216 - Overcoming Myeloid-Driven Resistance of High-Grade Gliomas to CAR T Cell Therapy: Mechanistic Determinants and Therapeutic Targeting
Presenter(s)
S. Gholamin1, H. M. Natri2, S. Xu1, M. Aftabizadeh1, A. L. Williams-Katek2, Y. Zhao1, C. Masia1, R. A. Wong1, L. Peter2, M. I. Chung2, B. Aguilar3, R. Starr1, D. Alizadeh3, A. Kalbasi4, X. Wu1, A. Ribas5, M. D'Apuzzo1, S. J. Forman6, B. Badie7, N. Banovich2, and C. Brown3; 1City of Hope, Duarte, CA, 2Translational Genomics Research Institute (TGen) affiliated to City of Hope, Phoenix, AZ, 3City of Hope National Medical Center, Duarte, CA, 4Department of Radiation Oncology, Stanford University School of Medicine, Stanford, CA, 5UCLA, Los Angeles, CA, 6Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Duarte, CA, 7Division of Neurosurgery, Department of Surgery, City of Hope National Medical Center, Duarte, CA
Purpose/Objective(s): Treatment outcomes for high-grade gliomas (HGG) and glioblastoma (GBM) remain poor despite decades of therapeutic effort. While CAR T cell trials in HGG/GBM are rapidly expanding, clinical efficacy remains limited, highlighting the need for improved mechanistic insight. Emerging evidence implicates the tumor microenvironment (TME) as a major determinant of resistance; however, the patient-specific intercellular programs that govern response versus resistance to IL13Ra2-targeted CAR T therapy remain incompletely understood. We integrated single-cell, spatial, and preclinical modeling approaches to identify dominant TME-driven mechanisms of resistance and to evaluate therapeutic strategies targeting myeloid-mediated suppression.
Materials/Methods: Human pretreatment tumors from a phase I IL13Ra2-targeted CAR T trial (NCT02208362) underwent single-cell RNA-seq (scRNA-seq; n = 41; progressive disease [PD], n = 23 vs stable disease/partial/complete response [SD/PR/CR], n = 18) with cell–cell communication analyses. Spatial profiling of baseline tumors from responders and non-responders included 10x Xenium (16 samples/13 patients) with adjacent high-plex in situ protein imaging using multiplex spatial proteomic imaging—CO-Detection by indEXing (CODEX). In vivo, two orthotopic syngeneic glioma models expressing luciferase and IL13Ra2 were treated with murine IL13Ra2-targeted CAR T cells ± anti-SPP1 antibodies, and follow-up bioluminescence imaging and survival analysis were used to assess efficacy. Kluc tumors (n = 12; untreated, CAR, anti-SPP1, CAR+anti-SPP1; n = 3/arm) were additionally profiled using scRNA-seq and 10x Visium HD pre- and post-therapy.
Results: scRNA-seq of patient samples identified responder tumors enriched for lymphocytes and MHC I/MHC II signaling, whereas non-responders exhibited extracellular matrix–dominant communication with SPP1 as a top pathway and increased SPP1+ myeloid–fibroblast interactions. Higher SPP1+ myeloid abundance was associated with inferior survival. Spatial transcriptomics and CODEX localized SPP1+ macrophages to fibroblast/ECM-rich cellular neighborhoods with distinct immune organization in non-responders. In both glioma models, anti-SPP1 antibody combined with IL13Ra2-targeted CAR T cells enhanced antitumor activity and prolonged survival compared with CAR T alone. scRNA-seq and Visium HD profiling further demonstrated that combination therapy remodeled spatial immune infiltration and signaling programs in preclinical glioma models.
Conclusion: A spatially organized SPP1+ myeloid–stromal network represents a dominant, actionable mechanism of CAR T resistance in HGG/GBM. Targeting SPP1 reprograms the TME and improves IL13Ra2-targeted CAR T efficacy, supporting combination strategies for future translational studies. These findings also provide mechanistic insight that may inform the development and optimization of next-generation CAR T cell engineering strategies.