P1‐139: the contribution of sex‐specific associations in genetic studies of alzheimer's disease pathology

Women are disproportionately affected by Alzheimer's disease (AD) and more likely to manifest clinical symptoms in the presence of AD neuropathology. Yet, sex differences are often ignored or modeled as a nuisance variable in genetic association studies of AD. We performed comprehensive sex-specific genetic analyses of AD biomarkers and autopsy measures of AD neuropathology to evaluate whether sex-specific genetic effects contribute to AD pathogenesis. First we performed a sex-stratified GWAS in 5,978 participants with autopsy measures of neurofibrillary tangles (Braak staging) and neuritic plaques (CERAD score). Then we used the innovative UTMOST (Unified Test for MOlecular SignaTures) method in 3,119 participants with CSF measures of Aβ42 and total tau to test for sex-specific associations with predicted gene expression. In these participants, gene expression profiles were jointly imputed in 44 tissues available in GTEx. Cross-tissue gene-level association results for 14,709 genes were then constructed using a generalized Berk-Jones test. Bonferroni-corrected significance thresholds for GWAS and UTMOST analyses were α=5.0x10−8 and α=3.4x10−6, respectively. In the GWAS, rs34331204 (intergenic; 7p21.1; MAF=0.07) was associated with presence of tangles in males (β=-0.74, p=9.1x10−9) but not females (β=-0.03, p=0.85). In the UTMOST analysis, we found a female-specific association between predicted expression of KLDR1 and CSF Aβ42 (pmales=0.13; pfemales=2.4x10−6). Male-specific associations were found for predicted expressions of PPP2R1B with CSF Aβ42 (pmales=1.5x10−7; pfemales=0.81) and NOTCH3 with CSF tau (pmales=5.6x10−7; pfemales=0.62). Results highlight how sex differences can enhance our understanding of the molecular basis and genetic architecture of AD. Among females, we identified KLDR1, an inhibitor of natural killer cells. This female-specific role of innate immunity is consistent with our previous findings. Among males, we identified rs34331204, a cis eQTL for BZW2 (BRAINEAC aveALL p=1.9x10−5), which has been implicated in neuronal differentiation. Additionally, we identified NOTCH3 and PPP2R1B, a regulatory subunit of PP2A, which are involved in Notch and Wnt signaling cascades. Combined with our past results, these findings suggest the innate immune pathway may be particularly relevant to AD pathogenesis in females, while the neurogenesis pathway may be particularly relevant in males. Future work will functionally assess these sex-specific candidate genes.