Summary: Researchers discovered that the three-dimensional architecture of the genome is fundamentally altered in specific brain cells of individuals with Alzheimer’s disease.
The study integrates single-cell GAGE-seq technology, spatial tissue mapping, and a novel deep learning framework named Hicformer to connect DNA spatial folding directly to gene expression and tissue pathology in the prefrontal cortex. The research reveals that Alzheimer’s is marked by “increased compartment mingling,” a breakdown where active and inactive genomic regions lose their distinct spatial boundaries.
This spatial genome degradation weakens gene-regulatory contacts, leading to reduced synaptic function in neurons, metabolic stress, and cellular senescence in microglia. The findings establish higher-order chromatin reorganization as a primary layer of Alzheimer’s pathology alongside traditional amyloid-beta plaques and tau tangles.
Key Facts
- Increased Compartment Mingling: Large active and inactive chromatin domains (A/B compartments) lose clear spatial separation in Alzheimer’s brain cells, leading to lower overall gene expression and disrupted cell function.
- Weakened Regulatory Connections: Contacts between genes and nearby enhancer regulatory elements weaken, while abnormal midrange and long-range genomic contacts increase across multiple brain cell types.
- Hicformer AI Framework: The team developed Hicformer, a deep learning transformer model that combines DNA sequences, spatial folding features, and local 3D contact maps to predict cell-type-specific gene activity.
- Cellular Pathology Link: Genomic structural decay directly correlates with impaired neuronal synaptic programs, altered metabolic and stress pathways, and senescence-associated programs in microglia.
- Multi-Scale Spatial Integration: By combining single-cell GAGE-seq (measuring gene expression and 3D contacts in the same cell) with spatial transcriptomics, researchers mapped genomic spatial changes directly onto intact prefrontal cortex tissue architecture.
Source: University of Pittsburgh
Researchers from Carnegie Mellon University’s School of Computer Science, the University of Pittsburgh School of Medicine and the University of Washington shed new light on Alzheimer’s disease that could point to new directions for treatment.
In a paper published in Science, researchers from SCS’s Ray and Stephanie Lane Computational Biology Department, Pitt’s Department of Neurobiology and collaborating institutions shows that the 3D genome architecture is organized differently in certain brain cells from people with Alzheimer’s disease, uncovering a previously underexplored layer of this disease’s biology.
The research team linked genome folding to gene activity and brain tissue organization in Alzheimer’s disease. The research team did this with single-cell technology, spatial mapping of brain tissue, and a new deep learning model.
“Alzheimer’s disease cannot be understood one layer at a time,” said Jian Ma, the Ray and Stephanie Lane Professor of Computational Biology who led and supervised the study.
“The genome’s 3D structure is a fundamental regulatory layer that helps to connect DNA sequence to gene activity. By integrating genome folding, cell state and tissue context, we can move beyond cataloging disease-associated changes toward understanding how they fit together and which mechanisms to test next.”
To build this multi-scale view, researchers analyzed postmortem tissue from the prefrontal cortex, a region at the front of the brain, obtained from individuals with and without Alzheimer’s disease who had participated in a long-term study on dementia and donated their brains to science after death. The team used GAGE-seq, which measures gene expression and 3D genome contacts in the same cell.
They also integrated these measurements with spatial transcriptomic maps of intact tissue. Together, these complementary data allowed the researchers to relate 3D genome organization to gene regulation and place disease-associated molecular and cellular changes within their broader tissue context.
“Our study represents a major advance in understanding what goes wrong in Alzheimer’s disease,” said Hansruedi Mathys, assistant professor of neurobiology at Pitt’s Department of Neurobiology, who directed the Pitt arm of the study.
“We know the classic hallmarks of Alzheimer’s disease – accumulation of amyloid-beta plaques and tau tangles – but our results establish higher-order chromatin alterations as a component of the molecular pathology associated with the disease, which currently affects seven million Americans, a number that continues to grow.”
A key computational advance was Hicformer, an AI model that combines DNA sequence, broad genome-folding features and local 3D contact maps to predict gene activity in different kinds of cells. Xinyue Lu, a doctoral student in Computational Biology who co-led the research, described Hicformer as a computational test bed for asking how altered genome folding may change gene activity.
“Measuring gene activity and genome folding in the same cell allows us to directly connect chromosome structure with disease-related gene programs,” said Yang Zhang, a project scientist in the Computational Biology Department who co-led the research. “Across several kinds of brain cells, this paired view revealed a consistent signature of 3D genome reorganization in Alzheimer’s disease and helped us prioritize regulatory regions for future mechanistic and therapeutic investigation.”
Researchers found that large active and inactive regions of the genome, known as compartments, were less clearly separated in cells from people with Alzheimer’s, a pattern the team calls “increased compartment mingling.”
Multiple kinds of brain cells showed fewer short-range contacts and more long-range contacts, and greater compartment mingling was associated with lower overall gene activity. Contacts between genes and nearby regulatory elements that help control gene activity also weakened, while some midrange contacts strengthened. These architectural alterations were linked to reduced neuronal and synaptic programs, altered metabolic and stress responses, and senescence-related programs in microglia.
By mapping these changes across intact brain tissue, the researchers found that this reorganization of the genome was linked to changes in gene activity and in how brain cells were organized. The findings identify 3D genome organization as an important layer of Alzheimer’s biology and provide a framework for future experiments to determine which changes in genome structure contribute directly to the disease and whether they could reveal new therapeutic targets.
Funding: This research was supported by grants from the National Institutes of Health.
Other CMU authors included doctoral students Shahul Alam and Shike Wang and postdoctoral research associate Junjie Tang. Other Pitt authors include doctoral students Alexander K. Kunisky and Jude Baroudi, post-baccalaureate research fellows Sahar and Sahel Ghorbanikalateh, and visiting scholar Shihan Wang.
The team included researchers from the Broad Institute of MIT and Harvard; the University of California, Los Angeles; the University of Washington; and the Rush Alzheimer’s Disease Center.
Key Questions Answered:
A: In healthy cell nuclei, active regions of DNA (A compartments) and inactive regions (B compartments) are spatially segregated. In Alzheimer’s brain cells, these boundaries blur, allowing active and inactive DNA regions to physically mix, which suppresses overall healthy gene expression.
A: Hicformer is a deep learning AI model created by the computational team. It combines DNA sequence data, broad genome-folding patterns, and local 3D contact maps to predict gene activity, serving as a virtual test bed to determine how changes in chromosome structure alter specific cell functions.
A: While amyloid plaques and tau tangles are classic hallmarks of Alzheimer’s, they do not fully explain how or why specific cell programs shut down. Mapping 3D genome architecture reveals an underlying regulatory layer that controls how genes are turned on or off, pointing to new therapeutic targets to preserve neural circuits.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- Journal paper reviewed in full.
- Additional context added by our staff.
About this Alzheimer’s disease and genetics research news
Author: Allison Hydzik
Source: University of Pittsburgh
Contact: Allison Hydzik – University of Pittsburgh
Image: The image is credited to Neuroscience News
Original Research: Open access.
“Single- cell multiomics connects 3D genome and transcriptome alterations in Alzheimer’s disease” by Yang Zhang, Xinyue Lu, Alexander K. Kunisky, Shahul Alam, Junjie Tang, Ruochi Zhang, Shike Wang, Han Zhang, Jude Baroudi, Walid Ichcho, Deyong Jia, Sahar Ghorbanikalateh, Sahel Ghorbanikalateh, Shihan Wang, David A. Bennett, Hansruedi Mathys, Zhijun Duan, Jian Ma. Science
DOI:10.1126/science.adz1652
Abstract
Single- cell multiomics connects 3D genome and transcriptome alterations in Alzheimer’s disease
INTRODUCTION
Alzheimer’s disease (AD) is the most common cause of dementia and is marked by progressive loss of brain function. Many studies have cataloged changes in gene activity across brain cell types; however, the molecular mechanisms underlying these changes remain elusive. Gene activity is controlled not only by DNA sequence and chemical marks on DNA but also by how the genome is folded in three dimensions inside the nucleus. How this three-dimensional (3D) genome organization changes in AD and how such changes relate to cell type–specific gene dysregulation in the human brain remain poorly understood.
RATIONALE
We aimed to determine whether changes in 3D genome folding are linked to the gene expression programs disrupted in AD and whether these links can be detected at single-cell resolution in human brain tissue. To do this, we used GAGE-seq (genome architecture and gene expression by sequencing), a technology that measures, in the same single cell, both gene expression and physical contacts within the genome. We integrated these measurements with chromatin accessibility data from the same donors and with spatial transcriptome maps in intact tissue sections, enabling analyses across molecular, cellular, and tissue scales. We also developed a transformer-based predictive model that integrates DNA sequence with 3D genome features to test when genome structure is necessary to explain AD-related gene expression changes.
RESULTS
Across major brain cell types, we observed a reproducible shift in genome contact patterns in AD, including reduced short-range interactions and increased longer-range interactions. Although overall compartment patterns were broadly preserved, active and inactive genome regions showed increased mixing, consistent with weakened compartment segregation. These architectural changes were linked to broad, cell type–specific transcriptional remodeling of disease-relevant pathways, and we further related these programs to the current landscape of AD clinical trial targets.
At regulatory elements identified by chromatin accessibility, promoter-proximal interactions weakened, while midrange regulatory interactions became relatively more prominent, particularly at sites associated with chromatin loop organization. In parallel, we observed AD-related changes in key cellular programs, including senescence-related activation in microglia and sex-dependent dysregulation of X-linked genes in females, accompanied by corresponding 3D genome changes at implicated loci.
Our predictive model showed that 3D genome features provide information beyond DNA sequence alone for explaining AD-relevant gene expression, enabling prioritization of distal regulatory elements whose effects are mediated through chromatin contacts. Integrating these molecular features with spatial transcriptomics further placed them in tissue context, revealing altered cell neighborhoods and disrupted spatial coordination of gene programs in AD. Overall, the study connects genome structure, gene regulation, and tissue organization through a unified multimodal analysis.
CONCLUSION
These results provide a multiscale map linking 3D genome remodeling to cell type–specific gene expression changes and spatial tissue organization in AD. The study establishes genome folding as a key regulatory layer associated with AD pathology and provides a framework and resource for mechanistic hypothesis generation, including prioritization of regulatory elements and AD-relevant gene programs for future functional testing and therapeutic exploration.
