Integrated Single-Cell and Spatial Multi-Omics Decipher the Myeloid Landscape and Reveal the Myelin-Induced SPP1-CD44 Axis as a Therapeutic Target in Glioblastoma

Here, we addressed these challenges by constructing a comprehensive multi-omics atlas of the GBM myeloid landscape. By integrating single-cell RNA sequencing (scRNA-seq) from a massive cohort of 231 samples with spatial transcriptomics (ST), we moved beyond cell type enumeration to define conserved "metaprograms"—fundamental functional states adopted by myeloid cells regardless of their ontogeny. We uncovered a continuous developmental trajectory wherein myeloid cells are reprogrammed towards a terminal Phagocytic Suppressive state, a process driven spatially by the ingestion of myelin debris and functionally by the activation of the SPP1-CD44 signaling axis. We show that this axis serves as a master regulator of myeloid-malignant crosstalk, linking tissue damage (demyelination) to immune evasion. Furthermore, we demonstrate that a myeloid-derived gene signature (MeySig) can stratify patients for treatment, and that disrupting this specific myeloid-tumor circuit restores the efficacy of anti-PD-1 therapy. Our study provides a high-definition roadmap of myeloid evolution in GBM and establishes a mechanistic rationale for next-generation combination immunotherapies.