Aβ presence in the caudate nucleus (Ca) partially defines Thal stage III in Alzheimer's disease (AD), but little is known about AD's cellular impact on the region. Leveraging a public basal ganglia taxonomy of cellular populations, we generated a cellular resolution atlas of AD-associated pathological changes in Ca. Unlike cortex, we found that Ca AD pathology is dominated by two key features: phosphorylated tau (pTau)-containing neuropil threads enriched near oligodendrocytes in white matter tracts and amyloid-β diffuse plaques enriched in gray matter. Although AD pathology in affected cortical regions results in neuronal loss, we find no AD-driven reductions in neuron proportions in Ca. However, there were observable changes in multiple cellular populations. Protoplasmic astrocytes and FLT1+/IL1B+ microglia increased in abundance with global pTau levels. We also observe gene expression changes in fast-spiking PTHLH-PVALB interneurons indicative of disrupted signaling pathways and altered intrinsic physiological properties. This work provides a cellular-resolution framework for understanding AD pathology in Ca.

Aggregation of the protein tau defines tauopathies, the most common age-related neurodegenerative diseases, which include Alzheimer’s disease and frontotemporal dementia. Specific neuronal subtypes are selectively vulnerable to tau aggregation, dysfunction, and death. However, molecular mechanisms underlying cell-type-selective vulnerability are unknown. To systematically uncover the cellular factors controlling the accumulation of tau aggregates in human neurons, we conducted a genome-wide CRISPRi screen in induced pluripotent stem cell (iPSC)-derived neurons. The screen uncovered both known and unexpected pathways, including UFMylation and GPI anchor biosynthesis, which control tau oligomer levels. We discovered that the E3 ubiquitin ligase CRL5SOCS4 controls tau levels in human neurons, ubiquitinates tau, and is correlated with resilience to tauopathies in human disease. Disruption of mitochondrial function promotes proteasomal misprocessing of tau, generating disease-relevant tau proteolytic fragments and changing tau aggregation in vitro. These results systematically reveal principles of tau proteostasis in human neurons and suggest potential therapeutic targets for tauopathies.

Numerous studies have identified AD-associated molecular and cellular changes to the cortex using single nucleus RNA sequencing (snRNA-seq) and, to a lesser extent, single nucleus ATAC-seq (snATAC-seq), applied to millions of cells across hundreds of donors. It has proven challenging, however, to determine whether changes are consistent because of differences in cohort selection, reported clinical metadata, data pre-processing, cellular taxonomy construction/mapping, and analytical strategies across studies. We uniformly re-processed 10 publicly available datasets (Table 1) that had applied snRNA-seq to 4.3 million cells from the dorsolateral pre-frontal cortex (DLPFC) of 780 donors. We integrated each of them with SEA-AD’s multi-modal, single cell atlas of AD, which contains 4.5 million cells from the middle temporal gyrus (MTG) and DLPFC that were profiled with snRNA-seq, snATAC-seq, snMulitome or spatial transcriptomics (MERFISH). We then harmonized clinical data across donors and mapped all cells to the same 139 fine-grained cell supertypes in SEA-AD’s shared MTG and DLPFC cellular taxonomy. With common clinical data and cellular labels, we then compared cohorts and data quality across studies as well as identified cellular changes consistently associated with AD. Cohorts from nearly every study sampled across the spectrum of plaque and tangle pathology, even in those that performed case-control analyses. With rare exception, the fraction of donors in each cohort with an APOE4 allele, clinically diagnosed dementia, and severe co-morbidities were also similar. Comparing datasets, SEA-AD was uniquely successful in profiling both many donors and many nuclei per donor, while deeply sequencing libraries to drive high gene detection. Most importantly though, of the 34 supertypes that were significantly changed in the DLPFC in SEA-AD, 8 were also significantly changed across external studies with large enough cohorts to power discovery. This included 5 selectively vulnerable Sst interneuron supertypes and 1 Microglia supertype that were changed early in AD in SEA-AD. Consistent and early selective loss of Sst interneurons and increase in disease associated Microglia underscore their importance to AD progression. Our integrated dataset provides a valuable common starting point for future comparative analyses from the community and, we hope, will accelerate discovery of novel therapeutic targets.

Alzheimer’s disease (AD), an age-associated neurodegenerative disorder, is characterized by progressive neuronal loss and the accumulation of misfolded proteins such as amyloid-β and tau. While neuroinflammation, mediated by microglia and brain-resident macrophages, plays a pivotal role in AD pathogenesis, the intricate interactions among age, genes, and other risk factors remain elusive. Somatic mutations, known to accumulate with age, instigate clonal expansion across diverse cell types, impacting both cancer and non-cancerous conditions. Utilizing molecular-barcoded deep panel sequencing, which enables sensitive detection of somatic mutations with allele fractions as low as 0.1%, we profiled clonal somatic mutations among 149 cancer driver genes in 311 prefrontal cortex samples from AD patients and matched controls. Fluorescence-activated nuclei sorting and single-nucleus RNA sequencing were further used to study the cell-type composition and transcriptomic impact of somatic mutations. Our study unveiled an elevated occurrence of somatic single-nucleotide variants and insertions/deletions within cancer driver genes in AD brains. Recurrent somatic mutations, often multiple, were observed in genes associated with clonal hematopoiesis (CH). Remarkably, these somatic mutations were specifically enriched in CSF1R+ microglia and exhibited signals of positive selection, suggesting mutation-driven microglial clonal expansion (MiCE) in AD brains. Single-nucleus RNA sequencing of temporal neocortex samples from an additional 62 AD patients and matched controls revealed a nominal increase in mosaic chromosomal alterations (mCAs) associated with CH in AD microglia, with microglia carrying mCA exhibiting upregulated pro-inflammatory genes, resembling the transcriptomic features of the disease-associated state in AD. Our findings indicate that proliferation-related somatic mutations in microglia are prevalent in normal aging but further enriched in AD, driving MiCE and promoting inflammatory, disease-related microglial signatures. This study provides crucial insights into microglial clonal dynamics in AD, potentially paving the way for novel approaches to AD diagnosis and therapy.

Previously, we developed a co-calibrated and harmonized brain pathology score (BPS) across prospective cohort studies with research brain donation that incorporates multiple forms of postmortem neuropathology, using confirmatory factor analysis. We sought to identify genetic loci associated with BPS using a systems-biology approach, combining data from participants in the Adult Changes in Thought (ACT), the Religious Orders Study, and Rush Memory and Aging Project (ROSMAP) autopsy cohorts. We used PLINK in each cohort separately for genome-wide association studies (GWAS) of BPS using HRC imputed data from European ancestry participants, adjusting for age at death, sex, and population substructure. We performed meta-analysis using the adaptively weighted Fisher’s approach in METAL. We performed gene-wide analysis using the meta-analyzed results which we then integrated into the human protein-protein interaction (PPI) network using a dense module searching (DMS) method to identify network hub genes for BPS. We interrogated the Seattle Alzheimer’s Disease Brain Cell Atlas (SEA-AD) dataset on the middle temporal gyrus to determine which cell types both hub genes were expressed in and how they differed across donors with higher degrees of AD pathology (i.e. along AD pseudo-progression). The sample size consisted of 1,848 brain donors ( Table 1 ). The quantile-quantile plot and genomic inflation (λ = 1.005) for GWAS meta-analyses showed no bias ( Figure 1a ), with the Manhattan plot in Figure 1b . Apart from significant SNPs around the APOE region, we identified two candidate loci a) (Chr 9: rs1332179; MAF = 0.1; P_meta = 8.7 × 10 -8 ) and b) (Chr 17: rs11078196; MAF = 0.34; P_meta = 1.9 × 10 -7 ). Regional association plots for these two loci are shown in Figure 1c . The PPI network analysis identified VCP and IQCB1 as hub genes ( Figure 2a ). While both hub genes were expressed broadly across cell types, IQCB1 was specifically higher with higher degrees of AD pathology in Microglia and VCP was lower with higher degrees of AD pathology in several neuronal populations ( Figure 2b ). We identified two potentially useful candidate loci associated with BPS using a systems-biology approach. Further functional enrichment analysis is needed to determine whether these novel loci may identify targets for interventions to ameliorate AD.

Single-cell multiomic technologies have allowed unprecedented access to gene profiles of individual cells across species and organ systems, including >1000 papers focused on brain cell types alone. The Allen Institute has created foundational atlases characterizing mammalian brain cell types in the adult mouse brain and the neocortex of aged humans with and without Alzheimer's disease (AD). With so many public cell type classifications (or 'taxonomies') available and many groups choosing to define their own, linking cell types and associated knowledge between studies remains a major challenge. Here, we introduce Annotation Comparison Explorer (ACE), a web application for comparing cell type assignments and other cell-based annotations (e.g., donor demographics, anatomic locations, batch variables, and quality control metrics). ACE allows filtering of cells and includes an interactive set of tools for comparing two or more taxonomy annotations alongside collected knowledge (e.g., increased abundance in disease conditions, cell type aliases, or other information about a specific cell type). We present three primary use cases for ACE. First, we demonstrate how a user can assign cell type labels from the Seattle Alzheimer's Disease Brain Cell Atlas (SEA-AD) taxonomy to cells from their own study and compare these cell type mappings to existing cell type assignments and cell metadata. Second, we extend this approach to ten published human AD studies which we previously reprocessed through a common data analysis pipeline. This allowed us to compare brain taxonomies across otherwise incomparable studies and identify congruent cell type abundance changes in AD, including a decrease in abundance of subsets of somatostatin interneurons. Finally, ACE includes translation tables between different mouse and human brain cell type taxonomies publicly accessible on Allen Brain Map, from initial studies in individual neocortical areas to more recent studies spanning the whole brain. ACE can be freely and publicly accessed as a web application (https://sea-ad.shinyapps.io/ACEapp/) and on GitHub (github.com/AllenInstitute/ACE).

Applying single-cell RNA sequencing (scRNA-seq) to the study of neurodegenerative disease has propelled the field towards a more refined cellular understanding of Alzheimer’s disease (AD); however, directly linking protein pathology to transcriptomic changes has not been possible at scale. Recently, a high-throughput method was developed to generate high-quality scRNA-seq data while retaining cytoplasmic proteins. Tau is a cytoplasmic protein and when hyperphosphorylated is integrally involved in AD progression. The relationship between pTau accumulation and the molecular changes underlying cell type specific selective vulnerability remains poorly understood and has not been assessed in the Middle Temporal Gyrus (MTG), a critical transition zone in AD neuropathologic progression. Human brain tissue was collected at rapid autopsy (postmortem interval (PMI) <12hrs). One hemisphere was embedded in alginate for fresh coronal slicing (4mm). Slabs were frozen in a dry ice isopentane slurry. Sampled MTG was processed by mechanical dissociation using a Potter-Elvehjem tissue grinder without enzymes or detergents, followed by iodixanol gradient centrifugation, and immunolabeling (AT8/pTau and MAP2/neurons) for FACS. For each donor, MAP2+/pTau+ and MAP2+/pTau- fractions were collected and sequenced using the 10x Genomics Chromium Single Cell Assay. Analysis was performed using standard pipelines in R (Seurat) and Python. Nine donors from Seattle Alzheimer’s Disease Brain Cell Atlas (SEA-AD) were selected for advanced pTau pathologic distribution without end stage neurodegeneration by including Braak Stages IV-VI (Table 1). After QC and filtering, 125,505 cells were confidently mapped to the SEA-AD taxonomy with approximately half being pTau+ (Figure 1). While the pTau+ population primarily consisted of supragranular excitatory neurons (L2/3_IT), several less abundant supertypes also had large fractions of pTau+ cells (Figure 2). Differential expression analysis within L2/3_IT identified several upregulated AD-related genes such as NEFM, NPTXR, and NAP1L5 in the pTau+ population. Disentangling the relationship between cellular vulnerability and pathology is critical to understand early AD pathogenesis. By adopting a standardized cell taxonomy and integrating datasets, we refined the definition of vulnerable cells to distinguish highly resolved pTau-prone cell types and identified a pTau-specific gene expression signature in the MTG.

Throughout an organism’s life, a multitude of complex and interdependent biological systems transition through biophysical processes that serve as indicators of the underlying biological states. Inferring these latent, unobserved states is a goal of modern biology and neuroscience. However, in many experimental setups, we can at best obtain discrete snapshots of the system at different times and for different individuals. This challenge is particularly relevant in the study of Alzheimer’s Disease (AD) progression, where we observe the aggregation of pathology in brain donors, but the underlying disease state is unknown. This paper proposes a biophysically motivated Bayesian framework (B-BIND: Biophysical Bayesian Inference for Neurode-generative Dynamics), where the disease state is modeled and continuously inferred from observed quantifications of multiple AD pathological proteins. Inspired by biophysical models, we describe pathological burden as an exponential process. The progression of AD is modeled by assigning a latent score, termed pseudotime, to each pathological state, creating a pseudotemporal order of donors based on their pathological burden. We study the theoretical properties of the model using linearization to reveal convergence and identifiability properties. We provide Markov chain Monte Carlo estimation algorithms, illustrating the effectiveness of our approach with multiple simulation studies across various data conditions. Applying this methodology to data from the Seattle Alzheimer’s Disease Brain Cell Atlas, we infer the pseudotime ordering of donors. Finally, we analyze the information within each pathological feature to refine the model, focusing on the most informative pathologies. This framework lays the groundwork for continuous pseudotime modeling in the analysis of neurodegenerative diseases.

Single-cell genomics is a powerful tool for studying heterogeneous tissues such as the brain. Yet little is understood about how genetic variants influence cell-level gene expression. Addressing this, we uniformly processed single-nuclei, multiomics datasets into a resource comprising >2.8 million nuclei from the prefrontal cortex across 388 individuals. For 28 cell types, we assessed population-level variation in expression and chromatin across gene families and drug targets. We identified >550,000 cell type–specific regulatory elements and >1.4 million single-cell expression quantitative trait loci, which we used to build cell-type regulatory and cell-to-cell communication networks. These networks manifest cellular changes in aging and neuropsychiatric disorders. We further constructed an integrative model accurately imputing single-cell expression and simulating perturbations; the model prioritized ~250 disease-risk genes and drug targets with associated cell types.

Early stages of deadly respiratory diseases including COVID-19 are challenging to elucidate in humans. Here, we define cellular tropism and transcriptomic effects of SARS-CoV-2 virus by productively infecting healthy human lung tissue and using scRNA-seq to reconstruct the transcriptional program in "infection pseudotime" for individual lung cell types. SARS-CoV-2 predominantly infected activated interstitial macrophages (IMs), which can accumulate thousands of viral RNA molecules, taking over 60% of the cell transcriptome and forming dense viral RNA bodies while inducing host profibrotic (TGFB1, SPP1) and inflammatory (early interferon response, CCL2/7/8/13, CXCL10, and IL6/10) programs and destroying host cell architecture. Infected alveolar macrophages (AMs) showed none of these extreme responses. Spike-dependent viral entry into AMs used ACE2 and Sialoadhesin/CD169, whereas IM entry used DC-SIGN/CD209. These results identify activated IMs as a prominent site of viral takeover, the focus of inflammation and fibrosis, and suggest targeting CD209 to prevent early pathology in COVID-19 pneumonia. This approach can be generalized to any human lung infection and to evaluate therapeutics.

Alzheimer's disease (AD) is an age-associated neurodegenerative disorder characterized by progressive neuronal loss and pathological accumulation of the misfolded proteins amyloid-β and tau1,2. Neuroinflammation mediated by microglia and brain-resident macrophages plays a crucial role in AD pathogenesis1-5, though the mechanisms by which age, genes, and other risk factors interact remain largely unknown. Somatic mutations accumulate with age and lead to clonal expansion of many cell types, contributing to cancer and many non-cancer diseases6,7. Here we studied somatic mutation in normal aged and AD brains by three orthogonal methods and in three independent AD cohorts. Analysis of bulk RNA sequencing data from 866 samples from different brain regions revealed significantly higher (~two-fold) overall burdens of somatic single-nucleotide variants (sSNVs) in AD brains compared to age-matched controls. Molecular-barcoded deep (>1000X) gene panel sequencing of 311 prefrontal cortex samples showed enrichment of sSNVs and somatic insertions and deletions (sIndels) in cancer driver genes in AD brain compared to control, with recurrent, and often multiple, mutations in genes implicated in clonal hematopoiesis (CH)8,9. Pathogenic sSNVs were enriched in CSF1R+ microglia of AD brains, and the high proportion of microglia (up to 40%) carrying some sSNVs in cancer driver genes suggests mutation-driven microglial clonal expansion (MiCE). Analysis of single-nucleus RNA sequencing (snRNAseq) from temporal neocortex of 62 additional AD cases and controls exhibited nominally increased mosaic chromosomal alterations (mCAs) associated with CH10,11. Microglia carrying mCA showed upregulated pro-inflammatory genes, resembling the transcriptomic features of disease-associated microglia (DAM) in AD. Our results suggest that somatic driver mutations in microglia are common with normal aging but further enriched in AD brain, driving MiCE with inflammatory and DAM signatures. Our findings provide the first insights into microglial clonal dynamics in AD and identify potential new approaches to AD diagnosis and therapy.

Alzheimer's disease (AD) is the most common cause of dementia in older adults. Neuropathological and imaging studies have demonstrated a progressive and stereotyped accumulation of protein aggregates, but the underlying molecular and cellular mechanisms driving AD progression and vulnerable cell populations affected by disease remain coarsely understood. The current study harnesses single cell and spatial genomics tools and knowledge from the BRAIN Initiative Cell Census Network to understand the impact of disease progression on middle temporal gyrus cell types. We used image-based quantitative neuropathology to place 84 donors spanning the spectrum of AD pathology along a continuous disease pseudoprogression score and multiomic technologies to profile single nuclei from each donor, mapping their transcriptomes, epigenomes, and spatial coordinates to a common cell type reference with unprecedented resolution. Temporal analysis of cell-type proportions indicated an early reduction of Somatostatin-expressing neuronal subtypes and a late decrease of supragranular intratelencephalic-projecting excitatory and Parvalbumin-expressing neurons, with increases in disease-associated microglial and astrocytic states. We found complex gene expression differences, ranging from global to cell type-specific effects. These effects showed different temporal patterns indicating diverse cellular perturbations as a function of disease progression. A subset of donors showed a particularly severe cellular and molecular phenotype, which correlated with steeper cognitive decline. We have created a freely available public resource to explore these data and to accelerate progress in AD research at SEA-AD.org.

The Seattle Alzheimer’s Disease Brain Cell Atlas (SEA-AD) is a multifaceted open-data resource that is designed to identify cellular and molecular pathologies that underlie Alzheimer’s disease. Integrating neuropathology, single-cell and spatial genomics, and longitudinal clinical metadata, SEA-AD is a unique resource for studying the pathogenesis of Alzheimer’s disease and related dementias.

ABSTRACTOrgan- and body-scale cell atlases have the potential to transform our understanding of human biology. To capture the variability present in the population, these atlases must include diverse demographics such as age and ethnicity from both healthy and diseased individuals. The growth in both size and number of single-cell datasets, combined with recent advances in computational techniques, for the first time makes it possible to generate such comprehensive large-scale atlases through integration of multiple datasets. Here, we present the integrated Human Lung Cell Atlas (HLCA) combining 46 datasets of the human respiratory system into a single atlas spanning over 2.2 million cells from 444 individuals across health and disease. The HLCA contains a consensus re-annotation of published and newly generated datasets, resolving under- or misannotation of 59% of cells in the original datasets. The HLCA enables recovery of rare cell types, provides consensus marker genes for each cell type, and uncovers gene modules associated with demographic covariates and anatomical location within the respiratory system. To facilitate the use of the HLCA as a reference for single-cell lung research and allow rapid analysis of new data, we provide an interactive web portal to project datasets onto the HLCA. Finally, we demonstrate the value of the HLCA reference for interpreting disease-associated changes. Thus, the HLCA outlines a roadmap for the development and use of organ-scale cell atlases within the Human Cell Atlas.

Alzheimer’s disease (AD) is the most common form of dementia. The progression of AD throughout the brain follows a stereotypical pattern that can be described by the quantification of histopathological changes in several brain regions. As part of The Seattle Alzheimer’s Disease Cell Atlas (SEA-AD, https://sea-ad.org ), we previously quantified pathological proteins in the middle temporal gyrus (MTG) and used them to describe disease progression, vulnerable cell types, and the MTG molecular changes related to disease In this work, we expanded our results and undertook a comparative approach. We measured a battery of neuropathological proteins (staining NeuN, AT8, IBA1, GFAP, 6e10, pTDP43, and a-Syn immunoreactivity, and analyzing them with machine learning) in the middle temporal gyrus, the middle frontal gyrus, and medial entorhinal cortex in 84 donors. In addition, we undertook a multimodal single-cell analysis of such regions, profiling cells using single-cell RNA-seq, ATAC-seq, Multiome, and spatial transcriptomic. Borrowing from our hierarchically resolved cell types in MTG, based on the BRAIN initiative cell type reference, we used machine learning algorithms to create multimodal cell type maps across regions. We used Bayesian statistics to identify differentially accessible regulatory elements, link them to the affected genes and the regulatory transcriptional machinery binding in such elements. We identified regional vulnerable neuronal populations and a core set of transcriptionally similar cell types susceptible to disease across the cortex. We created a cortex-wide disease progression map combining neuropathological data from our three profiled regions with latent Bayesian statistical algorithms. In each area, we developed an AD pseudo-progression timescale that describes the burden of disease in the area and, next, combined all areas to create a cortex-wide holistic pseudo-progression. We used the mapped cell types and our developed scale to synchronize changes occurring in each patient, describing abundances, gene expression, and chromatin accessibility modifications. The progression scale developed in this study serves as a reference to align changes across brain regions and, in the future, to reconcile differences across cohorts.

While traditional histologic neuropathological assessment of neurodegenerative disease has tremendous diagnostic value, insights into neurodegeneration’s pathophysiological underpinnings can be significantly enhanced with large-scale quantitative analyses. As part of the Seattle Alzheimer’s Disease Brain Cell Atlas (SEA-AD), we present a neuropathology dataset across 84 brains spanning the spectrum of AD neuropathologic change donated by UW Alzheimer’s Disease Research Center and Adult Changes in Thought study participants. Human brain tissue was collected at rapid autopsy (postmortem interval <12 hours). One hemisphere (randomly selected) was embedded in alginate for uniform coronal slicing (4mm), with alternating slabs fixed in 10% neutral buffered formalin or frozen in a dry ice isopentane slurry. Superior and Middle Temporal Gyrus (MTG) was sampled from fixed slabs and subjected to standard processing, embedding, sectioning (5µm), histochemical (H&E/Luxol fast blue), and immunohistochemical (duplex pTau and pTDP-43, Aβ and Iba1, and monoplex GFAP, a-synuclein, and NeuN) labeling. HALO image analysis software (Indica Labs, Albuquerque NM) was used to define and quantify relevant cellular and pathological features within each cortical layer (Figures 1-3). 84 individuals were included in these analyses (Table 1). While quantitative metrics of pTau and Aβ aligned with traditional neuropathological measures, they showed substantial variability within discrete diagnostic stages (Figure 4). These findings were robust to quantification (discrete count vs. total area stained) and normalization (total nuclei vs. total area analyzed) methods, and highlight the importance of deploying highly quantitative, continuous measures of pathology in the study of neurodegeneration. This unique resource provides spatially preserved measures of pathologic burden across the spectrum of AD with potential for correlative analysis across pathologies, clinical and biomarker parameters, and omics applications to better support the identification of vulnerable cell types in early stages of AD progression, the principal goal of SEA-AD. Here we present a large-scale quantitative neuropathology dataset, which enables analysis of pathological proteins and cell types at unprecedented scale and accuracy. Little is known about the interactions of AD-relevant cell types with pathologic peptides across disease progression. The presented analyses exemplify the potential of this unique publicly-accessible resource to identify and quantify novel disease-relevant relationships.

The Seattle Alzheimer’s Disease Cell Atlas (SEA-AD, https://sea-ad.org ) is a large-scale effort to identify, at single cell resolution, the cellular and molecular mechanisms that cause AD and facilitate its progression. SEA-AD brings together quantitative neuropathology (QNP) derived from classic histopathological and cell type stains with single nucleus transcriptomics, epigenetics, and spatial transcriptomics on a cohort of 84 aged donors that span the pathological and cognitive disease spectrum, including unaffected controls. We ordered SEA-AD donors based on QNP using a Bayesian latent space model to construct a model-estimated “pseudo-progression” (PS). We then used a general linear mixed effects model to test for gene expression changes along the PS in each of the BRAIN initiative’s 139 cell types present in SEA-AD’s single nucleus RNA sequencing dataset (∼1.2 million nuclei) from the middle temporal gyrus. By separately fitting models to donors with lower and higher pathology, we obtained two beta coefficients for each gene that enabled distinction between early, late, and consistent up- and down-regulation along PS. We then tested whether the betas for each gene were different across all cell types or in only certain ones to measure specificity. We curated gene modules from biological processes potentially relevant to AD, identified in gene-set enrichment analyses and prior studies, and categorized each gene within them based on their dynamics and specificity. Roughly one-third of genes had differential expression across neuronal cell types (mean betas > 0.2), split evenly between up- and down-regulation, while a smaller subset were changed in specific cell types. Modules enriched in broadly changed genes that had increased expression early in PS included those involved with the microtubule cytoskeleton, vesicle adaptors, and neurotransmitter receptors, while those with decreased expression included those involved with cholesterol synthesis, glycolysis, mitochondrial function, and neurofilaments. Among the genes changed only in specific cell types were NOTCH2 in Somatostatin and Parvalbumin interneurons selectively vulnerable to pathology, PTPRG in disease associated microglia, and COL21A1 in disease associated astrocytes. Broadly affected modules identified by SEA-AD suggest most neurons are affected by disease pathology and cell type-specific differences may provide clues to selective vulnerability.

Variation in cytoarchitecture is the basis for the histological definition of cortical areas. We used single cell transcriptomics and performed cellular characterization of the human cortex to better understand cortical areal specialization. Single-nucleus RNA-sequencing of 8 areas spanning cortical structural variation showed a highly consistent cellular makeup for 24 cell subclasses. However, proportions of excitatory neuron subclasses varied substantially, likely reflecting differences in connectivity across primary sensorimotor and association cortices. Laminar organization of astrocytes and oligodendrocytes also differed across areas. Primary visual cortex showed characteristic organization with major changes in the excitatory to inhibitory neuron ratio, expansion of layer 4 excitatory neurons, and specialized inhibitory neurons. These results lay the groundwork for a refined cellular and molecular characterization of human cortical cytoarchitecture and areal specialization.

Alzheimer’s disease (AD) is the most common form of dementia, representing ∼70% of all dementia cases, and affecting much of the aged population. Currently our understanding of AD neuropathology is largely centered on characteristic deposition of certain neuropathological proteins and identification of histologically defined neuronal populations. Molecular tools to characterize the transcriptome, epigenome, and spatial organization of single cells in complex brain tissues allow a more refined look at how the neurotypical brain changes in AD than had been possible previously. Linking molecular, cellular, genomic, cognitive, and pathological indicators of AD progression will be essential for understanding disease mechanisms. The Seattle Alzheimer’s Disease Brain Cell Atlas (SEA-AD) brings together experts in AD research and large-scale molecular/anatomical brain mapping to modernize AD tissue banking methods, and to combine traditional and quantitative neuropathology with emerging genotyping, single nucleus transcriptomics, single nucleus epigenomics, and spatial transcriptomics technologies. Applied to multiple brain regions in donors spanning different stages of AD pathology, these techniques will identification of cell type specific molecular pathways and yield valuable insights into selective vulnerability or resistance to pathology. Two well characterized cohorts (Adult Changes in Thought (ACT) and the University of Washington ADRC) provided 84 subjects of varying degrees of AD severity, including cognitively normal subjects. The initial focus of SEA-AD was on integrating and interpreting relationships between all collected multimodal data in middle temporal gyrus (MTG). Identified cell types and molecular pathways of interest are presented separately and are both consistent with and expand upon results found from ROSMAP and other cohorts. The complete SEA-AD MTG atlas is now freely and publicly available at brain-map.org and includes access to all data and associated donor metadata, tools for visualization and exploration of molecular data, exploration of pathology images, a tool for assigning cell types to community-based -omics datasets, comprehensive documentation, and user support. SEA-AD has developed a high resolution, high-quality, publicly accessible atlas of aging and AD, and is designed to better understand the complex and likely heterogenous nature of AD.

Single-cell transcriptomic studies have identified a conserved set of neocortical cell types from small postmortem cohorts. We extended these efforts by assessing cell type variation across 75 adult individuals undergoing epilepsy and tumor surgeries. Nearly all nuclei map to one of 125 robust cell types identified in the middle temporal gyrus. However, we found interindividual variance in abundances and gene expression signatures, particularly in deep-layer glutamatergic neurons and microglia. A minority of donor variance is explainable by age, sex, ancestry, disease state, and cell state. Genomic variation was associated with expression of 150 to 250 genes for most cell types. This characterization of cellular variation provides a baseline for cell typing in health and disease.