Welcome to the AD/PD™ 2024 Interactive Program
The conference will officially run on Western European Standard Time (Lisbon, UTC+0)
To convert the conference times to your local time Click Here
WHOLE GENOME-WIDE SEQUENCE ANALYSIS OF LONG-LIVED FAMILIES (LLFS) IDENTIFIES MTUS2 GENE ASSOCIATED WITH LATE ONSET ALZHEIMER’S DISEASE
Abstract
Aims
Late Onset Alzheimer’s disease (LOAD) has a strong genetic component. Participants in the Long-Life Family Study (LLFS) exhibit a delayed age at onset of dementia, offering a unique opportunity to investigate the genetics of LOAD. Characterizing the genetic factors in LOAD can identify diagnostic biomarkers and therapeutics. This study aimed to examine the association between sequence variants and LOAD in the LLFS cohort, and to investigate whether the LLFS genetic associations generalize to cohorts with different risk of dementia.
Methods
The LLFS is a family study designed to assess genetic and environmental risk factors associated with exceptional longevity. Additional studies with a wide range of LOAD risk included high risk cohorts (familial LOAD and Down Syndrome), Alzheimer’s disease referral cohort, and population-based cohorts. We conducted a genome-wide analysis of 3,475 LLFS members. Genetic associations identified in LLFS were examined in six independent studies (N=14,260). Association analysis in a sub-sample of the LLFS cohort (N=1,739) evaluated the association of LOAD variants with beta amyloid levels.
Results
We identified SNPs in tight linkage disequilibrium within the MTUS2 gene associated with LOAD (rs73154407, p=7.6 x10-9). The MTUS2 gene-based test reached genome-wide statistical significance (p=2.6 x 10-8). The association of MTUS2 with LOAD was observed in the five independent cohorts (NIA-LOAD p=7.7 x 10-4, ABC-DS p=0.009, omicsADDS p=0.009, ROSMAP p=4 x 10-4, ADGC p=3.7 x 10-5, and WHICAP p=0.002). The association of MTUS2 with LOAD in LLFS appeared significantly stronger within high levels of Aβ42/40 ratio (p=3 x 10-6).
Conclusions
Our results identified MTUS2 as a novel LOAD locus. The gene encodes a microtubule scaffold protein implicated in the development and function of the nervous system, thus making this a plausible candidate to further investigate LOAD functional biology.
LIFESTYLE FOR BRAIN HEALTH SCORE, GENETIC RISK AND INCIDENCE OF DEMENTIA
Abstract
Aims
Evaluating whether genetic susceptibility modifies the association between lifestyle-related factors with the risk of dementia is primordial for prevention. We leveraged a large cohort to investigate the association between increasing number of unhealthy lifestyle-related factors, reflected by the lifestyle for brain-health (LIBRA) score, and the incidence of dementia and cognitive decline, studying effect modification by genetic risk.
Methods
We included participants from the Three-City study, aged >=65 years and free of dementia at baseline and followed for up to 17 years. The LIBRA was constructed including twelve factors and genetic predisposition to dementia was reflected by the epsilon 4 allele of the apolipoprotein E gene (APOEε4) and a genetic risk score (GRS). We used Cox models and linear mixed models to investigate the incidence of dementia and cognitive trajectories, respectively, according to the LIBRA, genetic instruments and their interactions.
Results
Among 5,170 participants (mean age 74 years) followed-up for 8.4 years on average, 652 incident dementia cases were adjudicated. The hazard ratios for dementia per +1point of LIBRA (range -5.9 to 11.2) were: 1.09 (95%CI, 1.05; 1.13) in APOEε4 non-carriers and 1.15 (1.08; 1.22) in carriers; 1.10 (1.04; 1.16) in low and 1.12 (1.06; 1.18) in high GRS tertile; with no evidence of effect modification by genetics (P>0.15 for interactions). In all genetic strata, increasing LIBRA was associated with worse initial cognition and steepest cognitive decline (P ≥ 0.11 for interactions).
Conclusions
In this large cohort of older persons, increasing number of lifestyle-related factors was associated with higher dementia risk and greater cognitive decline in all levels of genetic risk. These findings suggest prevention programs with targeted lifestyle modifications may be efficient even in those with highest genetic risk for dementia.
TRANSCRIPTIONAL ANALYSIS OF MOUSE MODELS HARBORING CODING AND NONCODING HUMAN GENETIC VARIANTS FOR LATE-ONSET ALZHEIMER’S DISEASE
Abstract
Aims
Determining the precise genetic mechanisms that contribute to LOAD, both in coding and noncoding variants, will enable a deeper understanding of pathogenesis and advance preclinical models for the testing of targeted therapeutics.
Methods
We have introduced candidate genetic variants in the EPHA1, BIN1, CD2AP, SCIMP, KLOTHO, PTK2B, ADAMTS4, IL1RAP, and PTPRB loci into a sensitized mouse model already harboring humanized amyloid-beta, APOE4, and Trem2.R47H alleles knocked in to a C57BL/6J background. Variants were selected based on predicted function, cross-species conservation, increased risk of LOAD, and allele frequency. Coding variants were chosen in PTPRB and KLOTHO while promoter and/or enhancer variants were modeled for the remaining loci. Genome editing with CRISPR-Cas9 was performed and mouse cohorts were aged to four and 12 months. Homogenized brain hemispheres were assayed from both male and female mice with RNA-seq. Transcriptomic changes were compared to postmortem human brain data to determine disease relevance.
Results
Transcriptomic effects from these genetic variants recapitulated a variety of human gene expression patterns observed in LOAD study cohorts. By 12 months of age, PTPRB*D57N mice exhibited neuroimmune signatures that correlate with postmortem LOAD cases relative to controls. The BIN1 promoter variant exhibited both neuroimmune and oligodendrocyte-related changes that correlated with LOAD modules. A second, potentially microglial-specific enhancer in BIN1 altered immune processes. These changes were more pronounced with age, supporting their role in age-related dementia.
Conclusions
We have characterized in vivo signatures of nine genetic candidates for LOAD, identifying alterations in specific LOAD-related pathways in each variant on a sensitized genetic background. These results provide animal models for preclinical testing of therapeutics designed to correct specific molecular alterations that contribute to LOAD pathology and progression.
SINGLE-NUCLEUS EPIGENOMIC CHARACTERIZATION OF ALZHEIMER’S AND PICK’S DISEASE.
Abstract
Aims
Tauopathies represent a class of neurodegenerative disorders distinguished by the presence of tau-positive inclusions within neurons or glial cells. These conditions can be pathologically categorized based on the predominant tau isoforms found within these inclusion bodies. In primary tauopathies such as Pick's disease (PiD), the prevailing tau isoform is typically 3R tau, whereas secondary tauopathies like Alzheimer's disease involve both 3R and 4R isoforms. Scientific investigations have unveiled that alterations in tau, such as its overexpression or hyperphosphorylation, can exert influence on chromatin structure. In the context of our study, we embarked on characterizing epigenomic changes within the human brain afflicted by Alzheimer's disease (AD) and Pick's disease (PiD) at a remarkably precise single-cell resolution.
Methods
We generated single-nucleus ATAC-seq (10x genomics) from pre-frontal cortex of human AD, PiD, and respective control cases (n=8-12/group). We used Signac and ArchR pipelines for peak-calling and downstream processing. We used Cicero for candidate cis-reulatory elements (cCREs). Finally we used our data to predict an unknown human-gained enhancer of a critical GWAS gene and gained deep insights into disease biology.
Results
We applied UMAP dimensionality reduction and Leiden clustering to the batch-corrected epigenomic datasets, identifying distinct cell-type clusters from 198,722 nuclei in snATAC-seq. We established enhancer-promoter regulatory links and subsequently constructed disease-specific and shared transcription factor (TF) regulatory networks for AD and PiD.
Conclusions
We have performed a genome-wide fine-mapping of GWAS variants of AD and FTD, which will be critical for functional validation of non-coding GWAS variants and understand disease biology. Our analysis reveals shared and distinct changes in open chromatin changes in AD and PiD. Furthermore, the experiential validation provided compelling evidence of the enhancer's functional relevance, underscoring its potential significance in the regulatory landscape.
DNA METHYLATION ASSOCIATES WITH REGIONAL TAU PATHOLOGY BEYOND AGE AND APOE
Abstract
Aims
Alzheimer’s Disease (AD) is partially characterized by neocortical dissemination of neurofibrillary tangles (NFTs) while Primary Age Related Tauopathy (PART) has NFTs that are confined to hippocampus only. Thus, PART and AD represent extremes of a spectrum of NFT spread in the brain: from relatively protected to high risk, respectively. Given inter-individual variation in NFT spread, only partially accounted for by age and APOE, we investigated the extent to which age-independent epigenetic mechanisms contribute to NFT severity.
Methods
We evaluated DNA methylation (DNAm) using the SeSAMe package from dorsolateral prefrontal cortex of Religious Orders Study and Memory and Aging Project (ROSMAP; N=707). We trained age-adjusted elasticnet regression models to predict NFT burden in middle frontal cortex, hippocampus, and a summary score from 5 regions (“total”) using 5-fold cross-validation. Multiple regression models then assessed variance explained and error in prediction of NFTs in models that include age, sex, APOE with and without elastic net identified CpGs.
Results
Age-adjusted elasticnet models of frontal cortex DNAm associated with total (cor=0.25, p<2.2e-16) and midfrontal (cor=0.26, p<2.2e-16), but less so with hippocampal (cor=0.08, p=7.4e-6) NFTs. Multiple regression models with age, sex and APOE were inferior at predicting NFTs (total: adj R2=0.155, rmse=0.327; midfrontal: adjR2=.069, rmse=0.788) relative to models that also included DNAm (total: adjR2=0.682, rmse=0.175; midfrontal: adjR2=0.594, rmse=0.479). Frontal cortex DNAm was least associated with hippocampal NFTs (adjR2=0.193, rmse=1.20)
Conclusions
DNAm provides an additional mechanism to account for variance in NFT burden amongst aging individuals. Additionally, frontal cortex DNAm is more highly associated with frontal cortex NFTs than hippocampal NFTs suggesting region-specific DNAm may support drivers of p-tau spread across regions, but future research is necessary to determine whether DNAm differences across AD and PART are causal or consequential.
NEUROBOOSTER ARRAY: A GENOME-WIDE GENOTYPING PLATFORM TO STUDY NEUROLOGICAL DISORDERS ACROSS DIVERSE POPULATIONS
Abstract
Aims
Our primary goal was to design an innovative genome-wide array for the Global Parkinson's Genetics Program in collaboration with the Center for Alzheimer's and Related Dementias. This array aimed to comprehensively capture diverse genetic variations in neurological disorders. Additionally, we sought to promote inclusivity in genetic studies and expedite the discovery of pivotal genetic factors influencing Parkinson’s disease.
Methods
To achieve our objectives, we introduced the Illumina NeuroBooster array (NBA). This state-of-the-art, high-throughput, and cost-effective platform is specially tailored to detect nuanced genetic variations across diverse global populations. Intricately designed, the NBA seamlessly merges 1,914,934 variants with a curated custom content that sheds light on over 70 distinct neurological conditions.
Results
A deep dive into the NBA's content reveals 1,873,290 autosomal variants, 79,994 variants from sex chromosomes, and an additional 1,509 mitochondrial variants. In a side-by-side comparison with the renowned NeuroChip, we identified a significant overlap of 126,220 variants. Moreover, NBA features over 10,000 multi-ancestry GWAS locus tagging variants, an essential feature facilitating intricate analyses across diverse population subsets. The figures below show counts of variants included on the chip within AD, PD, and ALS loci:
Conclusions
In essence, the NeuroBooster Array is a state-of-the-art tool in standardizing genetic studies in neurological disorders across diverse ancestral backgrounds. By leveraging the NBA, the global scientific community is now poised to undertake inclusive genetic research, promising to unravel the mysteries of neurological disease genetics.
AD RISK GENES AND DYSFUNCTION OF THE CHOROID
Abstract
Aims
Recent studies indicated that individuals with Alzheimer’s disease (AD) showed an increase in cerebrospinal fluid (CSF) concentration of proteins with a high expression in the Choroid Plexus (CP) (PMID: 35346299;36825691). The underlying mechanism of this increase remains unclear. Here we studied which AD genetic risk variants from a large GWAS (PMID: 35379992) were associated with increased cerebrospinal fluid levels of proteins with a high expression in the CP.
Methods
We included 600 individuals with CSF proteomics (200 controls with normal cognition and normal CSF abeta, average age 64 years; 400 individuals with AD defined by abnormal CSF abeta, average age 66 year) from the Amsterdam Dementia Cohort and associated cohorts. For each AD risk gene we tested the association between the number of risk alleles and increased protein levels. We considered an AD risk gene associated with the CP if CSF proteins with an high expression in the CP were overrepresented according to the Allen Brain Atlas.
Results
Risks alleles of ABCA7,TNIP1, SEC61G, PICALM and UNC5CL were enriched for proteins with a high expression in the CP (table). ABCA7 risk alleles were enriched for CP proteins in both AD and controls; TNIP, SEC61G and PICALM risk alleles were enriched for CP proteins in AD only; UNC5CL risk alleles were enriched for CP proteins in controls only. GO biological process enrichment analyses revealed that proteins with high expression in the CP were associated with TGF-beta signaling (TNIP and UNC5CL) and antigen presentation (ABCA7).
Conclusions
A number of AD risk genes are associated with CP functioning. These risk genes may affect AD through alternations in TGF-beta signaling and antigen presentation.
BRAIN AND BLOOD TRANSCRIPTOME-WIDE ASSOCIATION STUDIES IDENTIFY SEVEN NOVEL GENES ASSOCIATED TO ALZHEIMER’S DISEASE
Abstract
Aims
We leveraged the largest available bulk cis-eQTL meta-analysis summary statistics from brain tissue (MetaBrain Brain-Cortex; N=2,547) and whole blood (eQTLGen; N= 31,684) and applied them to largest clinically diagnosed AD meta-analysis summary statistics from Kunkle et al. 2019 (AD Cases = 21,982; Controls = 44,944) to identify novel Alzheimer’s disease (AD) genes and uncover key functional genetic variants driving GWAS associations using newly developed OTTERS transcriptome-wide association pipeline.
Methods
We removed genes from the APOE region, filtered genes to have concordant effect size directions across all five methods within OTTERS, and were nominally significant in three or more methods (p < 0.001). We identified novel genes by excluding genes with models that overlapped a 1MB window around GWAS hits identified by Kunkle et al. 2019 or Bellenguez et al. 2022. We also validated our hits using one-tailed T-tests applied to RNA-seq expression data from an independent NHW whole-blood dataset (AD=119; Controls=117) and a partially independent brain RNA microarray dataset called KRONOSII that was curated for high AD pathology and low secondary pathology (AD=168; Controls=177).
Results
We identified 38 novel genes within blood and 22 novel genes within brain-cortex. We validated four novel hits within blood (ERVK13-1, SLC1A3-AS1, PRDM15, CNDP2) and three novel brain-cortex hits (CD8A, LYSMD4, PTGR1). We fine-mapped eQTLs with CD8A and found one of the key eQTLs driving the TWAS signal is within repressive H3K27me3 chromatin peaks. We also identified 15 genes within blood and 16 genes within brain-cortex that harbored known GWAS loci.
Conclusions
By combining multiple large scale eQTL summary statistics with AD GWAS results, we have identified multiple new gene associations to AD and elucidated the functional relevance of several previously associated SNPs.