One-off Huntington's gene therapy still slows decline at 4 years

People with early Huntington's disease who received a single high dose of an experimental gene therapy were still declining more slowly than a comparison group four years later, according to new data from uniQure. The company released the figures on 29 September 2026 from its phase 1/2 studies of AMT-130, now given the international name ifezuntirgene inilparvovec.

The picture is encouraging but mixed. At three years, the high dose was linked with an 80% slowing of decline on the main combined rating scale. At four years, the slowing on that same scale was 44% and was not statistically significant. A second measure of everyday function held up better, with a 61% slowing at four years. All of these comparisons are against an external group of untreated people drawn from a large observational study, not against a placebo group within the trial, and that shapes how far the numbers can be trusted.

The data matter because AMT-130 is already under review in the United States. uniQure submitted its licence application before these four-year results were in, after the US Food and Drug Administration (FDA) agreed that three-year data from 12 high-dose patients could form the basis of a request for accelerated approval.

What Huntington's disease is

Huntington's disease is an inherited brain disorder caused by an expanded stretch of repeated DNA letters (CAG) in the huntingtin gene, known as HTT. The faulty gene produces an abnormal huntingtin protein that gradually damages nerve cells, particularly in the striatum, a deep part of the brain made up of the caudate and putamen. The hero image above, an MRI scan, shows the shrunken caudate typical of the condition.

Symptoms usually start in adult life, often between the ages of 30 and 50. They include involuntary movements (chorea), problems with balance and coordination, difficulties with planning, memory and concentration, and changes in mood and behaviour such as depression, irritability and apathy. The disease progresses over 15 to 20 years or so, and people eventually need full-time care.

Each child of a parent with Huntington's has a one in two chance of inheriting the gene. uniQure estimates that about 75,000 people live with the disease across the US, the European Union and the UK, with many more at risk. There's no treatment that slows or stops it. Current medicines, such as tetrabenazine for chorea, ease some symptoms only.

How AMT-130 works

AMT-130 is a gene therapy that uses a harmless adeno-associated virus (AAV5) to carry a piece of genetic code into brain cells. That code makes a short strand of RNA, called a microRNA, which latches on to the messenger RNA produced by the huntingtin gene and leads to its breakdown. The aim is to lower the amount of huntingtin protein, including the harmful version, inside cells in the striatum.

It's given once, by neurosurgery. Surgeons use MRI guidance to inject the therapy directly into the caudate and putamen on both sides of the brain. The idea is that the cells keep producing the silencing microRNA for years, so no repeat dosing is needed. That one-off approach is attractive, but it also means the treatment can't be switched off if problems emerge, and the operation itself carries risk.

Earlier attempts to lower huntingtin with drugs given into the spinal fluid had setbacks. Roche's tominersen, an antisense drug, was halted in phase 3 in 2021 after the higher dosing schedule did worse than placebo, although later analyses suggested younger people with less advanced disease might benefit. AMT-130 takes a different route, delivering its payload straight to the most affected brain region.

The studies

The data come from two open-label phase 1/2 studies. The US study (NCT04120493) began as a randomised trial against an imitation (sham) surgery, then moved to open-label dosing. The European study (NCT05243017) was open-label from the start.

Across the programme, 51 people with early Huntington's disease have been dosed. In the US study, 26 people took part; 10 had sham surgery, and 4 of those later crossed over to receive the therapy. The European study treated 13 people. A third cohort of 12 people received the therapy alongside a course of immunosuppressive drugs, and a fourth cohort includes 6 people.

The latest analysis used a data cutoff of 30 June 2026 and focused on 29 people treated in the first two cohorts: 17 at the high dose and 12 at the low dose.

How the comparison was made

Because the early studies didn't keep a placebo group long term, uniQure compared treated patients with people from ENROLL-HD, a large international observational study that follows people with Huntington's disease. For the new analysis, 1,337 people from ENROLL-HD were matched to the high-dose group using a statistical method called propensity scoring, which picks untreated people who looked similar at the start on factors such as age, CAG repeat length and baseline symptom scores.

This approach has well-known weaknesses. The two groups weren't randomised, so unmeasured differences can remain. And by four years, data were missing for 53% of the external control group. uniQure says this missing data and a survivor effect, where people who decline fastest drop out of observational studies, likely made the untreated group look as though it declined more slowly than it really did. If so, the treatment effect would be underestimated. That's plausible, but it's an argument the company is making about its own data, and it can't be checked without the full analysis.

The results

The two main measures were the composite Unified Huntington's Disease Rating Scale (cUHDRS), which combines movement, thinking and functional scores, and Total Functional Capacity (TFC), a 13-point scale that rates how well a person manages work, money, household chores, daily activities and their level of care. On both, a fall means things are getting worse.

At three years, high dose

Among 15 high-dose patients with three-year data:

• cUHDRS fell by 0.28 points, against 1.39 points in the matched external group. That's an 80% slowing of decline (nominal p=0.005).
• TFC fell by 0.27 points, against 0.82 in the external group, a 67% slowing (nominal p=0.011).

These figures are close to the three-year results the company reported in 2025, which drew wide attention.

At four years, high dose

Among 12 high-dose patients with four-year data:

• cUHDRS, the main measure, fell by 0.90 points, against 1.61 in the external group. That's a 44% slowing, but the difference wasn't statistically significant (p=0.144).
• TFC fell by 0.37 points, against 0.94, a 61% slowing (nominal p=0.008).

In a post hoc analysis, meaning one decided after the data were seen, uniQure re-ran the four-year comparison using the external control group from its earlier analyses. On that basis the slowing was 53.5% on cUHDRS (nominal p=0.041) and 68.3% on TFC (nominal p=0.001). Post hoc tests carry less weight than planned ones, because picking a comparison after seeing the results raises the chance of a flattering answer.

High dose against low dose

The dose comparison offers a different kind of support, since it doesn't rely on the external group. At four years, people on the high dose declined by 0.91 points on cUHDRS, against 1.94 on the low dose. On TFC, the figures were 0.30 and 0.70. A dose response like this fits with the treatment having a real effect, although the groups are small and weren't randomised against each other.

A marker of nerve damage

Neurofilament light chain (NfL) is a protein released when nerve cells are damaged, and it tends to rise steadily in Huntington's disease. In the spinal fluid of high-dose patients, NfL was 6% below its starting level at three years (14 people) and 4% above it at four years (11 people). In untreated Huntington's disease, NfL would usually be expected to rise over that time. These figures suggest nerve damage was at least held back, although there's no matched comparison for this measure in the release.

What the numbers mean in practice

The cUHDRS and TFC changes are small in absolute terms. A fall of 0.9 points on cUHDRS over four years, against 1.6 in the comparison group, isn't something a family would notice from day to day. The point is the slope: if a treatment can halve the rate of decline over many years, people could keep their independence, their jobs and their ability to manage at home for considerably longer.

Experts have urged caution. The HDBuzz team, a science news service for the Huntington's community, described the new analyses as "nuanced", and BioPharm International's headline noted that the primary measure had missed significance at four years. With so few patients and an external comparator, small shifts in who is counted can swing the result.

Safety

The main safety concern is inflammation in the brain. Five high-dose patients (17%) had serious central nervous system inflammation, which resolved in all cases. This is one reason uniQure tested a cohort given immunosuppressive drugs around the time of surgery.

One person in the low-dose group died by suicide about five years after treatment. Investigators judged this unrelated to AMT-130. Depression and suicidal thoughts are sadly common in Huntington's disease itself, and suicide rates are higher than in the general population.

The neurosurgery also carries the usual risks of brain surgery, including bleeding and infection. Anyone considering future treatment would need to weigh these against the potential benefits.

Where regulators stand

The FDA has granted AMT-130 breakthrough therapy designation and agreed that three-year data from 12 high-dose patients, compared against an external control, could support an application for accelerated approval. Under that pathway, a drug can be approved on the basis of a likely benefit, but the company must confirm the benefit later.

uniQure submitted its Biologics Licence Application before the four-year data were available. The FDA will now weigh the new results alongside the original submission. Whether the weaker four-year cUHDRS figure affects the review isn't yet known. In Europe and the UK, no application has been announced.

Limitations

• Small numbers. Four-year findings for the high dose rest on 12 people.
• No placebo group. Comparisons are with an external observational group, which can't account for every difference between patients.
• Missing data. More than half the external group lacked four-year data, which could bias the comparison in either direction.
• Post hoc analyses. The more favourable four-year figures come from a comparison chosen after the data were in.
• Open-label treatment. Patients and assessors knew who had been treated, which can affect ratings on scales that involve judgement.
• Not yet peer reviewed. The results come from a company announcement and conference call, not a journal paper.

What it means for patients and families

For now, AMT-130 is available only in clinical studies. If the FDA grants accelerated approval, it would be the first treatment shown to slow Huntington's disease, but access would initially be limited to the US and to specialist neurosurgical centres. It's likely to suit people at an early stage of the disease, since the aim is to protect brain cells that haven't yet been lost.

People with Huntington's disease, or at risk of it, can ask their neurologist or Huntington's disease clinic about current trials. Patient organisations such as the Huntington's Disease Association in England and Wales, Scottish Huntington's Association and the European Huntington's Disease Network publish updates on research and how to take part.

These four-year results don't settle the question of how well AMT-130 works. They show that the high dose continues to be associated with slower decline on everyday function, while the more sensitive combined scale is less clear than it was a year ago. A larger, controlled study will be needed to give a firm answer.