A polygenic risk score for mosaic loss of chromosome Y susceptibility is associated with higher risk of MCI to AD conversion

Background Mosaic loss of chromosome Y (mLOY) is a highly common somatic variant among men, and has been associated to higher risk in overall mortality and several types of disease, including Alzheimer’s disease (Dumanski et al. 2016). In the present study, we aimed to replicate these findings in both a cross‐sectional and longitudinal setup by determining mLOY phenotype and its associated polygenic risk score (PRS). Method MADloy R package (González et al. 2020) was used to determine mLOY in germline blood DNA from males in the GR@ACE cohort, composed of AD patients recruited in Fundació ACE (Barcelona, Spain) and population‐based controls recruited from several Spanish centres. Additional MCI patients recruited in Fundació ACE were used to study disease progression. PRS was calculated by additive multiplication of the beta values of 114 genome‐wide significant variants reported in the most relevant mLOY GWAS published to date (Thompson et al. 2019) and their genotypes in our dataset. To ease interpretation of results, PRS per SD units were used. Only individuals with ages 65‐85 were kept for the case‐control and PRS analyses (see Table 1). Statistical analysis was performed by fitting logistic regressions and Cox proportional‐hazards models adjusted by age, APOE and PCs (when needed). Result mLOY phenotype did not show a significant association with AD in the case‐control setup and showed suggestive results for higher risk of conversion to AD (HR=1.45; p=0.09). However, the mLOY PRS yielded borderline significance in the case‐control dataset (OR=1.08; p=0.09), and statistical significance in MCI to AD conversion (HR=1.22; p=0.01). Importantly, PRS was independent of age, and effects were only observed in male samples, with no effect in conversion to AD or case‐control in the female subset, supporting the validity of our approach. Conclusion The age‐dependent nature of mLOY and lack of age at sampling information for most controls limited our power to study mLOY phenotype’s effect on AD. The mLOY PRS allowed us to overcome these limitations, acting as a male‐specific AD risk factor and providing further evidence of mLOY’s impact on AD.