An anonymous senior FDA official told CNBC in March 2026 that uniQure is "the latest company to make a failed therapy for Huntington's patients." That is a brutal sentence. It is also, methodologically, the correct one to say out loud.
AMT-130 works by using an adeno-associated virus to deliver a microRNA that silences the gene producing mutant huntingtin protein, the malformed molecule that spends decades dismantling neurons from the inside out. The biological premise is genuinely exciting. A one-time delivery that instructs the brain to stop making the protein that kills it: that is the kind of idea that makes you want to declare victory before the data arrives. But the data has arrived, and it is not what the excitement promised.
The Phase 1/2 results showed cerebrospinal fluid levels of mutant huntingtin and neurofilament light chain, two key biomarkers of brain target engagement, varying over time without a consistent signal. Think of it like tracking a fever with a thermometer that gives different readings every hour. You cannot conclude the fever broke. You conclude the thermometer needs replacing, or the study design does.
The Methodology Is the Crisis
Here is what makes this genuinely hard to watch: the sample size in Phase 1/2 trials is small by design, the follow-up is short, and there is no sham control group, meaning no one received a fake procedure to isolate the placebo effect. Without that control, you cannot disentangle real biological change from the improvement patients feel simply because something was done to them. The FDA is asking for a randomized, sham-controlled trial. That request is not cruelty. It is the minimum standard for knowing whether a treatment works.
I want to grant uniQure's defenders a fair point: for a disease with no cure, an average diagnosis delay of seven years, and a 100 percent fatality rate, demanding perfect trial design before any patient accesses a therapy carries its own moral weight. That tension is real. But the answer to insufficient evidence is not lowered standards; it is faster, better-designed trials. Approving AMT-130 on inconsistent biomarkers would not accelerate the path to a real treatment. It would muddy it, consuming regulatory bandwidth, patient trust, and research resources on a therapy we cannot yet confirm works.
The broader HD pipeline makes this clearer. Branaplam lowered mutant huntingtin levels in the VIBRANT-HD trial and was still halted because of serious safety issues. A molecule can do exactly what you measured it doing and still harm the person you measured it in. That is why the sham control exists. That is why replication exists. The history of Huntington's drug development is a history of promising biomarker results that did not translate into patient benefit, and AMT-130's Phase 1/2 data has not yet proven it is different.
What Should Actually Happen Now
Skyhawk Therapeutics is already hiring for commercial launch of SKY-0515. Roche's GENERATION-HD2 trial is expected to complete this year. CRISPR-based approaches corrected mutant alleles in mouse models in 2023 and are moving toward human trials. The field is not standing still. What it needs is not lower evidence thresholds; it needs methodologically rigorous trials that run faster, recruit broader populations, and produce data that actually answers the question.
The FDA should hold the line on sham-controlled Phase 3 data for AMT-130. uniQure should design that trial with the urgency the disease demands. And the rest of us should stop confusing the desire for a breakthrough with the evidence of one. Those are very different things, and Huntington's patients deserve to know which one they are getting.
