Long-Term Study Confirms Efficacy of AAV9-Based Gene Therapy for Spinal Muscular Atrophy (SMA)
Results of a multi-year analysis in mouse models, published in Nature Communications, showed that optimized AAV9 gene therapy provides sustained expression of the SMN1 gene. This leads to long-term preservation of motor neurons and significant functional improvement, paving the way for safer treatment protocols for patients.
SMARD1 vs. SMA: Why Long-Term Mouse Data Changes the Game for Zolgensma
Industry Insider Analytical Review
May 25, 2026
\[The Gist]: What's Really Happening
Colleagues, pay attention. The publication in Journal of Biomedical Science on January 3, 2026, is not just "another mouse study." It is a quiet but fundamental paradigm shift in how we think about the long-term efficacy of AAV9 therapy.
The study, conducted by the Pagliari group from Centro Dino Ferrari (Milan) in collaboration with Nationwide Children's Hospital, compared two optimized versions of the AAV9 vector for treating SMARD1 — a rare but conceptually important disease. But the key insight that is being overlooked lies not in the result itself (both vectors worked), but in why one of them (P546) proved superior in the long term.
The news is tied to SMA? Yes, but indirectly. This study is not about Zolgensma, but about how to improve the platform. And the improvement targets the most vulnerable aspect of gene therapy for the nervous system: inefficiency and inflammation.
\[Timeline and Context]
Why 2026? Because we have accumulated enough data on the problems of the first wave of AAV therapies, such as Zolgensma (approved in 2019).
Here's the crux of the study. Scientists compared two promoters — Chicken β-Actin (CBA) and truncated MeCP2 (P546). CBA is a "powerful beast," providing high expression of the IGHMBP2 gene, but quickly and intensely. P546 is a "smart regulator," providing expression at a physiological level close to healthy norms.
The result? Both saved mice from death (median survival 162 days vs. 24 in controls). But:
- Inflammation. The study showed for the first time that SMARD1 has a pronounced neuroinflammatory component (astrogliosis, microgliosis). The CBA vector, with its overexpression, handled this worse in the long term. P546 completely normalized the inflammatory profile of the spinal cord.
- Safety. P546 had no toxicity peaks. From a clinical translation perspective, it is a "clean" candidate.
And most importantly: this study laid the groundwork for an already ongoing human clinical trial (NCT05152823) for SMARD1. That is, science is not catching up to the clinic; the clinic is catching up to science at full speed.
\[Who Wins and Who Loses]
Winners:
- Stefania Corti and Elisa Pagliari (Dino Ferrari Center). They have just created the "gold standard" for preclinical validation of AAV9 therapies. This is a strong argument for grants and patents.
- Nationwide Children's Hospital. They are leading this clinical protocol. Their reputation as a gene therapy translation center is strengthened.
- Companies developing "physiological" promoters. While everyone chases potency, P546 shows that slow and steady wins the race. This is a blow to the "higher dose, better result" logic.
Losers:
- Novartis (Zolgensma). Yes, Zolgensma is a blockbuster. But its Achilles' heel is hepatotoxicity and the need for long-term immunosuppression with corticosteroids. The SMARD1 study shows that using smarter promoters can avoid the severe side effects associated with overexpression. New players (GEMMABio with GB221) are already using intracisternal administration and other promoters to bypass Zolgensma's problems.
\[What the Media Isn't Saying]
Here's the main insight that won't make it into press releases.
Problem: Not all AAV9 are equal. And not all neurons are saved.
Zolgensma (intravenous) saves motor neurons, but recent data show that some patients, especially older ones, have ongoing neurodegeneration. The new ASO drug nusinersen (Biogen) shows a 75% reduction in neurofilaments (NfL) in patients after Zolgensma, indicating that some neurons were not transduced.
The SMARD1 study gives us a clue: perhaps the problem is not only delivery but the vector itself. The CBA promoter (similar in "aggressiveness" to that used in some constructs) may cause chronic inflammation that undermines neuronal health in the long term. P546 works "quietly but surely," supporting the cell without "overheating."
Second insight: This study is a triumph of local (intracerebroventricular/intrathecal) delivery. We are moving toward systemic IV injection (like Zolgensma) becoming a thing of the past. Intrathecal administration allows a 10-100 fold dose reduction, avoids hepatotoxicity, and enables the use of "gentle" promoters like P546. The CHMP has already approved intrathecal Itvisma (Zolgensma for intrathecal administration) for patients over 2 years old. This is a direct signal: the market is shifting to local delivery.
\[Forecast: Next 30 Days and 90 Days]
Next 30 days:
Watch for registration data on GB221 (GEMMABio) and intrathecal Itvisma. If P546 data are extrapolated to SMA, we will see a wave of "redesign" of existing AAV constructs. Small biotechs with patents on promoters like MeCP2 will receive buyout offers. Startups with "heavy" vectors will lose investors.
Next 90 days:
Pay attention to FDA statements regarding long-term monitoring of inflammation in AAV therapies. This study provides a strong argument that neuroinflammation should become a mandatory secondary endpoint for all CNS AAV therapies. If this happens, many existing programs will have to revise safety protocols, adding long-term anti-inflammatory therapy.
Forecast: The P546 study is not just "mice"; it is a blueprint for second-generation therapy. Zolgensma was a "truck" — powerful but crude. The next generation of AAV9 therapies will be a "sports sedan with smart suspension" — precise dosing, minimal inflammation, maximum durability. Clinical data for SMARD1 will emerge in 2026-2027, and if they replicate the preclinical success, a therapy based on the P546 promoter will quickly receive accelerated approval. And Novartis will either have to buy this technology or catch up.
— Editorial Team
