Huntington's is a fatal inherited brain disease with no approved cure. Roche's drugs lowered the toxic huntingtin protein but did not slow decline, a gap reshaping brain targeted gene silencing.
Roche has ended two Huntington's disease drug programs built around gene-silencing, an approach whose genetic logic has never been in doubt but whose clinical translation has now failed twice inside the same company. The terminations were announced in a community letter from Roche to patients on 2026-07-10 (HDBuzz, drawing on Roche's letter), taking out a Phase 2 antisense oligonucleotide that had reached late-stage testing and a first-in-human candidate that had not. Huntington's is a fatal inherited neurodegenerative disease with no approved therapy that modifies its course, and the patient community that has tracked these programs now watches its treatment pipeline thin out further.
The two programs died for different reasons. The first, GENERATION HD2, tested tominersen, a total huntingtin-lowering antisense oligonucleotide (a synthetic strand of DNA-like chemistry that binds the huntingtin protein's messenger RNA and marks it for destruction), developed with Ionis. It failed to meet a primary endpoint on clinical progression, the standard scale used to capture walking, cognition, and independence (Endpoints News, BioPharma Dive). The second, POINT-HD, tested RG6496, a newer ASO that selectively lowers only the expanded form of huntingtin produced by the mutant allele. Roche halted it because animal studies suggested the molecule would not tolerate the chronic dosing a Huntington's patient would need over a lifetime, raising doubts about whether intrathecal injections could be sustained across decades of treatment. Roche has stressed the two decisions were independent, drawn from different datasets, and unrelated to patient safety (HDBuzz).
Tominersen worked at the molecular level. In earlier studies it became the first drug ever shown to lower huntingtin in people. In GENERATION HD2, cerebrospinal fluid measurements of mutant huntingtin fell, and a neurodegeneration marker called neurofilament light chain dropped in both spinal fluid and blood. The clinical scores, the ones that capture walking, cognition, and independence, did not move enough to declare benefit.
This is the second time tominersen has failed at this translation step. Roche halted the Phase 3 GENERATION HD1 in 2021 after the trial's data and safety monitoring board judged the benefit-risk balance unfavorable; a post-hoc analysis later suggested younger patients with lower CAG repeat counts, the genetic expansion that drives the disease, might still benefit. GENERATION HD2 was built to test that narrower hypothesis. It has now returned a negative result on the primary endpoint, weakening the case that patient selection alone could rescue what the broader population could not (Endpoints News, STAT+).
The wider pattern is recalibration, not refutation. Huntington's is caused by a single dominant toxic gene, which is why silencing it has anchored the field's drug-development work since the disease-causing mutation was identified. That genetic logic still anchors the work. What is now in play is the harder clinical question: how to get a silencing drug into enough neurons, at the right dose, for long enough, and measure benefit on timelines that match a disease unfolding over decades. Other sponsors are still pushing. uniQure and Wave Life Sciences remain the field's most watched alternatives still standing: uniQure is testing a one-time gene therapy, Wave a next-generation allele-selective ASO. Whether either can clear the same clinical translation bar that tominersen has now failed to clear twice is the question now hanging over the field.
For the patient community, the wait continues. The next data from uniQure and Wave will be the live tests of whether lowering huntingtin can move the clinical needle, not just the biomarker one. Until then, huntingtin-lowering will remain the field's most biologically certain bet and its least clinically proven one.