Abstract
Alzheimer's disease (AD) represents a substantial unmet need, due to increasing prevalence in an ageing society and the absence of a disease modifying therapy. Epidemiological evidence shows a protective effect of non steroidal anti inflammatory (NSAID) drugs, and genome wide association studies (GWAS) show consistent linkage to inflammatory pathways; both observations suggesting anti-inflammatory compounds might be effective in AD therapy although clinical trials to date have not been positive. In this study, we use pathway enrichment and fuzzy logic to identify pathways (KEGG database) simultaneously affected in both AD and by NSAIDs (Sulindac, Piroxicam, Paracetamol, Naproxen, Nabumetone, Ketoprofen, Diclofenac and Aspirin). Gene expression signatures were derived for disease from both blood (n = 344) and post-mortem brain (n = 690), and for drugs from immortalised human cell lines exposed to drugs of interest as part of the Connectivity Map platform. Using this novel approach to combine datasets we find striking overlap between AD gene expression in blood and NSAID induced changes in KEGG pathways of Ribosome and Oxidative Phosphorylation. No overlap was found in non NSAID comparison drugs. In brain we find little such overlap, although Oxidative Phosphorylation approaches our pre-specified significance level. These findings suggest that NSAIDs might have a mode of action beyond inflammation and moreover that their therapeutic effects might be mediated in particular by alteration of Oxidative Phosphorylation and possibly the Ribosome pathway. Mining of such datasets might prove increasingly productive as they increase in size and richness.