Cellular senescence associated with aging and chemotherapy exposure primes bone marrow for immune dysfunction and myeloma persistence

Abstract

Abstract Age-related decline in immune function increases susceptibility to hematologic malignancies. Although bone marrow (BM) orchestrates immunity, age-related changes in the marrow microenvironment are not well understood, partially due to challenges of sample accessibility. Here, we created a comprehensive multi-omics atlas of aging human bone marrow using single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics (ST) as part of the Cellular Senescence Network (SenNet). We collected iliac crest, femoral, and tibial BM samples from 65 donors (aged 25-90) who underwent orthopedic surgery, enabling molecular characterization of over 300,000 cells from 55 sc/snRNA-seq samples and Xenium ST profiling of 2.3 million cells across 32 samples. These data captured over 20 cell types within myeloid, lymphoid, erythroid, and stromal lineages, and revealed the spatial compartmentalization of distinct progenitor- and immune-rich neighborhoods. Both data types showed that hematopoietic stem/progenitor cell (HSPC) subpopulations undergo age-associated increase of cell cycle arrest markers including CDKN1A (P21). Stromal cells, particularly endothelial cells, were frequently found in close proximity with stem cells and exhibited age-associated elevation of cellular senescence, senescence-associated secretory phenotype (SASP), and interferon response signatures. CODEX protein imaging revealed P21+ adipocytes in older individuals, corroborating the occurrence of age-related stromal senescence. Microenvironmental stromal dysfunction has been previously demonstrated to impair stem cell function and may contribute to systemic immune decline. Additionally, we identified megakaryocytes and lymphoid-dense neighborhoods with enriched expression of senescent markers. As chemotherapy is frequently used to induce cellular senescence in animal models, and BM immune surveillance is particularly important in the context of BM-resident malignancies, we extended our spatial analysis to 82 multiple myeloma samples (38 untreated and 44 chemo-treated) to see how cellular senescence relates to disease. We observed a striking increase in stromal expression of senescence markers including CDKN1A, TP53, TP53BP1, and the SASP markers DKK1, and IGFBP7 in treated samples relative to both age-matched donors and untreated disease. We further observed plasma cell tumor-dense regions infiltrated by monocytes and CD8+ T cells with elevated expression of exhaustion markers in treated samples. As DKK1, IGFBP7, and other SASP members disrupt immune responses at high levels, our findings provide new insights into cellular senescence as a previously under-studied source of immune evasion in myeloma pathogenesis. Our atlas of donor and patient data helps elucidate the cellular and spatial mechanisms underlying immunosenescence, provides potential therapeutic targets for age-associated immune dysfunction, and is a resource for further investigations of how the aging microenvironment contributes to hematologic malignancy.