Between the ages of fifty and seventy-five, the human brain's original immune cells are overthrown. Not by disease. By the body's own blood.
A study published in Science by researchers from UC San Diego, the New York Genome Center, and UC Irvine analyzed postmortem hippocampal tissue from 40 neurologically healthy adults aged 20 to 95. The team combined gene expression analysis with epigenetic profiling and three-dimensional genome mapping to build one of the most detailed pictures of brain aging ever assembled. The finding that matters most: a regime change.
The Sentinels
Microglia are the brain's resident immune cells. They emerge during embryonic development and were long assumed to renew throughout the human lifespan, maintaining the cellular environment by clearing debris, pruning synapses, and guarding against infection. The study found that between approximately ages 50 and 75, these original microglia progressively decline. In their place appear cells whose molecular signatures resemble immune cells from the blood, carrying elevated inflammatory characteristics.
"Gene expression tells us what a cell is doing today, but epigenetic signatures preserve information about where a cell came from," said Nathan Zemke, the study's first author and Director of Single-cell Genomics at UC San Diego's Center for Epigenomics. "We uncovered a major shift in the identity and lineage of immune cells in the aging human brain that gene expression data alone would not have revealed."
The blood-brain barrier, the protective layer that normally shields the brain from blood-borne substances, also showed substantial deterioration. And across many brain cell types, the genome's three-dimensional architecture eroded, disrupting the regulatory structures that control which genes are switched on or off.
The Wrong Target
The pharmaceutical industry's response to Alzheimer's disease has focused on two downstream targets: amyloid plaques and neuroinflammation. Both approaches assume the brain's existing cellular machinery is still in place and needs help.
Eisai and Biogen's Leqembi, the leading anti-amyloid antibody, generated $168 million in the first quarter of 2026, up 74% year over year. At $26,500 per patient annually, it slows cognitive decline by 27% in clinical trials. Eli Lilly's Kisunla showed a 35% slowing in patients with low to intermediate tau levels. Real effects, but modest. Neither drug reverses the disease.
The neuroinflammation approach has fared worse. Alector and AbbVie's AL002, a TREM2 agonist antibody designed to activate microglial function, failed its Phase 2 INVOKE-2 trial in 381 patients with early Alzheimer's disease. No treatment benefit on the primary endpoint. No reduction in brain amyloid. Alector laid off 41 workers after the results.
But AL002 was designed to activate existing microglia. If those microglia are being replaced wholesale by blood-derived cells between ages 50 and 75, a drug that targets resident microglial biology may be targeting cells that are no longer there.
The Clock
"Aging is the single largest risk factor for dementia, but our understanding of how it drives disease is still incomplete," said Richard Hodes, Director of the National Institute on Aging. "This previously hidden microglial shift, now uncovered by innovations in technology and thinking, may be an important clue to help us complete the puzzle."
The study is one of six papers published in Science through the NIH's 4D Nucleome program, a decade-long effort that mapped how the genome is organized in three-dimensional space and how that organization changes over time. Together, the papers represent the most comprehensive genome architecture resource ever assembled.
About 7.4 million Americans aged 65 and older live with Alzheimer's disease. Every drug on the market targets what happens after the brain's immune system has already been overrun. This study suggests the overrun itself, the progressive replacement of embryonic microglia by inflammatory blood-derived cells through a weakening barrier, may be the event that matters.
The brain doesn't simply age. It gets infiltrated.