AAV2-GAD sham-controlled Phase 2

The research record

This is the project's file for this landmark, as written and checked. Sources are listed at the end; anything not yet confirmed is marked "Not yet verified".

Who

Peter LeWitt, Andrew Feigin, Michael Kaplitt, Matthew During and colleagues at seven US centres; sponsor Neurologix. Lancet Neurology 10(4):309–319, 2011. Later: MeiraGTx, which now runs the programme (how it came to hold it is not verified).

What happened

GAD makes GABA, a calming signal. In Parkinson's the subthalamic nucleus (a movement-control hub) becomes overactive. The idea was to use gene therapy to make the nucleus produce GABA and quiet itself.

• 2008–10. Randomised, double-blind, sham-controlled Phase 2 (NCT00643890). Of 66 assessed, 23 were assigned to sham surgery and 22 to AAV2-GAD (one infusion into each nucleus). After exclusions (failed infusion or catheter off target) 21 and 16 were analysed. At 6 months, the off-medication UPDRS motor score fell 8.1 points (23.1%) on AAV2-GAD and 4.7 points (12.7%) on sham. The difference over time was p = 0.04. One serious adverse event (bowel obstruction) was judged unrelated.
• Stopped. The registry says the trial ended for financial reasons.
• 2022–24. MeiraGTx's NCT05603312: 14 people, randomised 5 low dose, 5 high dose, 4 sham, 26 weeks. The company reported that the high dose improved MDS-UPDRS part 3 by an average of 18 points (p = 0.03), with no significant change for low dose or sham. The registry reports no treatment-related serious adverse events.

Why it moved us toward a cure

This treats symptoms and does not slow the disease. It matters because it showed a gene therapy could be delivered to deep brain targets and judged against sham surgery, the design needed for later disease-modifying gene therapy such as landmark 160. The sham arm also improved by 4.7 points, so any open-label gene therapy result needs this check. The earlier AADC gene therapy (landmark 112) was open label, and the neurturin sham trials (landmark 147) did not beat sham.

Outcome

Advanced. The registry record (completed; primary completion 6 Sep 2024; results posted 5 Sep 2025) lists safety as the only primary measure, with one treatment-related adverse event (mild confusion and headache, high dose) and no treatment-related serious events; the 18-point gain is the company's analysis. The FDA gave the programme RMAT (regenerative medicine advanced therapy) designation, announced 9 May 2025 (NeurologyLive), and MeiraGTx signed a Hologen AI deal in March 2025. The company says it is discussing a Phase 3 with regulators, and none is confirmed. The 2011 result was a modest gap with many post-randomisation exclusions, and the MeiraGTx numbers come from company statements.

Sources

• LeWitt et al., Lancet Neurol 2011, DOI 10.1016/S1474-4422(11)70039-4, PMID 21419704
• Niethammer et al., JCI Insight 2017, DOI 10.1172/jci.insight.90133 (12-month follow-up)
• ClinicalTrials.gov NCT00643890, NCT05603312
• CGTLive on MeiraGTx results, 21 Oct 2024
• Not yet verified: the MeiraGTx acquisition of Vector Neurosciences and its link to Neurologix (press release blocked); MeiraGTx's own releases on RMAT and Hologen (trade press only); any peer-reviewed 2022–24 results; the 2022–24 doses beyond the CGTLive figures (7.0 and 21 x 10^10 vector genomes).