We utilized UCSC XENA (https://xenabrowser.net/datapages/) to analyze processed RNA-seq data from TCGA (https://portal.gdc.cancer.gov) and Genotype-Tissue Expression (GTEx) database (https://www.gtexportal.org/home/), which were uniformly processed using Toil methods, for comparing the mRNA levels of TMEM115 in glioma (n = 689) and non-tumor tissues (n = 1157). Additionally, clinical data were retrieved to investigate whether TMEM115 expression is linked to glioma patients’ Overall Survival (OS) and Disease-Free Survival (DFS). In addition, we extracted data and corresponding clinical information from the GBM and LGG datasets of the TCGA database as well as from the mRNAseq_325 (CGGA 325) and mRNAseq_693 (CGGA 693) datasets of the CGGA database (http://www.cgga.org.cn) for clinical characterization.

TMAs comprising 214 tissue samples were collected from the Nantong University Affiliated Hospital between 2013 and 2017 (16). These samples included 189 glioma tissues and 25 non-tumor brain tissues from individuals with benign brain diseases. Patient data, including gender, age, histological classification, molecular classification, WHO grading, and whether they had undergone radiotherapy or chemotherapy, were collected according to the 2021 WHO Classification of Tumors of the Central Nervous System (20). This study was approved by the ethics committee of our hospital (Grant No. 2018-K020).

The Opal 7-color manual IHC kit (NEL797B001KT; PerkinElmer, MA, US) was used for mIHC staining as previously described (15, 21). Briefly, tissue slices were dewaxed and hydrated, followed by heat-induced antigen retrieval using an AR6 solution (AR600, PerkinElmer) in a microwave oven. Next, the slides were incubated with indicated primary antibodies, followed by the polymer HRP-conjugated secondary antibody (ARH100EA, AKOYA). Fluorescent dyes were added for signal visualization, and multiple staining was accomplished by repeating the process with microwave heating treatments performed between each step. Lastly, FluoroshieldTM with DAPI (F6057, Sigma, NY, USA) was utilized to stain the nucleus for visualization. In this experiment, the following antibodies were used: anti-TMEM115 antibody (1:5000, NBP1-80898, Novus, USA), anti-CD68 antibody (1:1500, 76437S, Cell Signaling Technology, USA), anti-CD163 antibody (1:200, 93498S, Cell Signaling Technology, USA), anti-CD86 antibody (1:500, 91882S, Cell Signaling Technology, USA), anti-PD-L1 antibody (1:500, 13684S, Cell Signaling Technology, USA).

We utilized an automated imaging system (Vectra 3.0, Perkin Elmer/Akoya, USA) to scan and quantify all samples, followed by analysis and scoring using InForm software (Akoya, USA). The fluorescence intensity threshold of each marker used in the analysis was assessed to confirm the presence of a specific cell phenotype. The final score (range: 0-100) was determined based on the ratio of positively stained cells to nuclei times 100%.

Cell lines (U87MG, U251, and HEK293T) were obtained from the Cell Bank/Stem Cell Bank at the Chinese Academy of Sciences (Shanghai, China). They were cultured in a complete growth medium consisting of the appropriate base medium (Gibco, Grand Island, NY, USA), supplemented with 10% fetal bovine serum (Shuangru, Jiangsu, China), and 1% penicillin-streptomycin (NCM Biotech, Jiangsu, China). Additionally, U251 required the inclusion of Non-Essential Amino Acids Solution (C0332; Beyotime) and Sodium Pyruvate (C0331; Beyotime). The cells were incubated at 37°C with 5% carbon dioxide in an incubator.

On Merck’s website (https://www.sigmaaldrich.cn/CN/zh/semi-configurators/shrna?activeLink=productSearch), designing two different short hairpin RNA (shRNA) sequences and a targeting TMEM115 control sequence (shRNA negative control). They were synthesized by the company of Tsingke Biotechnology (Jiangsu, China). The shRNA forward sequences targeting TMEM115 were shRNA (NC): (TTCTCCGAACGTGTCACGT) shRNA (TMEM115 1^#) (GCCCTCATCTGAGGTAAGAAT) and shRNA (TMEM115 2^#) (GAAGGTAAAGATATGCCAGAA). After annealing and purification, the fragments were integrated into the pLV-shRNA vector, reseparately. This vector was co-transfected with pMD2.G and psPAX2 plasmids into HEK293T cells to generate recombinant lentiviral particles. After 72 hours, the supernatant containing lentiviral particles was collected and infected with glioma cell lines. After appropriate culture screening, target cells were collected and analyzed by Western blotting to verify transfection efficiency, followed by cell proliferation, migration, and invasion assays.

Total proteins from glioma cell samples were extracted using RIPA lysis buffer (p0013B; Beyotime) with protease and phosphatase inhibitors (P1260; Solarbio, Beijing, China). Equal amounts of proteins were added to a 10% SDS-PAGE gel (PG112; Epizyme Biotech, Shanghai, China) for electrophoretic separation and transferred to a 0.45µm PVDF membrane (Millipore, USA). Afterward, the membrane was blocked with 5% skim milk and incubated overnight with primary antibodies such as TMEM115 antibody (1:1000, 25536-1-AP, protein tech) or β-Actin antibody (1:5000, AC026, ABclonal). The next day, it was detected again with Peroxidase-conjugated AffiniPure Goat Anti-Rabbit IgG (H+L) antibody (1:10000, 111-035-003, Jackson ImmunoResearch). Protein bands were visualized using an ECL kit (BL520B; Biosharp, Beijing, China) and captured by a chemiluminescence imaging system (Tanon 4600SF, Nanjing, China). Finally, optical density was calculated using Image J software to determine the TMEM115 protein expression level.

2× 10³ U87MG and U251 cells were seeded into 96-well plates, and after the cells were attached to the wall, the medium was discarded, and the cells were treated according to the instructions of Cell Counting Kit-8 (CCK-8) (C6050; NCM Biotech, Suzhou, China) reagent, and the absorbance was detected after incubating the cells at 37°C for 2h to reflect the cell proliferation. After cell seeding, the same steps were repeated 24, 48, 72, and 96h.

The migration and invasive ability of glioma cells were evaluated and analyzed by Transwell assay. U87MG and U251 glioma cells (5 × 10⁴ cells) were placed into the chambers without Matrigel coating for migration analysis and to simulate an invasive environment in the chambers coated with Matrigel for invasion analysis. In the upper chamber, medium without FBS was added. In the lower chamber, a medium containing 10% FBS was added. After 24 hours of incubation, migrating or invading cells were fixed with 4% paraformaldehyde solution and stained with crystal violet. Finally, quantitative analysis.

To assess the correlation between TMEM115 expression and tumor-infiltrating immune cells, we used a CIBERSORT-based core algorithm using markers for 22 immune cells provided on the CIBERSORT website (CIBERSORTx, http://cibersort.stanford.edu/) to calculate TMEM115 expression and immune infiltration of tumor-infiltrating immune cells.

T-test and Cox regression were performed to compare the mRNA expression and prognosis of TMEM115 in glioma and non-tumor brain tissues. Experiments involving Western blot analysis and cell function were repeated independently three times. Comparisons between two groups were performed using unpaired or paired two-tailed t-tests, and the significance of differences between more than two groups was analyzed by Two-way ANOVA. The optimal cut-off point for TMEM115 protein expression was determined based on patients’ OS using X-tile (Yale University, USA) (22). Pearson’s test was used to determine TMEM115’s association with immune markers. TMEM115 relationship with clinicopathologic features was assessed using χ² test test. The Cox proportional hazard model was also utilized to identify prognostic factors for TMEM115. Statistical analyses and data visualization were conducted using SPSS (Version: 22.0, IBM, USA) and GraphPad Prism software version 9.0.0 (GraphPad Software, La Jolla, CA), and significance was set at P < 0.05.

We first assessed whether TMEM115 is a marker in glioma. The results showed a significant increase in TMEM115 mRNA expression in glioma tissues compared to nonmalignant brain tissues (P < 0.05) ( Figure 1A ). To further determine the potential of TMEM115 as a prognostic biomarker for gliomas, correlation analysis with OS and DFS was performed. In glioma patients, elevated TMEM115 expression was strongly correlated with adverse prognosis (HR = 1.9, P < 0.05; HR = 1.6, P < 0.05) ( Figures 1B, C ). Furthermore, given the new definition of adult gliomas in the 2021 WHO classification, we investigated the expression of TMEM115 in IDH mutation astrocytoma, IDH mutation, and 1p19q codeletion oligodendroglioma, and IDH wild-type glioblastoma (P < 0.05) ( Figure 1D ). The significant upregulation of TMEM115 expression in IDH wild-type glioblastoma. In addition, differences in TMEM115 between other clinical variable groupings (IDH status, WHO grade) were directly analyzed based on the TCGA database (P < 0.05) ( Figures 1E, F ). These results were also validated in the CGGA database ( Figures 1G-I ). These suggest the clinical significance of TMEM115 and its potential utility as a prognostic biomarker for glioma patients.

Given the limitations of predicting protein expression only based on transcription levels (16), we conducted mIHC staining of glioma TMA sections to assess TMEM115 protein expression and localization. TMEM115 protein was found to be primarily localized in glioma cells plasma membrane ( Figure 2A ). Consistent with mRNA levels, elevated levels of TMEM115 were observed in tumor tissues ( Figures 2A, B ).

The optimal cut-off value for TMEM115 protein expression in gliomas was determined based on the OS of glioma patients and categorized into high and low-expression groups. TMEM115 protein levels were classified as a low and high-expression group based on a threshold of 64. Chi-square test results revealed a significant correlation between TMEM115 protein expression in tumor cells and WHO grading (χ² = 17.134, P < 0.001), but no significant associations were found with gender, age, histological classification, molecular typing, or the presence of radiotherapy and chemotherapy ( Table 1 ).

In the univariate analysis, TMEM115 protein expression (HR = 2.450, P < 0.001), age (HR = 2.261, P < 0.001), gender (HR = 1.588, P = 0.008), histologic classification (HR = 0.796, P = 0.024) and grade (HR = 2.462, P < 0.001) were found to significantly impact patient survival. In the multivariate analysis, TMEM115 protein expression (HR = 1.840, P = 0.001), age (HR = 1.522, P = 0.019), and grade (HR = 2.066, P < 0.001) were independently associated with 5-year OS in glioma patients ( Table 2 ). Survival analyses demonstrated that high TMEM115 levels led to poorer OS in glioma patients (HR = 2.51, P < 0.001) ( Figure 2C ). Collectively, these findings support TMEM115 as a potential indicator for glioma patient survival.

To elucidate the effect of TMEM115 on glioma phenotype and to determine its oncogenic role, we designed and established shRNAs and their respective lentiviral systems. The U87MG and U251 cell lines, which are relatively malignant, were selected to be infected with shRNAs, firstly, using lentiviral infection to generate stable knockdown cell lines, thus promoting the knockdown of TMEM115. Knockdown of TMEM115 significantly reduced the protein expression of glioma cells compared to those infected with negative controls ( Figures 3A, B ) The results of the CCK-8 assay showed that the proliferative capacity of U87MG and U251 cells was significantly decreased with the reduction of TMEM115 levels ( Figures 3C, D ) Subsequently, using Transwell assay assessment, after TMEM115 downregulation, the migration and invasion abilities of U87MG and U251 glioma cells were inhibited ( Figures 3E-J ). Thus, the effects of TMEM115 on glioma proliferation, migration, and invasion were demonstrated.

Considering that TME plays a crucial role in determining the response of gliomas to immunotherapy, we extracted RNA-seq data from gliomas in TCGA. We evaluated the correlation between TMEM115 expression and TIICs. The results showed that TMEM115 was positively correlated with M2 macrophages ( Figure 4A ). Additionally, we confirmed this result by incorporating clinical information from the GTEx databases based on the TCGA database. Since immune checkpoints can fuel tumor immunotherapy, we also evaluated the correlation of TMEM115 with the common immune checkpoints, the results showed that TMEM115 was correlations with CD274 (PD-L1) (R = 0.15, P = 2.7e-11) ( Figures 4B-D ). To verify this, we performed mIHC staining and analyzed the expression of TMEM115 protein in the glioma TME using Pearson’s test ( Figure 4E ). We found that TMEM115 protein levels in tumor tissues were positively associated with CD68⁺ (R = 0.283, P < 0.001) and CD68⁺CD163⁺ (R = 0.145, P = 0.046) macrophages ( Figures 4F, G ). And interestingly, protein expression also correlated with the immune checkpoint PD-L1 (R = 0.247, P = 0.001) ( Figure 4I ).