Five asymptomatic spinal findings, a halted gene therapy, and a warning for anyone putting biologics into the spinal canal
Regenxbio said on August 24 that the FDA had put RGX-121 back on clinical hold. The trigger was five MRI scans showing small nodules or cystic masses in patients’ spines. None of those patients had symptoms. All five are doing fine. Those scans only exist because of an earlier problem.
In January the FDA halted RGX-111, Regenxbio’s MPS I candidate, after an intraventricular tumor turned up in a treated patient. It was spotted on a routine brain MRI four years after dosing, in a child with no complaints. RGX-121 caught a partial hold at the same time on grounds of similarity, which Regenxbio disputed publicly. That one came off in May.
But the RGX-111 tumor was not just an odd picture. Surgeons removed it, and sequencing found an AAV vector genome integrated near PLAG1, a proto-oncogene prone to rearrangement. Regenxbio called the case single and inconclusive, and the patient has developed well since. Still, that is a mechanism, not a curiosity. It explains why nobody at the agency was going to shrug at masses in the spinal canal of a sister program.
Meanwhile Regenxbio put its RGX-121 patients under expanded MRI monitoring, brain and spine. Which is where the five findings came from. Dosing had been three to six years earlier, intracisternal or intraventricular. Investigators called them nonserious, radiologists think they are probably benign, no brain lesions turned up. There is no tissue and no pathology. Nobody knows what they are. Nobody knows how common they are either, and that is the whole trap. Spine MRI is not part of routine MPS care or MPS trials. There is no baseline.
Set the two cases beside each other. With RGX-111 there was something to cut out and sequence, and sequencing gave an answer. With RGX-121 there are five images and five children with no reason to be operated on. Regenxbio has been handed the job of proving what these lesions are not, using a comparison population that has never been scanned.
Spine surgeons know this shape of problem. Boden published it in 1990: scan enough asymptomatic lumbar spines and out come the disc degeneration, the annular tears, the Modic changes, all in people who feel perfectly well. Thirty-six years of replication since. The finding is real. What it means is not.
In a clinic that produces an awkward conversation, and sometimes an unnecessary operation. In a regulatory file it produces a hold. Start looking where nobody has looked before, and everything you find belongs to your drug until you prove it doesn’t. To be fair to the FDA, the presumption is not arbitrary here. Asymptomatic masses in the spinal canal, a CSF-delivered AAV, and a sibling program carrying vector integration next to a proto-oncogene. That is a reasonable place to start from suspicion.
The refiling is off. Regenxbio and NS Pharma are gathering more imaging and longer follow-up, and waiting on the full hold letter, which has not arrived yet.
For anyone in spine, the lesson transfers, and it lands hardest on CSF delivery. Intrathecal and intracisternal programs target the exact compartment that produced this. Intradiscal injectates do not distribute that way, so read across with care. But the design flaw is shared, and it is easy to build in without noticing.
Layer imaging surveillance onto a compartment with no published natural history and you will generate findings you cannot interpret. That is not an argument for looking away. After January, looking away was never on the table, and it never should be. It is an argument for building the comparison first: baseline scans before dosing, imaged controls, whatever prevalence data exists.
That costs money up front. Skipping it has left Regenxbio unable to defend a filing it had already dragged back from the dead once.
COVER IMAGE: https://www.magnific.com
###
Legal Disclaimer:
Content published on SPINEMarketGroup (thespinemarketgroup.com) is provided for informational purposes only and on an “as is” basis, without warranties of any kind. We do not guarantee the accuracy, completeness, legality, or reliability of any information, images, videos, links, or third-party materials featured on the site. Any views or claims expressed in contributed content or press releases belong to their respective authors or sources. For copyright concerns, factual corrections, or content-related complaints, please contact us directly for review.