Machine Learning Identifies Key Cells and Therapeutic Targets in Intervertebral Disc Degeneration: SASP-Driven Matrix Catabolism, Inflammation Amplification, and Metabolic Collapse
摘要
Intervertebral disc degeneration (IVDD) is a major contributor to low back pain, yet its cellular and molecular mechanisms remain incompletely understood. In this study, we integrated single-cell and bulk transcriptomic data to uncover the role of senescence-associated secretory phenotype (SASP) signaling in IVDD progression. Single-cell RNA sequencing delineated the heterogeneity of nucleus pulposus cells (NPCs) subtypes and revealed significant differences in senescence levels and SASP activity. Bulk RNA-seq integration across multiple datasets further confirmed widespread SASP activation and defined a core regulatory network centered on Bone Morphogenetic Protein 2 (BMP2) and Matrix Metalloproteinase 3 (MMP3), identified via WGCNA and machine learning algorithms (LASSO, Random Forest, Boruta). A SASP scoring model based on these two genes showed strong diagnostic performance. Drug screening identified Simvastatin as a high-affinity dual inhibitor of BMP2/MMP3, with molecular docking supporting its therapeutic potential. In vitro, Simvastatin treatment reduced NPCs senescence and apoptosis, while in vivo studies demonstrated that Simvastatin preserved disc structure, decreased pro-inflammatory cytokine levels, and mitigated degenerative changes in a murine IVDD model. Collectively, this study establishes a regulatory framework of SASP in IVDD and proposes BMP2/MMP3 as promising targets for intervention. Our findings offer insights and a potential translational path for slowing disc degeneration.
Graphical Abstract