of Biomedicine Research at the Tongji University. Written informed consent was obtained from all patients. The selection criteria were previously described (18). Briefly, the selection criteria were as follows: i) The subject had a diagnosis of primary GBM and no history of other tumors; ii) the subject had complete clinical data, including age, sex, clinical manifestations, mean tumor diameter (defined as the geometric mean of the three diameters by MRI scan), extent of resection and adjuvant therapy; and iii) the subject underwent evaluation by enhanced head MRI scans for tumor relapse or progression after surgery at least once every six months. A total of 124 patients who received glioma-resection surgery were recruited at the Department of Neurosurgery, Shanghai Tongji Hospital of Tongji University and of Changzheng Hospital, the Second Military Medical University. The subjects were recruited from July 2016 to December 2018. The follow-up was carried out by telephone or mail every 6 months and the survival time was evaluated until March 2019. GBM was diagnosed according to the 2016 WHO classification of tumors of the CNS by two independent experienced pathologists. The clinical characteristics of the patients with GBM are listed in Table I.

The U87 cell line (glioblastoma of unknown origin) was purchased from the American Type Culture Collection (cat. no. HTB-14™). The cells were grown in DMEM (Sigma-Aldrich; Merck KGaA) supplemented with 8% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.), penicillin G (100 U/ml) and streptomycin (100 g/ml) and maintained at 37°C in humidified air with 5% CO₂. The cells were incubated in anoxic and/or hypoxic (3–5% O₂) environments. Hypoxic conditions were obtained by replacing oxygen with N₂, using a Heracell 150i CO₂ incubator (Thermo Fisher Scientific, Inc.).

To examine the expression of RLIP76 in GBM, a LinkedOmics analysis was performed of The Cancer Genome Atlas (TCGA) GBM databases according to their IDH1 status. LinkedOmics (http://www.linkedomics.org) is a publicly-available portal comprising multi-omics data from all 32 types of cancer in the TCGA database.

To examine the hypoxia-induced gene expression profiles of IDH1^wt and IDH1^Mut glioma stem cell lines, a previously published microarray study was downloaded from the Gene Expression Omnibus (GEO) database (accession no. GSE118683) (21). Differentially expressed genes (DEGs) were identified using the edgeR package (http://www.bioconductor.org/packages/release/bioc/html/edgeR.html) in R software; P<0.05 and fold change >2 were considered as statistically significant. The expressions of all DEGs were presented in the heatmap.

The DAVID bioinformatics resources 6.8 (https://david.ncifcrf.gov) were used to perform gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses on the differentially expressed genes (22,23). Following the analyses for significance and false discovery rate (FDR), the GO terms were selected from the significantly enriched gene sets (P<0.05 and FDR <0.05).

The IDH1-Flag and IDH1 R132H-Flag plasmids were constructed into the pCMV-Tag2B vector as previously described (24). Lipofectamine™ 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) was used for transfection according to the manufacturer's instructions. The efficacy of the transfection was tested by examining the expression of the Flag protein using western blotting. IDH1 siRNA (cat. no. sc-60829) and siRNA reagent system (cat. no. sc-45064), and RLIP76 siRNA (cat. no. sc-36376) and control siRNA (cat. no. sc-37007), were purchased from Santa Cruz Biotechnology, Inc., and all transfections were conducted according to the manufacturer's instructions. On the 3rd day following transfection, protein expression was examined by western blotting in order to evaluate the knockdown efficacy.

Total RNA, including miRNA, was extracted from cells/tissues with TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.). RNA was reverse transcribed into cDNA using the ReverTra Ace qPCR RT kit (FSQ-101; Toyobo Life Science), according to the manufacturer's protocol. cDNA was amplified using the Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen; Thermo Fisher Scientific, Inc.) on a 7500 Fast Real-time PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.). The PCR conditions were: 10 min at 95°C, 1 min at 55°C, followed by 40 cycles of 15 sec at 95°C and 30 sec at 55°C. Relative fold expression of the target gene was normalized to β-actin and calculated according to the 2^−∆∆Cq method (25). The sequences for the forward and reverse primers of RLIP76 and β-actin were previously described (18).

Cell apoptosis was assessed using the ApoScreen Annexin V Apoptosis Kit (Bender MedSystems, GmbH). After washing twice with PBS, cells were collected and stained with FITC-Annexin V and propidium iodide, as per the manufacturer's instructions. All cells were analyzed by flow cytometry (FACScan; BD Biosciences). Data analyses were performed using CellQuest software Version 5.0 (BD Biosciences). Experiments were conducted in triplicates.

Western blot analyses were performed as described previously (26). The primary antibodies used were the following: Anti-Flag antibody (cat. no. LT0420; LifeTein, LLC), IDH1 (Cell Signaling Technology, Inc.; cat. no. 8137; 1:500), RLIP76 (Abcam; cat. no. ab56815; 1:1,000), caspase-9 (Santa Cruz Biotechnology, Inc.; cat. no. sc-17784; 1:500), Bcl-2 (Santa Cruz Biotechnology, Inc.; cat. no. sc-509; 1:500), p53 (Abcam; cat. no. ab-32389; 1:1,000) and β-actin (Abcam; cat. no. ab-8227; 1:1,000). Western blot analysis was quantified by normalizing the signal intensity of each sample to that of β-actin.

Cell viability was assessed by the Cell Counting Kit-8 (CCK-8) assay. Briefly, all cells were seeded in collagen-coated 96-well plates at a density of 5×10³ cells/ml and incubated with 10 µl CCK-8 solution for 4 h at 37°C. The optical density of each sample was recorded with a microplate reader (Bio-Rad Laboratories, Inc.) at 450 nm. The assays were performed in triplicate.

The immunohistochemical assay (IHC) was conducted as previously described (18). The RLIP76 monoclonal antibody (Abcam; cat. no. ab-133549) was used at a dilution of 1:1,000. The assessment of RLIP76 expression was conducted as described previously (18). Briefly, RLIP76 expression was divided into ‘high’ (++, final score 4–6 and +++, final score >6) vs. ‘low’ (+, final score 1–3 and -, final score 0).

GSH was quantified using a GSH-Glo™ Glutathione Assay kit (cat. no. V6912; Promega Corporation), according to the manufacturers' instructions, following incubation of 1×10⁴ cells/well in 96-well plates for 1 h at 37°C with 5% CO₂ in serum-free HBSS with added Ca²⁺ and Mg²⁺.

The experimental data were presented as mean ± standard deviation. Statistical analysis was performed by the Student's t-test. Kaplan-Meier survival analysis was used to compare the overall survival (OS) in glioma patients. Univariate survival analysis was carried out by the Kaplan-Meier method and analyzed via the log-rank test to assess differences in survival between the groups. The Cox proportional hazard model for multivariate survival analysis was used to assess predictors of survival. P<0.05 was considered to indicate a statistically significant difference.

The expression levels of RLIP76 in GBM were investigated from the TCGA database using LinkedOmics (http://www.linkedomics.org). RLIP76 mRNA expression levels were significantly higher in the IDH1^Wt (n=131) group compared with the IDH1^Mut (n=7) group (Fig. 1A; P=0.0406). Subsequently, the protein expression levels of RLIP76 were investigated in 124 human GBM specimens by IHC staining. The GBM tissues were divided into IDH1^Mut (n=26) and IDH1^Wt (n=98) groups, according to their WHO grading. The IHC results revealed that the IDH1^Wt group exhibited higher immunoreactivity for RLIP76 in their GBM tissues compared with the IDH1^Mut group (Fig. 1B). By RT-qPCR analysis, it was also revealed that the IDH1^Wt group exhibited significantly higher RLIP76 mRNA expression levels compared with the IDH1^Mut group (Fig. 1C). Similar differences were also observed by western blotting, with regard to the RLIP76 protein expression levels between the two glioma groups (Fig. 1D). These results indicated that the expression levels of RLIP76 were significantly upregulated in IDH1^Wt gliomas compared with the IDH1^Mut gliomas.

To further explore the prognostic value of RLIP76 in GBM tissues with different IDH1 mutational status, the patients (n=124) were divided into the IDH1^Mut (n=26) and IDH1^Wt (n=98) groups. The results confirmed that IDH1^Wt was significantly associated with poor OS (P=0.0241; Fig. 2A) and progression-free survival (PFS) in patients with GBM (P=0.0178; Fig. 2B). Kaplan-Meier analysis further revealed that increased RLIP76 expression levels were associated with poor OS and PFS in the IDH1^Wt group (P=0.0297 and P=0.0374, respectively), while no significant difference was noted with regard to the prognostic value of RLIP76 in the IDH1^Mut group (P=0.162 and P=0.219, respectively; Fig. 2C and D). These results suggested that RLIP76 may exert a specific role in IDH1^Wt GBM that is different from that noted in IDH1^Mut GBM.

Univariate survival analysis indicated that the expression levels of RLIP76 were a prognostic factor for OS and for PFS in patients with IDH1^Wt glioma (Table II). Multivariate analysis further confirmed that high expression levels of RLIP76 at diagnosis were a critical and independent prognostic factor for OS and PFS in patients with IDH1^Wt glioma (Table III).

To identify novel oncogenic mRNAs in GBM tissues with different IDH1 mutational status, the present study analyzed a previously published microarray study of hypoxia-treated IDH1^wt and IDH1^Mut glioma stem cell lines (full data available at GEO, accession no. GSE118683) (21). The present analysis identified 2,928 mRNAs that were upregulated in IDH1^Wt glioma stem cells compared with IDH1^Mut glioma stem cells (fold change >2 and P<0.05; Fig. 3A) under hypoxia conditions. In addition, 1,263 mRNAs were downregulated in IDH1^Wt glioma stem cells compared with IDH1^Mut glioma stem cells (fold change >2 and P<0.05; Fig. 3A).

GO analysis on the targeted genes was conducted using DAVID 6.8 (https://david.ncifcrf.gov). Based on GO analysis, ~1,234 differentially expressed genes (|fold change|>4 and P<0.05) were classified (Fig. 3B). GO analysis revealed that specific biological processes were enriched, including ‘DNA replication’, ‘cell division’, ‘cell proliferation’ and the ‘apoptotic process’. In addition to the biological processes, the differentially expressed genes were also enriched in the GO terms associated with ‘cellular component’ and ‘molecular function’, such as ‘protein binding’, ‘DNA binding’, ‘ATP binding’ and ‘nucleoplasm’ (Fig. 3B). KEGG pathway enrichment analyses were also performed. The differentially expressed genes were significantly and predominantly associated with the ‘metabolic pathway’, a significant process in the progression of tumor proliferation and apoptosis. Other enriched pathways involved the ‘cell cycle’, ‘purine metabolism’, the mTOR and p53 signaling pathways, the ‘long-term potentiation’ and the ‘pyrimidine metabolism’ (Fig. 3C).

U87 cells that overexpressed either empty vector or pCMVtag-2B containing IDH1^Wt or IDH1^Mut were used to investigate the effects of IDH1^Wt and IDH1^Mut proteins on glioma cell growth and apoptosis under hypoxia. U87 cells with stable overexpression of IDH1 and IDH1 R132H mutant proteins were successfully established (Fig. 4).

Subsequently, the contribution of the IDH1 mutational status to the growth of U87 cells under hypoxia was explored by the CCK-8 assay. Overexpression of the IDH1 R132H mutant protein significantly suppressed cell proliferation in U87 cells compared with cells transfected with empty plasmid or control parental cells (Fig. 5A). By contrast, overexpression of the IDH1^Wt protein did not influence cell proliferation compared with cells transfected with empty plasmid or control parental cells (Fig. 5A). In addition, overexpression of the IDH1^Mut protein resulted in enhanced apoptosis of U87 cells (Fig. 5B). No differences were noted with regard to cell apoptosis in IDH1^Wt cells compared with the parental control cells (Fig. 5B). These findings suggested that overexpression of IDH1^Mut, but not IDH1^Wt, suppressed cell proliferation and promoted cell apoptosis in GBM cells under hypoxic conditions.

To further explore the mechanism by which the IDH1 R132H mutant protein mediated cell growth inhibition and apoptosis, the expression levels of RLIP76, p53, caspase-9 and Bcl-2 proteins were detected by western blotting. The bioinformatic analysis of the GSE118683 database performed in the present study indicated that the expression levels of the aforementioned genes were significantly altered between IDH1^Mut and IDH1^Wt GBM cells under hypoxic conditions (data not shown). Western blot analysis confirmed that the IDH1 R132H mutant-overexpressing cells exhibited higher protein expression levels of p53 and caspase-9, and lower protein expression levels of Bcl-2 and RLIP76, compared with the IDH1^Wt-overexpressing and the control cells (Fig. 5C). Notably, no difference was observed in the protein expression levels of RLIP76 between the control and IDH1^Mut groups. These results suggested that the antitumor effects of the IDH1 R132H mutant proteins on GBM were associated with an induction of the p53-dependent apoptotic pathway under hypoxia.

U87 cells were transfected with IDH1 siRNA and western blotting results demonstrated that the protein expression levels of IDH1 were successfully reduced (Fig. 5D). Following IDH1 siRNA transfection, U87 cell proliferation was significantly suppressed under hypoxic conditions compared with that of the control cells, as demonstrated by the CCK-8 assay (Fig. 5E). In addition, flow cytometric analysis indicated that the apoptosis rate was significantly increased in U87 cells following IDH1 silencing compared with that of the control cells (Fig. 5F). These results indicated that IDH1 functioned as a tumor oncogene in U87 cells under hypoxic conditions.

Since RLIP76 has a critical role in the prognosis of IDH1^Wt GBM (Fig. 2C), the expression levels of RLIP76, p53, caspase-9 and Bcl-2 were measured in order to investigate their ability to regulate tumor progression under hypoxia. Knockdown of IDH1 resulted in a significant decrease in RLIP76 expression levels without affecting p53 expression in U87 cells (Fig. 5G). This result was not noted in the tissues containing the IDH1 R132H mutant phenotype (Fig. 5C). The data demonstrated that although the IDH1 R132H mutant phenotype exerted similar antitumor effects on GBM with those noted in the presence of IDH1 knockdown, it promoted apoptosis under hypoxia via a distinct mechanism of action. The IDH1 R132H mutant protein promoted p53-induced apoptosis, whereas IDH1 knockdown inhibited the RLIP76-dependent apoptotic pathway. This may partially explain why RLIP76 was a better prognostic indicator in IDH1^Wt GBM compared with IDH1^Mut GBM.

Transfection of parental control, IDH1^Wt or IDH1^Mut U87 cells with RLIP76 siRNA significantly suppressed RLIP76 protein expression (Fig. 6A). RLIP76 knockdown significantly inhibited cell growth and promoted the induction of apoptosis in IDH1^Wt U87 cells, although it did not affect the proliferation and apoptosis of IDH1^Mut U87 cells under hypoxia (Fig. 6B and C). The results suggested that RLIP76 expression contributed to the malignant progression of IDH1^Wt glioma cells under hypoxic conditions.

The present bioinformatic analysis revealed that the differential gene expression noted in IDH1^Mut glioma cells was mainly associated with the metabolic pathway (Fig. 3). In addition, a previous study has reported that RLIP76 has a critical role in the regulation of the metabolic pathway required for GSH detoxification (27). In light of the aforementioned findings, the expression levels of GSH were investigated. GSH is considered the main antioxidative defense mechanism that prevents intracellular damage under hypoxia. IDH1^Mut cells exhibited a dramatic decrease in GSH levels compared with those of the control and IDH1^Wt cells under hypoxia. In addition, knockdown of RLIP76 in IDH1^Wt glioma cells significantly decreased GSH levels. However, GSH levels were not significantly different in IDH1^Mut cells following transfection with either control or RLIP76 siRNA. These results suggested that RLIP76 may influence the metabolic pathway of IDH1^Wt glioma cells by regulating GSH levels in the hypoxic microenvironment.