Abstract Body

Alzheimer’s disease (AD) is currently untreatable, and therapeutic strategies targeting the amyloid cascade have not yet been successful, indicating that novel treatment strategies are required. Recent genome wide association studies have identified a number of risk factors in genes expressed in microglia, including the Triggering receptor expressed on myeloid cells 2 (TREM2). TREM2 is essential for the transition of homeostatic microglia to disease associated microglia. TREM2 loss of function locks microglia in a homeostatic state, and affects a multitude of microglia functions such as chemotaxis, phagocytosis, cell survival, lipid- and energy metabolism. Biomarker studies revealed that TREM2 may protect humans from AD. To enhance TREM2 activity, we selectively increased the full-length protein on the cell surface via reducing its proteolytic shedding by ADAM proteases. We generated a panel of monoclonal antibodies against the stalk region of TREM2, which encompasses the cleavage site, with the aim to compete for α-secretase mediated shedding. Monoclonal antibody 4D9, which binds to an epitope close to the ADAM10/17 cleavage site, stabilized TREM2 on the cell surface, reduced its shedding and concomitantly activated phospho-SYK signaling in a dose dependent manner. Moreover, 4D9 stimulated survival of cultured macrophages, increased myelin debris uptake of primary microglia and reduced the amyloid burden in a mouse model for AD pathology. Thus, our findings demonstrate that antibodies elevating full-length TREM2 on the cell surface allow selective modulation of TREM2 dependent functions in microglia and macrophages, which may be of potential therapeutic benefit.

Aims

Therapeutic modulation of microglial functions may provide an additional strategy to slow progression of Alzheimer disease (AD). Although animal models suggest that TREM2 function is protective, we do not know if that is the case in AD patients. We therefore studied the dynamics of soluble TREM2 (sTREM2) in cerebrospinal fluid (CSF) as a surrogate marker of TREM2 signalling and its relationship with AD evolution.

Methods

We measured sTREM2 in longitudinal CSF samples from 261 participants in the Dominantly Inherited Alzheimer Network (DIAN) study, including 161 mutation carriers (MC) and 100 non-carriers (NC). We assessed the association between the CSF sTREM2 dynamics and the longitudinal changes in amyloid-β (Aβ), tau and neuronal damage markers along with the cognitive decline (cognitive composite).

Results

Higher Aβ burden at baseline (lower CSF Aβ42) was associated with a higher subsequent rate of sTREM2 increase (β=-3.704x10^-2, p=0.004) in MC. This increase was independent of tau-related pathology at baseline (CSF t-tau and p-tau). A higher rate of sTREM2 increase was associated with slower Aβ deposition either measured by CSF Aβ42 (r=0.56, p=0.01, presymptomatic MC) or PIB-PET (r=-0.38, p=0.05, all MC). Furthermore, higher sTREM2/p-tau at baseline predicted slower hippocampal volume and precuneus cortical shrinkage (β=1716.05, p=0.004; β=2.032x10^-2, p=0.03, respectively). Strikingly, we found a strong association between higher rate of sTREM2 increase and slower cognitive decline (r=0.62, p=0.003) in presymptomatic MC.

Conclusions

CSF sTREM2 increase in autosomal-dominant AD is mainly driven by Aβ accumulation. Increased TREM2 expression and signalling appears to protect against Aβ deposition and cognitive decline.

Abstract Body

Genetic variants nearby, and in some cases, rare missense mutations in a variety of genes expressed in microglia associated with either increased or decreased risk for late onset Alzheimer’s disease (AD). These AD-associated genes include the triggering receptor expressed in myeloid cells 2 (TREM2), phospholipase C gamma 2 (PLCG2), Abl Interactor member 3 (ABI3), MS4A4 amongst others. We show here that Aβ oligomers, but not Aβ monomers, bind to TREM2 with nanomolar affinity, and induce Aβ-dose-dependent shedding of TREM2 ectodomain, DAP12 and SYK phosphorylation. Similar effects are induced by anti-TREM2 antibodies. We propose that these intracellular signalling pathway changes activate protective microglial mechanisms, and the AD related variants affect these mechanisms. In addition, we show that the shed soluble sTREM also has protective effects, inhibiting Aβ oligomer formation, disrupting preformed Aβ oligomers and reducing Aβ-induced neurotoxicity. Crucially, R47H sTREM2 failed to disrupt Aβ oligomerization, and in fact encouraged larger more neurotoxic oligomers. sTREM2 and TREM2 signalling pathways represent potential therapeutic targets.

Abstract Body

Microglia are increasingly implicated in age-related Alzheimer's disease (AD). Many late-onset AD-related genes are microglia-specific. However, interplay between the AD-related genes in microglia, Aß metabolism and aging remains largely elusive. Recently, we have established isogenic human ES-derived microglia (hMGLs) harboring AD-associated variants in APOE, TREM2, CD33, INPP5D, and SORL1 loci or knockout (KO) by CRISPR/Cas9 gene editing, and through induced differentiation and subsequent treatments with aged human serum. We carry out multi-omics studies including RNAseq, ATACseq, ChIPseq, and proteomics to systematically evaluate the correlations among aging, Aß metabolism and AD-related genes. Moreover, we functionally characterize hMGLs in vitro and in vivo, such as responses to various physio/pathological stimuli, Aβ phagocytosis, and AD-like phenotypes in AD model mice through xenotransplantation. In our preliminary study, we have established the microglial aging model by adding aged human serum in culture and generated hMGLs harboring the AD-associated TREM2 mutant R47H. AD-like expression signatures were observed in TREM2-R47H hMGLs, while integrative multi-omic analysis of combined epigenetic and expression datasets indicated that APOE was a convergent pathogenic node. AD-associated TREM2-R47H can enhance APOE expression and impair hMGLs Aβ uptake in an APOE-dependent manner. Xenotransplants of TREM2-R47H hMGLs reduced Aβ uptake/clearance in the brain of AD mice. Moreover, we found that AD-associated gene expression landscape significantly altered during microglial aging. Our study shall provide new insights into the contribution of microglial aging and the interplay between aging, Aß metabolism and AD genes in microglia to AD pathogenesis, and may identify new strategies for AD therapeutics.

Aims

The underlying aim of this study was to understand how the newly discovered risk factor for Alzheimer’s Disease, Frontotemporal Dementia and other neurodegenerative diseases, polymorphisms of the triggering receptor expressed on myeloid cells 2 (TREM2) contributes to disease. The interaction of TREM2 and tau (pathology) was of particular interest in this project.

Methods

TREM2 knockout (KO) mice were crossed onto TAU58/2 mice which express the P301S mutation and have been previously well characterised. Mice were subjected to functional tests including RotaRod, elevated plus maze, open field and Morris water maze at 1, 2, 3 and 6 months of age. After behavioural assessment mice were sacrificed and their brains collected for histological analysis.

Results

2 month old TAU58/2 mice present with moderate motor deficits on the rotarod, this phenotype was significantly aggravated upon TREM2 depletion. Additionally, loss of TREM2 exacerbated the disinhibition like phenotype of TAU58/2 mice on the elevated plus maze at 2 months. Interestingly, TREM2 KO itself induces a disinhibition phenotype at 6 month of age. Histological analysis revealed an increase in cells positive for tau phosphorylated at serine 422, a late stage pathology marker. Double labelling for NFT-like cells and microglia revealed an increase in the number of microglia surrounding these lesions in young mice.

Conclusions

Our results show that depleting TREM2 exacerbates the functional deficits of TAU58/2 transgenic mice. Furthermore, we found an accelerated progression of tau phosphorylation. Hence, our work suggests that loss of TREM2 could contribute to functional deficits in AD via augmenting microglia response to tau pathology.

Aims

The discovery of myeloid-specific risk loci for Alzheimer’s disease (AD) has renewed interest in the role immune cells play in the progression of AD. To produce a more complete picture of how these risk loci and nearby genes alter disease, it is crucial to understand their homeostatic function as well as interactions with other risk loci. In this study, we leveraged isogenic human iPS-microglia to understand the functions and interactions between two AD risk genes: TREM2 and MS4A6A.

Methods

We generated human microglia lacking TREM2 and/or MS4A6A. These isogenic lines were profiled by RNA-sequencing and results were corroborated by functional assay and xenotransplantation into a chimeric mouse model of AD.

Results

Previous work suggests that decreased TREM2 function incurs increased risk of AD, while decreased expression of MS4A6A is correlated with protection. We further find that TREM2 RNA is decreased in MS4A6A isogenic lines which suggests potential interaction between these risk mechanisms. However, MS4A6A expression is not altered in TREM2 isogenic microglia which may suggest MS4A6A acts upstream of TREM2.

To further profile the relationship between these two loci, we performed single-cell RNA-sequencing of xenotransplanted microglia. After exposure to disease pathology in vivo, TREM2 knockout microglia remain homeostatic. Conversely, MS4A6A knockout microglia exhibit increased expression of MHCII genes, indicative of a more activated and likely protective state.

Conclusions

These data highlight potential interactions between two AD risk genes that alter the balance between homeostatic and activated microglial phenotypes. Understanding the nuances of microglial activation is an important step towards development of microglia-targeted therapeutic agents.

Aims

Innate immune Aβ clearance may be involved in Alzheimer’s disease (AD) etiology and treatment response following anti-amyloid immunotherapy. Regulation of intracellular Aβ clearance is not measured in AD risk- or diagnostic assessments. We present assays for Aβ catalysis in myeloid cells (THP-1 cells and human iPSC microglia).

Methods

THP-1 cells were differentiated toward macrophages, and treated with various inhibitors including IDE inhibitor (6bK, Tocris), BACE inhibitor (LY2886721, Tocris), NEP inhibitor (Phosphoramidon, Tocris) and Bafilomycin (1h) before adding 100 ng/ml Aβ40 (2h), followed by 4h chase. For AD patient-derived microglia, reprogrammed iPSCs were differentiated through a myeloid progenitor stage. Mature microglia were characterized by qPCR, immunofluorescence and phagocytosis assay.

Two immunoassays were developed on the Simoa platform using proprietary sheep monoclonal antibodies (BioventixLtd, UK). Both assays employ an Aβ c-terminal antibody for capture (AB.4D7). Assay 20_x employs a specific Aβ20’ cut point antibody (NH2-20.2B8) and assay x-40 a linear mid-domain antibody (AB.2A9) binding full-length Aβ as detector.

Results

Innate immune cells (THP1, microglia) catabolize Aβ1-40 producing 20-x fragments.

Inhibition of Nep and BACE reduces the production of 20-x fragment of Aβ in THP1 cells.

Autophagy inhibitor BafA1 increases both 1-40 and 20-x fragments indicating the role of the endo-lysosomal system in the degradation of full length and 20-x Aβ.

Conclusions

We demonstrate a novel assay for Aβ degradation by Innate immune cells including THP-1 and iPSC derived microglia.

Inhibition of Nep and BACE reduces 20-x peptide generation indicating this as a potential marker for Aβ clearance.

Aims

We and others recently identified several Alzheimer disease (AD) risk variants in TREM2. Here we aim to elucidate the downstream effect of genes and functional mechanisms leading to AD through multi-tissue proteomics study of AD, autosomal-dominant AD (ADAD) and TREM2 risk variant carriers.

Methods

Deep proteomics profiling was obtained (SOMAscan; 1305 proteins) from brain, cerebrospinal fluid (CSF), and plasma tissue. These neurologically relevant tissues were from Knight-ADRC and DIAN cohorts with comprehensive clinical information about AD pathology and cognition. After stringent QC, we analyzed 1079 proteins in brain (n=370), 713 proteins in CSF (n=699), and 931 proteins in plasma (n=486).

Results

We identified 27, 38 and 69 TREM2-specific proteins in brain, CSF, and plasma, respectively (at Bonferroni-corrected significance). Twenty-three plasma proteins showed nominal differential levels in brain and CSF and led to a prediction model, which discriminates TREM2 carriers from controls (AUC=0.94) and other AD cases (AUC=0.91) well. We identified 371 ADAD-specific proteins, among which 225 were nominally associated with AD neuropath traits. Furthermore, 54, 89, and 85 proteins showed nominal differential levels in sporadic AD vs controls in brain, CSF, and plasma. TREM2-specific proteins are involved in growth factors including VEGF, PDGF, EGF and immunological response. ADAD-specific proteins converge in immunological response pathways including cytokine-mediated signaling and DAP12-mediated pathway.

Conclusions

Our multi-tissue proteomics study for genetically defined AD cases identified multiple novel AD biomarker candidates. These findings not only help create novel prediction models but also point to specific pathways implicated in AD, supporting its potential utility as a clinically useful biomarker.