When Heather Mefford, a physician scientist and clinical geneticist at St. Jude Children’s Research Hospital, realized that two of her patients carried a mutation that can also occur in healthy people, she saw the beginnings of a path to treatment. Both children had developmental and epileptic encephalopathy (DEE) with symptoms including treatment-resistant epilepsy, developmental delays, and motor and vision problems. They each had the same mutation in one copy of the gene ubiquitin like modifier activating enzyme 5 (UBA5), which is involved in protein modification. This mutation reduces, but doesn't eliminate, UBA5 function. Their other copy of the gene was non-functional.
A complete lack of UBA5 is lethal, but about 0.25 percent of healthy people carry the same UBA5 variant that Mefford’s two patients do — including some who have two copies of it. That suggested that people might only need a certain threshold of UBA5 production.
To investigate whether boosting patients with this UBA5 variant closer to or above a particular threshold could help, Mefford and her colleagues created the first human organoid model of UBA5-associated DEE, which they described in a new paper published in Science Translational Medicine (1).
Until now, researchers have only studied the condition in animal models or using non-neuronal human cells, but these systems can’t fully recapitulate the gene’s effects on human nervous system cells. Mefford and her colleagues generated stem cells from the two patients and their healthy parents and grew them into organoids. The patient organoids were 25 percent smaller and had fewer inhibitory neurons, which produce the neurotransmitter GABA, than the parent organoids. Neurons in the patient organoids also fired at unusual rates.
Those differences aligned with the fact that the children have epilepsy, Mefford said.
The brain is likely still going to respond in a way that, even if it’s not a cure, that it would improve quality of life.
- Heather Mefford, St. Jude Children’s Research Hospital
“Both the growth phenotype and the neuronal firing phenotype gave us something that told us we had a model that recapitulated at least parts of the condition,” Mefford said. This enabled them to test the disease’s origins and path to treatments.
Cells with one copy of the variant showed disruptions in a cellular process called ufmylation, Mefford and her team found. But cells with two copies of the variant functioned typically. In cell lines, Mefford and her colleagues used two different approaches to boost production of UBA5: CRISPR activation, which turned up expression of UBA5, and SINEUP, which is a long noncoding RNA that binds to mRNA at the start of the gene to increase translation. Both approaches normalized ufmylation in the cells.
Mefford and her colleagues found that patient organoids treated with SINEUP no longer had abnormal neuronal firing rates.
The effect wasn’t permanent — it faded within 12 days. But it served as a proof of principle that upping UBA5 production could be therapeutic, Mefford said.
The work is a strong example of how drug discovery research can be done using only human in vitro models, said Steven Sloan, a geneticist at Emory University School of Medicine who was not involved in the study.
“It’s definitely not that now we have a therapy, but it’s a really impressive first step that never touched an animal,” he said.
Mefford and her team treated the organoids after allowing them to develop for 100 days, which represented the very early stages of brain development that take place in utero — years before a person would likely be diagnosed and seeking treatment for DEE. But the approach could still help kids diagnosed with the condition, Mefford said.
“The brain is likely still going to respond in a way that, even if it’s not a cure, that it would improve quality of life,” Mefford said. “I hope that’s true for a lot of neurodevelopmental and epilepsy conditions.”
Mefford next plans to examine why inhibitory neurons are particularly susceptible to UBA5 mutations when the gene is active in cells across the body. She also wants to explore the best ways to deliver a future therapy and extend its efficacy.
Reference
- Chen, H. et al. Patient-derived models of UBA5-associated encephalopathy identify defects in neurodevelopment and highlight potential therapeutic avenues. Sci Transl Med 17, eadn8417 (2025).
