The brain has its own immune system, staffed for most of a lifetime by cells that arrive early in development and stay. A new study of the human hippocampus, the memory center, suggests that around age 50 this long-term staff begins to be replaced by newcomers: cells carrying the inflammatory signatures of blood-derived immune cells. The shift, visible only when gene expression is combined with epigenetic and three-dimensional genome data, may help explain why the risk of Alzheimer’s disease begins its steep climb in midlife.
The study, published in Science and led by Nathan Zemke of UC San Diego with Bing Ren of the New York Genome Center and Xiangmin Xu of UC Irvine, profiled postmortem hippocampal tissue from 40 neurologically healthy adults spanning ages 20 to 95. For each donor the team measured four molecular layers from the same cell types: gene expression, chromatin accessibility, DNA methylation, and 3D genome architecture, using single-nucleus techniques across more than 295,000 nuclei.
The centerpiece finding concerns microglia, the brain’s resident immune cells. Textbooks describe microglia as embryonically derived residents that renew themselves in place throughout life; the new data show that assumption was incomplete. By analyzing DNA methylation, which preserves a record of a cell’s origin, the researchers identified two distinct microglial states. In younger adults the hippocampus is dominated by homeostatic microglia; in the oldest donors, almost all microglia resemble a different lineage with a proinflammatory, monocyte-like identity characteristic of blood-derived immune cells. The transition occurs gradually between roughly age 50 and 75.
Gene expression alone would not have revealed this. The two data layers answer different questions: expression measurements show what a cell is doing in the moment, while epigenetic marks carry a longer memory of the cell’s origin and history, and the methylation record is what exposes the lineage switch. The brain’s immune environment is evidently not as static as once believed.
The microglial change accompanies a broader reorganization. Astrocytes, the star-shaped cells that support synapses and maintain the blood-brain barrier, decline with age more than any other cell type in the dataset, a loss confirmed by immuno-staining of hippocampal tissue. The mechanism appears to involve NRF1, a transcription factor that regulates mitochondrial energy production: with age, NRF1 binding sites lose accessibility, mitochondrial genes are downregulated, and aging astrocytes show signs of autophagic cell death. Endothelial cells lining blood vessels also decline, which may contribute to the age-related breakdown of the blood-brain barrier.
The study also found a global erosion of 3D genome architecture. Chromatin, the packaged form of DNA, loses its organized compartmentalization with age across many cell types, and these structural disruptions track closely with changes in gene regulation and cell identity. The physical organization of the genome is part of the aging process, not merely a bystander.
The timing of the microglial switch is what makes the finding relevant to dementia research. Aging is the single largest risk factor for Alzheimer’s disease, and midlife is when pathological changes begin accumulating in the brain, years before symptoms appear. A proinflammatory immune environment in the hippocampus during that window is a plausible contributor to later vulnerability, though the study cannot prove causation.
The study has clear limitations. The data come from postmortem tissue, a snapshot at the end of life rather than a longitudinal view of the same brains, and the donors were neurologically healthy, so the link to Alzheimer’s is inference from risk-factor timing, not direct observation. Forty brains is a large sample for this kind of deep multi-omic work but small for epidemiology. Whether the blood-derived cells actively cause damage or are a consequence of other aging processes remains unknown.
The finding also complicates the concept of the brain’s immune privilege. The brain has long been described as immunologically special, shielded from the peripheral immune system by the blood-brain barrier and populated by its own resident cells. If blood-derived immune cells enter the aging hippocampus and take up residence, that picture changes: the barrier and the boundary between brain and body immunity appear more dynamic in midlife and old age than the textbook version suggests. Whether those cells cross from the blood or arise from a different developmental route is among the questions the authors flag for future work.
The study offers a new description of brain aging: a coordinated reprogramming in which the hippocampal immune cell population changes identity, supportive cells are lost, and the genome’s architecture frays, all beginning in midlife. The work is part of the NIH Common Fund’s 4D Nucleome program, which studies how the genome’s three-dimensional organization shapes development, aging, and disease. If the microglial transition is confirmed as a driver rather than a marker of age-related vulnerability, it would give researchers a concrete cellular target for interventions aimed at preserving brain function in later life.
Sources: Zemke, N.R., Lee, S., Mamde, S. et al. Epigenetic and 3D genome reprogramming during the aging of the human hippocampus. Science 393, eadt8307 (2026). DOI: 10.1126/science.adt8307. Preprint: bioRxiv 10.1101/2024.10.14.618338.