A novel score system based on arginine metabolism-related genes to predict prognosis, characterize immune microenvironment, and forecast response to immunotherapy in IDH-wildtype glioblastoma.

Glioblastoma is one of the most lethal cancers and leads to more than 200,000 deaths annually. However, despite lots of researchers devoted to exploring novel treatment regime, most of these attempts eventually failed to improve the overall survival of glioblastoma patients in near 20 years. Immunotherapy is an emerging therapy for cancers and have succeeded in many cancers. But most of its application in glioblastoma have been proved with no improvement in overall survival, which may result from the unique immune microenvironment of glioblastoma. Arginine is amino acid and is involved in many physiological processes. Many studies have suggested that arginine and its metabolism can regulate malignancy of multiple cancers and influence the formation of tumor immune microenvironment. However, there is hardly study focusing on the role of arginine metabolism in glioblastoma. In this research, based on mRNA sequencing data of 560 IDH-wildtype glioblastoma patients from three public cohorts and one our own cohort, we aimed to construct an arginine metabolism-related genes signature (ArMRS) based on four essential arginine metabolism-related genes (ArMGs) that we filtered from all genes with potential relation with arginine metabolism. Subsequently, the glioblastoma patients were classified into ArMRS high-risk and low-risk groups according to calculated optimal cut-off values of ArMRS in these four cohorts. Further validation demonstrated that the ArMRS was an independent prognostic factor and displayed fine efficacy in prediction of glioblastoma patients' prognosis. Moreover, analyses of tumor immune microenvironment revealed that higher ArMRS was correlated with more immune infiltration and relatively "hot" immunological phenotype. We also demonstrated that ArMRS was positively correlated with the expression of multiple immunotherapy targets, including PD1 and B7-H3. Additionally, the glioblastomas in the ArMRS high-risk group would present with more cytotoxic T cells (CTLs) infiltration and better predicted response to immune checkpoint inhibitors (ICIs). In conclusion, our study constructed a novel score system based on arginine metabolism, ArMRS, which presented with good efficacy in prognosis prediction and strong potential to predict unique immunological features, resistance to immunotherapy, and guide the application of immunotherapy in IDH-wild type glioblastoma.