Distinct tumor immune microenvironmental (TIME) landscapes drive divergent immunotherapy responses in glioblastoma

Glioblastoma (GBM) exhibits significant molecular heterogeneity leading to variable treatment responses. Despite multimodal therapies, prognosis remains poor, highlighting the need for personalized approaches targeting the tumor immune microenvironment (TIME). Using single-cell RNA sequencing, multiplex immunohistochemistry, and orthotopic mouse models, we characterized distinct TIME subtypes and evaluated responses to anti-angiogenic immunotherapy and myeloid-targeting approaches. We identified three TIME subtypes: TIME-low (immune-excluded with abnormal vasculature), TIME-med (immune-infiltrated with functional T cells), and TIME-high (heavily infiltrated with immunosuppressive myeloid cells and anergic T cells). TIME-low GBMs responded transiently to anti-angiogenic immunotherapy with immunostimulatory T cell shifts, while anti-angiogenic therapy was ineffective in TIME-high GBMs due to immunosuppressive myeloid cells. CD40 agonist treatment worsened outcomes in TIME-high GBMs by increasing immunosuppressive cells and reducing NK recruitment. Conversely, PI3K?/d inhibition combined with anti-angiogenic immunotherapy modestly extended survival in TIME-high tumors. Our study reveals that GBM subtypes require tailored therapeutic strategies, with TIME classification potentially predicting treatment responses and TIME-high tumors requiring myeloid reprogramming to overcome immunosuppression. Overall design: We performed single-cell RNA sequencing to investigate how different murine GBM models can recapitulate distinct human GBM subtypes, and how tumor cells and the tumor immune microenvironment adapt to different immunotherapies. The study included NSCG (TIME-low) and NFpp10 (TIME-high) murine GBM tumor models under different treatment conditions: untreated (UT) or vehicle control (IgG), BP (B20 + PD-L1), CD40 agonist, and BPI (BP + IPI-145). Tissues were collected at approximately day 14 (after 3 rounds of treatment). Single-cell suspensions were prepared from fresh whole tumor tissues from at least two mice pooled within each treatment group using enzymatic digestion before 10X Genomics single-cell RNA sequencing.