Evaluating the autonomous and non-autonomous effects of interacting cell-types using novel stem cell co-culture models of Alzheimer’s disease

Evaluating the autonomous and non-autonomous effects of interacting cell-types using novel stem cell co-culture models of Alzheimer’s disease

Sarah E. Heuer, Alex Lish, Nancy Ashour, Pallavi Gaur, Richard V. Pearse, Vilas Menon, Tracy Young-Pearse

Alzheimer’s disease (AD) is a common neurodegenerative disorder that is predominant in the aged population. AD is clinically classified through accelerated memory loss, and the pathological presence of beta amyloid plaques and neurofibrillary tau tangles in the brain. While there are known genetic factors that increase risk for developing AD, little is known about how variants drive abnormal cellular endophenotypes and interactions, especially in highly affected brain cell-types: neurons, microglia, and astrocytes. Since the complexity of cellular interactions if often difficult to disentangle using traditional in vivo approaches, we leveraged a novel stem cell co-culture system to understand how donor genetic background and cellular environment affect cellular interactions from individuals with and without an AD diagnosis. We used induced pluripotent stem cell (iPSCs) lines from 10 members of the Religious Orders Study/Memory and Aging Project (ROSMAP) from whom we have gained detailed pathological, clinical, whole-genome sequencing, and post-mortem molecular data. Of the 10 iPSC lines in this study, 4 were derived from individuals that had an AD diagnosis (AD), 3 were from individuals that had a pathological but not clinical diagnosis of AD (HP-NCI), and 3 from individuals with no pathological or clinical AD (LP-NCI). Each line or “family” was differentiated into microglia, neurons, and astrocytes separately, then combined and cultured together as triple cultures. To test for non-autonomous impacts of cells derived from AD and non-AD individuals (as well as from individuals with varying AD-related pathologies), we generated an additional “village” culture containing microglia, neurons, and astrocytes from all 10 lines. The village and each of the families were co-cultured for 3 days, then cells were dissociated and harvested for single cell RNA sequencing. After demultiplexing to define the origin individual and culture-type of each cell, we defined clusters of microglia, astrocytes, and neurons from which we could perform cell type-specific analyses. Sub-clustering of each main cell type identified diverging states including microglia defined by activation markers such as TREM2, proliferating cells, and a small population of phagocytosing cells. The state of activated microglia defined by TREM2 were more abundant in families derived from AD individuals, suggesting autonomous genetic drivers cause a predominance of this population. In addition to autonomous changes in families, we investigated how inferred ligand-receptor interactions were altered in families compared to cells derived from the same donors in villages. Several key ligands were predicted to be altered in villages such as microglia-secreted C1QB and TNF, and astrocyte-secreted IL-6, identifying molecules with a greater non-autonomous role in AD cellular dysfunction. Future work will look to validate these altered ligand-receptor interactions using proteomics-based analyses which will prioritize molecules with non-autonomous significance to AD cellular dysfunction.