Single-cell multiregion dissection of central nervous system aging

Abstract Central nervous system (CNS) aging is a major risk factor for many disorders, including cerebrovascular disease, neurodegeneration and amyotrophic lateral sclerosis, yet the cellular pathways driving its progression across CNS regions remain poorly defined. Here we present a single-nucleus transcriptomic atlas spanning seven human CNS regions, comprising ∼1.0 million nuclei from 235 post-mortem samples derived from 200 neurologically and psychiatrically normal donors aged 19–101 years. Across regions, we delineate both shared and region-specific features of CNS aging, integrating analyses of transcriptional noise, programmed cell-death signatures, disease associations, metabolic reprogramming and transcriptomic remodeling. We identify cross-regional vulnerability of astroglia, oligodendrocytes, and excitatory and inhibitory neurons, and show that microglial and astrocytic activation represents a broadly conserved aging response across the CNS, highlighting potential targets for intervention. Finally, we present within-dataset proof-of-concept predictive modeling use cases based on aging-associated gene signatures, providing a resource for within-atlas prioritization and hypothesis generation of candidate biomarkers of CNS aging. Together, this work offers a region-by-region map of the aging human CNS and informs the selection of specific cell types and/or regions for future anti-aging strategies.