Ó³»­´«Ã½

BACKGROUND: The amygdala is involved in early stages of multiple neurodegenerative disorders such as Alzheimer's disease (AD), limbic-predominant age-related TDP-43 encephalopathy neuropathologic change (LATE-NC), and Lewy body disease (LBD). However, the single-cell architecture of the aging human amygdala is not well characterized.

METHOD: The basolateral amygdala was dissected from 9 brain donors recruited at Rush Alzheimer's Disease Center. We generated 10x single-nucleus multi-ome (RNA + Assay for Transposase-Accessible Chromatin [ATAC]-seq) data: we pooled 3 samples per batch and then demultiplexed them using freemuxlet. The following QC parameters were used: mitochondrial RNA content<10%, RNA>1000 UMIs, and ATAC>100 UMIs. We used Harmony and Weighted Nearest Neighbor (WNN) algorithms to integrate multi-omic data across batches. We identified major brain cell types, cell states, and their marker genes, and visualized the clusters using uniform manifold approximation and projection (UMAP).

RESULT: All 9 participants (6 females; average age of death 89.7±4.95) had at least some AD pathology, 6 had LATE-NC, and 2 had LBD pathology. Our pipeline yielded 31,662 high-quality single-nucleus multi-omes. We observed well-defined clusters of astrocytes (Ast), excitatory (ExN) and inhibitory (InN) neurons, microglia (Mic), oligodendrocytes (Oli), oligodendrocyte progenitor cells (OPC), and T cells. We identified 3 Ast, 7 ExN, 4 InN, 4 Mic, 5 Oli, 3 OPC, and 1 T cell states. In our focused analyses of Mic, APOE was highly expressed in two abundant Mic states, one of which also highly expressed TREM2; markers of human AD-associated microglia states (Green et al., Nature 2024) and mouse disease-associated microglia (DAM) (Keren-Shaul et al., Cell 2017) were also highly expressed in those two states. Notably, the observed proportion of these states (73.9±11.2% of total microglia) were much higher than that from the frontal cortex of AD patients (10.9±9.0%). In addition, we observed a homeostatic-like state (high P2RY12, CX3CR1) and a state highly expressing interferon (IFN) response genes (IFIT2, IFIT3).

CONCLUSION: Our pilot study demonstrated the feasibility of generating high-quality, large-scale single-cell multi-omic data from the human amygdala. We observed abundant AD-associated microglia in amygdala, suggesting the possibility that the regional vulnerability of amygdala to multiple neurodegenerative disorders might be in part due to different glial composition.