GM1 gangliosidosis is one of those diseases that most people never hear of, because the children who have it rarely grow up. A single broken gene leaves cells unable to make a working copy of an enzyme called beta-galactosidase. Without it, a fatty molecule called GM1 ganglioside piles up inside neurons and slowly poisons the brain. There is no approved treatment. In the type II form, kids lose skills they had already learned, and families watch the reversal happen.
A team led by researchers at the National Human Genome Research Institute has now tested a way to put the missing gene back. According to PubMed, their phase 1-2 trial appeared in the New England Journal of Medicine in February 2026. Nine children with late-infantile or juvenile type II GM1 gangliosidosis each received a single intravenous infusion of AAV9, a harmless virus engineered to carry a functional copy of the beta-galactosidase gene into cells throughout the body, including the brain.
What showed up in the spinal fluid
The clearest signal came from chemistry. In every one of the nine participants, the enzyme the disease had stripped away came back. Beta-galactosidase activity in the cerebrospinal fluid rose after treatment, and the toxic GM1 ganglioside that had been accumulating went down. That is the biochemical loop the disease breaks, and the therapy appeared to close it.
Brain scans pointed the same direction. Neuroimaging showed patterns consistent with a slower rate of cerebral atrophy and more favorable changes in myelination compared with each child's own baseline. On a standardized clinician rating of overall change, the median score was 3 (minimal improvement) at two years and 4 (no change) at three years. That may sound modest until you consider the comparison: in historical controls, scores drift the other way, toward worsening, as the disease runs its course.
The therapy is not gentle
Restoring a gene this way asks a lot of the body, and the trial documented the cost. Over three years the nine children logged 124 adverse events. Thirty were judged possibly, probably, or definitely related to the gene therapy, including gastrointestinal problems and laboratory signs of inflammation. Five serious adverse events occurred. One child was hospitalized for severe vomiting that the investigators attributed to the treatment. Liver enzymes rose in all nine participants, a known hazard when high doses of AAV vector reach the liver, though those levels returned to baseline by 18 months.
The clinical picture was also mixed rather than uniformly hopeful. Expressive communication and gross motor scores looked stable, which in a relentlessly progressive disease is itself notable. But fine motor and receptive communication scores declined. The gene therapy seems to have bent the curve on some abilities while others kept slipping.
Reading the result honestly
This is a small, open-label study. Nine children is not a number that settles anything, and without a randomized placebo group, some of the apparent stability could reflect who was enrolled or the natural variability of the disease. Safety, not efficacy, was the primary endpoint, so the biochemical and developmental findings are secondary readouts that need confirmation in larger, controlled trials. The follow-up runs three years, which is meaningful but still short for a condition measured across a childhood.
What the trial does establish is a proof of principle for a disease that had none. A one-time infusion reached the brain, switched the missing enzyme back on, cleared the substrate it is supposed to break down, and left several measures of decline flatter than they would otherwise be. For a fatal inherited disease of early childhood, moving from nothing to that is a real, if early, step. The authors are clear that the goal now is to test the approach in more children, under conditions strict enough to say how much of the benefit is the therapy and how much is chance.
Comments