Søren Hough,Editor, Drug Modalities

May 23, 2026

11 Min Read

BioXconomy/Søren Hough

Few moments in the history of medicine have been as earth-shaking as what scientists achieved in Philadelphia last year. For the first time, doctors delivered an mRNA-based CRISPR therapy directly to a human patient. The recipient was an infant named KJ born with a debilitating and potentially fatal metabolic disorder called carbamoyl-phosphate synthase 1 (CPS1) deficiency.

Until base editing became an option, the only available treatment would have been a liver transplant – a procedure KJ was too young to undergo.

The project brought together an interdisciplinary team of physicians, researchers, and manufacturers to develop and deliver their base editing system to the patient. The goal was to correct the CPS1 gene in KJ’s liver cells to restore healthy function. Today, the University of Pennsylvania and Children’s Hospital of Philadelphia (CHOP) report that KJ has seen “meaningful clinical benefits” resulting from this treatment, including the ability to eat a normal protein-filled diet. His doctors are hopeful that if he eventually needs a transplant, they can delay the surgery until he’s old enough.

Related:Infant treated with first in vivo CRISPR therapy

The case study was undoubtedly a breakthrough, earning lead University of Pennsylvania investigators Kiran Musunuru and Rebecca Ahrens-Nicklas a spot on the 2026 TIME100 list of most influential people in the world. Their work has only fueled the global race to get CRISPR into the clinic as companies like Verve Therapeutics and YolTech Therapeutics push into clinical trials with their own base editing therapies.

Yet Musunuru, a physician scientist who also serves as the co-director of Penn and CHOP’s Orphan Disease Center, doesn’t quite see it that way.

“That was not a very scientifically meaningful effort,” he told audiences at TIDES USA in Boston, Massachusetts. “In fact, it wasn't really science at all. It was an expanded access IND [Investigational New Drug] application for a single patient as part of clinical care.”

“It was not a clinical trial. It was not clinical research. It was unambiguously not a cure.”

For Musunuru, the intervention was necessary and warranted – but it was only a demonstration of what is possible.

“The best we can say is that we hope we've turned a devastating life-threatening disease into a milder or manageable condition,” he said. “But honestly, it's really too early to even say that.”

If it seems like Musunuru is down on his own efforts, he isn’t. In fact, he looks forward to a world where personalized genome editing is the norm rather than the exception. What he has in mind involves many more case studies and an overhaul in the US regulatory framework.

Related:Epigenetic editing therapy may reduce cholesterol and coronary heart disease risk

“I would rather be forward-looking at what comes next.”

The LNP Express

Musunuru and his colleague Ahrens-Nicklas, who is also a physician scientist and is an assistant professor of pediatrics, are already bringing their personalized CRISPR approach to a range of conditions. The team recently published a preprint demonstrating in vitro success correcting a genetic variant responsible for another rare disease called methylmalonic acidemia. In other work, they are tackling elevated cholesterol levels by targeting the PCSK9 gene.

Two key breakthrough technologies underpin these putative therapies: mRNA and lipid nanoparticles (LNPs). This approach builds from the success of vaccines against COVID-19 and immunotherapies against solid tumor cancers to deliver CRISPR machinery in patients. mRNA-LNP is increasingly the tool of choice for researchers looking to develop nonviral in vivo cell and gene therapies.

“Viral vectors, such as AAV, are disfavored for a variety of reasons,” Musunuru said, referring to issues around immunogenicity, safety, and redosing. Nonviral proponents believe mRNA-LNPs will avoid some of these pitfalls.

Related:Deceiving the immune system to cure type I diabetes with cell therapy

One of the characteristics of LNPs is that they tend to end up in the liver following injection. In Musunuru and Ahrens-Nicklas’ efforts to edit CPS1, PCKS9, and other targets, homing in on the liver was part of the appeal, as each of these genes is expressed primarily in hepatocytes.

In just 48 hours, the phenylalanine levels dropped to normal – just 48 hours to correct a lifelong genetic condition.

“Fortuitously, most of lipid regulation is centered in the liver, and we have a very effective delivery method into the liver: LNPs,” Musunuru explained. “You can formulate LNPs to contain an mRNA encoding a gene editor and a guide RNA that tells the editor which gene to target, intravenously infuse trillions of LNPs into the bloodstream, and they'll go straight into the liver.”

The whole process is short-lived but effective. After the mRNA-LNP makes its way to the liver, hepatocytes internalize the mRNA-LNP, translate the mRNA, and the base editor alters the target gene. Then, the editing components are filtered out of the body, reducing the risk of repeat base changes or off-target effects.

“That's the job of the liver – to clean things out of the blood,” Musunuru said.

He and Ahrens-Nicklas are applying this base editing mRNA-LNP approach to other disease areas. He highlighted a case study looking at phenylketonuria (PKU), a severe metabolic condition caused by genetic variants in the PAH gene.

“PKU is a fairly well-known disease marked by very high levels of the amino acid phenylalanine in the blood, which, if inadequately treated, can cause a host of neurological problems, seizures, intellectual impairment, and so forth,” he said.

The fact that the disease is monogenic and liver-based made it ideal for base editor mRNA-LNP treatment. The team first tested their design in animal models and published their results in HGG Advances.

“We adapted an LNP drug with an adenine-based editor mRNA and a guide RNA optimized to target one or the other of the [PAH] variants and gave the appropriate versions to the mice,” he said. “To our surprise and delight, in just 48 hours, the phenylalanine levels dropped to normal – just 48 hours to correct a lifelong genetic condition. And it's not just phenylalanine; a lot of the neurological phenotypes resolved within a few weeks, as well.”

Carving new paths

One area the baby KJ story certainly shifted was regulatory perception. In BioXconomy’s coverage of the cell and gene therapy roundtable US Food and Drug Administration (FDA) discussion last year, we highlighted top administrators who referred to the news as a “miracle.” This created a fertile environment for those working in rare disease to push forward with new frameworks, including the “plausible mechanism pathway,” for faster approval in rare diseases.

Why can't we just change the guide RNA and give it to a healthier person for a more common disease? What possible downside is there?

Musunuru and Ahrens-Nicklas are working on expanding that crack in the door to set precedents for conditions with small patient populations. In his talk, Musunuru called it an “umbrella clinical trial,” where his team would invite PKU patients to participate and stratify them by PAH variant. Each of the patients would then receive the relevant mRNA-LNP base editor for their variant and, depending on the outcome of constant safety and efficacy monitoring, would contribute to the evidence needed to show that this approach is a viable method to address PKU.

“Becca and I think the answer is an adaptive real-time clinical trial design,” Musunuru said. “The boundaries between Phases I, II, and III become quite blurred. Instead, we think of there being a proof-of-concept phase and a validation phase.”

“We test therapies for proof of concept of efficacy using more lenient phase-appropriate standards. Once we have persuasive evidence, then we advance the therapies to the validation phase,” he continued. “And if we're unable to show efficacy convincingly in the proof-of-concept phase, that's it. It doesn't advance and we'll have to find some other strategy for tackling that particular disease.”

Musunuru and Ahrens-Nicklas’ idea is that the cumulative evidence of smaller case studies will eventually build confidence that the platform itself is worthy of regulatory approval. So far, the pair have been in talks with the FDA about this approach since February 2024. Musunuru announced that they have filed an official master IND application for their umbrella study following years of dialogue and anticipate needing to file additional INDs for each variant.

In the long term, they hope this pilot of the umbrella study will pave the way for future trials where investigators do not need to apply again for every new variant affecting the same gene for the same condition.

“The interesting question is once you get enough experience with a drug product and you're starting to deploy it widely across many rare diseases – because that's going to be easier first – then you can start to ask a very provocative question,” he said. “Why can't we just change the guide RNA and give it to a healthier person for a more common disease? What possible downside is there?”

“I don't know what the FDA will say, but I think that would be a pretty strong argument.”

A question of safety

Early in the days of base editing, researchers believed off-target changes to the genome were diminished relative to classical CRISPR-induced edits. Part of this comes down to using a Cas9 nickase rather than its double-strand break-inducing counterpart, because single-strand breaks are generally much safer than fully severing our genome. However, more recent studies suggest that base editors, like all genome editors, still need to be carefully evaluated with specific tools to ensure safety.

“Base editors can have off-targets. Any editor can have off-target edits,” Musunuru said to BioXconomy. “The nature of the off-target editors or edits is different than, say, nucleases. With nucleases, you can get chromosomal rearrangements and so forth. You see a lot less of those larger-level structural variants [with base editors] because you're not necessarily cutting all the way through the genome.”

Still, even without the danger of full breaks, base editors can change the subject’s DNA at unintended locations. In extreme cases, this could lead to disease outcomes.

“The off-target edits tend to be more along the lines of single nucleotide edits,” Musunuru said. “Some of them are predictable from sequence homology. Some of them are so-called guide RNA-independent editing. That just kind of happens stochastically if the deaminase bumps into an RNA or DNA.”

Kiran Musunuru presenting at TIDES USA. Photo by Søren Hough.

He emphasized careful design and monitoring are key for in vivo base editing studies.

“You de-risk. And the FDA, believe me, makes you go through all the studies to de-risk,” he said. “You don't have to show that it's 100% clean, because that's probably impossible.”

Musunuru said that they were particularly conscious of the potential for nonspecific editing in the KJ case, as well.

“[When] we dosed this single patient with CPS1 deficiency, we found an off-target,” he said. “It was an intron of a copper transport gene. It didn't seem to have any meaning whatsoever in terms of oncogenic risk for the patient.”

Perhaps more importantly, the urgency of the case meant that the potential safety concerns were outweighed by the potential that KJ might experience if the therapy worked.

“This was a very, very, very sick kid. It's horrible,” Musunuru stressed. “It's all about benefit-risk, right?”

This argument works well in the context of a fatal disease, but questions remain about how this calculation will play out when considering in vivo base editing for other conditions.

“Win-win-win”

Given the extraordinary resources that go into cutting-edge bespoke treatments, there remain outstanding questions about access to in vivo CRISPR therapies. BioXconomy asked Musunuru if he had any thoughts on whether these emerging drugs for CPS1 or PKU will be affordable if they finally receive approval.

“That's a long conversation, but these are ultra-rare diseases,” Musunuru told us. “There's not a commercial model. I just don't think there is. Right now, it's all supported by federal funding and philanthropy, and I think we have enough resources to actually get the clinical trial done.”

Now, how do you extend that to many diseases? I haven't figured that part yet. But one step at a time.

Musunuru hinted at plans to expand access through charity efforts, potentially mirroring the work being done at the n-Lorem Foundation, which provides free custom antisense oligonucleotide therapies to patients with nano-rare conditions.

“If it truly requires as few subjects as the FDA is saying that it should and everything works well – that's a big if – [and] if we can get to an approval, then the idea would be to spin it off into a nonprofit entity that's mission-driven.”

Ultimately, Musunuru believes that this is the smart play even at a financial level for the industry. He argues that the cost of care ends up falling on hospitals and other healthcare providers in the end, anyway. Treating them upfront could avoid these issues.

“I think you can make an argument to payers because the costs for urea cycle disorder patients run into the millions over a lifetime,” he continued. “If you come up with a gene editing drug that's hundreds of thousands of dollars, you'll save money, and the patients will have a better outcome. It will be a win-win-win.”

“Now, how do you extend that to many diseases? I haven't figured that part yet. But one step at a time.”