Abstract
Background
Alzheimer's disease (AD) is a multifactorial disease, with pathological characteristics that are difficult to identify in the blood. While the defining mutations and gene expression (GE) signatures linked to familial AD have been characterized, the complex relationship between low risk common single nucleotide polymorphisms (SNP), GE profiles, and late onset AD is still to be elucidated.
Methods
In the current study, we assessed allelic association of a targeted selection of 2,084 AD related SNPs, GE (Affymetrix HuEx-1_0-st-v2, 22,011 probe-sets) and eQTL profiles along with enriched network decomposition (END) with amyloid beta (Ab) status in a sub-cohort (N=193) of the Australian Imaging, Biomarkers and Lifestyle (AIBL) study of ageing.
Results
From 14,825 annotated probe sets, we identified 530 that were able to separate Aβ- from Aβ+ participants (p<0.05). Using pathway (KEGG) information for these 530 genes, we identified 5,073 unique genes within 249 pathways for further analyses via graphical network analyses. Using the extra biological information from the networked set of genes, analyses with END increased the precision of the top genes to discriminate Aβ- from Aβ+ participants. Assessing the allele frequencies from 1,984 SNPs for the same participants, we identified 87 SNPs associated with Aβ status (p <0.05); 83 of these aligned within 59 genes. eQTL analyses of all SNPs across the entire sample (N=182) identified 130 SNPs associated with GE of 2,242 annotated genes (<0.05 [FDR-adjusted level]), eight of which were acting in cis and 123 in trans. Investigating the relationship in stratified Aβ status groups, and adjusting for APOE ε4 allele status, age and gender, we identified 47 genes where the relationship was altered in Aβ- as compared with Aβ+ participants. Top relationships include a SNP in the PLD3 gene (rs145999145) modifying expression in the ADGRG7 gene P=2.1*10−39, and two SNPs in LD in the FYCO1 gene modifying expression in the RBM46 gene (P=1.06*10−15).
Conclusions
eQTL analyses of blood-based expression profiles elucidated trans SNP-expression relationships to help understand the complex genomic architecture of AD. Future work will assess the downstream effects on protein levels, and possible targets for therapeutic intervention.