DNA repeat expansion drives Huntington’s development

Usually, most of us can zip a zipper without incident. But sometimes, an unwary dresser makes an alignment error that goes undetected until the ends of a garment don’t match up.

That’s not such a terrible problem. But when matching mistakes happen in repetitive DNA sequences, they can lead to expansion of the repeating region—and the long-term result can be dire. In a pair of studies of Huntington’s disease published this week, researchers report that this process, called somatic repeat expansion, is a key driver in the onset of disease. The studies cement a shift in how Huntington’s researchers think about disease development—and point to possible treatments.

Huntington’s is caused by variation in a single gene, huntingtin (Htt). The gene has a variably sized swath of one repeating codon, CAG, tucked into its sequence. People who inherit more than about 35 repeats are very likely to develop the disease in middle age, when cell death, concentrated in certain parts of the brain, disrupts movement and cognition.

The longer the inherited repeat, the earlier symptoms develop. For decades researchers thought this was because the Htt protein encoded by the long repeat clumps together and disrupts cellular function, just as proteins aggregate in other neurodegenerative diseases.

But genetics research in the past decade has linked the timing of symptom development to expansion of the repeat region and to the DNA repair enzymes that can cause a repeating region to grow from the length a person was born with to much greater lengths.

Repeat expansion within a cell happens because there are many possible ways to match two repetition-rich DNA strands, but only one is correct. Occasionally, after the two strands of DNA in the Htt gene separate to enable transcription, they match back up incorrectly—like when a zipper pull latches on at the wrong spot. DNA mismatch repair enzymes solve this problem by making new DNA to match the unpaired ends, adding extra repeats. Those errors can compound to make a longer and longer gene, says Steve McCarroll, a professor at Harvard Medical School and the Broad Institute. “The longer the repeat gets, the more frequently you’ll get these misadventures.”

In a study published Jan. 16, researchers in the laboratories of McCarroll and Harvard collaborator Sabina Berretta used single-cell RNA sequencing to study brain tissue donated postmortem by people with Huntington’s disease (Cell 2025, DOI: 10.1016/j.cell.2024.11.038). They compared the number of repeats in a cell’s Htt RNA to its overall gene expression. The technique let them match cell type with repeat number, and they found that a certain subset of neurons in the striatum showed repeats much longer than most other cells—as many as hundreds of CAG codons. To their surprise, these cells could rack up scores of repeats before showing any ill effects. But after a cell crossed a threshold of around 150 repeats, it began to lose expression of important functional genes and upregulate genes for programmed cell death.

“Technically, it’s terrific,” says Harvard Medical School geneticist Jim Gusella, who was not involved in the Cell study. “The combining of single-cell measurement of repeat length with expression is a major advance in the field.”

In the second study, published today in Nature Medicine, scientists led by Sarah Tabrizi, a physician-scientist at University College London, linked repeat expansion with very early signs of neurodegeneration in a cohort of young adults with a mutant Htt gene (2025, DOI: 10.1038/s41591-024-03424-6). The study began as a search for early signs of neurodegeneration in people long before they show symptoms. After the genetic association studies pointed to the importance of repeat expansion, Tabrizi says, they investigated the phenomenon when the participants returned for follow-up testing.

“Obviously, we can’t do brain biopsies of living persons carrying the gene,” says Tabrizi. Instead, her team looked at blood cells, which also undergo Htt repeat expansion. To her team’s surprise, they found “that the rate of somatic expansion in blood was absolutely predictive of the earliest striatal neurodegeneration.”

Even decades before symptoms are predicted to arise, MRI scans of people with the Htt risk gene show reductions in striatal volume over time compared to controls. The number of CAG repeats in blood cells correlates with the atrophy researchers can see by brain imaging.

They also identified some signs of inflammation in the cerebrospinal fluid, proteins that might be used as biomarkers for studies of future drug candidates.

The two studies dovetail so well that the scientists worked with the journals to coordinate their publication. “All the data from human genetics . . . really has been realized more in these studies,” says Tabrizi, adding that they show in postmortem brain and living people “that actually, yes, somatic expansion is appearing to drive the rate of the disease.” That’s not to say that the mutant protein is unimportant; researchers think the drastically expanded Htt protein, and perhaps its RNA, are the reason that expanded DNA repeats eventually kill cells. “There has to be a toxicity driver because the DNA itself is not toxic,” Tabrizi says.

The emerging model suggests that cells with the disease-related Htt gene start out without protein problems—only a ticking time bomb in their DNA. A fairly small number of extra inherited CAG codons make the repeat region unstable. At random, different cells accumulate enough additional DNA repeats in the Htt gene to make the protein toxic. That point hits at different times for different neurons. Early in life, the brain loses only a few neurons and can compensate, but eventually, enough neurons succumb that people start to lose function.

When Tabrizi presented her results to the study participants, she says, some of them wept, realizing that the new research suggests that blocking repeat expansion could delay the onset of Huntington’s disease.

Drug candidates that aim to block Htt expression, modulate RNA splicing, or edit the gene have been in development for some years. But McCarroll and Tabrizi say therapies that slow repeat expansion might have a better shot at blocking the dangerous protein from being produced in the first place. Triplet Therapeutics went broke trying to stop repeat expansion in 2022, but Gusella, who was a scientific cofounder, blames that failure on pandemic financing problems. He says that pharmaceutical companies and new startups such as LoQus23 Therapeutics have recently expressed renewed interest in targeting repeat expansion. Tabrizi expects that the first such drug candidates may reach early-stage clinical trials by 2027.