Podcast recap: Christopher Walsh on how somatic mosaicism rewrites the story of neurodegeneration

For decades, the dominant framework for understanding neurodegenerative disease has centered on inherited risk genes. A growing body of single-cell genomics research is challenging that assumption. The mutations that matter most may be the ones your neurons acquire after you are born, accumulating steadily across a lifetime and driving disease through mechanisms that look more like cancer biology than classical neurodegeneration.

In this episode of The Genetics Podcast, host Dr. Patrick Short speaks with Dr. Christopher Walsh, Bullard Professor of Pediatrics and Neurology at Harvard Medical School, Chief of the Division of Genetics and Genomics at Boston Children's Hospital, and Howard Hughes Medical Institute investigator. Walsh's lab pioneered the use of single-cell whole-genome sequencing to map the genetic mosaicism of the human brain, and his recent work draws a direct mechanistic line from somatic mutations to Alzheimer's disease, ALS, frontotemporal dementia (FTD), and chronic traumatic encephalopathy (CTE).

The disease is the somatic mosaicism

The conventional model treats inherited, germline variants as the primary cause of neurodegeneration, with environmental exposures as modifiers. Walsh's research inverts that framework entirely. Using single-cell whole-genome sequencing of hundreds of individual neurons, his team has shown that somatic mutations, acquired over a lifetime, accumulate in patterns that correlate directly with disease pathology.

As Walsh puts it: "The situation is completely inverted. The disease is the somatic mosaicism."

In this model, the germline genome still matters, but its role shifts. Rather than causing disease outright, inherited variants act as modifiers that shape susceptibility to the somatic damage that accumulates with age.

Microglia with cancer-driver mutations become neuron killers

One of the most striking findings from Walsh's lab involves the brain's resident immune cells: microglia. The team found that microglia in Alzheimer's brains carry significantly more somatic variants in well-known cancer-driver genes, including TET2, ASXL1, KMT2D, ATRX, and CBL.

These mutations do not produce tumors. Instead, they drive microglia into chronically inflamed, hyperactivated states that damage surrounding neurons. Walsh described the phenomenon with characteristic directness: "Instead of forming cancers, those cells just turn into angry murderers and start killing the neurons nearby."

This finding opens a potential therapeutic avenue. Drugs already developed to target these same genes in blood cancers could, in principle, be repurposed to suppress the expansion of pathogenic microglia and slow neurodegeneration.

A shared DNA damage signature across four neurodegenerative diseases

Walsh's team performed single-cell whole-genome sequencing on 469 neurons from affected and control brains and found two consistent patterns of DNA damage across Alzheimer's disease, ALS, FTD, and CTE.

The first is an oxidative damage signature, consistent with reactive oxygen species attacking neuronal DNA. The second is a pattern of small insertions and deletions linked to topoisomerase 1 (TOP1)-mediated mutagenesis.

The presence of a shared damage mechanism across conditions that have traditionally been studied in isolation suggests a common upstream vulnerability in neuronal DNA maintenance.

The brain's lifelong Darwinian competition

Walsh also described the elimination of genetically damaged neurons through a form of biological quality control. During early development, the brain produces a substantial excess of neurons. A significant fraction are eliminated perinatally through mechanisms that selectively remove cells with damaged genomes.

This process does not end at birth. As Walsh described it: "Every day our brain becomes a little more perfect, albeit at the cost of containing fewer neurons."

Neurons as biological clocks, and a path toward drug repurposing

Single-cell sequencing has revealed that neurons accumulate somatic single-nucleotide variants at a remarkably consistent rate of approximately 17 to 18 per year. This linear accumulation makes neurons one of the most reliable biological clocks in the human body.

Combined with the finding that cancer-driver mutations in microglia fuel neuroinflammation, this work raises a concrete translational possibility: repurposing existing cancer therapeutics to target the specific mutated genes driving pathogenic microglial expansion.

Listen to the full episode of The Genetics Podcast to hear the complete conversation: