After One Baby’s Custom CRISPR Rescue, a Startup Bets on Made-to-Order Gene Editing

On a June morning last year, after nearly 10 months in the neonatal intensive care unit, Kyle “KJ” Muldoon Jr. left Children’s Hospital of Philadelphia in his mother’s arms instead of on life support.

Born Aug. 1, 2024, KJ had a lethal disorder called carbamoyl phosphate synthetase 1 (CPS1) deficiency. His liver could not process ammonia, which built up in his blood and threatened to irreversibly damage his brain and other organs. Most infants with the condition die within days or weeks.

Doctors tried extreme measures—a near-protein-free diet, drugs to scavenge nitrogen, dialysis. By five months, KJ was on the waiting list for a liver transplant. That was when his care team and a group of gene-editing scientists decided to try something medicine had never attempted before: design a CRISPR treatment for one child, and build it fast enough to save his life.

Six months later, that one-off rescue is becoming a business plan.

In January, a new company called Aurora Therapeutics, co-founded by CRISPR pioneer Jennifer Doudna, emerged from stealth with $16 million in seed funding and a mission to industrialize the kind of bespoke gene-editing therapy that helped KJ. The startup’s launch comes as the Food and Drug Administration rolls out a new “plausible mechanism” pathway that could allow drugs for ultra-rare diseases to be approved on the strength of data from only a handful of patients.

Taken together, they signal a shift in how drugs may be developed and approved in the era of personalized gene editing—from medicines designed for millions to treatments tailored to one mutation at a time.

A baby, a lethal mutation and a six-month sprint

KJ’s disease belongs to a group of urea cycle disorders that prevent the body from safely disposing of nitrogen. The incidence of the neonatal-onset form he has is estimated at about 1 in 1.3 million births.

“He was critically ill from the very beginning,” Dr. Rebecca Ahrens-Nicklas, a pediatric geneticist at Children’s Hospital of Philadelphia, said in an interview released by the hospital when his case was made public. “We were running out of options.”

Genetic testing identified a severe, ultra-rare mutation in KJ’s CPS1 gene. With conventional therapies, his prospects were poor even with a transplant. The CHOP team reached out to collaborators at the University of Pennsylvania and the Innovative Genomics Institute, a research center led by Doudna at the University of California, Berkeley.

The group decided to design an in vivo CRISPR base-editing therapy that would correct KJ’s specific mutation in his liver cells. Base editors are modified CRISPR systems that swap single DNA letters without cutting both strands of the double helix—a feature researchers hope will reduce the risk of dangerous errors.

Under an emergency expanded-access authorization from the FDA—a process sometimes called compassionate use—the team moved with unusual speed. Within about one month, according to scientific and institutional accounts, they had engineered cell models carrying KJ’s mutation. By two months, they had selected an optimal adenine base editor and guide RNA combination. Nonhuman primate safety studies were completed by around five months.

By late February 2025, about six months after KJ’s birth, the group had manufactured a clinical-grade therapy, sometimes referred to as “k-abe,” and secured FDA permission to infuse it into the infant.

Delivered intravenously in lipid nanoparticles that carried the editing components to the liver, the first dose was followed by additional infusions in March and April. A case report in The New England Journal of Medicine described marked reductions in ammonia levels, fewer metabolic crises, improved tolerance of dietary protein, and no serious adverse events in the short term. After nearly a year in the hospital, KJ went home in early June.

He has since been described by his doctors and parents as “thriving,” hitting developmental milestones such as walking. Clinicians stress that he will need lifelong monitoring to assess how durable the gene edit is and whether any late complications emerge.

“We were able to design and deploy a personalized CRISPR therapy in the time frame of a human pregnancy,” Fyodor Urnov, a genome-editing scientist at Berkeley’s Innovative Genomics Institute who helped lead the effort, told Wired in a feature published Jan. 9. “That changes what you can even imagine is possible.”

From N-of-1 to a platform

The treatment KJ received is not an approved drug. It was given under an individual expanded-access investigational new drug application, a mechanism that lets patients with life-threatening illnesses try experimental therapies outside formal clinical trials when no alternatives exist.

But regulators have signaled that such N-of-1 cases will no longer be isolated acts of medical heroism.

In November, FDA leaders outlined a new plausible mechanism pathway in The New England Journal of Medicine. The pathway, aimed at severe, ultra-rare conditions, would allow the agency to grant marketing approval based on data from as few as several patients when traditional randomized trials are not feasible—so long as there is strong mechanistic evidence.

Under the new framework, conditions must have a well-defined molecular cause and natural history. The therapy must directly target that molecular abnormality, its effects on the biological pathway must be measurable, and the clinical changes in patients must align with what the mechanism predicts.

According to public descriptions by the agency and outside legal analyses, single-patient and very small-patient expanded-access experiences like KJ’s are expected to form the evidentiary foundation for later marketing submissions. Companies would still have to collect real-world data after approval to confirm benefit and monitor safety.

“We must adapt our evidentiary standards for diseases where it is impossible to run large trials, while still insisting on scientific rigor,” FDA Commissioner Martin Makary wrote in November, arguing that current requirements were “onerous and unnecessarily demanding” for families with no time to wait.

It is into this evolving regulatory landscape that Aurora Therapeutics is stepping.

Aurora’s bet: a CRISPR factory for rare mutations

Aurora, headquartered in the Boston–Cambridge biotech corridor, was legally formed in 2025 and publicly launched Jan. 9. Its co-founders are Doudna, who shared the 2020 Nobel Prize in chemistry for her work on CRISPR, and Urnov, a longtime genome-editing researcher instrumental in KJ’s case.

Edward Kaye, a pediatric neurologist and former chief executive of Sarepta Therapeutics, leads the company as CEO. Menlo Ventures, a Silicon Valley venture firm with experience in life sciences, led the $16 million seed round and chairs the board.

Aurora’s premise is that the tools used to design KJ’s CPS1 editor can be turned into a platform for many diseases caused by single-gene mutations. Instead of spending years building one gene-editing drug for a common condition, the company plans to refine a small number of base-editing “chassis”—combinations of editor and delivery system—and then customize them patient-by-patient or mutation-by-mutation by swapping the guide RNA that steers the editor to its DNA target.

“Our vision is to transform personalized gene editing from a one-patient breakthrough into a scalable model capable of bringing therapies to millions of patients with rare diseases,” Kaye said in Aurora’s launch announcement.

The company is starting with phenylketonuria (PKU), an inborn error of metabolism that affects an estimated 13,500 people in the United States. Caused by mutations in the PAH gene, PKU prevents the body from breaking down the amino acid phenylalanine, which can build up and cause brain damage if not controlled through strict diet and, in some cases, drug therapy.

There are more than 1,000 known PAH mutations. Under the traditional drug-development model, each might require a separate product and trial. Aurora intends to develop a single PKU base-editing platform and tailor it to many of those variants by altering the guide RNA sequence.

The company plans to run so-called umbrella or basket trials that enroll patients with different PAH mutations under one protocol. In the long term, Aurora and its investors hope that regulators will treat multiple guide RNA-based variants as part of the same product family—a natural fit, they argue, with the FDA’s plausible mechanism framework.

“A single program could potentially help dozens of mutations, rather than starting from zero every time,” Johnny Hu, a Menlo Ventures principal and chair of Aurora’s board, told Chemical & Engineering News.

Promise, pressure and unanswered questions

The approach raises scientific, ethical and economic questions that extend beyond one company.

Because CRISPR editing alters DNA permanently in living cells, the long-term risks are not yet fully understood. Even with base editors that avoid cutting both DNA strands, off-target changes and unintended on-target effects remain concerns. Delivering editors to infants adds another layer of uncertainty about lifetime risk.

Bioethicists have also pointed to the pressures on families asked to consent to an unproven, bespoke therapy when the alternative is a high probability of death or severe disability.

“For parents, saying no may feel impossible,” one commentary in the AMA Journal of Ethics noted last year in a discussion of individualized genetic treatments. The authors urged clearer standards for institutional review boards overseeing such cases.

Access is another issue. The cost of developing and manufacturing KJ’s customized editor has been likened by those involved to that of a liver transplant, and much of the expense was offset by partner companies and philanthropy. Without efficiencies of scale, such interventions risk being limited to patients treated at major academic medical centers with researchers willing and able to assemble ad hoc development teams.

Aurora argues that by standardizing the underlying editor, automating guide design with artificial-intelligence tools and building modular manufacturing processes for small batches, it can drive down the marginal cost and time needed for each new mutation.

Outside analysts say it is too early to tell whether those efficiencies, combined with the FDA’s new regulatory pathway, will be enough to make personalized gene-editing therapies sustainable in routine practice—or to convince insurers and national health systems to pay for them.

For now, the visible proof that any of it is possible is a toddler in Pennsylvania whose genome has already been rewritten once.

KJ’s doctors caution that they do not yet know how long his corrected CPS1 gene will keep functioning, or whether new problems could appear years from now. But his case has already helped reshape expectations for what medicine might one day offer families facing a devastating genetic diagnosis.

As more children like him are treated under expanded access and in small trials, regulators will have to decide how much weight to give their stories when considering drugs that may benefit only a few patients worldwide.

Those decisions will help determine whether KJ remains a singular figure in the history of gene editing—or the first of many children whose medicines are made to order.