Causal association and mediating effect of blood biochemical metabolic traits and brain image-derived endophenotypes on Alzheimer's disease

Abstract

Background: Recent genetic evidence supports that circulating biochemical and metabolic traits (BMTs) play a causal role in Alzheimer's disease (AD), which might be mediated by changes in brain structure. Here, we leveraged publicly available genome-wide association study data to investigate the intrinsic causal relationship between blood BMTs, brain image-derived phenotypes (IDPs) and AD.

Methods: Utilizing the genetic variants associated with 760 blood BMTs and 172 brain IDPs as the exposure and the latest AD summary statistics as the outcome, we analyzed the causal relationship between blood BMTs and brain IDPs and AD by using a two-sample Mendelian randomization (MR) method. Additionally, we used two-step/mediation MR to study the mediating effect of brain IDPs between blood BMTs and AD.

Results: Twenty-five traits for genetic evidence supporting a causal association with AD were identified, including 12 blood BMTs and 13 brain IDPs. For BMTs, glutamine consistently reduced the risk of AD in 3 datasets. For IDPs, specific alterations of cortical thickness (atrophy in frontal pole and insular lobe, and incrassation in superior parietal lobe) and subcortical volume (atrophy in hippocampus and its subgroups, left accumbens and left choroid plexus, and expansion in cerebral white matter) are vulnerable to AD. In the two-step/mediation MR analysis, superior parietal lobe, right hippocampal fissure and left accumbens were identified to play a potential mediating role among three blood BMTs and AD.

Conclusions: The results obtained in our study suggest that 12 circulating BMTs and 13 brain IDPs play a causal role in AD. Importantly, a subset of BMTs exhibit shared genetic architecture and potentially causal relationships with brain structure, which may contribute to the alteration of brain IDPs in AD.

Keywords: Alzheimer's disease; Blood biochemical and metabolic traits; Brain image-derived phenotypes; Mediating effect; Mendelian randomization.

Fig. 1

Study design flowchart of the Mendelian randomization study.

Fig. 2

Mendelian randomization analysis of blood BMTs (IVW or Wald ratio FDR＜0.05) and AD. Glutamine from three data aggregations with different sources was classified as a type of trait in this study. We identified 12 blood BMTs that support a causal association with AD. AD, Alzheimer's disease; 1-PGPE, 1-palmitoyl-glycero-phosphoethanolamine; S_HDL_CE, cholesteryl esters in small high density lipoprotein; IDL_PL_pct, Phospholipids to total lipids ratio in intermediate density lipoprotein; XXL_VLDL_C_pct, cholesterol to total lipids ratio in chylomicrons and extremely large very low density lipoprotein; XXL_VLDL_CE_pct, cholesteryl esters to total lipids ratio in chylomicrons and extremely large very low density lipoprotein; IVW, inverse variance weighting; SNP, single nucleotide polymorphism.

Fig. 3

Mendelian randomization analysis of brain IDPs (IVW or Wald ratio p-value＜0.05) and AD. We identified 13 brain IDPs that support a causal association with AD. Cortical thickness: odds ratio per one standard deviation increase of cortical thickness. Subcortical volume: odds ratio per one standard deviation decrease of volume. AD, Alzheimer's disease; lh, left; rh, right; IVW, inverse variance weighting; SNP, single nucleotide polymorphism. We visualized the alteration of IDPs based on the anatomical automatic labeling (AAL) brain atlas [63]. However, some IDPs (left choroid plexus, bilateral white matter, fourth ventricle, and hippocampus subarea) are not included in the AAL brain atlas, so we didn't show them.

Fig. 4

Mendelian randomization analysis of 12 blood BMTs and 13 brain IDPs. “+” indicates that there is a causal effect between the two traits (IVW or Wald ratio P-value＜0.05). Glutamine from three data aggregations with different sources was classified as a type of trait in this study. Our results revealed a causal relationship between 5 blood BMTs and 4 brain IDPs.