Researchers uncover distinct cell vulnerabilities in Alzheimer’s progression with novel multiomics approach

By Hugo Francisco de SouzaReviewed by Susha Cheriyedath, M.Sc.Oct 16 2024

Scientists reveal how distinct brain cells contribute to Alzheimer’s progression, unlocking new insights for developing personalized treatments and improving the accuracy of diagnosis across disease stages.

Study: Integrated multimodal cell atlas of Alzheimer’s disease. Image Credit: illustrissima / Shutterstock

In a recent study published in the journal Nature Neuroscience, researchers combined single-nucleus RNA sequencing (snRNA-seq), spatial genomics, single-nucleus assay for transposase-accessible chromatin with sequencing (snATAC-seq), multiomics, and preexisting reference atlases to evaluate the molecular and cellular alterations in the middle temporal gyrus (MTG) across the spectrum of Alzheimer’s disease (AD) progression.

They further used quantitative neuropathology in conjunction with a machine learning model to devise a patient-specific pseudoprogression score (CPS), a continuous metric that orders donors along a neuropathological continuum.

Study findings revealed the presence of two distinct major disease phases (early/slow and late/exponential), each with unique cell physiology.

A small subset of donors in the study also exhibited a third ‘terminal’ disease phase characterized by more severe pathology.

Notably, the paper provides a framework for integrating previously confounding lines of evidence, allowing for cross-validation of Alzheimer’s disease observations across studies, thereby increasing the robustness and consistency of study findings across seemingly disconnected investigations.

Background

Alzheimer’s disease (AD) is a neurological disorder characterized by the progressive accumulation of amyloid beta (Aβ) plaques and hyperphosphorylated Tau (pTau) in the cerebral, brainstem, and limbic systems. This results in the shrinkage of brain cells, loss of neural connections, and even cell degradation and death, leading to a loss of memory and routine functional capabilities.

AD is a global public health risk, currently estimated to impact more than 55 million patients and their families. Alarmingly, epidemiological projections expect AD prevalence to increase to 78 million by 2030 and 139 million by 2050, making it the fastest-growing neurological disease in today’s world.

AD is the leading risk association with dementia, further spurning neuroscience research aimed at determining the risk factors, pathophysiological mechanisms, and severity of the disease.

Cell morphology and physiology changes during AD progression have been extensively characterized, resulting in the formulation of ‘aggregate scores’ (e.g., Braak, CERAD, Thal, and ADNC) to describe AD severity.

Unfortunately, conventional studies routinely describe brain location-specific changes but fail to specify the vulnerable and disease-associated cell-type-specific alterations accompanying AD progression.

Recent advances in spatial and single-cell genomics technologies, alongside the widespread adoption of multiomics analyses, have given rise to ‘brain cell atlases’ – detailed, high-resolution, brain-wide cell physiology references of cellular properties across genomics, transcriptomics, and patch sequencing data approaches.

These provide a standardized knowledge base of AD- and dementia-associated brain cell types, substantially augmenting our understanding of the processes underpinning AD risk and severity.

In particular, these approaches allow for mapping cellular changes to highly curated brain reference atlases, improving our understanding of which cell types are most vulnerable in the disease’s early and late stages.

About the study

The present study combines high-resolution single-nucleus- and multiomics approaches with quantitative neuropathology-inspired temporal disease modeling, thereby highlighting the myriad of unique cell types and the alterations they experience as AD progresses. It focuses on the middle temporal gyrus (MTG), the brain region associated with semantic retrieval and language processing.

It underscores cell characteristics and location and the morphological or gene expression changes that occur during various stages of AD.

Since AD stages have been poorly defined (dependent on conventional aggregate scores), the present study employs quantitative neuropathology coupled with immunohistochemistry (IHC) and Bayesian inference models to identify discrete AD progression stages.

Study data was obtained from 84 postmortem donors (51 females, ages – 65 to 102) from two independent studies (University of Washington ADRC and Kaiser Permanente Washington Health Research Institute ACT Study) consolidated in the UW BioRepository and Integrated Neuropathology (BRaIN) dataset.

Participants were screened to include those undergoing ‘precision rapid procedure’ – a standardized methodology of optimized tissue collection and preservation – while excluding those with histories of confounding degenerative disorders.

Experimental procedures included immunohistochemistry (IHC; single and duplex) for quantitative neuropathology analyses. IHC outcomes were used within a Bayesian inference framework to compute a novel’ continuous pseudoprogression score (CPS),’ an objective measurement ordering study participants along a neuropathological continuum of AD progression.

Single-nucleus cell isolations were obtained from MTG cortical areas using cryo-dissections, flow cytometry, and snRNA-seq libraries built from these nuclei. Genomic reads were compared to reference atlases using platforms like ChromA, providing unprecedented resolution on cell-type-specific vulnerabilities. Spatial transcriptomics (MERSCOPE platform) and patch-seq data (publicly available) completed the dataset, allowing for validation of the CPS and estimation of the electrophysiological features of different cell types across varying AD phases.

Study findings

The present study updates the BRAIN Initiative Cell Census Network (BICCN) with novel MTG-specific cell and disease phase information.

The quantitative neuropathological-derived CPS metric computed here demonstrated the presence of two typical (early/slow progression and late/exponential progression) epochs in AD progression. A rare subset of older patients further demonstrated a third ‘terminal’ epoch.

“In the early epoch, donors had sparse Aβ plaques (albeit increasing in size) and pTau⁺ tangle-bearing neurons, accompanied by early increases in inflammatory or reactive microglial and astrocytic states and associated gene expression changes in relevant inflammatory and plaque-induced genes. In the later epoch, there is an exponential rise in Aβ and pTau pathology, continued increases in inflammatory microglia and astrocyte states, and a decrease in the expression of both the OPC differentiation program and oligodendrocyte expression of myelin-associated proteins (previously characterized using quantitative PCR).”

The study identified multiple vulnerable cell types, including excitatory neurons in layer 2/3 (L2/3 IT), somatostatin (Sst) inhibitory neurons, and oligodendrocytes. These cells showed early vulnerability, while other types like Pvalb+ interneurons declined later in disease progression.

Extensive, previously hidden cascades of cell-type-specific activation and excitation were observed, suggesting that microglial activation during early (low severity) AD stages triggers losses of astrocytes, oligodendrocytes, and corticocortical (L2/3 IT) neurons, in turn contributing to cognitive dysfunction.

Additionally, spatial transcriptomics confirmed the correlation between specific vulnerable cell populations and AD severity, particularly in supragranular layers of the cortex.

These observations were strongest in participants exhibiting the most extreme later-life cognitive decline, suggesting a biological underpinning. CPS analyses further identified neuronal and non-neuronal subtypes at increased risk of AD and dementia (n = 58).

Most importantly, the present work provides a platform and methodologies allowing for integration, direct comparisons, and standard annotations (cell states and types), thereby enhancing the consistency and robustness of future AD research.

Conclusions

The present study employs multiple cutting-edge neuropathological techniques, including single-cell nucleus genomics, multiomics, and quantitative neuropathology analyses, to reveal changes in individual MTG cell types across different stages of AD progression. It identifies discrete epochs of AD progression and the physiological changes accompanying these stages.

The study also highlights specific vulnerable neuronal subtypes, such as Sst and L2/3 IT neurons, and their critical role in cognitive decline associated with AD. It unravels genetic, demographic, and behavioral risk associations, potentially exacerbating AD severity.

Most importantly, the present work provides a database and standardization suggestions to improve future AD research.