A roadmap to human hippocampal neurogenesis in adulthood, aging and AD

In the rodent, hippocampal neurogenesis plays critical roles in learning and memory, is tightly regulated by inhibitory neurons and contributes to memory dysfunction in Alzheimer's disease (AD) mouse models. In contrast, the mechanisms regulating neurogenesis in the adult human hippocampus, the dynamic shifts in the transcriptomic and epigenomic profiles in aging and AD and putative niche interactions within the cellular environment, remain largely unknown. Using single nuclei multi-omics of postmortem human hippocampi we map the molecular mechanisms of hippocampal neurogenesis across aging, cognitive decline, and AD neuropathology. Transcriptomic and epigenetic profiling of neural stem cells (NSCs), neuroblasts and immature neurons suggests that the earliest shift in the characteristics of neurogenesis takes place in NSCs in aging. Cognitive impairment was associated with changes in neuroblast profile. In AD, there was a widespread cessation of the transcription machinery in immature neurons, with robust downregulation of genes regulating ribosomal and mitochondrial function. Further, there was substantial loss of parvalbumin+ inhibitory neurons in the hippocampus in aging. The number of the rest of inhibitory neurons were reduced as a function of age and diagnosis. Notably, a similar system-level effect was observed between immature and inhibitory neurons in the transition from aging to AD, manifested by common molecular pathways that were ultimately lost in AD. The numbers of neuroblasts, immature and GABAergic neurons inversely correlated with extent of neuropathology. Using CellChat and NeuronChat, we inferred the ligands and receptors by which neurogenic cells communicate with their cellular environment. Loss of synaptic adhesion molecules and neurotransmitters, either sent or received by neurogenic cells, was observed in AD. Together, this study delineates the molecular mechanisms and dynamics of human neurogenesis, functional association with inhibitory neurons and a mechanism of hippocampal hyperexcitability in AD. Overall design: Nuclei isolated from the hippocampus of young adults (YA, n=2), aging with no cognitive impairments (NCI, n=4), mild cognitive impairments/early dementia (MCI, n=4) and Alzheimer's disease dementia (AD, n=4) were sequenced for multi-ome single nuclei RNA sequencing (snRNAseq) and single nuclei Assay for Transposase-Accessible Chromatin (snATACseq) analysis. Machine learning label transfer algorithm scANVI was used to transfer labels from two scRNAseq datasets, a human developmental forebrain and an adult human hippocampal dataset to ensure unbiased cell annotation. Single-cell trajectory analysis ( CytoTRACE, Pseudotime) were used t infer and confirm the different developmental paths through the cell populations. Differential gene and peak expression as well as Motif analysis were used to assess the alterations in transcriptomic profiles of diffenet neurogenic niches accross diagnosis. Lastly, CellChat and NeuroChat to infer putative interaction of the neurogenic niches with their cellular niche in the human brain and the ligand-receptor pathways that may play a role in a decline in cell-cell interaction as a function of pathology in immature neurons