Integrated single-cell multi-omics profiling reveals a senescence-associated hematopoietic landscape and regulatory network in aging bone marrow.

Abstract

Bone marrow exhibits functional decline, yet cellular heterogeneity and molecular mechanisms remain unclear due to limitations of traditional research methods. This study aims to characterize age-related changes and identify key drivers in bone marrow. Integrated multi-omics analysis was performed using scRNA-seq, proteomics, pseudo-bulk transcriptomics, weighted gene co-expression network analysis (WGCNA)-based transcription factor (TF) network modeling, and CellChat analysis. Samples included 6 young and aged bone marrow specimens. Statistical validation involved differential expression analysis, Cox regression modeling, and receiver operating characteristic (ROC) curve analysis. A novel hematopoietic subpopulation (3.19% of aged samples) was identified, activating the cellular senescence pathway (KEGG) and enhancing inflammatory crosstalk with CD8⁺ T cells via NMU signaling (|avg_log2FC|> 0.58, p < 0.001). Pseudo-bulk and proteomic analyses identified CAPN1, MAP2K1, and JUND as potential signal modules. Immunohistochemistry and Western blot confirmed their co-expression, while molecular docking revealed interaction interfaces. In two independent bulk-RNA cohorts (n = 58), a Cox model based on the CAPN1-MAP2K1-JUND module showed robust predictive value for aging, with AUCs of 0.7507 (p = 0.0154) and 0.90 (p = 0.0274). This study identifies a pivotal molecular module linking single-cell dynamics to tissue-level senescence in bone marrow, providing new insights into aging mechanisms and potential therapeutic targets.