A scientist with Huntington's is racing to outrun the disease he carries

When Jeff Carroll's genetic test came back positive in 2003, the same year his mother died of Huntington's, the rational response, he decided, was to become one of the scientists trying to beat it. Twenty-three years later, from a new lab at the Allen Institute in Seattle, he is leading a strand of research that aims to slow the disease before the first symptom appears, by stopping a runaway DNA change inside the very neurons Huntington's kills.

Huntington's is a fatal inherited neurodegenerative disease. Each child of an affected parent faces a 50% chance of inheriting it; about 1 in 20,000 people worldwide carry the mutation. The disease typically unfolds over a decade or two, from mood and cognitive changes to the loss of movement, speech, and swallowing.

For most of that history, therapies have tried to lower the mutant huntingtin protein that accumulates in the brain. The leading example is uniQure's gene therapy, which in 2024 became the first treatment to slow Huntington's progression in a clinical trial. In June 2026, the U.S. Food and Drug Administration agreed that the trial data could support a new-drug application, despite the limits of a small, uncontrolled study that requires hours of brain surgery to deliver.

The new generation of drugs targets something more upstream: the DNA change itself. The Huntington's mutation sits in a stretch of three letters, C, A, G, repeated dozens of times inside the HTT gene. In vulnerable neurons, that repeat keeps getting longer over a patient's lifetime, as the cell's own DNA-repair machinery turns against itself. Once a neuron's repeat count crosses roughly 80, growth accelerates. Past about 150, the gene's normal activity collapses and the cell dies within months.

Steven McCarroll's group at the Broad Institute, Harvard Medical School, and McLean Hospital mapped that picture in a single-cell study in Cell in 202501379-5). The team measured CAG-repeat lengths in more than 500,000 individual cells taken from 53 post-mortem Huntington's brains and 50 controls. The runaway expansion was concentrated in striatal projection neurons, which are exactly the population that dies first in the disease.

A companion genome-wide association study in Nature Genetics in 2025 pointed at the same lever. Variation in a DNA-repair gene called MSH3 turned out to be the strongest genetic modifier of when symptoms begin, a finding reinforced in a 2025 review of genetic modifiers of the disease. When Anastasia Khvorova's lab at UMass Chan partially suppressed MSH3 in mice, runaway expansion slowed or nearly stopped.

Together, those papers describe Huntington's as a process that can be slowed, not a fixed sentence handed down at birth. A two-stage model, slow expansion followed by a distinct toxic phase, was described in a 2025 review by Heintz and colleagues at Rockefeller University and now anchors the new drug strategy.

The first somatic-expansion-targeting drug candidates are slated to enter clinical testing later in 2026. Triplet Therapeutics, a Cambridge, Massachusetts, antisense-therapy company founded in 2018, had built a candidate aimed at MSH3 and described itself as ready to take it into the clinic, but the program did not advance to a human trial. Other programs are now in motion. Carroll's Huntington's strand of a new Allen neurodegenerative-disease initiative, launched in June 2026, is meant to keep the basic biology moving in step with the pipeline.

Carroll, who learned at 28 that he carried the same mutation that killed his mother, is one of a small group of carrier-scientists now driving the new approach. The next wave of drugs is aimed at the mutation's source rather than its consequences, and the people running the trials are increasingly reading their own genotypes alongside their protocols. Whether the expansion can be slowed enough to delay symptoms by years remains the question the 2026 trials are designed to answer.