Aged brain multi-omic integration captures immunometabolic and sex variation

Abstract

Resolving the molecular basis for heterogeneous aging in the human brain requires integrating its diverse molecular layers, including the transcriptional, epigenetic, and proteomic states. To provide such a view, we applied a tensor decomposition framework to jointly analyze single-nucleus RNA-seq, DNA methylation, histone acetylation, and proteomic data from 276 postmortem human dorsolateral prefrontal cortex samples from the ROSMAP aging cohort. Our analysis revealed two dominant themes defined by components with strong multi-omic coherence. First, we identified a robust immunometabolic axis characterized by microglial activation, suppression of PI3K-Akt-mTOR signaling, and epigenetic signatures suggesting the repurposing of developmental transcription factors. Second our analysis resolved the effect of sex into a mosaic of distinct, cell-type-specific molecular programs within glia. These included male-biased mitochondrial programs in microglia, a shift toward precursor-like states in the female oligodendrocyte lineage, and bidirectional epigenetic remodeling in astrocytes. This work provides a high-resolution atlas of glial aging, demonstrating that broad risk factors manifest as complex, cell-specific vulnerabilities; a critical insight for developing targeted therapeutic strategies.