Protective Microglia Found in Atlas of 1,607 Human Brains
Icahn School of Medicine at Mount Sinai mapped 832,505 immune cells from the prefrontal cortex of 1,607 donors and found a microglia subtype that expands as Alzheimer's progresses.
A microglia subtype that becomes more abundant as Alzheimer’s disease advances appears to defend brain tissue rather than damage it. That is the finding of a human brain immune cell atlas published by a team at the Icahn School of Medicine at Mount Sinai in New York on August 11 in the journal Nature Genetics, built from 832,505 cells taken from the prefrontal cortex of 1,607 donors.
Microglia are the brain’s resident immune cells. Alongside them, the team also examined perivascular macrophages, cells that regulate the immune response around blood vessels. That the brain’s immune system plays a role in Alzheimer’s disease was already known, mainly because inherited variants in the immune genes TREM2 and APOE raise disease risk. What was not known was how these cells change as the disease advances, since answering that requires data from many brains, not a few dozen.
How the atlas was built
The material consisted of two datasets covering the same brain region, the dorsolateral prefrontal cortex. The first included 137 donors and 543,012 cells isolated from fresh autopsy tissue, the second covered 1,470 donors and 289,493 cell nuclei from frozen tissue. The donors spanned a range from younger to the oldest individuals and covered different stages of Alzheimer’s pathology, from brains with no changes to advanced cases.
From this material, the researchers identified 13 transcriptionally distinct myeloid cell subtypes, grouped into 6 subclasses. The classification is not based on how a cell looks but on which genes it is reading at a given moment, so the same microglia can fall into two different subtypes depending on its state.
What the GPNMB-high subtype does
One microglia subtype, marked by elevated expression of the gene GPNMB, becomes increasingly abundant as pathology advances. Rather than contributing to damage, this subtype shows heightened activity in engulfing and clearing harmful material from tissue. The same subtype turned out to be enriched for polygenic Alzheimer’s risk, meaning it sits exactly where genetic risk accumulates.
MITF was identified as the regulator that keeps the cell in this state, with the effect depending on signaling through TREM2. Disease progression was also linked to the signaling pairs APOE and SORL1, and APOE and TREM2. The findings were validated in both human tissue and mouse models.
The study’s first and corresponding author is Donghun Li, assistant professor of genetics and genomic sciences and of psychiatry at the Icahn School of Medicine at Mount Sinai. The work was overseen by John F. Fullard and Panos Roussos.
Our study provides the clearest picture yet of how the brain’s immune cells adapt during aging and Alzheimer’s disease.
In the Mount Sinai statement, Li also explained why the team believes the finding has practical value.
By identifying the specific immune cells that appear to protect the brain, and the molecular signals they rely on, we have uncovered potential new targets for therapies aimed at slowing the progression of Alzheimer’s disease.
The data did not stay in the lab. Single-cell profiles, clinical metadata and analysis results are available through the AD Knowledge Portal and the CELLxGENE browser, and the analysis code has been posted to GitHub and Zenodo, letting other teams replicate the analysis on the same material.
What this does not prove
The atlas is a snapshot, not a film. Tissue was collected after death, once from each brain, so changes over time are reconstructed by comparing donors at different disease stages rather than by following the same person. For that reason, the protective role of this microglia subtype is for now described as an association with slower pathology progression, backed by experimental confirmation in mouse models, not as a proven cause in humans.
The statement does not mention any clinical trial, nor any drug targeting the TREM2, MITF and GPNMB pathway. A new target is a target, not a therapy, and the path from a genomic atlas to an applicable drug is measured in years. What can be said from this study is that the brain mounts its own defense in Alzheimer’s disease, one that strengthens as the disease advances, and that the cells making up that defense are now known.
