Search

Background: To date, Alzheimer's disease (AD) research has principally focused on neurons. In contrast, recent studies suggest that genetic mechanisms drive microglia towards prolonged inflammation in AD brains, exacerbating neurodegeneration. Indeed, many of the 70 disease‐associated loci uncovered with genome‐wide association studies (GWAS) reside near genes related to microglial function, such as TREM2. However, GWAS cannot directly identify causal variants or effector genes, as GWAS only reports the sentinel variant at a given locus, which tags all variants in high linkage disequilibrium. We used a combination of ATAC‐seq and high‐resolution promoter‐focused Capture‐C in two human microglial cell models to map interactions between GWAS‐implicated variants and their putative effector genes. We then validated an observed interaction between SNP rs6024870 at the 'CASS4' AD locus and the promoter of RTFDC1, a DNA damage response gene, using CRISPR‐Cas9. Method: We employed Capture‐C to characterize the genome‐wide interactions of all promoters in the HMC3 microglial cell line and iPSC‐derived microglia. We confirmed enhancer activity of the region harboring rs6024870 using dual‐luciferase assays. To validate the regulatory interaction with the Capture‐C implicated effector gene RTFDC1, we used lentiviral CRISPR‐Cas9 to delete an ∼400bp region containing the SNP in HMC3 cells. We confirmed changes in RNA and protein expression using RNA‐seq and immunoblotting. Furthermore, we assessed DNA damage repair deficiencies in enhancer knockout cells using an immunofluorescent assay for the histone mark GH2AX, and confirmed the role of RTFDC1 in the DNA damage phenotype using siRNA knockdown and lentiviral overexpression of RTFDC1, respectively. Result: Proxy SNP, rs6024870 (r2 = 0.93 to sentinel rs6014724), at the 'CASS4' locus, coincided with open chromatin and contacted the promoter of RTFDC1. Deletion of the putative enhancer region harboring rs6024870 by CRISPR‐Cas9 in HMC3 reduced the expression of both RTFDC1 mRNA and protein. Knockout cells had reduced ability to clear DNA double‐strand breaks (DSBs), and siRNA knockdown and lentiviral overexpression of RTFDC1 confirmed this phenotype was due to reduction of RTFDC1. Conclusion: We implicate rs6024870 and RTFDC1 as a SNP‐effector gene pair at the AD 'CASS4' GWAS locus, and the gene product is directly involved in driving a reduction in DNA DSB clearance. [ABSTRACT FROM AUTHOR]

Background: Recent studies suggest genome‐wide‐association‐studies (GWAS) loci confer their effects on microglia in late‐onset Alzheimer's disease (LOAD) brains. Relatively fewer studies have investigated the effects of other genome‐wide significant loci (p<5e‐8) using human neurons. Method: GWAS itself cannot directly identify causal variant‐(effector)gene‐pairs as GWAS only reports the sentinel variant at a given locus. To map the physical interactions between GWAS‐implicated putative causal non‐coding variants and their corresponding effector genes in LOAD, we leveraged the post‐GWAS analytical framework, INFERNO, coupled with the largest harmonized functional genomics data catalog to date, FILER. Our goal was to unravel the effects of all 75 genome‐wide significant genetic loci reported in Bellenguez et al in human brain cell types. We searched for plausible evidence of causality for LOAD candidate variants in any brain‐specific variant‐gene pair through a total of 3,873 publicly curated datasets (123 enhancer, 3,153 epigenetics, 105 QTL) across 12 resources. We then validated the top variant‐gene pairs using high‐resolution promoter‐focused Capture C, ATAC‐seq and RNA‐seq datasets from several neuronal cell types and performed CRISPR interference. Result: The computational variant‐gene analyses identified a novel candidate causal SNP, rs74504435 (r2 = 0.94 to sentinel SNP rs76928645) at the SEC61G locus, located between SEC61G/EGFR. With a CADD score of 15, it is predicted as "regulatory" by both ENSEMBL and regulomeDB2. rs74504435 is found in ChromHMM and EpiMAP brain enhancers, and perturbs the transcriptional factor OCT6, which controls Schwann cell myelination. The minor allele T is protective to LOAD and downregulates EGFR in GTEx brain cortex (p = 2.3e‐13), CommonMind (FDR p = 5.8e‐5), and ROSMAP DLPFC (FDR p = 8.6e‐9) datasets with the same effect direction. In our Capture C and ATAC‐seq datasets, the SNP resides in open chromatin and directly contacts EGFR in human iPSC‐derived cortical neurons and ES‐derived hypothalamic neurons. We further validated a regulatory role for the enhancer region containing rs74504435 on EGFR using CRISPR interference in human microglial cells. Conclusion: Together, we identified a novel LOAD‐associated neuronal enhancer that drives EGFR gene expression. Further research is needed to investigate if EGFR inhibitors approved as anti‐cancer drugs can serve as potential candidates for repurposing as LOAD therapeutics. [ABSTRACT FROM AUTHOR]

We previously demonstrated that in Alzheimer's disease (AD) patients, European apolipoprotein E (APOE) ε4 carriers express significantly more APOE ε4 in their brains than African AD carriers. We examined single nucleotide polymorphisms near APOE with significant frequency differences between African and European/Japanese APOE ε4 haplotypes that could contribute to this difference in expression through regulation. Two enhancer massively parallel reporter assay (MPRA) approaches were performed, supplemented with single fragment reporter assays. We used Capture C analyses to support interactions with the APOE promoter. Introns within TOMM40 showed increased enhancer activity in the European/Japanese versus African haplotypes in astrocytes and microglia. This region overlaps with APOE promoter interactions as assessed by Capture C analysis. Single variant analyses pinpoints rs2075650/rs157581, and rs59007384 as functionally different on these haplotypes. Identification of the mechanisms for differential regulatory function for APOE expression between African and European/Japanese haplotypes could lead to therapeutic targets for APOE ε4 carriers. [ABSTRACT FROM AUTHOR]

Background: Previously, we demonstrated that the ancestry‐related risk for Late Onset Alzheimer's Disease (LOAD) is driven by a local genomic region (termed Local Ancestry; LA) around APOEε4. Furthermore, we showed that in the brain of individuals bearing European LA there is higher expression of APOEε4 compared to those with African LA. In a follow‐up study, utilizing reporter assays and Capture‐C data we located two intronic regions within the European LA, both in the TOMM40 gene (named B10 and B13), that increased APOE expression in microglia and astrocytes. In this study, we sought to validate their regulatory role in APOE expression using CRISPR interference/activation (CRISPRi/a). Method: Human Microglial Clone 3 (HMC3) CRISPRi/a lines were produced by transducing inducible dCas9‐VP64 (Activation), dCas9‐KRAB (Interference) or dCas9 (control) using lentiviral vectors. To direct the dCas9 constructs to our regions of interest, we generated multiplex vectors that encode 4 short‐guide RNAs (sgRNAs) targeting either B10 or B13. We used 4 different sgRNAs in each case to ensure full‐length coverage of the tested regions (∼850bp size). An empty multiplex vector was used as a control. We then transduced either of the multiplex vectors into the HMC3 CRISPRi/a lines. We induced expression of the dCas9 constructs for 2 or 6 days with Doxycycline (2ug/ml). RNA was extracted and the expression of APOE and TOMM40 was measured by qRT‐PCR. Result: APOE expression significantly increased when targeting B10 or B13 (p=0.001; p=0.003 respectively) with dCas9‐VP64 after 2 days of Doxycycline treatment. Six days after treatment the significance persisted only when targeting B10 (p=0.01). No significant changes in APOE expression were observed in the cells bearing the dCas9‐KRAB presumably due to low endogenous APOE levels. Expression of TOMM40 did not vary under any treatment. Conclusion: These preliminary results support our previous findings that regions B10 and B13 may act as regulators for APOE expression as demonstrated by the elevation of ApoE expression when targeting an activator to these regions. The expression of TOMM40 did not vary across cell lines in the evaluated time points supporting that the effect observed is specific for APOE. [ABSTRACT FROM AUTHOR]

Background: The risk for late‐onset Alzheimer disease (AD) in APOEε4 carriers differs between ancestral groups, where APOEε4's odds ratio for AD risk is lower in African (AFR) homozygous carriers than in non‐Hispanic White (NHW) or Japanese (JPT) carriers (odds ratio ∼2‐5 vs >15). Local ancestry (LA) analyses in APOEε4 carrier populations have shown the protective effect in AFR relative to EUR/JPT is due to noncoding factors lying in the LA surrounding APOEε4. Thus, regulatory differences between risk and protective LA haplotypes are most likely involved in the differential risk effect seen for APOEε4 on different backgrounds. Methods: We identified 56 significant sequence differences between AFR and EUR/JPT APOEε4 haplotypes from the 1000 genomes in the immediate topologically associated domain surrounding APOE. We performed two different Massively Parallel Reporter Assay (MPRA) designs; one assessing small haplotype (∼900bp) effects and one based upon single variant effects. We supplemented these results with single fragment luciferase reporter assays. All assays were performed in at least duplicate in HMC3 (microglia), U118 (astrocytes) and SH‐SY5Y (neurons) cell lines. Additionally, we integrated chromatin interaction information from promoter capture C chromatin conformation assays in the same cell types. Results: We identified a region in the first introns of TOMM40 with increased EUR/JPT enhancer activity, supported by both MPRA analyses and APOE promoter interaction in astrocytes and microglia. Two additional regions with differential enhancer activity in neurons, but no promoter interaction, were identified; downstream of APOE and in PVRL2 introns upstream of APOE presenting with higher EUR or higher JPT haplotype variant enhancer activity compared to AFR, respectively. Conclusions: Our results indicate several areas of differential regulation in this LA region on APOEε4 haplotypes dependent on cell type. As APOE is mostly expressed in glial cells, the data in TOMM40 introns points to this region as having the biggest impact on APOE expression in our study and thus highly supports the involvement of this region in the differential risk effects seen for APOEε4. Follow‐up of the identified regulatory regions is currently ongoing using in‐house iPSC derived cell lines. [ABSTRACT FROM AUTHOR]

Background: Alzheimer's disease (AD) research has principally focused on neurons due to their role in neurodegeneration. In contrast, recent studies suggest that genetic mechanisms drive microglia to prolonged inflammation in AD brains, exacerbating neurodegeneration. Indeed, recent genome‐wide association studies (GWAS) of AD have identified multiple loci near genes related to microglial function, such as TREM2 and CR1. However, GWAS does not have the sensitivity to identify causal variants or effector genes. We used a combination of ATAC‐seq and high‐resolution promoter‐focused Capture‐C in two human microglial cell models to map interactions between GWAS‐implicated variants and their putative effector genes. We then validated an observed interaction at the 'CASS4' locus using CRISPR‐Cas9 genome editing. Method: Improving on the relatively low resolution of available Hi‐C data, we employed high resolution Capture‐C to characterize the physical genome‐wide interactions of all human promoters. We performed this in the human microglial cell line HMC3 and human iPSC‐derived microglia (iMg). We confirmed the enhancer activity of a SNP we elected to pursue and its associated regulatory element using dual‐luciferase assays. To confirm the regulatory interaction with the Capture‐C implicated effector gene, we used lentiviral CRISPR‐Cas9 to delete an approximately 300bp region containing the SNP in HMC3 cells. We then confirmed changes in RNA and protein expression using RNA‐seq and Western blotting, respectively. Result: Variant‐to‐gene mapping in both microglial cell models identified 67 putative effector genes (51 coding) across both cell types, with 14 observed in both models. We identified a novel proxy SNP, rs6024870 (r2 = 0.93 to sentinel SNP rs6014724), at the 'CASS4' locus, which coincided with open chromatin and directly contacted the promoter of RTFDC1, a gene not previously implicated in AD. Deletion of the putative enhancer region harboring rs6024870 by CRISPR‐Cas9 in HMC3 reduced the expression of RTFDC1 at both the mRNA and protein level. We also note that CASS4 levels were modestly influenced by this CRISPR‐mediated perturbation. Conclusion: We implicate RTFDC1 as a novel effector gene at the AD 'CASS4' GWAS locus. Further efforts will characterize the phenotypic effect of this variant in microglial cell models, including on inflammation and phagocytic activity. [ABSTRACT FROM AUTHOR]

Background: Previous studies of Alzheimer's Disease (AD) have primarily implicated neurons and microglia as disease‐conferring cell types due to obvious links with degeneration and inflammation, respectively. However, astrocytes may also play a role in AD pathogenesis, but have been less studied in this context. Indeed, genome‐wide association studies (GWAS) have identified multiple genomic variants that reside near genes with astrocyte‐related functions, such as lipid processing and synaptic function. However, GWAS only reports genomic variants associated with a given trait and not necessarily the precise localization of culprit genes. High resolution chromatin conformation capture‐based techniques detect 3D genomic interactions between GWAS‐implicated signals and their effector genes, and allow for the characterization of non‐coding variants in the context of their regulatory activity. Method: To improve on the low resolution of available Hi‐C data, we employed a high resolution DpnII‐based Capture‐C method to simultaneously characterize the physical genome‐wide interactions of 36,691 baited regions covering human promoters genome‐wide (including non‐coding transcripts) ‐ which we applied to a human primary astrocyte line (NHA, Lonza). Following sequencing, we investigated significant interactions at two different resolutions (1‐fragment and 4‐fragment resolutions; median fragment size = 265bp and 1,441bp, respectively). In parallel, we generated ATAC‐seq open chromatin maps to filter for informative (r2>0.7) proxy single nucleotide polymorphisms (SNPs) for each of the 45 AD independent signals reported to date. Result: ATAC‐seq fine‐mapping yielded 100 candidate SNPs in open chromatin for 30 of these loci. By further constraining on our promoter Capture C data in astrocytes, at both one and four DpnII fragment resolution (median distance between interacting regions 9kb and 71kb, respectively), we observed contacts to "open" promoters for 6 putative target genes, corresponding to 6 of the original GWAS loci. These included PICALM, FERMT2, CASS4, and CLU. Conclusion: We observed informative contacts between proxy SNPs and putative effector genes in the human astrocyte context for ∼13% of AD GWAS loci. Further efforts in other relevant cell types, where many of the other loci may in fact be principally operating, should shed light on additional signals. Follow‐up functional studies are warranted to validate these findings. [ABSTRACT FROM AUTHOR]