Glioblastoma multiforme (GBM) is the most aggressive form of brain tumor. In general, tumor growth requires disruption of the tissue microenvironment, yet how this affects glioma progression is unknown. We studied program death-ligand (PD-L)1 in neurons and gliomas in tumors from GBM patients and associated the findings with clinical outcome. Remarkably, we found that upregulation of PD-L1 by neurons in tumor-adjacent brain tissue (TABT) associated positively with GBM patient survival, whereas lack of neuronal PD-L1 expression was associated with high PD-L1 in tumors and unfavorable prognosis. To understand the molecular mechanism of PD-L1 signaling in neurons, we investigated PD-L1 function in cerebellar and cortical neurons and its impact on gliomas. We discovered that neuronal PD-L1-induced caspase-dependent apoptosis of glioma cells. Because interferon (IFN)-β induces PD-L1 expression, we studied the functional consequences of neuronal Ifnb gene deletion on PD-L1 signaling and function. Ifnb-/- neurons lacked PD-L1 and were defective in inducing glioma cell death; this effect was reversed on PD-L1 gene transfection. Ifnb-/- mice with intracerebral isografts survived poorly. Similar to the observations in GBM patients, better survival in wild-type mice was associated with high neuronal PD-L1 in TABT and downregulation of PD-L1 in tumors, which was defective in Ifnb-/- mice. Our data indicated that neuronal PD-L1 signaling in brain cells was important for GBM patient survival. Reciprocal PD-L1 regulation in TABT and tumor tissue could be a prognostic biomarker for GBM. Understanding the complex interactions between tumor and adjacent stromal tissue is important in designing targeted GBM therapies.