Aims

Risk factors for Parkinson’s disease (PD) can be more or less relevant to a population due to population-specific genetic architecture, local lifestyle habits, and environmental exposures. Therefore, it is essential to study PD at a local, regional, and continental scale in order to increase the knowledge on disease etiology. We aimed to investigate the contribution of genetic and environmental factors to PD in a new Swedish case-control cohort.

Methods

PD patients (n=929) and matched population-based controls (n=935) from the southernmost county in Sweden were included in the cohort. Information on environmental exposures was obtained using questionnaires at inclusion. Genetic analyses included a genome-wide association study (GWAS), haplotype assessment, and a risk profile analysis using cumulative genetic risk scores.

Results

The cohort is a representative PD case-control cohort (64% men, mean age at diagnosis=67 years, median Hoehn and Yahr score=2.0), in which previously reported associations between PD and environmental factors, such as tobacco, could be confirmed. We describe the first GWAS of PD solely composed of PD patients from Sweden, and confirm associations to well-established risk alleles in SNCA. In addition, we nominate an unconfirmed and potentially population-specific genome-wide significant association in the PLPP4 locus (rs12771445).

Conclusions

This work provides an in-depth description of a new PD case-control cohort from southern Sweden, giving insights into environmental and genetic risk factors of PD in the Swedish population.

Aims

Family-based genetic studies in early-onset Alzheimer’s disease (EOAD) are powerful to identify novel genes in AD. In a long-lasting effort to collect clinicopathological data and biomaterials from EOAD patients for genetic studies, we identified a multigenerational Flanders-Belgian family with autosomal dominant AD, average onset age 57.8 years (range 58-64), negative for mutations in known causal and high-risk AD genes.

Methods

DNA, biomaterials and clinicopathological data in 2 consecutive generations: 5 AD patients, 2 with definite neuropathological diagnosis and 9 unaffected or at-risk individuals. Whole exome sequencing in 4 patients for gene identification. Functional modelling of mutation and pathway.

Results

WES analysis identified a novel mutation, p.E317D, in TLR9 co-segregating with AD. TLR9 is a DNA-sensing receptor expressed in immune cells and p.E317D falls in the DNA-sensing domain. The mutation caused >50% decrease in TLR9 signaling. Cytokine profiling of peripheral blood mononuclear cells (PBMC) upon TLR9 activation revealed a complex signaling with predominance of anti-inflammatory cytokines including IL-1RA and IFNβ. This cytokine pool reduced the levels of pro-inflammatory cytokines IL1β and TNFα and promoted phagocytosis of Aβ42 in human iPSC-derived microglia. We also demonstrated that phagocytosis was attributed to IFNs in the cytokine pool. Transcriptome analysis identified several genes including AXL and RUBICON as potential drivers for IFNs-induced phagocytosis.

Conclusions

We presented that PBMC mediated TLR9 signaling exerts a protective role in AD pathophysiology, promoting microglial anti-inflammation and phagocytic clearance of pathological proteins, functions that could be lost due to TLR9 genetic mutations.

Aims

The SorLA protein, encoded by the SORL1 gene, is a major player in Alzheimer’s disease (AD) pathophysiology. Functional and genetic studies demonstrated that SorLA deficiency results in increased production of Aβ peptides, and thus a higher risk of AD. A large number of SORL1 missense variants have been identified in AD patients, but their functional consequences remain largely undefined.

Methods

In an initial screening, we overexpressed 71 SORL1 variants, identified in AD patients, in HEK293 cells to analyze the maturation profile and subcellular localization of SorLA. Five of these variations were further studied in details in CRISPR/Cas9-modified hiPSCs.

Results

Overexpression analyses revealed that 15 of the 71 SORL1 variants induced a maturation and trafficking-deficient phenotype. Validation studies on endogenous SORL1 expression in hiPSCs confirmed that maturation-defective variants were largely retained in the endoplasmic reticulum, resulting in a reduction in the delivery of SorLA to the plasma membrane and endosomal system. Importantly, this was associated with an increase of Abeta secretion, demonstrating a loss-of-function effect of these SorLA variants regarding this ultimate readout, and a direct link with AD pathophysiology. Structural analysis suggested that impaired cellular trafficking of SorLA could be due to subtle variations of the protein 3D structure resulting from changes in the interatomic interactions.

Conclusions

We here identify a new pathophysiological mechanism, by which a subset of rare SORL1 missense variants identified in AD patients induce a likely misfolded protein, thus altering the protein maturation and trafficking, and leading to a loss of the protective function of SorLA towards Abeta secretion.

Aims

Rare missense and truncating nucleotide variants in SORL1, TREM2 and ABCA7 are moderate-to-high Alzheimer disease (AD) risk factors, to which ATP8B4 and ABCA1 were recently added thanks to the aggregation of exome sequencing data from the ADES and ADSP consortia.

Methods

To assess the role of copy number variants (CNVs, deletions and duplications), we applied a highly-sensitive bioinformatics pipeline based on the CNV caller CANOES to 20,661 exomes (8941 controls, 3770 early-onset AD [EOAD], 7950 late-onset AD [LOAD]) from ADES-ADSP. We focused on rare CNVs (frequency<1%) and aggregated CNV counts at the gene level.

Results

We detected 49,460 rare CNVs. Among previously mentioned genes, we identified 6 deletions of ABCA7 in cases and 3 in controls and 3 ABCA1 deletions in three cases (all EOAD, none in controls). To assess the effect of ABCA1 loss of function on EOAD and LOAD risk simultaneously, we gathered data from high-confidence truncating nucleotide variants with CNVs in an ordinal regression model. While the previously reported aggregation of truncating and missense variants (OR=1.9 [1.5-2.5]) hid a heterogeneous effect of missense and truncating variants considered separately (OR=1.7 [1.3-2.2] and OR = 4.7 [2.2-10.3] respectively), the joint analysis of the full spectrum of rare high-confidence loss-of-function ABCA1 variants (p=3.1x10-5) confirmed their strong association with EOAD risk (OR=8.1 [2.9-29.4]).

Conclusions

Our data suggest that extremely rare CNVs may contribute to AD risk, some of them could be associated with a large effect, as the loss of ABCA1 function. Additional genes are currently being analyzed and will be presented.

Aims

Rare premature termination codon (PTC) mutations and expansion of a variable number of tandem repeats (VNTR) polymorphism in the Alzheimer’s disease (AD) risk gene ABCA7 likely increase risk of AD through haploinsufficiency. ABCA7 has suggested functions in lipid metabolism and the immune system, possibly affecting APP metabolism, but the exact mechanisms through which ABCA7 reduction causes AD remain unclear. Studying CSF biomarkers reflecting different pathological processes in mutation carriers could shed light on these mechanisms.

Methods

We conducted ELISA assays for eight AD-related biomarkers on the CSF of 229 AD patients for whom ABCA7 VNTR length determination was performed, including 28 PTC mutation carriers and 16 pathogenic expansion carriers. For reference, 75 controls were included. We investigated the effect of expression-reducing mutations and VNTR length on amyloid-beta (Aβ) 40 and 42, the amyloid 42/40 ratio, t-tau, p-tau, sAPP-α and -β, YKL-40 and hFABP.

Results

Carriers of expression-reducing mutations had significantly lower Aβ42 levels (p=0.0075) and amyloid ratio (p=0.0039) compared to non-carriers. The effect on Aβ42 levels was limited to non-APOE4 carriers (p=0.020). An increased VNTR length was associated with increased p-tau levels (p=0.036) and decrease in amyloid ratio (p=0.029). Other biomarkers, reflecting neuroinflammation and vascular pathology, were not altered in relation to ABCA7 variants.

Conclusions

Our results highlight that ABCA7 PTC and repeat expansion carriers show differences in CSF amyloid beta markers, but not in CSF biomarkers reflecting other pathophysiological aspects of AD. This suggests that reduced ABCA7 expression leads to increase in amyloid pathology, especially in non-APOE4 carriers.

Aims

Polygenic scores (PGS) are a powerful tool to combine the predictive value of GWAS-identified genetic variants, but the value to predict the rate of progression in tau pathology underlying cognitive worsening is unclear.

To test whether:

higher PGS is associated with longitudinal rates of tau PET.

PGS effects on tau progression were cell-type specific.

risk stratification by PGS is of utility for saving sample size in clinical trials on tau.

Methods

We leveraged the yet largest number of lead SNPs (n = 85) from two recent GWAS including up to 1.1 million participants (Bellenguez et al., 2020; Wightman et al., 2020) to estimate the PGS in over 1800 participants assessed by longitudinal composite memory scores and global cognition (average FU-duration = 4 yrs) and tau-PET (N = 231, average FU-duration = 2 yrs) from ADNI. In linear mixed effects models, we tested PGS to predict the rate of tau-PET in major Braak-stage ROIs and on cognitive decline.

Results

Higher PGS was associated with faster rates of increases in tau-PET, which was pronounced in amyloid-PET positive participants independent of APOE genotype. Faster tau-PET increases mediated PGS effects on cognitive decline. Cell-type specific PGS analysis implicated microglia and oligodendrocytes alterations in faster tau-PET progression. Risk stratification (> median PGS) reduced by 15% the required sample size to detect a 20%-treatment effect on tau-PET progression in amyloid-positive participants.

Conclusions

PGS is associated with faster increases in tau-PET and provides utility for risk stratification in disease-modifying treatments on tau.

Aims

We aimed to investigate differences in gene expression of linear and circular transcripts from known Autosomal-dominant Alzheimer disease (ADAD) genes (PSEN1, APP and, PSEN2) in controls, sporadic AD, and ADAD.

Methods

We obtained and sequenced RNA from brain cortex using standard protocols. Linear counts were obtained using the TOPmed pipeline and circular counts using DCC. After stringent QC, we obtained the counts corresponding the ADAD genes. Only circPSEN1 passed QC. We used DESeq2 to compare the counts across groups correcting by biological and technical variables. We performed in-silico functional analyses using the Circular RNA interactome website and DIANA mirPath software.

Results

Our results show significant differences in gene counts of circPSEN1 between ADAD cases and controls but not with sporadic AD (ADAD=18, AD=59, Controls=10–p<0.001 log2FC=0.556, Fig.1A). The same trend is found in the replication dataset (ADAD=4, AD=197, Controls=13–p=0.017, log2FC=0.518, Fig.1B). This finding is specific to circular PSEN1; no significant differences were observed for linear PSEN1 . In addition, the high circPSEN1 levels do not seem to be specific to PSEN1 carriers but to ADAD. In-silico functional analyses suggest the involvement of circPSEN1 in the several pathways such as axon guidance (p=3.39×10^-07), hippo signaling pathway (p=7.38×10^-07), lysine degradation (p=2.48×10^-05) or Wnt signaling pathway (p=5.58×10^-04) among other KEGG pathways. Additionally, circPSEN1 counts were able to discriminate the brains from ADAD from the other with an AUC above 0.70.

Conclusions

circPSEN1 show differences that are unique to ADAD that might be related to neuroinflammatory events that lead or are caused by the accumulation of amyloid-beta.

Aims

Methods

Results

Conclusions