Multi-scale transcriptomic integration reveals cell-type immune networks and lncRNA remodeling in Alzheimers disease

Abstract

Alzheimers disease (AD) displays pronounced regional heterogeneity, yet how transcriptional changes across brain regions converge into coordinated cellular and molecular programs remains unclear. Here, we integrated bulk and single-cell transcriptomics with network modeling to characterize gene expression remodeling across cortical and hippocampal subregions in APP/PS1-21 mice. We show that amyloid pathology follows distinct regional trajectories, with early cortical activation, delayed but robust remodeling in CA1, and a late-stage shift toward widespread transcriptional repression in the dentate gyrus. Despite these differences, cross-region analyses revealed a conserved immune activation core spanning cortical and hippocampal circuits. Network-level modeling further demonstrated that disease-associated transcriptional changes organize into immune-enriched modules that map onto specific cellular compartments, predominantly associated with microglia in cortex, astrocytes in CA1, and coordinated multi-lineage remodeling in the dentate gyrus. Notably, long noncoding RNAs were consistently embedded within disease-associated networks despite weak single-cell differential expression signals, suggesting their involvement in coordinated regulatory programs. Together, these findings link regional transcriptomic remodeling to cell-type-resolved network architecture and identify convergent immune-driven programs underlying amyloid-associated neurodegeneration.