Cellular dynamics across aged human brains uncover a multicellular cascade leading to Alzheimer’s disease

Abstract

Alzheimer's Disease (AD) is a progressive neurodegenerative disease seen with advancing age. Recent studies have revealed diverse AD-associated cell states, yet when and how they impact the causal chain leading to AD remains unknown. To reconstruct the dynamics of the brain's cellular environment along the disease cascade and to distinguish between AD and aging effects, we built a comprehensive cell atlas of the aged prefrontal cortex from 1.64 million single-nucleus RNA-seq profiles. We associated glial, vascular and neuronal subpopulations with AD-related traits for 424 aging individuals, and aligned them along the disease cascade using causal modeling. We identified two distinct lipid-associated microglial subpopulations, one contributed to amyloid-β proteinopathy while the other mediated the effect of amyloid-beta in accelerating tau proteinopathy, as well as an astrocyte subpopulation that mediated the effect of tau on cognitive decline. To model the coordinated dynamics of the entire cellular environment we devised the BEYOND methodology which uncovered two distinct trajectories of brain aging that are defined by distinct sequences of changes in cellular communities. Older individuals are engaged in one of two possible trajectories, each associated with progressive changes in specific cellular communities that end with: (1) AD dementia or (2) alternative brain aging. Thus, we provide a cellular foundation for a new perspective of AD pathophysiology that could inform the development of new therapeutic interventions targeting cellular communities, while designing a different clinical management for those individuals on the path to AD or to alternative brain aging.

Psychosocial experiences are associated with human brain mitochondrial biology

“…Knowing that different cell types contain vastly different mitochondrial phenotypes (50, 53, 77), we therefore quantified mitochondrial phenotypes in specific brain cell types using single-nucleus RNA sequencing of ROSMAP DLPFC samples (n=424). Of the 87 distinct cell types and subtypes previously identified (78), we first examined mitochondrial psychobiological associations with transcript levels in cell types where the same MitoPathways as above were sufficiently well represented (>50% of participants and >50% coverage of mitochondrial genes). Based on these criteria, we included the following six major cell types: excitatory and inhibitory neurons, microglia, astrocytes, oligodendrocytes and oligodendrocyte progenitor cells (OPCs) (Fig.…”

“…Single-nucleus RNAseq was performed from frozen DLPFC specimens (N=424) as described previously (78, 125). Gray matter was extracted and dissociated into nuclei suspension.…”

Associations of cortical SPP1 and ITGAX with cognition and common neuropathologies in older adults

“…Previous publications have extensively described the snRNASeq dataset used in this work 38,39 . Briefly, nuclei were isolated from 479 DLPFC tissues of ROSMAP.…”

“…The Seurat FindAllMarkers function was used for differential gene expression per cluster, followed by gene set enrichment analysis and comparison with previously published papers. Each subpopulation was characterized by the expression profile of marker genes (eg, P2RY12 marker of homeostatic microglia, APOE damage‐associated microglia [DAM] stage 1 and ITGAX DAM stage 2) followed by pathway analysis 38 . For downstream analysis, pseudo‐bulk matrices were created by summing counts per donor.…”

“…Finally, the thresholds by cell type of nUMI and nGene, respectively, included for downstream analysis were ITGAX DAM stage 2) followed by pathway analysis. 38 For downstream analysis, pseudo-bulk matrices were created by summing counts per donor. The genes were filtered by each cell type, keeping genes with CPM > 1 in 80% of samples.…”