The Food and Drug Administration on Monday placed a clinical hold on Regenxbio's gene therapy for Hunter syndrome, STAT's Jason Mast reported, after small masses were found on the spines of five patients in the company's trial, three to six years after each received a single dose. The company said the patients "did not have any symptoms related to the masses." The finding followed an earlier, more serious event: in January, a patient treated with a closely related therapy for a similar disease developed a brain tumor, the first ever conclusively linked to the kind of gene therapy the field has used for twenty years.
For families living with Hunter syndrome, a rare, progressive disorder that damages the brain and the body from early childhood, the hold is another delay in a line of delays. For the field, it is something more specific: the moment when the defining feature of a one-time gene therapy, its permanence, came under formal examination.
The promise and the risk are the same property
The entire premise of a one-time gene therapy is that it does not need an off switch, because it is designed to be permanent. The therapy delivers a working copy of a gene inside a delivery vehicle, an adeno-associated virus, or AAV, that carries the corrected instructions into cells, where they are meant to keep working for the patient's lifetime. A child treated once is supposed to be treated forever. That is the promise, and for diseases with no adequate treatment for the brain, it is a profound one.
The risk is the same property viewed from the other side. What persists for a lifetime also persists to be observed for a lifetime, and anything that goes wrong inside a permanent treatment has no easy way to be turned off. The masses found on the spines of five patients are, by the company's account and the investigators' judgment, asymptomatic and likely benign. Their significance is unknown, and unknown is exactly what the FDA will not allow before it decides the therapy's fate. The point is not that the findings are dangerous. The point is that a permanent treatment concentrates its risk in exactly the timescale, years after dosing, that a fast-moving disease leaves no time to watch.
The first tumor changed the field
The expanded monitoring that found these masses exists because of what happened in January. A patient in the trial of RGX-111, Regenxbio's therapy for the related disorder MPS I, developed a central nervous system tumor, and researchers concluded it was the first tumor conclusively linked to AAV gene therapy. AAV is the most common delivery vehicle in gene therapy, the workhorse of a field that has treated thousands of patients over two decades. A first of this kind was always the field's nightmare scenario, and the response to it was, appropriately, more looking.
Regenxbio began expanded brain and spine MRI monitoring across its programs. The new imaging found what earlier, brain-focused monitoring had not: the spinal findings in five Hunter syndrome patients. That sequence matters. The field's safety record is partly a record of what it has chosen to measure, and the moment the measurements widened, the findings followed. Nobody knows yet whether that means the therapy is less safe than believed or simply more observed.
The scans ask a question no one has answered
Here is the complication at the center of the hold: spinal MRI is not routine for Hunter syndrome patients, in ordinary clinical care or in clinical trials. The underlying prevalence of asymptomatic spinal findings in this population has never been established, because the scans were never done. Radiologists consider the masses likely benign, investigators deemed them nonserious, no brain masses were found, and all five patients continue to do well clinically, with stable or improving scores on neurocognitive and neurobehavioral assessments. Some experts have suggested the findings may be inherent to Hunter syndrome itself, the disease showing up in places nobody had previously looked.
A finding without a baseline is not evidence. It is a question. The FDA's job in that situation is to answer the question before approving a permanent treatment, and answering it requires time, more imaging, and comparison data that may take years to collect. The hold is not a verdict. It is the safety system doing the thing it exists to do, at the speed at which safety questions are actually answerable.
What is actually known about the five findings
The facts are narrower than the headlines and worth stating precisely. The five patients were treated three to six years ago, through administration into the cerebrospinal fluid space, the route designed to reach the central nervous system where Hunter syndrome does its damage. The findings were small nodules or cystic masses on the spine, spotted by an MRI program that scans both brain and spine, which Regenxbio put in place only after the January tumor case. Investigators deemed the findings nonserious, radiologists consider them likely benign, and there is no clinical or pathological evidence confirming what they are or what caused them. No brain masses were found. All five patients continue to do well clinically, with stable or improving scores on neurocognitive and neurobehavioral assessments.
The company's own view is that the findings are specific to the Hunter syndrome program. Its Duchenne muscular dystrophy therapy and its eye disease therapy use different delivery constructs and routes, and both programs remain on schedule, with a Duchenne application planned and eye disease topline data expected late this year. Whether the distinction holds is part of what the FDA will now examine. The agency's question is not whether the five patients are sick. It is what the masses mean for everyone else who might receive the therapy, and the honest answer, as of this week, is that the data to answer that question do not yet exist.
What the hold actually stops
The practical consequences are concrete. Regenxbio's application for approval of the Hunter syndrome therapy received a complete response letter from the FDA in February 2026, after which the agency agreed to review a resubmission based on existing studies. The company had planned to refile in the third quarter. The hold, Fierce Biotech reported, has derailed that plan, and Regenxbio now says it does not expect to resubmit in the near term. Its shares fell roughly 35 percent. The company notes that its Duchenne muscular dystrophy and eye disease programs use different constructs, and both remain on track. The company and NS Pharma will evaluate additional imaging and longer-term follow-up data and incorporate the FDA's full clinical hold letter into whatever comes next, and the agency will take its time: the question now is not whether the therapy works, but what a decade of permanence looks like.
For families, the accounting is simpler and harder. A therapy that might have been decided on this year is now undecided for longer, and Hunter syndrome does not pause while regulators look. The hold protects future patients from an unresolved question. It costs current ones the one thing their disease makes scarce.
Permanence is being measured on the disease's schedule
This is the honest difficulty, and it does not resolve neatly. A one-time gene therapy is being evaluated on the very timescale that makes it valuable, years after a single dose, and the evidence needed to understand that timescale accumulates more slowly than the disease it treats. The FDA is right to look. The monitoring program was right to widen. The families are right to feel every day of it. None of those truths cancels the others.
The field will learn from this what a permanent treatment looks like over a lifetime, because the therapy has no off switch and the question cannot be hurried. The learning will spread, as these lessons always do: monitoring programs that begin at one company become the standard the regulators expect from all of them, and the scans that found these masses are already the model for what the next company's program must include. That is how a field absorbs its first tumor and its first unexplained findings, and it is not a failure of the technology. It is the technology's safety system finally gaining the instrument that measures the timescale the technology itself created.
That is not a tidy resolution, and it should not be dressed up as one. It is the shape of the waiting itself: a safety system that refuses to unsee what it has started looking for, and patients for whom thoroughness and urgency pull in opposite directions.
Primary sources
- STAT's reporting by Jason Mast for the clinical hold, the five spinal findings, the January tumor case, the expanded monitoring program, and the regulatory timeline.
- Fierce Biotech's coverage of the hold and Regenxbio's regulatory update, including the company's statement on the patients' lack of symptoms.