FROM SYSTEMS LEVEL TRANSCRIPTOMICS TO NEW IMMUNE TARGETS FOR ALZHEIMER’S DISEASE

Todd E. Golde,Nilufer Ertekin-Taner,Paramita Chakrabarty,Mariet Allen,Minerva M. Carrasquillo,Dennis W. Dickson,Steven G. Younkin,Yona Levites,Nathan D. Price,Corey Funk,Benjamen D. Heavner,Ben Logsdon,Lara M. Mangravite,Hong-Dong Li,Jeremy D. Burgess,Curtis S. Younkin,Ronald C. Petersen,Pedro E. Cruz,Carolina Ceballos-Diaz,Awilda M. Rosario

DOI: https://doi.org/10.1016/j.jalz.2016.06.383

2016-01-01

Abstract:We have been funded by the NIH/NIA (AG046139) to use systems approaches to identify new therapeutic targets for Alzheimer’s Disease (AD). We have focused on large-scale transcriptomic analysis as an entry point for guiding discovery. AD has a progressive nature, thus a comparative transcriptomic analysis across AD and different neurodegenerative pathologies in humans and in model systems can identify network-specific differences. We have transcriptome measurements from human post-mortem cohorts representing AD, n=86, Progressive Supranuclear Palsy (PSP) n=84, Pathological Aging (PA) n=30, and Elder Control brains, n=80 (temporal cortex and cerebellum). We also have mouse models of amyloidosis and tauopathy in aging series for comparison, including APP CRND8 and PS1/APP mouse brains (3-20 months of age), as well as Tau P301L mouse brains and spinal cords (rTG4510 and JNPL3, 2.5-12 months of age). To understand the transcriptional landscape we have analyzed these RNA-seq data using a variety of bioinformatic methods. Predictions from these transcriptome-based models are currently being validated and then used to guide modeling studies to manipulate key nodes with a focus on innate immune networks. To better understand the molecular drivers of differences between AD, controls and other neurodegenerative conditions, we built a variety of networks, including an initial transcriptional regulatory network based on prior knowledge of transcription factor binding from ENCODE, and expression data from our RNA-seq data and those from other groups. These networks are first steps in achieving a mechanistic understanding between transcription factors and target genes. We find that numerous networks are altered in AD and PSP brains and that these key networks include genes that regulate both immune and oligodendrocyte function. Analysis of AD vs. PSP samples enables us to determine transcript changes that distinguish AD from a primary tauopathy. Comparisons with PA brains show many fewer transcriptomic changes. Analysis of transgenic APP and tau mice show modest overlap in differentially expressed genes (DEGs) between mouse model of amyloidosis and human AD brains. In contrast, tau mice show relatively minimal number of DEGs that overlap with those that are differentially expressed in AD or PSP versus controls. We will present our latest data on perturbed networks models, and also describe how we are using these data to guide modeling studies designed to evaluate immune networks as potential therapeutic targets. There are numerous perturbed networks in AD and PSP brains, many of which are not well conserved in our mouse models of Abeta or tau pathology. As perturbed networks do not necessarily imply causality, we have begun modeling studies to determine how we will manipulate immune networks to gain insights into novel AD therapeutic targets.