For decades, the brain was treated as an immune-privileged organ — a walled-off compartment where the blood-brain barrier kept circulating immune cells out and left microglia, the brain’s resident immune cells, to self-renew locally for life. A new study published in Nature, led by researchers at Stanford University, upends that picture: as the human brain ages, it recruits reinforcements. Immune cells born in blood stem cells cross into brain tissue starting in middle age and differentiate into microglia-like cells, quietly reshaping the brain’s immune landscape from the inside.
What the study found
The paper, titled “Somatic mutations reveal the ontogeny of microglia in human aging,” reports that a meaningful fraction of the microglia found in aged human brains did not originate there. Instead, they trace back to blood stem cells and appear to have migrated into brain tissue over the course of a person’s life. “We usually think of the brain as a closed system,” said first author Julia Belk. “What we found is that actually a lot of immune cells enter the human brain during aging.” Senior author Siddhartha Jaiswal added that “the life history of blood stem cells could influence the risk of brain diseases by altering the microglia” — a direct link between blood cell biology, long studied in the context of clonal hematopoiesis and blood cancers, and neurological aging.
How they traced the cells: somatic mutations as a lineage map
The methodology is what makes the claim credible rather than speculative. As blood stem cells divide over a person’s lifetime, they randomly accumulate small DNA changes — somatic mutations. Every immune cell descended from a given stem cell inherits that same mutational signature, effectively a genetic barcode. By sequencing paired blood and postmortem brain tissue from the same individuals and matching mutation signatures between the two, the researchers could show that specific microglia-like cells in the brain shared ancestry with specific blood cell lineages. That shared signature is difficult to explain any other way, which is what elevates this from correlation to a genuine lineage-tracing result.
The ethics and data-provenance angle: how you get paired blood-and-brain tissue from humans
Studies like this are only possible because of long-running human-tissue donation infrastructure. The brain tissue came primarily from Stanford’s Rapid Autopsy Center, a program (with co-author Jody Hooper among those involved) that depends on rapid, consented postmortem tissue recovery from brain donors so that RNA and cellular structures remain intact enough for high-resolution sequencing. The second major tissue source was the University of Washington-led Alzheimer’s Disease Sequencing Project (ADSP), a federally supported genomic data resource that aggregates sequencing data and matched clinical and tissue information from people with and without Alzheimer’s disease. Neither resource is trivial to build: both depend on years of donor recruitment, informed consent from patients and families facing terminal or end-of-life circumstances, and careful chain-of-custody handling of tissue between clinical sites and sequencing labs. The result is a rare paired dataset — blood and brain from the same person — that most research programs simply do not have access to, and that underpins the entire lineage-tracing approach used here.
Why mice and monkeys didn’t show this
One of the study’s most consequential findings is what it did not find in standard animal models. Belk noted plainly that the blood-to-brain microglia replenishment observed in humans “doesn’t appear to happen in other species, such as mice or nonhuman primates.” That is a significant limitation for the field, because mouse models are the default tool for studying microglial biology and neuroinflammation, and macaques are often used as a closer proxy for human brain aging when mouse results need validation. If a core aging mechanism is human-specific, it cannot be discovered, and possibly cannot even be fully validated, in either standard model organism — it can only be seen by working directly with human donor tissue. That reframes how much weight the field should put on mouse-derived conclusions about microglia turnover and brain immune aging, and strengthens the case for continued investment in human tissue-donation infrastructure like the Rapid Autopsy Center and ADSP, rather than treating animal models as a full substitute.
Why it matters
Microglia are increasingly implicated in Alzheimer’s disease and other neurodegenerative conditions, both as first responders to amyloid and tau pathology and as potential drivers of chronic neuroinflammation. If a portion of the aging brain’s microglial population originates from blood stem cells rather than long-resident local cells, then anything that shapes blood stem cell health over a lifetime — including clonal hematopoiesis, a well-studied age-related expansion of mutant blood stem cell clones — could plausibly influence brain disease risk indirectly, through the immune cells it sends into the brain. That is a new, testable hypothesis rather than a settled conclusion, and the paper’s authors frame it that way, but it opens a research direction connecting hematology and neurodegeneration that did not have this kind of direct evidentiary link before.
Source
Primary source: Belk, J. et al., “Somatic mutations reveal the ontogeny of microglia in human aging,” Nature (2026), DOI: 10.1038/s41586-026-10939-0. Coverage: Medical Xpress, “Discovery reveals aging human brains receive immune cell reinforcements from blood,” published August 6, 2026.
