Genetic Associations with Age at Dementia Onset in the PSEN1 E280A Colombian Kindred

Background: Genetic associations with Alzheimer's disease (AD) age at onset (AAO) have revealed genetic variants with the potential for therapeutic application. Such studies in late onset AD are limited by population heterogeneity and co-morbidities associated with aging. Studies to date in ADAD (autosomal dominant AD) have been limited by their small sample size. The large Colombian kindred presented here provides a unique opportunity to discover AAO genetic associations. Methods: A genetic association study was conducted for AAO in 344 individuals with the PSEN1 E280A mutation via array imputation and whole genome sequencing in a subset of 80 individuals. Scrambled age conditions were used as a control, and replication was assessed using three studies on AD age of onset, age of onset survival, and meta-analysis. Results: The proportion of variance explained (+/- 95% CI) was 56% (+/-15%). 29 loci reached genome-wide significance, of which 23 had a GWAS hit from the NHGRI-EBI catalog within 500 kb relevant to neurodegeneration. Hits included a new variant in the CLU locus, rs35980966, (which replicated), a variant in a locus on a separate chromosome previously associated with plasma clusterin (rs138295139), and a missense variant in SORBS3 (rs34059820). Former GWAS index variants at the CLU locus also replicated in this cohort (rs4236673, rs9331896). APOE {epsilon}4 (rs429358) and APOE {epsilon}2 (rs7412)-associated variants exhibited modest (less than 3 years) associations with earlier and later age of dementia onset, respectively. Other nominal associations included coding variants in IL34 (rs4985556), TSPAN10 (rs6565617, rs7210026), STIM2 (rs1457401458), HTT (rs1473464204), and KCNT1 (rs557219607). Interpretation: A large proportion of the variability in AAO was explained by genetic variation at numerous loci. The unique demography of this population as a tri-continental admixture that passed through a bottleneck about 500 years ago might predict that drift would uncover rare variants with a large effect size on AAO. Indeed, candidates for large contributions from rare alleles were observed. Common variation, presumably ancestral to the bottleneck, was also associated with AAO. The detection of these effects in the presence of a strong mutation for ADAD reinforce the potentially impactful role of the identified variants.