This project recently established a more accurate and efficient model to identify fluorescent images generated by treating the Alzheimer’s in a Dish model with compounds that can clear accumulated phosphorylated tau. Meanwhile, 29 new candidate compounds recently going through clinical trials for various diseases were generated by in silico prediction and validated using the ADiD model. Some 23 of these 29 validated predictions cleared p-tau by more than 95%. These validated predictions include compounds originally designed for treating cancer, autoimmune diseases or metabolic diseases; we now are working to determine why these compounds can clear p-tau and how these candidates can lead us to novel therapeutic options for AD. Our mechanism study allows us to identify subgroups among known screening hits that may share similar ways of clearing p-tau.
This project seeks to advance drug discovery for Alzheimer’s disease through the 3-Dimensional Drug Screening consortium. Via collaboration with the Tanzi and Kim labs at Massachusetts General Hospital, this research intends to combine bioinformatics-based screening and modeling methods with the 3-D human neural cell culture system of Alzheimer’s in a Dish to test the drug candidate, Ebselen. Using the SMART framework (SysteMatic Alzheimer’s disease drug ReposiTioning), 2,640 carefully selected compounds were physically screened, 30 compounds were found to inhibit phosphorylated tau in 3-D culture, and Ebselen was found ready for animal studies to test long-term toxicity and efficacy. Ebselen is a drug molecule with anti-oxidant and anti-inflammatory activity.
Partnering with the research consortium led by Massachusetts General Hospital, we successfully developed the SysteMatic Alzheimer’s disease drug ReposiTioning (SMART) framework to identify candidate drugs for repurposing for Alzheimer’s in a previous grant from Cure Alzheimer’s Fund. A high-throughput screening using the Alzheimer’s in a Dish model and a library of more than 2,000 compounds identified 38 preliminary hits, three of which can achieve almost complete inhibition of accumulation of phosphorylated tau (p-Tau). Using these preliminary hits as “baits,” SMART took advantage of public available large cellular perturbation response data (more than 20,000 drugs and compounds) and predicted and validated nine clinically used compounds beyond the original library that can ignite a similar cellular phenotype as the original top three preliminary hits, i.e., almost complete inhibition of p-Tau accumulation. Compared with the high-content drug screening, SMART improved the success rate of hit identification by more than 50-fold en route to quadrupling the panel of candidates for fast-track drug repositioning study. Thus, the integrative neurobiology framework has shown its potential as a powerful, cost-effective platform for drug discovery and mechanism study in Alzheimer’s disease.
In this proposal, we will expand the modeling capability of SMART with a mechanism discovery module grounded on advanced machine learning techniques to investigate and validate novel mechanism for inhibition of p-Tau accumulation regarding the drug hits discovered in the prior funding period. The proposed study has two specific aims: Aim 1 will construct an image-omics workflow to uncover the molecular mechanism underlying representative FDA-approved compounds that block AD neuropathogenic events, while Aim 2 will evaluate the selected drug hits in cell assays with validation results feeding back to Aim 1 to ensure the efficacy of drug repositioning and mechanism discovery. The success of this research would enable better understanding of novel mechanisms of known drugs identified and lead to new, cost-effective treatment targets for Alzheimer’s.