ID #326 Targeting TAM and microglial states to potentiate B7H3 CAR Tcell therapy in Group 3 Medulloblastoma

Abstract Background Group 3 medulloblastoma (G3MB) is the most aggressive molecular subtype and is associated with poor survival in children. Although B7H3 is consistently expressed in G3MB, CAR Tcell therapy has shown limited efficacy in primary brain tumors due to suboptimal cellular persistence and profoundly immunosuppressive tumor microenvironments (TME). Myeloid cells, including microglia and tumorassociated macrophages (TAMs), represent a major barrier through local immune suppression, yet proinflammatory myeloid programs can enhance CAR Tcell activity. Understanding how CAR design and myeloid rewiring interact is essential for advancing immunotherapy for pediatric brain tumors. Aims To improve B7H3–directed CAR Tcell therapy for G3MB by optimizing CAR architecture and reprogramming myeloid compartments through systemic Resiquimod, a brainpenetrant TLR7/8 agonist. Methods B7H3 CAR constructs incorporating CD28, 41BB, or dual CD28/41BB costimulation were generated and evaluated using invitro cytotoxicity assays and immunocompetent G3MB models. Resiquimod was administered systemically to achieve intracranial activity and promote myeloid repolarization. CAR Tcell activity, myeloid phenotypes, and TME transcriptional programs were assessed by flow cytometry, bulk RNA sequencing, and pathway enrichment analyses. Blood-brain barrier penetration of Resiquimod was confirmed by immunofluorescence. Results All CAR constructs exhibited strong invitro cytotoxicity; dualcostimulatory CARs demonstrated faster killing kinetics. In NSG mice lacking TME-mediated suppression, CAR variants showed comparable tumor control. In contrast, in immunocompetent models, dual CAR T cells displayed enhanced cytotoxicity and prolonged survival. Adjunctive Resiquimod further improved outcomes in both NSG and B6 mice, while localizing to tumorbearing brain regions, and shifting TAM and microglial states toward proinflammatory programs associated with improved CAR Tcell function. Ongoing analyses are characterizing Resiquimod’s impact on myeloid subsets and TME remodeling. Conclusion Optimizing CAR design and reprogramming myeloid populations improve CAR Tcell efficacy against G3MB. This strategy addresses key barriers in pediatric brain tumor immunotherapy and supports the development of TMEinformed CAR Tcell approaches for highrisk medulloblastoma.