Brain's Immune Cells Undergo Wholesale Turnover Starting in Midlife, Study Finds

Researchers who spent a decade mapping how DNA folds inside the cell nucleus have found that the human brain undergoes a sweeping reorganization of its genome beginning in midlife, with the sharpest changes occurring in the hippocampus, the region that governs learning and memory. The findings, drawn from postmortem brain tissue spanning seven decades of adult life, challenge a long-standing assumption that the brain's resident immune cells persist unchanged from before birth into old age.

The study, led by scientists at the New York Genome Center together with the University of California, San Diego, and the UC Irvine Center for Neural Circuit Mapping, appears in the journal Science under the title "Epigenetic and 3D genome reprogramming during the aging of human hippocampus." It is one of the last major outputs of the National Institutes of Health's 4D Nucleome program, a ten-year Common Fund initiative that closed out in 2025 after building tools to chart how chromosomes are arranged inside living cells.

What the tissue showed

The team examined hippocampal samples from 40 neurologically healthy donors ranging in age from 20 to 95, using single-cell techniques that simultaneously capture which genes a cell is switching on, which stretches of DNA are chemically marked for silencing, and how the genome is folded into loops and neighborhoods inside the nucleus. That combination let the researchers watch cellular identity change from the inside out, rather than relying on gene activity alone.

The most pronounced shift emerged in microglia, the brain's resident immune cells, somewhere between roughly age 50 and age 75. Microglia that trace their origin to embryonic development declined substantially over that window and were replaced by cells carrying molecular signatures closer to immune cells found in blood. Alongside that, the researchers reported a weakening in cells that help maintain the blood-brain barrier, losses among astrocytes and oligodendrocyte progenitor cells, and a broad erosion of three-dimensional genome architecture that cut across multiple cell types rather than being confined to one.

Why microglia were thought to be different

Microglia colonize the brain before birth and have generally been understood to renew themselves locally for the rest of a person's life, sealed off from the blood-derived immune cells that patrol the rest of the body. That assumption underpins a lot of thinking about how neuroinflammation builds with age. The new results complicate it: if a large fraction of midlife and older brains are running on immune cells of a different lineage and disposition, models of brain aging built on the idea of a stable, self-renewing microglial population will need revisiting, and so will efforts to target microglia therapeutically.

The hippocampus was a deliberate choice of tissue. It is unusually vulnerable to age-related decline and is among the earliest regions to show pathology in Alzheimer's disease, which made it a natural place to look for cellular transitions that might explain why aging remains the single strongest risk factor for the condition, ahead of any known gene variant.

Who the findings speak to

The direct beneficiaries of the work are researchers studying dementia and neuroinflammation, who now have a single-cell reference map of what a hippocampus in its fifties, sixties and seventies actually looks like at the molecular level, compared with one in its twenties or nineties. But the implications reach further, to anyone in or approaching the age window the study flags. The researchers frame midlife not as a quiet interval before old age but as the period when several brain systems, immune, vascular and neuronal, begin shifting in a coordinated way. That reframing matters for how clinicians and public health officials think about when to intervene against cognitive decline, since it points toward the fifties and sixties, rather than old age itself, as a window worth watching.

"Microglia are critical for maintaining brain homeostasis. When these cells fail to perform their housekeeping duties, toxic materials accumulate that can trigger inflammatory processes that may contribute to neurodegenerative diseases," said Bing Ren, scientific director and chief executive of the New York Genome Center and a senior author of the study.

Reaction and the road ahead

Nathan Zemke, director of single-cell genomics at UC San Diego's Center for Epigenomics and the study's first author, said the layered approach was central to the discovery. "Gene expression tells us what a cell is doing today, but epigenetic signatures preserve information about where a cell came from," he said, according to the New York Genome Center's account of the findings, explaining why the team paired standard measures of gene activity with newer methods for reading the genome's physical structure and its chemical modifications.

Xiangmin Xu of UC Irvine's Center for Neural Circuit Mapping, another author on the paper, said the broader lesson was about the shape of aging itself rather than any single cell type. "Aging is not simply a gradual decline, but involves coordinated and dynamic remodeling of immune, vascular, and neuronal systems," he said, a framing that has been picked up in subsequent coverage from ScienceDaily's writeup of the paper and by a separate summary published by Neuroscience News.

The paper is one of six studies emerging from the 4D Nucleome program's final phase, an effort the NIH Common Fund describes as a decade-long push to map how the genome's spatial organization shapes health and disease across tissue types. With that program now wound down, the immediate next step for the hippocampus team is establishing whether the immune and genome changes they documented are a cause of cognitive decline, a downstream consequence of some other aging process, or both feeding into each other. The authors have not yet published a follow-up establishing causality, and outside researchers, including those quoted in a separate rundown of the findings at SciTechDaily, have noted that the current dataset is observational, built from tissue donated after death rather than tracked in living patients over time.

That leaves open questions that will shape where the research goes from here: whether the shift from embryonic to blood-derived microglia can be detected in living patients through blood or imaging markers, whether slowing the transition would blunt inflammation-linked decline, and whether the same genome erosion recorded in the hippocampus turns up in other brain regions the 4D Nucleome program's remaining papers have yet to fully detail. For now, the dataset itself, drawn from real human tissue across the full adult lifespan, stands as the most detailed single-cell map yet of what changes, and when, inside an aging hippocampus.