A newborn baby was diagnosed with a condition associated with a 50% early mortality rate. Just seven months later, the same patient received a therapy that didn’t exist when they were born. Musunuru et al. (2025) describe a custom base-editing treatment for a single infant with Carbamoyl-phosphate synthetase 1 (CPS1) deficiency, delivered via lipid nanoparticles under an expanded access Investigational New Drug (IND)¹. 

CRISPR base editors can rewrite single DNA bases without causing double-stranded breaks, offering a more precise alternative to traditional CRISPR editing. This precision makes them suitable for correcting the single-letter mutations that underlie many rare genetic disorders. In theory, they could address more than 90% of disease-causing variants. In this article, we look at a use case that brings that theory to the bedside.

What is CPS1 Deficiency?

CPS1 deficiency is a rare disorder of the urea cycle that prevents the removal of excess nitrogen. Without treatment, ammonia accumulates rapidly in the blood, leading to neurological injury or death. It affects around 1 in 1.3 million people and carries a 50% mortality rate in early infancy.

Management usually involves strict dietary control, nitrogen scavenging drugs, and, in many cases, liver transplantations. But timing is critical; many infants experience life-threatening crises before they are big enough, or stable enough, to undergo transplant. For this patient, the window to intervene was measured in weeks.

Patient-Specific Therapy

The newborn in this study was diagnosed with neonatal-onset CPS1 deficiency within 48 hours of birth. Genetic testing revealed two truncating variants, one on each allele. In response, a multidisciplinary team rapidly designed a corrective adenine base editor targeting the Q335X mutation. The therapy, named k-abe, combined this editor with a guide RNA (kayjayguran) and was delivered to hepatocytes using a lipid nanoparticle formulation.

Preclinical work, including in vitro testing in human liver cell lines and in vivo testing in mice and nonhuman primates, supported the decision to proceed under a single-patient, expanded-access IND protocol. Two infusions were given at 7 and 8 months of age, with careful monitoring and prophylactic immunosuppression.

Preclinical Validation

To test editing efficiency, researchers first developed proxy cell lines with the patient’s CPS1 variant inserted. These were used to screen multiple editor-guide RNA combinations. The most precise approach was selected based on editing accuracy and minimal bystander effects.

Mouse models carrying the patient’s mutation showed up to 42% corrective editing in liver tissue at clinically relevant doses. Safety studies in primates revealed no adverse effects, and off-target activity was minimal when tested in both engineered and primary human hepatocytes.

Clinical Outcomes

The patient tolerated both infusions without serious adverse events. Within weeks of the first dose, clinicians were able to increase dietary protein and reduce nitrogen-scavenger medication. After the second dose, the drug was halved. Notably, the patient recovered from two viral infections without a hyperammonemic crisis, something that had previously triggered severe biochemical instability.

Median ammonia levels dropped from 23 µmol/L pre-treatment to 13 µmol/L after the second dose. Weight gain improved, and neurological function remained stable. While follow-up is still limited, these initial results suggest a durable biochemical response and potential clinical benefit.

Implications and Outlook

This case shows what is now possible when urgency, infrastructure, and technical capabilities align and may point to a new way of responding to rare disease in real time. A therapy designed for one patient, tailored to a unique mutation, manufactured in months, and delivered safely, was once unimaginable. Now, it is a working reality.

While the approach is not ready for routine clinical use, it suggests a new way of treating severe and very rare diseases, where therapeutic design begins not with a population, but with an individual genome. If core elements like the lipid nanoparticle and editor mRNA can be standardised, then only the guide RNA needs to be adjusted, allowing faster and more accessible development in future cases.

It is early days. Long-term safety, scalability, and regulatory frameworks still need to be tested. But this case suggests something important: the one-patient model might not remain the exception.

References

1. Musunuru K, Grandinette SA, Wang X, et al. Patient-Specific In Vivo Gene Editing to Treat a Rare Genetic Disease. N Engl J Med. Published online May 15, 2025:NEJMoa2504747. doi:10.1056/NEJMoa2504747