Central nervous system (CNS) disorders, such as Alzheimer’s disease (AD), affect up to 1 billio people worldwide, and their incidence continues to rise with the aging of the population. Despite progress made toward understanding CNS disorders, the molecular mechanisms underlying disease development are still unknown and few effective treatments are currently available.

In fact, a major issue is the limiting effect of the blood–brain barrier (BBB), a physical and selective barrier between the brain and the systemic circulation. One approach to overcome the BBB bottleneck is to target endogenous BBB transport systems. Although previous studies have identified BBB-targets by transcriptomic and proteomic analysis in healthy mouse models, endothelial cells and pericytes exhibit the greatest transcriptional divergence between mouse and humans. Furthermore, several genes can vary upon neurodegeneration, therefore it is critical to first understand the molecular composition of the BBB in health and pathological conditions.

First, we need to obtain an accurate Gene Expression Profile (GEP) of these cells. Unfortunately, due to the low abundance of endothelial cells, their transcriptome may not be
well characterized using single-cell RNA sequencing. Instead, we will develop a novel computational method based on neural networks (NNs) to deconvolve bulk RNA-seq into the GEP of each cell type in the sample. To train the NN, we will use existing single-nucleus RNA sequencing data of AD patients and healthy individuals. We will then apply the method to deconvolve bulk RNA-seq brain data (healthy and disease) to extract the GEP of endothelial cells and compare between healthy and AD. The output will be a list of potential genes that will be validated in AD mice. The best candidate will be used to develop highly specific transmigrating antibodies fused with an anti-β- amyloid antibody that will be tested in the AD model.

With this approach, we aim to identify and characterize novel BBB targets that can be used for brain delivery for Alzheimer therapy and furthermore, provide precise cellular targets for AD therapeutic development.