Tumor Microenvironment Hijacks and Accelerates a Physiological Myeloid Senescence Signature Associated with...

Immunosenescence is a critical driver of tumor initiation and progression. In this study, we systematically characterized immune cell senescence by integrating transcriptomic profiles from 17 physiologically aged tissues with pan-cancer single-cell datasets, encompassing 206 samples across nine cancer types. Cross-tissue comparison of senescence-associated alterations, integrated with spatial transcriptomics, revealed that malignant cells triggered senescence in the core myeloid subpopulation designated Mac_DAB2 via a conserved MIF-CD74 signaling axis. By integrating shared myeloid differentiation programs across normal tissues and the tumor microenvironment (TME) with their transcriptional regulatory networks, we defined a myeloid senescence-associated gene (MSAG) signature. This signature successfully distinguishes a senescence-associated, immunosuppressive subtype linked to poor prognosis in pan-cancer cohorts. Finally, we established the MSAG.SIG prognostic model using an ensemble framework of 117 machine learning algorithms, which demonstrated robust and consistent predictive performance across multiple independent cohorts. Overall, this study elucidates the mechanisms underlying TME-driven myeloid senescence, establishes MSAG as a conceptual framework for characterizing myeloid immunosenescence, and provides a clinically relevant pan-cancer prognostic tool with translational potential.