Those with CPS1 deficiency typically die in their first year of life without liver transplantation. Researchers at Children's Hospital of Philadelphia have changed this prognosis for one specific child. They developed a gene therapy in six months that had never existed before, tailor-made for the exact mutations of Kyle "KJ" Muldoon Jr. One year after the first infusion, KJ walks, speaks, and requires only half his previous medication.
What CPS1 deficiency does to the body
CPS1 stands for carbamoyl phosphate synthetase 1, the first enzyme in the urea cycle. This cycle is the biochemical pathway through which the liver disposes of toxic nitrogen compounds from normal protein metabolism. Without the enzyme, ammonia accumulates in the blood. Within hours of birth, levels can reach life-threatening heights because newborns produce particularly large amounts of ammonia as they break down proteins.
The disease is extremely rare. The medical database Medscape estimates its frequency in the United States at approximately one in 1.3 million newborns. Nevertheless, according to the research team, roughly half of affected individuals die in their first year of life. Until now, the only curative treatment was liver transplantation, which carries significant risks and is often not possible until after the first birthday.
In KJ's case, geneticists identified two specific CPS1 gene mutations: Q335X and E714X, as the team described in the New England Journal of Medicine. Both variants completely inactivate the enzyme.
How base editing corrects DNA
The research team chose an advancement of CRISPR technology called base editing. The classical CRISPR-Cas9 mechanism cuts both strands of the DNA double helix to remove or insert sequences. Base editing, by contrast, performs a chemical conversion: it converts a single faulty DNA base without severing the strand, greatly reducing error rates in repair.
For KJ, the team developed two separate base editors, one for Q335X and one for E714X. This is the crucial difference from all previous gene therapies: not a standardized tool for a disease category, but a custom-made instrument for the individual mutations of a single patient.
Six months of development, three doses of effect
KJ was born in the summer of 2024. Doctors diagnosed CPS1 deficiency shortly after birth. According to the research team, gene therapy development typically takes ten to fifteen years. The team at Children's Hospital of Philadelphia compressed this timeline into six months. In February 2025, KJ received the first infusion of the personalized base editor, becoming the world's first patient.
After three doses, significant improvements emerged: KJ could tolerate more protein, and medication dosage was reduced by 50 percent according to the team. By March 2026, at eighteen months old, he was walking and speaking at age-appropriate levels. No serious adverse events were observed. Development costs for KJ's therapy ranged from 2.2 to 3.1 million dollars, since each individualized variant must be synthesized anew and cannot be reused for other patients.
In context: Gene therapy milestones before KJ
KJ's case differs in one essential way from all previous gene therapy milestones. On August 30, 2017, the FDA approved Kymriah (tisagenlecleucel) from Novartis as the first CAR-T cell therapy in history, for certain forms of leukemia in children and adolescents. This therapy is also manufactured individually for each patient from their own immune cells. However, the goal is fighting a cancer form, not correcting a specific genetic defect, and the therapy works for all patients with the corresponding diagnosis.
On December 8, 2023, the FDA approved Casgevy (exagamglogene autotemcel) as the first CRISPR-based therapy ever, developed by Vertex Pharmaceuticals and CRISPR Therapeutics for sickle cell disease. Casgevy corrects an epigenetic mechanism that applies to all patients with this condition. It is not a patient-specific construction.
KJ's therapy follows different logic. Development did not begin with a disease but with the exact gene variants of a specific person. The Innovative Genomics Institute at UC Berkeley termed this approach "on-demand gene therapy": medicine developed on an as-needed basis for a single patient.
Between 2.2 million dollars and the next child
The team plans clinical trials for at least five additional children with editable mutations in seven genes of the urea cycle. Whether this succeeds quickly depends on three conditions.
First, development time and costs must decrease. Three million dollars per therapy is not sustainable without redistribution mechanisms in the health system. Bioethicists have pointed to the distribution question in similar high-cost therapy cases: resources spent for one person are not available to others. Researchers are working on standardized development pipelines that should produce new base editors for known mutations faster and cheaper.
Second, the technology needs a growing mutation database. The more editable gene variants become known and classified, the more patients qualify as candidates. Newborn screening programs that currently identify CPS1 deficiency in the United States provide the raw material for this.
Third, regulatory classification remains open. The FDA treated KJ's therapy as an experimental single case. Once multiple patients receive similarly individualized therapies, regulatory agencies must define new procedures that differ fundamentally from classical drug trials.