ε4, with AD, to gain some insights into the APOE isoform's functional properties which are protective or deleterious in the process of AD pathogenesis. This work led to the discovery of two novel rare missense variants located near and within APOE’s lipid binding domain. We demonstrated that these are associated with a two-to-three-fold decreased risk of developing AD. This result was obtained by testing the association of these variants in 400,000 European ancestry individuals, with about 1 every 1,000 carrying either of the two variants (V236E and R251G) and was published in JAMA Neurology (PMID: 35639372). In addition, another part of this project focused on APOE variants specific to African ancestry individuals. This is essential to guarantee an equitable precision medicine, albeit ensuring that AD genetic risk is assessed with the same quality in all populations, and also enables us to gain additional insights on APOE by studying variants not present in European ancestry. This work led to the discovery of a missense variant within APOE’s receptor binding region (R145C) which was associated with a five-to-eight-fold increased AD risk when combined with ε4. This result, published in JAMA (PMID: 36809323) was obtained by testing the association of this variant in 32,000 African ancestry individuals, with about 4% carrying R145C and 1.5% carrying the combination of ε3[R145C]/ε4 and displaying a similar risk of developing AD as ε4/ε4 individuals with significantly earlier disease onset.
Additionally, this project also identified novel genetic variants associated with AD risk by developing new statistical methods and results were published in Alzheimer’s Research & Therapy (PMID: 33794991) and The American Journal of Human Genetics (PMID: 34767756). These results identified genetic associations in novel and known genetic pathways and will foster functional studies to identify the mechanism by which these genes modify AD risk. Given the increased risk of AD in women, this project is also investigating the association of X chromosome (females have two copies, while males have one) variants with AD, which has until now been excluded from AD genetic association. This X-chromosome-wide association analysis in AD did not lead to any significant results, additional efforts are ongoing to increase our sample size by including FinnGenn and the Million Veteran Project.
Last, this project tackled the fine-mapping of the Human Leukocyte Antigen (HLA) locus in AD, which is one of the genetic loci (a region of the genome) identified in AD genome-wide association studies but for which the causal gene(s) remained undetermined. First, this project demonstrated that the protective association observed at the HLA locus is common to some other neurodegenerative diseases, notably Parkinson’s disease (PD) and Amyotrophic lateral sclerosis (ALS). Second, we gathered data from across the globe to perform HLA fine-mapping, which not only increases statistical power but also ensures to cover of a large diversity of HLA alleles. By studying over 120,000 AD cases and 400,000 cognitively healthy older controls, we determined that the protective association was supported by HLA-DRB1*04 allele subtypes and similar associations were observed in PD by studying over 55,000 PD cases and 1,450,000 non-PD controls (PMID: 37643212). We then demonstrate that protective HLA-DRB1*04 subtypes strongly bound the aggregation-prone tau PHF6 sequence, however only when acetylated at a lysine (K311), a common posttranslational modification central to tau aggregation. This HLA-DRB1*04-mediated adaptive immune response, common to AD and PD, offers the possibility of new therapeutic avenues.
