Druggable genome-wide Mendelian randomization analysis identifies potential treatment targets in vascular dementia

There are currently no US Food and Drug Administration-approved treatments for vascular dementia (VaD). Genome-wide approaches have successfully identified druggable targets and treatments for various disorders. In this study, we performed druggable genome-wide two-sample Mendelian randomization (2SMR) analysis to identify possible treatment targets for VaD.

2SMR analyses were used to estimate the causal effects of druggable gene expression on VaD risk. The exposure variables were significant cis-expression quantitative trait loci (eQTLs) and cis-protein quantitative trait loci (pQTLs) in the cerebrospinal fluid (CSF), brain, and plasma. The main outcome variable was genetic VaD risk, based on the Mega Vascular Cognitive Impairment and Dementia Consortium genome-wide association study. 2SMR analysis examined the causal relationship between eQTLs/pQTLs and imaging markers of VaD. A phenome-wide 2SMR analysis explored the relationships between significant druggable genes and phenotype summary statistics derived from the UK Biobank. False discovery rate (FDR) P value corrections were applied to all analyses.

A total of 12,224 druggable genes were identified from the Drug-Gene Interaction Database (DGIdb) and associated papers. Of these, the 2SMR analysis identified four FDR-significant genes in the pQTL analysis, with none identified among the eQTLs. In the CSF, TOMM40 had a significant (P = 3.67E-36) effect on VaD outcomes as well as cerebral small vessel disease (cSVD), white matter hyperintensities (WMH; P = 0.0001) and fractional anisotropy (FA; P = 0.0028). In the brain, apolipoprotein E (APOE; P = 1.90E-54) was associated with VaD and three cSVD markers: WMH (P = 1.61E-06), FA (P = 0.0018), and mean diffusivity (P = 0.0244). ERAP1 (P = 0.0163), and SAA1-4 (P = 0.0163) showed weaker associations with VaD, did not show colocalization, and were not associated with cSVD imaging markers.

This study identified four potential drug targets for VaD, using a 2SMR analysis approach. Two genes, APOE and TOMM40, are well understood to be associated with both Alzheimer's disease and VaD, whereas the other two, ERAP1 and SAA1-4, are novel targets involved in immune system regulation and inflammation.