The coordinated effort that last spring saved the life of little KJ Muldoon earned widespread and enthusiastic media coverage. But between the invention of the treatment and its delivery to the patient lay a lesser-told story: an unprecedented manufacturing sprint. Genetic Engineering & Biotechnology News organized an online roundtable led by its deputy editor in chief, Julianna LeMieux, to discuss how therapeutic components were produced quickly, cost-effectively, and to clinical-grade standards.

The spotlight was not on the clinicians who treated KJ at the Children’s Hospital of Philadelphia, nor on the researchers from the Innovative Genomics Institute (IGI) who supported them. Instead, it turned to those who usually remain behind the scenes: the industrial scientists tasked with executing the most operational—and time-critical—part of the mission.

Gene-editing pioneer Fyodor Urnov – now director of the first pediatric CRISPR treatment center – contacted Sadik Kassim, Chief Scientific Officer and Chief Technology Officer at Danaher Corporation. Kassim immediately brought in Sandy Ottensmann, who at Danaher subsidiary Integrated DNA Technologies (IDT) oversees guide RNA for editing, and Mark Wetzel, who manages messenger RNA (mRNA) drug manufacturing services at Aldevron, another Danaher company. Their teams worked late into the night to keep the logistics running smoothly, ultimately producing a drug in less than half the time of the most optimistic forecasts.

Many others also played essential roles: lipid nanoparticles came from Acuitas Therapeutics; the base editors were invented by David Liu’s group, with the model used for KJ optimized by Ben Kleinstiver; and Cat Lutz at the Jackson Laboratory developed the mutant mouse models for safety testing. Beyond IGI, the University of Pennsylvania played a pivotal, central role, of course, with gene-editing specialist Kiran Musunuru and metabolic disease expert Rebecca Ahrens-Nicklas. But let’s stay focused on the manufacturing leads.

At IDT, the team began by asking questions. Does the guide RNA sequence needed to target the chosen site have any unusual features that might complicate production? How do we streamline this? How quickly can we assess off-target risks? “We needed both a first batch for toxicology and an actual final batch for the drug product, so that we could get that to Aldevron as quickly as possible,” explains Ottensmann. “While the science is fairly straightforward here, despite the novel base editor, it was the timeline that was really daunting,” adds Wetzel.

During planning, they had to consider further issues. “How do we take really thoughtfully calculated risks and risk-benefit for this, especially given the patients? What is appropriate for a patient who doesn’t have time to necessarily wait for a year or two?” says Ottensmann. To save time, IDT took the unusual step of shipping the guide RNA to manufacturing before quality-control testing was complete, a decision they would never normally make.

“We parallel processed a lot of things. We started some things earlier, when normally you would wait for contracts to be 100% finalized or certain data to come in. And you take that risk, which normally you wouldn’t take as part of a business, but we did because we had to look into every single way to save time. And I think that was something that I take away from this, of getting people to think differently and to realize, when are you willing to take a little bit of a risk to do that? Because having those extra days, those extra weeks count so much,” she stresses. It was never about cutting corners, it was about taking business risks, and the project benefited from the skill of the partners’ legal teams. “We also learned a lot about how to make things more efficient by pre-planning, queuing up materials, doing some practicing on equipment in advance, just so that we were ready when the material came and we could produce efficiently. So, those are the kinds of steps that we took. And those are the lessons that we can take away and reapply going forward,” confirms Wetzel.

Kassim highlights the excellent collaboration among different ecosystems, including industry, academia, and regulators. The FDA was involved from the outset, well before the Investigational New Drug (IND) application was filed. Federal funding also played a crucial role in supporting preliminary research at Penn, IGI, and other institutions. “When you zoom out, what we have here is two Nobel prize-winning technologies, mRNA, which won the Nobel prize in 2024, and gene editing, which won the Nobel prize in 2020, coming together in the formation of this drug product. But then, you also have these innovations like base editors, which can target a single mutation. You have lipid nanoparticles, which enable targeted delivery to the liver, and you have sequencing technology and next-generation sequencing to enable on-target. So it really required all of these technologies to come together, be stitched together in such a way that enabled a coherent workflow. I don’t think this could have taken place in any other ecosystem outside of the United States, nor at any other time. So it’s really the convergence of technologies, luck, and willing partners across the ecosystem that enabled it. And now I think the responsibility for us, or the mandate is, how can we recreate what was lucky the first time, but how can we actually double down and make it sustainable?” says Kassim.

From a regulatory standpoint, everybody agrees that editing should be treated as a platform. The therapy has three components, two of which – the editor and the vector- can remain the same, while the guide RNA is patient-specific, tailored to each individual’s mutation. With platform-based regulation for monogenic Mendelian disorders, treatments could be developed almost on demand, like they did with Baby KJ. “But I think with each new turn of the learning loop, you can potentially compress that timeline. So, although KJ’s drug has been called the fastest drug ever developed, I think there’s a world of possibility where we can even compress it to below six months. I mean, maybe even a month or two months or three months, theoretically, once you platform the drug product,” Karim suggests.

Still, the guide RNA must be customized for each patient. How much room is there to improve this step? “[At IDT] we have built our business on coming up with oligos, synthesizing them very quickly in large quantities. I’m actually sitting in our facility where we make literally thousands of guides a day. But those are not the quality you need in a therapy, so we have to work with regulators. How do we work on this platform with phase-appropriate material that is the correct risk-benefit for these patients?” says Ottensmann.

While much is said about scaling up production, in the case of ultra-rare patients (“N of 1” or “N of few”), the opposite is true: scaling down is required. “Like many in this space, we are not set up to operate at this small scale. We’re manufacturing volumes of RNA at a much larger scale. So I mean, we overproduced here because that’s what we had to do. So, there’s a lot that we can do in terms of innovation here, technology platforms, continuous RNA manufacturing, for example. But also we talked a lot about the application of AI, in the way that it can benefit the front office, in construct optimization, and also in our operation. Can we auto-create batch records? Can we detect deviations using AI?” he says.

Many have wondered how much it costs to save baby KJ. There is no clear answer. “We didn’t see this as a commercial opportunity, so we didn’t go track all the costs of exactly how much this would all be,” Ottensmann explains. Still, some cost-reduction strategies are emerging. “What is appropriate when you’re working on an N of one or an N of few versus when you’re working on a bulk that’s going to 100,000 people? I mean, there are very different risk calculations you need to have, and we need to work with them on it because if we can get those rules, you’ll be able to drive the cost down significantly,” she says.

Kassim points to the role of technological progress with striking examples: “Just in the last 20 years, the price of whole genome sequencing has dropped 100,000-fold, and the infrastructure that’s been built to support sequencing is very different today than it was 20 years ago. Rockets, for example, when we first sent a man to the moon, that was cost-prohibitive, but now we have reusable rockets that are being sent into space regularly”. Moving toward more synthetic, chemistry-based processes conducted outside living cells will also drive costs down, not only in manufacturing, but in analytical characterization of the drug product.

Some of the key watchwords are reusability, scaling down, technological innovation, internal efficiency, and adaptive regulatory and reimbursement frameworks. Pressed again on whether a dollar figure will eventually be set, Kassim points instead to the “ecosystem costs”, i.e. the investments over the years that laid the foundation for this achievement.

“Even if we did have the specific number, it would be just arbitrary for what’s possible. I mean, this is what was possible six months ago or a year ago. But even in that time, from last year to today, I think the cost would be significantly lower than last year, just because of operational efficiencies that Mark and Sandy referred to, among other ways of working. So, I think this is just going to continue to decline over time,” says Danaher’s CSO.

[translated and adapted from Osservatorio Terapie Avanzate]