RBBP8 expression was significantly elevated in glioma
Results of the RT-PCR showed that the mRNA expression level of RBBP8 in the glioma cell line, A172, was significantly higher than that in the human astrocyte cell line (Fig. 1a). We next investigated RBBP8 expression levels in 33 common tumor types and found that aberrantly high expression of RBBP8 was common in tumors such as cervical squamous cell carcinoma, endocervical adenocarcinoma, lymphoid neoplasm diffuse large B-cell lymphoma, and lung squamous cell carcinoma (Fig. S1). The box plot shows the expression of RBBP8 in GBM and LGG (Fig. 1b). Specifically, the expression level of RBBP8 in 163 GBM and 518 LGG patients was significantly higher than the expression of RBBP8 in 207 normal brain tissues. Furthermore, the high expression of RBBP8 was confirmed in two GEO datasets (GSE35493 and GSE50161) (Fig. 1c and d). Finally, we identified a trend of increasing RBBP8 protein expression by comparing the extent of protein staining between immunohistochemical sections of normal brain tissue, LGG, and high-grade glioma (Fig. 1e-g).
Association of RBBP8 expression with clinical features in glioma
The relationship between RBBP8 expression and various clinical features was investigated based on TCGA and CGGA databases. RBBP8 expression was significantly correlated with age, and high expression of RBBP8 was more evident in older patients, which was confirmed in three datasets (TCGA RNA-seq, CGGA RNA-seq, and CGGA microarray) (Fig. 2a, d, and g). In addition, RBBP8 expression was higher in male than in female patients (Fig. 2b and 2h). Furthermore, there was a significant association between RBBP8 expression and one parameter double-hybrid molecules in the CGGA RNA-seq dataset (Fig. 2e). We also found a significant correlation between RBBP8 expression and clinical grade in all three datasets, which increased with increasing malignancy (Fig. 2c, f, and i).
Survival outcomes and diagnostic value of RBBP8 in glioma patients
We compared the difference in survival outcomes between the RBBP8 high and low expression groups to clarify the association between prognosis and RBBP8 expression. Thousands of samples were analyzed from TCGA RNA-seq, CGGA RNA-seq, and CGGA microarray datasets. Surprisingly, the results from the three datasets were consistent, indicating that patients with high expression of RBBP8 had a poor prognosis compared to those with low expression of RBBP8 (Fig. 3a-c). Moreover, we used the ROC curve to verify the prognostic value of RBBP8 in predicting survival in glioma. The area under the curve was greater than 0.7 in years 1, 3, and 5, except for the 1-year area under the curve in CGGA RNA-seq. This suggests that RBBP8 overexpression had a high predictive value for poor prognosis of glioma patients (Fig. 3d-f).
High expression of RBBP8 is a dependent risk factor for glioma patients
In TCGA RNA-seq dataset, a univariate Cox hazard analysis identified that RBBP8 (p < 0.001; hazard ratio (HR) = 1.047; 95% confidence interval (CI) = 1.036-1.058), and several clinical characteristics such as age (p < 0.001; HR = 4.701; 95% CI = 3.524-6.271), WHO grade (p < 0.001; HR = 4.688; 95% CI = 3.769-5.832), and cancer status (p < 0.001; HR = 2.034; 95% CI = 1.649-2.510), as risk factors related to survival time (Fig. 4a). However, to eliminate the deviation caused by other factors, multivariate Cox regression models were applied to confirm that RBBP8 (p < 0.001; HR = 1.040; 95% CI = 1.016-1.064), age, WHO grade, and cancer status were strikingly and dependently related to the overall survival of patients (Fig. 4b). Overall, RBBP8 expression is related to the prognosis of glioma, and this independent correlation was confirmed in the CGGA RNA-seq and CGGA microarray datasets (Fig. 4c-f).
The prognostic significance of RBBP8 methylation in glioma
Methylation modifications of genes can regulate gene expression with the potential to become tumor markers. Thus, we explored the methylation of RBBP8 in gliomas. There were eight methylation sites in the RBBP8 gene (cg15770728, cg02863476, cg0818828, cg14157947, cg06299186, cg05305025, cg05513509, and cg16801093), among which cg15770728 had the highest methylation level (Fig. 5a). Furthermore, cg05513509, cg06299186, and cg15770728 had positive correlations with RBBP8; the remaining methylation sites had no relationship with the expression of RBBP8 (Fig. 5b-d). Moreover, cg05513509 affected the prognosis of LGG, patients; those with hypomethylation at the cg05513509 site had a longer survival time than those with hypermethylation at this site (Fig. 5e). Finally, the methylation level of RBBP8 was significantly correlated with age, KPS, person neoplasm cancer status, postoperative rx-tx, PRS type, and WHO grade based on TCGA RNA-seq dataset (Fig. 6a-f).
Relationships of RBBP8 with immune cells and immune checkpoints
The TIMER database was used to identify the relationship between RBBP8 expression and the six types of infiltrating immune cells in the tumor microenvironment. In GBM, there was a significant positive correlation between the expression of RBBP8 and the infiltration of dendritic cells. Notably, this positive correlation in LGG was present between all six immune cells (B cells, CD8+T cells, CD4+T cells, macrophages, neutrophils, and dendritic cells) and RBBP8 expression (Fig. 7a). Kaplan-Meier survival curves showed that patients in the RBBP8 low expression group had a significantly better prognosis than those in the RBBP8 high expression group in the setting of six immune cell infiltrates. However, in LGG, RBBP8 expression was closely associated with survival in the infiltration of dendritic cells (Fig. 7b). Finally, we explored the correlation between the mRNA expression levels of RBBP8 and genes coding immune checkpoints (i.e., PD-1 (PDCD1), PD-L1 (CD274), PD-L2 (PDCD1LG2), CTLA4, and (TIM-3) HAVCR2). The results showed that the expression level of RBBP8 was positively correlated with PD-1 in GBM (Fig.S2a-e), and RBBP8 was positively correlated with PD-1, PD-L1, PD-L2, CTLA4, and TIM-3 in LGG (Fig. S2f-j).
RBBP8-related signaling pathways in glioma
To explore the mechanism of RBBP8 in gliomas, we used GSEA to analyze two groups of samples with high and low RBBP8 expression. Analysis of TCGA data showed that the cell cycle P53 signaling pathway, homologous recombination, mismatch repair, leukocyte transendothelial migration, and antigen processing and presentation were the most significantly enriched signaling pathways (FDR < 0.25, P < 0.05, NES > 1.5) (Fig. 8). For a better understanding of RBBP8 functions, data from the CGGA were used to perform a GSEA analysis. The P53 signaling pathway, base excision repair, homologous recombination, mismatch repair, protein export, and DNA replication (FDR < 0.25, P < 0.05, NES > 1.5) were the top six signaling pathways in the CGGA microarray dataset (Fig. S3). Meanwhile, cell cycle, the P53 signaling pathway, DNA replication, homologous recombination, mismatch repair, and antigen processing and presentation were the top six signaling pathways in the CGGA RNA-seq dataset (Fig. S4).
Meta-analysis of RBBP8 expression on the prognosis of glioma
Results were acquired from TCGA, CGGA, and GEO databases and used in a meta-analysis because no previous studies had revealed the relationship between high RBBP8 expression and overall survival among glioma patients. No significant heterogeneity between the two databases (I2 = 0%, p = 0.98) was observed, and a fixed-effect model was applied. The pooled HR for the correlation between high RBBP8 expression and patient overall survival was 2.24 (95% CI: 1.37-3.65) (Fig. 9). Thus, we can conclude that RBBP8 high expression is an independent predictor of unfavorable overall survival in glioma patients.