Identification of conserved proteomic networks in neurodegenerative dementia

Abstract

SummaryData-driven analyses of human brain across neurodegenerative diseases possess the potential for identifying disease-specific and shared biological processes. We integrated functional genomics data from postmortem brain, including label-free quantitative proteomics and RNA-seq based transcriptomics in an unprecedented dataset of over 1000 individuals across 5 cohorts representing Alzheimer’s disease (AD), asymptomatic AD, Progressive Supranuclear Palsy (PSP), and control patients, as a core analysis of the Accelerating Medicines Project – Alzheimer’s Disease (AMP-AD) consortium. We identified conserved, high confidence proteomic changes during the progression of dementias that were absent in other neurodegenerative disorders. We defined early changes in asymptomatic AD cases that included microglial, astrocyte, and immune response modules and later changes related to synaptic processes and mitochondria, many, but not all of which were conserved at the transcriptomic level. This included a novel module C3, which is enriched in MAPK signaling, and only identified in proteomic networks. To understand the relationship of core molecular processes with causal genetic drivers, we identified glial, immune, and cell-cell interaction processes in modules C8 and C10, which were robustly preserved in multiple independent data sets, up-regulated early in the disease course, and enriched in AD common genetic risk. In contrast to AD, PSP genetic risk was enriched in module C1, which represented synaptic processes, clearly demonstrating that despite shared pathology such as synaptic loss and glial inflammatory changes, AD and PSP have distinct causal drivers. These conserved, high confidence proteomic changes enriched in genetic risk represent new targets for drug discovery.HighlightsWe distinguish robust early and late proteomic changes in AD in multiple cohorts.We identify changes in dementias that are not preserved in other neurodegenerative diseases.AD genetic risk is enriched in early up-regulated glial-immune modules and PSP in synaptic modules.Almost half of the variance in protein expression reflects gene expression, but an equal fraction is post-transcriptional or -translational.