First of all, we are on the Gene Expression Profiling Interactive Analysis (GEPIA) (19) website (http://gepia.Cancer-pku.cn/index.html) to analyze the differential expression of ITGB2mRNA in GBM and LGG tissues and normal tissues, and the relationship between the expression of ITGB mRNA and Overall Survival in patients with GBM and LGG. Then, we download the clinical and RNA-seq expression data from the mRNAseq_325 dataset in the CGGA database (http://www.cgga.org.cn/) for follow-up analysis. Download clinical and RNA-seq expression data from the Rembrandt database on the gliovis online website (http://gliovis.Bioinfo.cnio.es/). Clinical data, transcription, and methylation of GBMLGG patients were obtained from the TCGA database through the biological portal (20) (https://www.cbioportal.org/). The inclusion criteria were (1) patients with WHO grade II-IV and (2) patients with complete clinical and transcriptional data (CGGA:324, TCGA:607, Rembrandt: 139) were included in this analysis.

We use TIMER, a comprehensive website (https://cistrome.shiny-apps.io/timer/) (21), to automatically analyze and visualize the association between immune infiltration levels and a range of variables. The correlation of ITGB2 expression in six kinds of immune cells (CD4+T cells, CD8+T cells, B cells, neutrophils, dendritic cells, and macrophages) in GBM and LGG was analyzed. Using TISIDB (an integrated repository portal for tumor-immune system interactions), a database of tumor immunity database included 4176 records from 2530 publications and recorded 988 genes involved in anti-tumor immunity. The spearman correlation between ITGB2 and lymphocyte Immunostimulator and Immunoinhibitor related molecules in GBM and LGG was analyzed.

The genes significantly related to ITGB2 expression were retrieved by using Pearson correlation analysis. The associated gene set was analyzed by using the GenecoDis website (22) (https://genecodis.genyo.es/). Besides, the hot map package of R Language is used to list genes related to ITGB2 expression positive height.

Formalin-fixed samples were embedded in paraffin and sectioned at a thickness of 3 μm. The slides were deparaffinized and rehydrated, then incubated with a ITGB2 antibody (Sigma-Aldrich, HPA008877, 1:200). The mean density(integrated optical density(IOD) SUM/area) was calculated in five randomly selected fields using Image-Pro Plus 6.0 software.

According to the median of ITGB2 mRNA expression in different data sets, the high expression and low expression groups of ITGB2 were established. Similarly, based on the median of ITGB2DNA methylation in TCGA-GBMLGG data sets, two ITGB2 hypomethylation and hypermethylation groups are established. Kaplan-Meier curve was used to evaluate the prognostic value of ITGB2 expression and ITGB2DNA methylation. The correlation between ITGB2 expression and ITGB2DNA methylation level was analyzed by Pearson correlation coefficient. In addition, we used univariate and multivariate Cox regression models to explore whether ITGB2 expression is an independent prognostic indicator for patients with LGG. The prediction performance of ITGB2 in the mesenchymal molecular subtype was evaluated by ROC curve analysis. Student t-test was conducted to explore the distribution of expression in different groups. Pearson correlation analysis was used to identify genes related to ITGB2 expression. R language packs (dplyr, stringr, survival, survminer, plyr, pheatmap, proc, and corcrac) are used for other statistical calculations and drawing data. All the differences were statistically significant at the level of P < 0.05.

First of all, the clinical and RNA-seq expression profile data with DataSet ID as mRNAseq_325 were downloaded from the CGGA database. The phenotypic data were downloaded from the gliovis (http://gliovis.bioinfo.cnio.es/) database. Figure 1A showed that the expression of ITGB2 mRNA was positively correlated with tumor grade. In addition, the expression of ITGB2 in the isocitrate dehydrogenase 1 wild type (IDH1 Wt) group was higher than that in the IDH1 mutant (IDH1 Mut) group. Figure 1B ). These findings were subsequently validated in TCGA GBMLGG’s RNA-seq dataset and Rembrant’s Microarray dataset ( Figures 1C–E ). The results showed that ITGB2 mRNA expression has a high correlation with tumor grade and IDH1 mutation in the CGGA database (p < 0.05). Subsequently, we randomly selected five patients with different grades of glioma to perform immunohistochemical experiments, and all patients voluntarily signed the informed consent. The results of immunohistochemistry (IHC) showed that the expression of ITGB2 protein (mean density, IOD SUM/area) was higher in high-grade gliomas (P<0.05), and there were statistically significant differences among different grades of gliomas ( Figures 2A, B ).

In 2010, Verhaak’s team (23) divided glioblastoma into four subtypes using TCGA data sets. Through genomic mutation, copy number variation, and expression profile data, it was proved that epidermal growth factor receptor (EGFR), neurofibromin 1(NF1), and platelet-derived growth factor receptor alpha (PDGFRA)/isocitrate dehydrogenase 1 (IDH1) could be used as markers of classical, interstitial and proneuronal types, respectively. It provides a basis for targeted therapy of gliomas. So we evaluated the expression patterns of ITGB2 in different molecular subtypes. Results showed that the ITGB2 expression of mesenchymal subtypes was significantly higher than other subtypes in the RNA-seq set of CGGA and TCGA-GBMLGG and the Microarray data set Rembrandt ( Figures 3A, C, E ). We performed ITGB2 expression and receiver operating characteristic curve (ROC) analysis of mesenchymal subtypes in all grades of gliomas to verify this finding further. What is exciting is that in these three datasets, the area under the curves (AUC) are 0.948, 0.862, and 0.848, respectively ( Figures 3B, D, F ).

We used the GEPIA database to analyze the RNA sequencing data of 681 cases of TCGA tissues (518 LGG tissues and 163 GBM tissues) and 207 normal tissues of the GTEx plan. It was found that ITGB2 mRNA was highly expressed in both GBM and LGG tissues, while low expression was found in normal tissues ( Figure 4A ). Then, We used the median expression level of ITGB2 mRNA as the cutoff point to evaluate the relationship between the level of ITGB2 mRNA expression and survival time in different data sets. It was found that the patients with high expression of ITGB2 in CGGA RNA-seq had lower OS in all grades of gliomas ( Figure 4B P < 0.0001). Similar results were also obtained in TCGA-GBMLGG RNA-seq set and Rembrandt Microarray data set ( Figures 4C, D ; for all grades of gliomas, p < 0.0001). Then, univariate and multivariate Cox regression analysis was performed in CGGA RNA-seq to verify whether the expression of ITGB2 was an independent prognostic factor ( Table 1 ). Univariate regression analysis showed that ITGB2 (p < 0.001), grade (p < 0.0001), age (p < 0.0001), and IDH1 status (p < 0.0001) were each associated with OS. In multivariate regression, ITGB2(p < 0.0001)and grade(p < 0.0001)showed significant results. These results suggest that ITGB2 plays an important role in the occurrence and development of gliomas. Subsequently, biological function analysis should be carried out to verify our findings further.

In order to study the relationship between ITGB2 and other ITGB molecular families, we carried out Pearson correlation analysis on the CGGA RNA-seq set. It can be seen from Figure 4A that the expression of ITGB2 was significantly correlated with ITGB1, ITGB3BP, ITGB4, ITGB7, and ITGB8. To explore the biological process related to the expression of ITGB2 in gliomas, we conducted Pearson related analysis between ITGB2 expression and other genes in whole-genome gene profiling of 325 patients in the CGGA RNA-seq set. The results showed that 658 genes (R > 0. 6) were positively correlated with ITGB2 expression. Among them, 140 genes (R > 0. 8) were highly positively correlated with the expression of ITGB2 ( Figure 5A ). The Biological Process (BP) and KEGG analysis of the biological process of these 140 genes was carried out by using the GeneCodis website (https://genecodis.genyo.es). Results have been shown: The five most enriched biological process annotations were (1) Immune system process (p = 2.615e-52); (2) Neutrophil degranulation (p = 2.774e-36); (3) Innate immune response (p = 5.923e-34); (4) Inflammatory response (p = 7.096e-23); (5) Signal transduction (p = 1.743e-22), containing 44、33、32、23 and 35 genes from the query set, respectively. KEGG analysis includes Osteoclast differentiation, Staphylococcus aureus infection, Tuberculosis, Phagosome, and Neutrophil extracellular trap formation. In summary, all above results suggest that ITGB2 can affect glioma-related immune activity.

The TIMER algorithm (https://cistrome.shinyapps.io/timer/) (21) was used to determine the relationship between the expression of ITGB2 and immune cell infiltration. According to the LGG and GBM cohort of TCGA, the expression of ITGB2 in LGG was negatively correlated with tumor purity (r = 0.369, P < 0.05). The expression of ITGB2 was positively correlated with B cells (r = 0.700, P < 0.05), CD4+T cells (r = 0.921, P < 0.05), macrophages (r = 0.820, P < 0.05), neutrophils (r = 0.836, P < 0.05) and dendritic cells (r = 0.925, P < 0.05). The expression of ITGB2 was weakly positively correlated with that of CD8+T cells (r = 0.137, P < 0.05). The expression of ITGB2 in GBM was negatively correlated with tumor purity (r = 0.559, P < 0.05) and CD8+T cells (r = 0.414, P < 0.05). The expression of ITGB2 was positively correlated with CD4+T cells (r = 0.435, P < 0.05), neutrophils (r = 0.390, P < 0.05) and dendritic cells (r = 0.562, P < 0.05). It was weakly positively correlated with B cells (r = 0.234, P < 0.05) and macrophages (r = 0.215, P < 0.05) ( Figure 6 ). It is suggested that ITGB2 is closely related to immune cell infiltration in both LGG and GBM.

In order to further explore whether the methylation CpG site of ITGB2 also has clinical prognostic value. Using cBioPortal online database (http://www.cbioportal.org/), We found that there were significant differences in ITGB2 methylation in IDH WT and mutation gliomas (P < 0.0001, Figure S1A ), The mRNA expression of ITGB2 was also significantly different in MGMT methylated and unmethylated gliomas(P < 0.0001, Figure S1B ). Then, We evaluated whether ITGB2 methylation has clinical and prognostic value. We found 45 CpG sites of ITGB2 methylation, and there was a significant negative correlation between the expression of ITGB2 mRNA and ITGB2 methylation (r = 0.66, P < 0.0001, Figures 7A, B ). Through univariate and multivariate cox regression analysis of 45 ITGB2 methylation sites, nine methylation sites may be independent risk prognostic factors were selected ( Figure S2 ). Then we analyzed the prognostic value of these nine ITGB2 methylation CpG sites in patients with TCGA LGG. The Kaplan-Meier diagram shows that the high methylation level at the selected CpG site is associated with a more favorable OS in LGG patients. In order to better predict the prognosis of LGG patients, we used the R language predict function to obtain the risk score of these nine ITGB2 methylation sites(ITMRS). According to the median, we divided the ITMRS into two groups: high and low. Figure 4C shows that ITMRS low can be used as an independent risk prognostic factor. Then we analyzed the prognostic value of ITMRS in patients with TCGA LGG. Excitedly, ITMRS has a better prognosis in LGG patients than a single ITGB2 methylation CpG site predicts OS in LGG patients ( Figure 7D ). At the same time, we obtain the analysis of receiver operating characteristic curve(ROC) of the three-year, five-year, and ten-year survival predicted by ITMRS, and the area under the curve (AUC) was 0.884, 0.767 and 0.704, respectively( Figure 7E ). The Kaplan-Meier diagram and ROC curves show that the low level of ITMRS is associated with a more favorable OS in LGG patients.