Genetic and glial heterogeneity in Alzheimer disease

Abstract Background Alzheimer disease (AD) has substantial genetic, molecular, and cellular heterogeneity associated with its etiology. However, much of the downstream cell‐type‐specific transcriptional and functional ramifications of familial mutations and high‐risk variants for AD are still poorly understood. Methods We generated unsorted single‐nuclei transcriptomic profiles (snRNA‐seq) of parietal lobes from 67 donors from the Knight ADRC and DIAN cohorts [1] to investigate altered molecular pathways in healthy and AD individuals, including carriers of pathogenic mutations in APP and PSEN1 and risk variants in TREM2 . We validated our findings using snRNA‐seq from the DLPFC from the ROS‐MAP cohort [2]. We analyzed ∼294K high‐quality nuclei and identified six major cell populations. We performed deep subclustering analysis for the distinct cell types to reveal and characterize transcriptional states associated with AD genetic variants. Results Carriers of TREM2 AD risk variants p.R47H, p.R62H, and p.H157Y showed significantly increased proportions of nuclei in a reduced activation state ( p = 3·29×10 −2 ), consistent with in vitro experiments showing reduced microglial activation associated with these variants. [2, 3] ( Figure ). This state was replicated in the ROSMAP cohort (p = 7·4×10 −107 ), and TREM2 p.R62H carriers showed a higher proportion of nuclei ( p = 2·26×10 −2 ). TREM2 risk variant carriers also showed an increased proportion of oligodendrocytes ( Figure ) exhibiting upregulation of 1,124 genes including TFEB (Knight ADRC p = 4·66×10 −2 ; ROSMAP p = 2·48×10 −2 ). Altered TFEB expression may be driven by the interaction of TREM2 with mTOR [4, 5]. TFEB is a central regulator of lysosomal biogenesis[6, 7] and represses myelination at different developmental stages.[8] Altered TFEB signaling has been implicated in several neurodegenerative diseases.[6, 9] Conclusions AD pathogenic and risk variants are sufficient to alter human brains' transcriptional and cellular landscape. While TREM2 is mainly expressed in microglia, and its loss of function mutations cause Nasu‐Hakola disease, which is characterized by loss of myelination, suggesting TREM2 ‐linked oligodendrocyte‐microglia crosstalk.[10] Here, we found that AD patients with TREM2 variants showed upregulation of TFEB and lysosomal pathways in a subset of oligodendrocytes likely driven by altered microglia transcriptional states. The integration of genetic and single‐cell molecular data facilitates our understanding of the heterogeneity of pathways, biological processes, and cell types affected in AD.