Proteins are the key players in all life processes, in health and disease, and the vast majority of drug
targets are proteins. High throughput detection and quantification of the proteome in complex tissues
such as blood has historically been hampered by the limitations of available methods. Recent progress
allows proteomics to be measured in large populations. An aptamer-based proteomics platform was
recently developed to measure 4137 proteins in the serum of 5457 individuals from the phenotypically
and genetically well-characterized population-based AGES Reykjavik study, of which 3289 participated
in a five-year follow-up study.
We have shown that
serum proteins cluster into co-regulatory networks
that were aligned with common diseases and as intermediaries between genetic variation driving
disease and with the appearance of disease across the population. In the proposed study, w
e will
leverage genetics, proteomics and network-based data across multiple brain related outcomes, using
both cross-sectional and longitudinal study design, to describe the relationships of global serum
proteins to each other, to genetics and to late-onset Alzheimer's disease (LOAD). With increasing aging
of the population the prevalence of LOAD is on the rise, making effective LOAD treatment one of the
fastest growing unmet medical needs in global healthcare systems. We propose three different aims. In
Aim 1, we will examine the association of 4137 serum proteins as well as amyloid-ß peptides and
protein networks, at baseline and five years later, to prevalent and incident LOAD. In Aim 2, we will
determine association of the 4137 serum proteins to structural brain MRI biomarkers of
neurodegeneration for further links to LOAD and brain atrophy progression rates. In Aim 3, we will
address causality between proteins and LOAD through bi-directional Mendelian randomization analysis,
using variants that regulate serum proteins as genetic instruments. To our knowledge this is the largest
proteomics data set to date as regards number of proteins measured and samples screened. The
proposed project is significant for many reasons including for instance: i. the large-scale proteomics
data integrated with variation in the genome and deep phenome data facilitates systems approaches to
LOAD. ii. The application of longitudinal data to assess intra-individual changes in protein biomarker
levels over time on long-term LOAD related outcomes offers unique opportunities for robust biomarker
discoveries in LOAD. iii. Whole genome analysis of global serum proteins will provide the scientific
community with a novel source of genetic instruments for tests of causality and to reveal the causal
mechanism(s) underlying the risk of LOAD. In summary, the results of the proposed project will offer a
holistic model of the etiology of LOAD yielding novel biomarkers and point to actionable targets for
treating the disease.
