SUMMARY Glioblastoma (GBM) remains deadly. A major reason is a highly infiltrative GBM border, making complete surgical resection unattainable. Residual tumor cells inevitably spawn recurrent GBM, typically within 2-3 cm of resection cavity. Understanding GBM margins is thus critical to curb GBM invasion and relapse. Tumor- associated myeloid cells (TAMs) are the most abundant stromal cells in GBM, and they promote GBM expansion and infiltration through immunosuppressive properties, secretion of tumor-promoting factors (gliomagens), and preparation of invasion tracks. Previous research has largely focused on bone-marrow derived macrophages (MDMs), which reside mostly in GBM interior, less is defined for microglia (MG), resident innate immune cells in the brain that preferentially congregate at GBM margins. The current proposal intends to fill this gap by testing the central hypothesis that perturbing the supportive niche provided by margin-MG may help slow down GBM invasion and relapse. The premise of the approach rests on our compelling preliminary data establishing MG as one of the most abundant immune cell populations at GBM borders. Furthermore, our pilot study demonstrated that targeting MG proliferation (via IL-1/IL1R1 signaling) or MG motility/alignment (via axon guidance receptor Plexin-B2) resulted in significant reduction of GBM growth and invasion. However, major knowledge gaps remain: i) what are the signaling networks driving the persistence of tumor-promoting phenotypes of margin-MG, ii) does genetic make-up of the tumor play a role in dictating the immune landscape at GBM margin, iii) what is the nature of MG-derived gliomagens, iv) can we find a way to target margin-MG interactions with GBM cells, and v) how does chemoradiation therapy alter gene signatures and tumor-promoting gliomagens of margin-MG? To address these questions, in Aim 1, we will map the immune contexture and crosstalk with tumor cells at GBM margins using multiplex immunostaining, advanced spectral flow cytometry, single cell transcriptomics, and spatial multiomics. Genetically engineered mouse models (GEMMs) of GBM with defined driver mutations and human GBM tissues of matching subtypes will be used. We will identify distinct margin-MG populations and pro- tumorigenic signals originating from margin-MG. We will also conduct co-cultures and ex vivo organotypic brain slice studies for validation. In Aim 2, we will study the functional significance of margin-MG in promoting GBM expansion and invasion by examining the efficacy of targeting MG proliferation (via IL-1/IL1R1 signaling), MG motility/alignment (via Plexin-B2 signaling), or both, using genetic and pharmacological approaches in GEMMs of GBM. This will provide proof-of-principle that targeting key signaling pathways with different mechanisms of action may achieve synergistic effects for perturbing the supportive niche provided by MG at invasive margins. In Aim 3, we will examine the impact of chemoradiation...