Researchers found that the 3D folding of DNA in human brain cells is altered in Alzheimer’s, blurring boundaries between active and inactive genome regions. Using GAGE-seq, spatial mapping and a new AI model (Hicformer), the team linked these structural changes to reduced neuronal gene activity, weakened gene-regulatory connections and microglial aging signatures. The findings position genome architecture as an additional layer of Alzheimer’s biology and a potential avenue for future research into therapies, though causality and translational steps remain to be established.
Hidden Alzheimer’s Clue: Disrupted 3D DNA Folding Found in Human Brain Cells
An illustration of a human brain.
Scientists report a previously unrecognized layer of Alzheimer’s disease biology: the three-dimensional folding of DNA inside certain brain cells is disrupted, altering which genes are switched on or off. The finding points to genome architecture and epigenetic regulation as potential new directions for research into disease mechanisms and therapies.
What The Study Did
Researchers from Carnegie Mellon University, the University of Pittsburgh and the University of Washington analyzed postmortem prefrontal cortex tissue donated by participants in a long-term dementia cohort. Using GAGE-seq — a technique that measures gene expression and 3D genome contacts in the same cell — and spatial mapping of intact tissue, the team linked changes in genome folding to shifts in gene activity and to the physical arrangement of cells in the brain.
Key Findings
The study identified several consistent alterations in cells from Alzheimer’s-affected brains:
“We know, clearly, that we are very different from a mouse, so having more studies that use human biology is really important,” said Dr. Lucy Hooper, a medical doctor and co-founder of Coyne Medical, who was not involved in the study.
Under normal conditions, DNA segregates into active and inactive compartments. In Alzheimer’s cells the boundaries between these compartments were blurred — a pattern the team calls "increased compartment mingling." Many cell types also showed fewer short-range DNA contacts but more long-range interactions; cells with more mingling generally had lower overall gene activity.
Researchers observed weakened links between genes and their regulatory elements and decreased activity of genes involved in neuronal function, synaptic processes, metabolism and cellular stress responses. Microglia — the brain’s immune cells — displayed related changes associated with cellular aging signatures. By mapping these molecular alterations across tissue sections, the team showed that genome reorganization correlated with how cells are spatially arranged in the cortex.
Tools And Modeling
To interpret how genome folding affects cellular behavior, the authors developed a new AI model, Hicformer, trained on single-cell multiomic data. Combining GAGE-seq, spatial transcriptomics and Hicformer enabled the researchers to connect structural genome features directly to transcriptional and tissue-level changes.
Implications And Limits
These results establish genome architecture as an additional molecular layer in Alzheimer’s disease and provide a framework for testing which structural changes might contribute directly to disease progression. Because folding and other epigenetic features can be dynamic, they could represent more tractable therapeutic targets than DNA sequence changes. However, the work is descriptive and performed on postmortem tissue; further studies are required to determine causality, timing in disease progression and potential for clinical translation.
Citation: Zhang, Y. et al. (2026). Single-Cell Multiomics Connects 3D Genome And Transcriptome Alterations In Alzheimer's Disease. Science. https://doi.org/10.1126/science.adz1652.
Help us improve.
