Summary-data-based Mendelian Randomization Analysis Identifies Nominal Evidence for Association of N6-Methyladenosine Genetic Variation with Alzheimer’s Disease

<i>N</i>6-methyladenosine (m6A) is an abundant post-transcriptional RNA modification that critically regulates brain function. Dysregulation of m6A signaling has been implicated in several neurological diseases, including Alzheimer's disease (AD). However, whether genetic variation associated with the risk of AD is mediated via m6A-dependent gene regulation is currently unknown. Here we investigated the association of m6A with the risk of AD using the summary-data-based Mendelian randomization (SMR) approach. By integrating m6A quantitative trait loci (m6A-QTLs) and genome-wide association study (GWAS) summary data for AD, we identified six nominally significant m6A-AD associations (uncorrected P_SMR < 0.05, P_HEIDI ≥ 0.01 with ≥5 SNPs), although none remained significant after false discovery rate (FDR) correction. We performed targeted SMR analyses for AD using brain- and blood-based expression QTL summary data, restricting instrumental variables to a set of 18,606 single nucleotide polymorphisms (SNPs) previously identified as m6A-related sites. This analysis identified 75 FDR-significant genes associated with the risk of AD via changes in gene expression (FDR < 0.05, P_HEIDI ≥ 0.01 with ≥5 SNPs); however, the instrumental SNPs for these genes showed no enrichment for m6A-QTLs. In summary, we found limited evidence for the direct association of m6A genetic variation with the risk of AD. Larger m6A-QTL datasets will be required to establish whether m6A variation is associated with the risk of AD.