Assessing genetic overlap and causality between blood plasma proteins and Alzheimer's Disease

Background Blood plasma proteins are modifiable and have been associated with Alzheimers disease (AD), but understanding which proteins are on the causal pathway remains challenging. ObjectiveInvestigate the genetic overlap between candidate proteins and AD using polygenic risk scores (PRS) and interrogate their causal relationship using bi-directional Mendelian Randomization (MR). MethodsFollowing a literature review, 31 proteins were selected for PRS analysis. PRS were constructed for prioritised proteins with and without the apolipoprotein E region (APOE+/- PRS) and tested for association with AD status across three cohorts (n=6244). An AD PRS was also tested for association with protein levels in one cohort (n=410). Proteins showing association with AD were taken forward for MR. ResultsFor APOE e3, apolipoprotein B-100, and C-reactive protein (CRP), protein APOE+ PRS were associated with AD below Bonferroni significance (pBonf, p-value <0.00017). No protein APOE-PRS or AD PRS (APOE+/-) passed pBonf. However, vitamin D-binding protein (protein PRS APOE-, p-value=0.009) and insulin-like growth factor-binding protein 2 (AD APOE- PRS p-value=0.025, protein APOE-PRS p-value=0.045) displayed suggestive signals and were selected for MR. In bi-directional MR, none of the 5 proteins demonstrated a causal association (p-value<0.05) in either direction. ConclusionApolipoproteins and CRP PRS are associated with AD and provide a genetic signal linked to a specific, modifiable risk factor. Whilst evidence of causality was limited, this study was conducted in a moderate sample size and provides a framework for larger samples with greater statistical power.