Abstract
Treatments for Alzheimer's disease (AD) have to date met with limited success but potentially promising approaches involve attempting to reduce expression of one of the genes coding for key proteins involved in AD pathogenesis. Potential targets for such interventions include APOE, APP and MAPT but there is little published evidence to help assess whether reducing expression of one of these genes might cause adverse effects. We investigated 470,000 exome-sequenced UK Biobank participants to identify those carrying a naturally occurring loss of function (LOF) variant in these genes to see if they had impaired general health or educational attainment. For APOE we additionally tested for association with hyperlipidaemia and for APP and MAPT we tested for associations with mental health and epilepsy. For APOE, APP and MAPT there were respectively 56, 118 and 333 carriers of LOF variants. For each gene, there was no association between carrying a LOF variant and any of the phenotypes tested and for some phenotypes confidence intervals of effect size were small, ruling out any large effect. The findings provide some reassurance that interventions aimed at reducing expression of one of these genes may not necessarily be associated with significant adverse effects, though the small number of phenotypes examined means we cannot be confident that such treatments would be completely safe. Since LOF variants are very rare, further investigation of the effects of reduced gene expression can be undertaken within the context of clinical trials. This research has been conducted using the UK Biobank Resource.</p>