The Gene Expression Profiling Interactive Analysis (GEPIA, http://gepia.cancer-pku.cn) is an interactive online analysis platform developed by Peking University. The platform contains a large amount of RNA sequencing data from human tumor tissues and mutually matched normal tissues (14). Based on this platform, we analyzed the expression levels of GNG12 in some common tumors and then used box plots to compare the differences in GNG12 expressions between tumor tissues (n=163) and normal brain tissues (n=207). The Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) contains high-throughput gene expression data submitted by research institutions worldwide and provides a range of web-based interfaces and applications (15). To explore GNG12 expression levels in gliomas, we examined microarray data from two datasets: GSE4290 (glioma=77, normal=23) and GSE50161 (glioma=34, normal=13). RNA sequencing data from 1,018 gliomas and their corresponding clinical information were obtained from the Chinese Glioma Genome Atlas (CGGA, http://www.cgga.org.cn) database. From the 1,018 glioma samples, patients with complete relevant clinical information were selected for this study ( Supplementary Table 1 ) (16). Based on the Human Protein Atlas (HPA; https://www.proteinatlas.org/), an immunohistochemical database, the differential expression of GNG12 between glioma and control groups was explored at the protein level (17). The IVY-GAP database (http://glioblastoma.alleninstitute.org/), an immunofluorescence database, was used to explore the differential expression of GNG12 between the glioma and control groups at the nucleic acid level.

From June 2019 to September 2019, tissue samples were collected from 24 patients with glioma and seven patients with epilepsy at Henan Provincial People’s Hospital (Zhengzhou, China). All the samples contained complete clinical information about the patients. Samples were obtained surgically by dividing the tissue into 1 cubic centimeter sizes, and then placing them in liquid nitrogen for freezing and storage at -80°C until total RNA was isolated. Real-time quantitative polymerase chain reaction (RT-qPCR) was used to verify the expression levels of GNG12 in glioma and non-tumor brain tissues. The study protocol was approved by the Ethics Committee of Henan Provincial People’s Hospital (Zhengzhou, China). All experiments were performed according to the guidelines approved by Henan Provincial People’s Hospital.

Human glioma cell (LN229) were purchased from Wuhan Procell Biotechnology (Wuhan, China). Cells were incubated at 37°C in an incubator with a gas environment of 95% O2 and 5% CO2, and the medium used was Dulbecco’s modified eagle medium (DMEM) (Procell, Wuhan, China) containing 10% fetal bovine serum (Gibco, USA) and a 1% penicillin-streptomycin mixture (Procell Wuhan, China); next, the cells were passaged and reserved. Transient transfection was used in this study, and the transfection reagent was Lipo3000 (Thermo Fisher, USA). Cells were evenly inoculated into 6-well plates, and then a mixture of siRNA-Mate (GenePharma, Shanghai, China) and lipo3000 was added to each well separately. Transfection was performed using serum-free medium, and the complete medium was replaced after 6 h of action. The knockdown efficiency was determined via RT-qPCR after the siRNA was functional (approximately 36 h). The dishes with siRNA-NC were added as a blank control (NC), and those with siRNA-1, siRNA-2, and siRNA-3 were added as the experimental group (KD). Sequences with the highest siRNA knockdown efficiency were selected for subsequent experiments (primers and siRNA sequences are listed in Table S2 ).

After transfection with Si-RNA, total RNA was isolated from sample tissues and corresponding cell lines using TRIzol^® (Invitrogen, Thermo Fisher Scientific, USA). Then, the RNA’s concentration was determined using a NanoDrop One spectrophotometer (Thermo Fisher Scientific, USA), and reverse transcription was performed to obtain cDNA (Novoprotein). Finally, the expression level of GNG12 was determined via RT-qPCR using the NovoStart SYBR qPCR SuperMix Plus (Novoprotein). Primers for GAPDH and GNG12 were purchased from Henan Shangya Biotechnology Co. Ltd., with the following sequences: GAPDH-F:5’-CAAGGTCATCCATGACAACTTTG-3’, GAPDH-R:5’-GTCCACCACCCTGTTGCTGTAG-3’, GNG12-F:5’-GAGCCCTTAGAGACCGAG -3’, GNG12-R:5’-AGACTTTGTGTGGTCCAATGT-3’. The thermal cycling conditions were as follows: initial denaturation at 95°C for 10 min, denaturation at 95°C for 10 s, and annealing and extension at 60°C for 30 s for a total of 40 cycles.

Untreated LN229 cells were inoculated in 96-well plates (1000 cells/well), and after waiting for wall attachment and interference with siRNA, the absorbances at 0, 24, 48, 72, and 96 h after transfection were measured. The absorbance at 450 nm was measured using an enzyme marker after incubation in a 37°C incubator for 4 h prior to each measurement.

For immunohistochemical staining (IHC), paraffin sections with a thickness of 4 μm were first dewaxed by placing them in an oven at 55°C for 1 h and then in xylene and concentration gradient ethanol for dewaxing and hydration. Antigen repair was performed in an ethylenediaminetetraacetic acid (EDTA) buffer (pH 8.0) with microwave heating for 18 min. Blocking was then performed using 10% goat serum to reduce nonspecific staining. GNG12 (1:100, Bioss, Beijing, China) primary antibody working solution was added dropwise to the slides and placed in a wet box overnight at 4°C. The following day, the secondary antibody was washed with PBST and incubated for 1 h. The exposed GNG12 protein was then labeled with 0.01% DAB chromogenic solution and the nuclei were stained with hematoxylin. Finally, the staining results were observed under a light microscope at 400x magnification, and five fields of view were selected for photography. The IHC results were processed using ImageProPlus (version 6.0). To verify the effect of GNG12 on cell proliferation, we performed cellular immunofluorescence analysis using Ki67. An equal number of LN229 glioma cells were first inoculated uniformly in 3 cm culture dishes, and the culture medium was discarded after 36 h of transfection, fixed with 4% paraformaldehyde for 10 min at room temperature, and then permeabilized with 0.5% Triton X-100 for 30 min. After washing thrice with PBS, the Ki-67 primary antibody (1:200, Abcam, China) was incubated at 4°C for 24 h. Following this incubation, the cells were washed thrice with PBS for 5 min each, and then incubated with the DyLight 594Ig G (1:200, Invitrogen, USA) secondary antibody at room temperature and protected from light for 1 h. Subsequently, the nuclei were stained with DAPI for 10 min. Finally, the Ki67 expression levels in the experimental and control groups were observed by fluorescence microscopy and photographed.

Experiments were performed to verify whether GNG had any effect on glioma cell migration. Equal amounts of cells were uniformly inoculated into 6-well plates. The experimental and control groups were set, and three wells were used for parallel experiments. After the cells reached 70% confluency, the experimental group was transfected with siRNA-GNG12, and the control group was transfected with siRNA-NC. Once the cells in the culture dish were fully grown, three parallel vertical lines were drawn in each of the six wells with a 200 μL sterile spiking gun tip, and the detached cells were washed with PBS. A 2 mL sample of serum-free medium was added to each well; next, a 0 h sample was taken with an inverted microscope at 200x magnification as the first experimental data, and then placed in a 37°C incubator for 48 h. Scratches were then taken at the same position as the previous one.

Gene Set Enrichment Analysis(GSEA) is an ideal bioinformatics analysis tool developed by the research team of MIT and Harvard University’s Broad Institute and is used to analyze cell signaling pathways. The RNA-sequencing data obtained from the CGGA database were batch corrected and normalized using SVA and LIMMA, and then divided into the “H” (high expression) or “L” (low expression) groups according to GNG12 expression levels. The GSEA (v.4.0.3) software was used for enrichment analysis; the number of permutations was set to 1000, and the “KEGG cell signaling pathway” was selected as the gene set database.

The transfected cells were added to RIPA lysate and protease inhibitor to extract total protein (EpiZyme, Shanghai, China). After lysis on ice for 30 min, the proteins were centrifuged at 12000 rpm for 15 min at 4°C. The loading volume of each sample was measured using a BCA kit (GenStar, Beijing, China). Identical masses of proteins were separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Subsequently, the proteins were transferred to PVDF membranes (Bio-Rad, UK) and sealed with skim milk powder. PVDF membranes were incubated overnight at 4°C with primary antibodies against VCAM-1 (1:500, Proteintech, USA), ICAM-1 (1:2500, Proteintech, USA), CDH2 (1:1500, Proteintech, USA), and GAPDH (1:10000, Proteintech, USA). Finally, the HRP-labeled goat anti-mouse secondary antibody (1:5000, Proteintech, USA) was incubated with the filter membrane for 1 h. GNG12 protein expression levels were detected using an imager and a chemiluminescent substrate (ECL) kit (Thermo Fisher Scientific, USA).

Gene co-expression is an analytical method that uses a large amount of gene expression data to construct correlations between genes and gene functions. The co-expression analysis of GNG12 was performed using Pearson’s method. According to the correlation coefficients, the P-value obtained ten genes were positively and negatively correlated with GNG12. Based on the positive and negative related top 10 genes, the complex algorithm of CMap (Connectivity Map, https://portals.broadinstitute.org/cmap) was used to obtain related genes that may downregulate the expression level of GNG12 and screen GNG12 gene therapy drugs in the PubChem database (https://pubchem.ncbi.nlm.nih.gov); this information included the name and chemical formula of the drug and 2D and 3D structures.

The Tumor Immune Estimation Resource (TIMER, https://cistrome.shinyapps.io/timer) is a rich database of tumor immunology and genetics, including gene expression, mutation, and copy number variation (18). In this study, we evaluated the association of GNG12 expression with the infiltration of six different immune cell types (B cells, CD4 + T cells, CD8 + T cells, macrophages, neutrophils, and dendritic cells).

R software (v.3.6.1) was used to perform the statistical data analysis. Survival and clinical characteristic data were obtained from the CGGA database. The overall survival of GNG12 was determined via COX regression and the Kaplan-Meier(KM) method. In addition, Wilcox or Kruskal tests were used to test the relationship between the expression level of GNG12 and clinical characteristics. A time-dependent receiver operating characteristic (ROC) curve shows the evaluation value of the clinical prognosis of GNG12 for glioma. Univariate and multivariate analyses were used to analyze the factors affecting the prognosis of glioma patients. The expression differences between the experimental and control groups of GNG12 were tested using GraphPad Prism software (9.1.0) with a Mann-Whitney test, Chi-square test, or Fisher’s exact test (P<0.05, considered statistically significant).

GEPIA was used to analyze the expression of GNG12 in different tumors. GNG12 was highly expressed in various tumors, including glioblastoma multiforme (GBM), lymphoid neoplasm diffuse large B-cell lymphoma (DLBC), pancreatic adenocarcinoma (PAAD), and thymoma (THYM) ( Figure 1A, B ). The GSE4290 and GSE50161 datasets from the GEO database revealed that the expression level of GNG12 was higher in gliomas than in normal brain tissue ( Figures 1C, D ). Moreover, GNG12 protein expression levels were similarly elevated in HPA immunohistochemistry and IVY-GAP in situ hybridization data ( Figures S1 and S2 , respectively). More importantly, we obtained the same results as those predicted by the database through further RT-qPCR and IHC experimental validations ( Figures 1E , F ). Combined with the above database analysis and experimental results, GNG12 was overexpressed in glioma.

Subsequently, the relationship between GNG12 expression levels and patients clinical characteristics with glioma was analyzed. Wilcoxon and Kruskal-Wallis tests were used to analyze the relationship between clinically relevant patient information and the expression levels of GNG12. We observed that the expression level of GNG12 significantly correlated with age, WHO classification, histological type, primary recurrence classification, 1p19q coding data, and IDH mutation status. As shown in Figure 2 , the expression of GNG12 in glioma tissues from patients aged >41 years was significantly higher than that in patients aged ≤ 41 years (p =0.009). The expression level of GNG12 was positively correlated with the WHO grade of glioma (P < 0.001). Regarding the 1p19q co-deletion status, there were lower gene expression levels in patients with a codeletion of 1p19q than in patients with non-codeletion of 1p19q (p < 0.001). Moreover, for the IDH mutation status, there was a higher level of gene expression in the wild type than in the mutant (p < 0.001). In addition, recurrent gliomas had higher GNG12 expression levels than primary gliomas, which explains why patients with recurrent gliomas had a worse prognosis (p<0.05). The strong association between GNG12 expression levels and the clinical characteristics of glioma patients suggests that GNG12 may be related to the survival prognosis of glioma.

We further explored the relationship between GNG12 expression levels and the survival prognoses of patients with glioma. First, the survival analysis results showed a significant correlation between high GNG12 expression and reduced survival rates of patients with glioma. Subsequently, the survival rates of the different molecular subtypes in each grade were analyzed separately. In grade II and III gliomas and in all samples, the results showed that the survival rate of the GNG12 high expression group was lower than that of the low expression group, regardless of the presence of IDH mutations or 1p/19q co-deletions ( Figures 3A –C ). In grade IV gliomas, the results were not statistically significant, which probably resulted from the poor prognoses of high-grade gliomas and the associated lower survival rate of the patients ( Figure 3D ). However, we nonetheless observed an overall trend that was consistent with the results of the analysis of grade II and III gliomas. Moreover, ROC analysis showed that the area under the ROC curve (AUC) was 0.701, 0.766, and 0.803 for the one, three, and five-year Overall Survival (OS), respectively ( Figure 3E ). The AUC data were meaningful for different WHO classifications ( Figures 3F –H ). Therefore, our results indicate that GNG12 may serve as a biomarker for glioma, especially in the five-year OS group. Univariate and multivariate analyses showed that high levels of GNG12 expression, PRS grading (p < 0.001; HR = 2.032; 95% CI, 1.724-2.393), WHO grading (p < 0.001; HR = 2. 623; 95% CI, 1.918-3.586) and age (p < 0.003; HR = 1.351. 95% CI, 1.105-1.652) may represent an independent risk factor for poor prognoses in patients with glioma. IDH mutation status (p < 0.019; HR = 0.750; 95% CI, 0.590-0.953) and 1p/19q co-deletion (p < 0.005; HR = 0.596; 95% CI, 0.415-0.856) may represent protective factors ( Figures 3I, J , respectively). These data indicate that GNG12 may serve as a predictive biomarker for poor prognosis.

The correlation between GNG12 expression levels and tumor immune cell infiltration was analyzed using the TIMER database. After adjusting for purity, the expression levels of GNG12 in GBM and LGG correlated positively with the degree of infiltration of CD4⁺ T cells (P = 2.71e-02, GBM; P = 1.42e-28, LGG), dendritic cells (P = 2.57e-12, GBM; P = 3.99e-49, LGG), and neutrophils (P = 5.96e-03, GBM; P = 6.70 e-46, LGG) ( Figure 4 ). In addition, GNG12 expression was associated with the degree of infiltration of B cells (P = 3.43e-40, LGG), CD8⁺ T cells (P = 1.09e-16, LGG), and macrophages (P = 1.99e-37, LGG), but no significant differences were found in GBM. In glioma patients, a loss of the GNG12 copy number resulted in decreased infiltration of CD4⁺ T cells, macrophages, neutrophils, and dendritic cells ( Figure S5 ). In addition, the infiltration of B cells and CD8⁺ T cells also decreased in the LGG. Therefore, we further used Kaplan-Meier curves to verify our hypothesis ( Figure S4 ), and the results indicated that high GNG12 expression level in enriched B cells, CD8⁺ T cells, CD4⁺ T cells, macrophages, neutrophils, and dendritic cells were associated with worse prognoses in patients with LGG, whereas high GNG12 expression level in enriched dendritic cells were associated with poor overall survival outcomes. In addition, owing to the broad prospects of immunotherapy, we further determined the relationship between the GNG12 expression and PD-1, PD-L1, and PD-L2 expressions. Encouragingly, GNG12 was positively correlated with PD-L1 (r = 0.293, P = 5.23e-04) and PD-L2 (r = 0.22, P = 9.73e-03) in GBM. We also found that GNG12 was positively correlated with PD-1 (r = 0.347, P = 5.76e-15), PD-L1 (r = 0.474, P = 4.39e-28), and PD-L2 (r = 0.684, P = 4.01e-67) in LGG ( Figure S4 ). In conclusion, it is possible that a high GNG12 expression leads to reduced immune cell infiltration and may be an important factor contributing to poor prognoses in patients with glioma.

To explore the related genes that could have positive or negative regulatory effects on GNG12, we performed a co-expression analysis. The genes that positively regulated GNG12 were SH3GLB1, MSN, CD58, VIM, CRYZ, FPGT, HS2ST1, CTBS, ANXA1, and CNN3. The negatively regulated genes were RASL10A, ACTL6B, HRH3, CXXC11, RP5-1119A7.17, FBXW4, AC062021.1, KCNJ11, RP1-293L6.1, and AMER3 ( Figures 5A , B ). We also screened for gene therapy drugs for GNG12 in the CMap and PubChem databases. Four possible gene therapy drugs were identified for GNG12: anisomycin, chloroquine, levodopa, luteolin ( Figures 5C –F ).

To further validate the effect of GNG12 on glioma, we performed a series of in vitro experiments. First, we designed three small molecule-interfering RNAs to inhibit the expression of GNG12 in glioma cells. The results showed that siRNA-1 was screened with the highest knockdown efficiency using RT-qPCR; therefore, we selected siRNA-1 for the subsequent target downregulation of GNG12 ( Figure 6A ). Knocking down the expression level of GNG12 clearly affected the proliferation and migration abilities of glioma cells. The results of the CCK-8 assay showed that the proliferation efficiency of glioma cells in the KD group was significantly lower than that in the NC group ( Figure 6B ). In parallel, the Ki-67 immunofluorescence assay suggested that the Ki-67 expression level was higher in the NC group than in the KD group, and the differences between the groups were statistically significant ( Figures 6C, D , respectively). The diminished migratory capacity of glioma cells within 48 h of GNG12 downregulation was ultimately verified using a cell scratch healing assay ( Figures 6E , F ). In summary, our targeted downregulation of GNG12 expression levels in glioma cell lines resulted in a certain degree of diminished proliferation and migration; therefore, we speculate that a high GNG12 expression may be an important factor in the poor prognoses of glioma patients.

We performed a GSEA to further explore the potential mechanisms through which GNG12 affects the malignant biological behavior of glioma cells. The results showed that GNG12 was enriched in tumor-related pathways, including the cell adhesion molecule signaling pathway, JAK-STAT signaling pathway ( Figure 7A ), TOLL-LIKE receptor signaling pathway, focal adhesion, VEGF signaling pathway, and MAPK signaling pathway ( Figure S6 ). These cellular signaling pathways showed significant different enrichment rates in samples from patients showing the GNG12 high-expression phenotype based on NES, NOM P-values, and FDR values ( Table 1 ), thus indicating a potential role for GNG12 in developing glioma. To probe the specific mechanism by which GNG12 leads to poor prognosis of glioma, we verified the predicted results of GSEA in WB experiments. After the GNG12 expression level was knocked down and the NC group was used for comparison, the assay revealed that the protein expression levels of VCAM-1 and CDH2 in the cell adhesion molecule pathway were significantly decreased in the KD group, and the ICAM-1 protein expression level was also decreased to an extent ( Figure 7B ). This suggests that GNG12 may help regulate the cell adhesion molecule pathway involved in the malignant process of gliomas.