Why Our Memory Slips After Age 50 - by Eric Topol

Ground Truths

SubscribeSign in

News and Analyses<br>Why Our Memory Slips After Age 50<br>Landmark independent studies challenge longstanding dogma for brain aging<br>Eric Topol<br>Aug 12, 2026

258

36

Share

There has been a big shake-up in the understanding for how our brain ages. Until now, the prevailing belief was that our brain tissue shrinks with age, exemplified by the hippocampus, the region tied to memory and learning, as seen via MRI and other imaging modalities. That the microglia immune cells we were born with (embryonic) were the only types to be found in our brain. And the brain aging process was accepted to be a linear, steady deterioration. It turns out that all of that was wrong!<br>In this edition of Ground Truths I will go through the completely revamped understanding for what happens to our brain as we age from recent breakthrough studies.

Subscribe

A Major Inflection around Age 50

From donor postmortem brains studies in people ranging from age 20 to 100 years, Zemke and colleagues in Science demonstrated the midlife inflection in the hippocampus tissue via multiple omics: gene expression, methylation, chromatin access, and 3D chromosomal architecture. You can easily appreciate the sudden shift in the Figure below (this and the other Figures made with the help of Gemini Notebook) characterized by cytokine production (inflammation) and an energy crisis for astrocyte mitochondria, starving for fuel because of impaired ATP synthesis. This was not what was expected: that with aging astrocyte cells would display senescent markers (“zombie cells”). They were nowhere to be found.

The Invasion, A Hostile Takeover

As astrocytes, the cells responsible for housekeeping and brain security functions, undergo attrition, the blood brain barrier loses its integrity and enables an invasion of monocytes from the blood. The original brain microglia (Micro1) are replaced by transformed monocytes (Micro 2) that are highly inflammatory. As soon as they take residence in the brain, they change their identity to microglial.

Share Ground Truths

The recent paper by Zemke and colleagues was fully replicated and extended by Belk et al proving the source of the Micro 2 cells—from the blood monocytes, derived from the bone marrow. And this replacement began in midlife. The two groups proved that via different tracking techniques. Zemke used the 4-way multiomic approach (summarized above), particularly methylation (single-nucleus methyl-3C sequencing, snm3C-sequencing), to define the origin of the replaced microglia, whereas Belk’s group used somatic mutations of the cells to identify their lineage (what they called Passenger Assisted Clone Tracking or PACT). This is completely opposed to what had been believed, with the blood brain barrier expected to prevent a massive infiltration of peripheral immune blood cells. Previously studies using just gene expression could not differentiate the 2 origins of the microglia, which can be considered the “transcriptomic illusion".m RNA-seq just scratches the surface. It took single-cell and multiomics to get this straight. Three other points were notable: (1) it’s a unique feature in humans, not seen even in non-human primates, (2) the process of replacement is more aggressive in men compared with women, and (3) the Belk study demonstrated the replacement in other regions beyond the hippocampus, such as the cerebellum and pre-frontal cortex.

Another new preprint study of the microglia derived from peripheral monocytes (produced in the bone marrow) identified the brain-meningeal reservoir for the incoming migration of cells.<br>Back to the hippocampal astrocytes, in contrast to the speed of monocyte invasion, they are dying off at a slow rate via lysosomal autophagy (waste clearance), about 0.2% per year in a steady pattern from age 20, as tracked by methylation.

Share

Then What Happens

The epigenetic approach (Zemke et al) gets us to 3D architecture thinking to explain the pro-inflammatory process unleashed properties of the Micro 2 cells. With the change in 3D folding, a weakened genome organization, there is new contact between gene components that were previously operated by tens of thousands of base pairs. As shown below for Micro2, the looped DNA sets up interleukin-15’s promoter and enhancer elements to be in contact, and the result of massive release of cytokines

On the other hand, the genetic approach (Belk et al) unraveled how these peripheral immune blood cells could be protective of Alzheimer’s. I’ve written a lot in Ground Truths about clonal hematopoiesis of indeterminate potential (CHIP): here and here . These mutations of blood stem cells are generally thought to be harmful, associated with an increased risk of cardiovascular diseases, clotting, and some cancers. But when CHIP clones were in the brain microglia, there was the paradoxical salutary impact of a marked increase in clearance of amyloid plaques and ~50%...