Rationale:
Accumulating evidence suggests neurodegeneration is a systemic metabolic failure. We propose an “Eye-Brain-Body” axis hypothesis: a specific “pro-inflammatory lipid” phenotype-characterized by Small VLDL particles (high endothelial permeability) and GlycA (chronic inflammation)-acts as a “dual-hit” stressor, causing synchronous Neurovascular Unit (NVU) failure in both the retina and brain.
Objectives:
1. Define Metabolic Risk: Utilizing Nightingale NMR data (N~500k), we will isolate the “Small VLDL-GlycA” interaction phenotype and validate its predictive power for dementia using Cox models, moving beyond traditional lipid profiles.
2. Map Eye-Brain Synchronization: We will apply Sparse Canonical Correlation Analysis (sCCA) to the paired Eye-Brain imaging cohort (N~100k). Crucially, to prevent head-motion artifacts from mimicking cortical atrophy, we will implement the “Euler Ladder” topological quality control strategy. We aim to identify structural covariance between retinal layer thinning (OCT) and hippocampal atrophy (MRI).
3. Uncover Molecular Mechanisms: Using UKB-PPP Olink proteomics (N~54k), we will employ High-dimensional Mediation Analysis (HIMA) to identify circulating proteins (e.g., GDF15, NfL) that mediate the pathway from metabolic stress to NVU damage.
4. Establish Causality: We will use Network Mendelian Randomization (Network MR) with Steiger filtering to strictly infer causal directions, controlling for the high pleiotropy of lipid genes.
Scientific Value:
We aim to validate retinal imaging as a non-invasive “holographic window” for brain health. By integrating metabolomics, proteomics, and imaging, this study will reveal specific targets for early intervention in Alzheimer’s disease.