State of HD research: an honest outlook

What's genuinely promising, what's overhyped, and where this is heading. Grounded in the live pipeline on this site.

This page is an opinionated outlook, not established fact. It reflects a data scientist's read of the field, with confidence noted. Predictions are guesses and timelines slip. Nothing here is a prognosis for any individual, that is a conversation for an HD neurologist.

What's actually good (ranked, with why)

1. Somatic-instability targeting (MSH3, FAN1, PMS1)

best science

Rated highest on evidence quality. Human genetics (GWAS) shows natural variation in these DNA-repair genes shifts HD onset by years, the strongest kind of evidence, because nature already ran the experiment. The idea: the CAG repeat keeps expanding in neurons over life; stop that and you slow the disease at its root. Confidence the mechanism is real and druggable: high. That a drug works soon: medium (early stage).

2. Huntingtin-lowering with real delivery — AMT-130

furthest along

A one-time gene therapy delivered to the brain to lower the toxic protein, with some encouraging early Phase 1/2 signals. If it holds up, it's the first proof a single treatment can bend the curve. Confidence it lowers HTT: high. That it meaningfully helps patients: medium, which is exactly what the next readouts test.

3. Biomarkers — the quiet, highest-leverage bet

enabler

Neurofilament light (a blood marker of neuron damage) plus wearable and digital gait sensors (via Enroll-HD, ~25k participants). Unglamorous, but this is what makes trials smaller, faster, and possible in gene carriers before symptoms. It multiplies the value of everything above. Confidence it reshapes the field: high.

Where I'm more skeptical (the honest ceiling check)

Oral splice modifiers (PTC518/votoplam, SKY-0515): promising and convenient (a pill), but so far they show they lower HTT in blood. Lower blood protein is an improvement metric, not proof of brain benefit. Watch the clinical readout, not the biomarker.

ASOs (tominersen): the high-dose trial harmed patients; the low-dose restart is "prove it," not "it's working."

In-vivo CRISPR and stem cells: exciting, but delivering to enough of the right neurons is unsolved, a decade-scale problem.

AI drug-repurposing hypotheses (including this hub's): triage aids, not evidence. Cross-experiment convergence is interesting, not validated. We rate these lowest on the evidence ladder, and want to be the first to say so.

The honest counterweight: Roche recently ended two HD programmes. Huntingtin-lowering is not a slam dunk, and more failures are coming.

Predictions (with confidence)

1–2 years: More mixed-to-negative readouts alongside pivotal AMT-130 data. NfL becomes a standard trial endpoint. At least one more program pauses. (High on the mix; low on which wins.)

3–5 years: First credible evidence a therapy slows HD, most likely from gene therapy or a somatic-instability drug. Trials shift toward pre-symptomatic gene carriers guided by biomarkers. (Medium.)

5–10 years: HD managed by a combination, not a magic bullet: lower the protein, stop the somatic expansion, support neurons. Plausibly a first approval for slowing, not curing. (Medium-low; timelines slip.)

The structural bet I'm most confident in: somatic instability becomes the dominant target class of the next decade. The genetics point there too strongly to ignore.

Wildcards: a cheap repurposed drug (an NAD+ booster or metformin-class) shows a modest, real signal; AI-designed ASOs/splice-modulators compress discovery timelines.

The one-line synthesis

HD has quietly become one of the best-understood, most "druggable" neurodegenerative diseases, with multiple independent shots in the air. For the first time, "slowing it" is a realistic few-year prospect, not a fantasy. The honest caveats: setbacks are part of this, the biomarker infrastructure matters more than any single drug, and none of this is a prognosis for a specific person.

See recruiting trials Ask about any of this

An opinionated outlook for education, grounded in the live pipeline on this site (papers, trials, news). Not established fact, not medical advice, and not a prognosis for any individual. We are data scientists, not doctors. For HD care, contact HDSA (hdsa.org) or your clinician.