How the human hippocampus ages: An immune shift after 50

How the human hippocampus ages cannot be understood from gene activity alone. In tissue from middle-aged and older adults, several layers of genome regulation changed together: chemical marks on DNA, access to regulatory regions, the three-dimensional folding of chromosomes, and the balance of cells that support memory and immune surveillance.

The hippocampus is involved in learning, memory formation, and spatial navigation. It is also particularly vulnerable to aging and neurodegenerative disease. Most previous studies, however, have focused mainly on gene expression. This reveals what a cell is doing at the time of measurement, but it may not disclose where the cell came from or which regulatory state it has retained over time.

Four layers of the same genome

The researchers examined postmortem hippocampal tissue from 40 neurologically healthy adults between 20 and 95 years of age. The samples covered four age groups and were balanced by sex. Gene expression and chromatin accessibility were profiled in 295,033 cell nuclei. In another 22,240 nuclei, the team simultaneously measured DNA methylation and three-dimensional contacts between genomic regions.

Each measurement captured a different aspect of cellular identity. Gene expression showed the cell’s current activity. Chromatin accessibility identified DNA regions available to regulatory proteins. DNA methylation provided a more stable epigenetic record that can retain information about cellular lineage. Three-dimensional contact maps revealed how the long DNA molecule was folded inside the nucleus and which distant regions were brought together to regulate genes.

The most pronounced nonlinear transition appeared in microglia, the immune cells that reside in the brain. Most microglia are thought to originate in the embryonic yolk sac, colonize the brain early in development, and maintain their population locally. Between roughly 50 and 75 years of age, however, the proportion of cells carrying epigenetic features of resident microglia declined, while a population with a profile resembling blood-derived monocytes became more prominent.

A transition hidden from RNA alone

The two microglial states were difficult to distinguish using gene expression alone. DNA methylation separated them much more clearly. In this context, methylation acted as a form of long-term cellular memory: two cells may temporarily activate similar inflammatory genes while retaining different origins and regulatory histories.

The monocyte-like population showed features of a state more strongly primed for inflammation. Its DNA methylation, regulatory accessibility, and three-dimensional genome contacts were altered near immune-response programs, including genes involved in interferon signaling. The authors propose that this epigenetic configuration may enable a faster and stronger inflammatory response. The study does not show, however, that these cells continuously release inflammatory molecules or directly damage neurons.

The proportion of astrocytes also declined with age. Astrocytes provide metabolic support to neurons, regulate signaling at synapses, and help maintain the blood–brain barrier. The decline was particularly apparent in a subgroup associated with tripartite synapses, structures in which an astrocyte actively interacts with the two neurons forming a synaptic connection. Remaining astrocytes showed reduced activity of mitochondrial energy-production programs and stronger cellular stress signatures. These findings point to possible metabolic dysfunction, but they do not establish why astrocytes become less abundant.

When DNA loses part of its internal order

Aging was also associated with broader changes in genome organization. In several cell types, the boundaries separating three-dimensional genomic domains became weaker, while contacts between different chromosomes became more frequent. The sequence of DNA letters remained unchanged, but some of the orderly packaging inside the nucleus appeared to deteriorate.

This packaging is not merely structural. It helps regulatory elements reach the appropriate genes while keeping unrelated gene programs apart. The weakening of three-dimensional architecture therefore coincided with changes in gene expression and cellular identity. That association does not yet show that altered DNA folding is the primary cause of the other age-related changes.

The study has important limitations. It compared postmortem samples from different people at different ages rather than following the same individuals over decades. The researchers did not directly trace cell lineages, so the monocyte-like epigenetic profile does not prove that blood cells entered the brain and replaced resident microglia. Cognitive performance during life was not measured, and the study did not analyze tissue from people with Alzheimer’s disease.

The results therefore do not explain the origin of dementia or identify a treatment. Instead, they suggest that hippocampal aging may involve more than a gradual accumulation of damage. It may also include coordinated changes in immune-cell identity, astrocyte support, and genome architecture, with a particularly prominent transition during midlife and later adulthood.