Characterizing malignant prognostic signatures in primary glioma based on single-cell and bulk transcriptome sequencing

Abstract

Glioma, characterized by its highly invasive nature, presents significant challenges in prognosis and treatment resistance. The advent of single-cell RNA sequencing (scRNA-seq) has facilitated a more nuanced understanding of the cellular and molecular landscapes of glioma cells. In this study, conducting a comprehensive analysis of scRNA-seq and bulk RNA-seq data, employing Cox regression and Least Absolute Shrinkage and Selection Operator (LASSO) methods, we identified three malignant prognostic signatures: IGFBP2, MDK, and RARRES2. The predictive accuracy of this model was validated across both The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA) cohorts. Additionally, we explored the correlation of these signatures with drug responsiveness and immune cell infiltration. Differential expression validation and functional analyses of RARRES2 were performed using external Gene Expression Omnibus (GEO) datasets and in-house samples. In both glioma and pan-cancer contexts, RARRES2 expression is significantly positively correlated with the infiltration of M2-like macrophages, NK cells, and CD8 + T cells. Given that RARRES2 receptors are predominantly found in myeloid and glioma cells, we hypothesize that RARRES2 may regulate tumor progression through autocrine pathways and influence macrophage recruitment and differentiation via paracrine pathways. Collectively, our findings provide valuable insights into potential novel prognostic markers for glioma, potentially enhancing the accuracy of prognostic predictions and serving as promising therapeutic targets.