TREM2 Genotype-Informed Drug Repurposing and Combination Therapy Design for Alzheimer’s Disease

PROJECT SUMMARY The cumulative evidence indicates neuro-inflammation play crucial roles in Alzheimer’s disease (AD) and anti- inflammatory agents (i.e., microglia-targeted therapies) show potential treatments for AD. The treatment window for microglia-targeted therapies may at least be open later in AD due to the nature of microglial biology. If amyloid-β is the trigger for AD, and tau is the executioner, then microglia are accelerators of disease progression. Microglia-targeted approaches have a better chance of success in mild to moderate disease compared to anti-amyloid therapies that have to prevent the trigger early in AD. Among the microglial genes, triggering receptor expressed on myeloid cells 2 (TREM2) has received much attention because a hemizygous R47H variant of TREM2 (TREM2R47H) increases the AD risk by 2-7 folds in various populations, including African American. Our preliminary single-nucleus RNA-sequencing (snRNAseq) study identified disease-associated microglia (DAM) with enhanced AD pro-inflammatory signatures (i.e., activation of Akt-signaling) associated with the TREM2R47H variant and pharmacological Akt inhibition reversed the TREM2R47H induced inflammation, showing proof-of-concept of TREM2R47H targeted drug discovery in AD. Using multimodal snRNAseq analysis, we identified two approved anti-inflammatory asthma drugs (fluticasone and mometasone) that were predicted to modulate DAM molecular networks are associated with a reduced incidence of AD in electronic health record (EHR) data of 7 million patients. We therefore posit that understanding how the AD-linked mutations (i.e., TREM2R47H) enhance microglial toxicity could lead to understanding how microglia become maladaptive/toxic and development of microglia-targeted therapeutic strategy for late-onset sporadic AD in general. This application calls for novel microglia-targeted drug repurposing and combination therapies for AD using our well-established multi-omics and deep learning-based EHR approaches, and functional observations in AD patient-induced pluripotent stem cells (iPSCs), cerebral organoids, and mouse models, with three specific Aims. Aim 1 will test the TREM2R47H informed microglia-targeted therapeutic hypothesis for identifying cell type-specific molecular drivers/networks, repurposable drugs, and combination therapies for AD using multi-omics evidence aggregation. These analyses will leverage our existing snRNAseq data (n = 24 TREM2R47H and n = 23 common variant (CV)-TREM2 with matched age, AD pathology and APOE genotypes) and public genomic data from the AD knowledge portal. Aim 2 will identify repurposable drugs and combination therapies using high-throughput EHR-based hypothesis generation and sequential deep learning-based propensity score matching approaches. We will leverage de-identified EHRs from the INSIGHT networks (~11 million patients across New York City’s 5 health systems and the greater metropolitan area). Aim 3 will screen, test and validate drugs using AD patient- derived iPSC lines carrying TREM2R47H in conjunction with cerebral organoid and mouse models. The successful completion of this project will offer a viable strategy to move AD drug repurposing from bench to bedside rapidly. 1