At Children’s Mercy Hospital in Kansas City, researchers have created something extraordinary: tiny, beating lab-grown “hearts.” Visible only under a microscope, the diminutive innards are called organoids. They can be grown in a matter of days from a patient's own stem cells, and their doctors use them to screen for the best medicine for their condition, sparing months of trial and error.
They’re also core to the future of drug testing, and someday perhaps the end of the lab rat.
Animal testing has been mandated by law since 1937, when a new formulation of a common antibiotic had a poisonous new ingredient — and killed more than 100 people. Nearly a century later, drugs are still being pulled from shelves because they have toxic effects, even though animal testing showed they were safe. Now, politicians, scientists and entrepreneurs are pushing for new, more accurate ways to test drugs before they get to human clinical trials — potentially saving lives and billions of dollars in the process.
In 2022, a group of scientists ran an experiment with 27 known drug compounds that animal studies had shown to be safe. Some of them had turned out to have toxic side effects and had been pulled from the market after they’d killed people. The researchers tested the 27 compounds on a new technology called “organ-on-a-chip”: similar to organoids, “organ chips” have clusters of cells embedded in a diminutive electronic device that can simulate an organ’s behavior. The researchers found that liver organs-on-a-chip accurately predicted which compounds were dangerous, an advancement that might someday lead to significant cost savings in the extremely expensive drug development process. More accurate testing using organ chips could save the industry over $3 billion a year, the study’s authors calculated.
On top of safety, cost is another reason to move away from animal testing. Today, pharma companies often spend more than $2 billion to bring a single drug to market, with the industry spending nearly $300 billion a year on research & development. But despite these vast R&D expenditures, more than 90% of drug candidates fail. It’s a wasteful process, contributing to the flabbergasting prices of drugs that do make it to market.
Animal testing, a first step in the process for many drugs, is a key factor here. It simply isn’t as accurate as it needs to be, leading researchers down a multitude of costly rabbit holes and dead ends. A common joke among them is that we’re capable of curing nearly every disease–in mice.
“Clearly we’re not getting realistic information from animals, because everything that gets to the clinical trial stage has first gone through animal trials and succeeded, right?” said Ali Afshar, CEO of London-based Mytos, which is developing a new, automated way to grow cell cultures, replicating human cells in a petri dish so you can then test drugs on them.
Conventionally, cells are often grown manually, which can lead to inconsistencies from one culture to another, making it harder to replicate experiments. Automating the process provides faster and more reliable data, while freeing up researchers’ time to do more important work. Mytos, which was founded in 2016 and has raised a total of nearly $29 million, is selling its cultures to pharmaceutical customers to test treatments for diseases where animal models don’t match what happens in people, Afshar said.
Organoids and cell cultures are a few of the ways the FDA has proposed eliminating animal testing, starting with a class of drugs called monoclonal antibodies, which mimic the immune system’s natural antibodies and are used to treat everything from cancer to Crohn’s disease to COVID. Testing these drugs is difficult because they don’t often work in mice and must be trialed on larger, more closely related to humans–such as monkeys, which can cost tens of thousands of dollars each. But even then, animal tests often produce misleading results on how these drugs will work in humans. The FDA issued guidance in April suggesting that drug developers use these alternatives to prove these compounds are safe. The idea is to rely on actual human data to determine what drug candidates are the most promising.
That’s only possible because President Joe Biden signed the FDA Modernization 2.0 Act in 2022 after it unanimously passed the Senate. The law dropped the requirement to test drugs on animals for FDA approval when there’s other safety data from computer simulations or mini-organs satisfactory to regulators. It also eliminates the requirement in medicines that are biologically similar to drugs already on the market. Bipartisan consensus on reducing animal testing has continued during the second Trump Administration, with the National Institutes of Health also announcing a new initiative to reduce the use of animals in research by prioritizing funding and inter-agency coordination for these new technologies.
But each new form of testing has tradeoffs. Organoids, for instance, have limitations, said Julie Frearson, chief scientific officer at drug R&D company Charles River Laboratories, which last year launched a $500 million initiative to reduce reliance on animal-based research. They give you a picture of how a drug impacts one specific area of the body, but they don’t reveal how it affects a patient systemically, she said. Testing a treatment on a heart organoid, for instance, doesn’t tell you how it will impact the liver or kidneys. Organoids are also relatively short lived, making it harder to understand a drug’s long-term effects.
San Francisco-based startup Gordian Biotechnology hopes to reduce the number of animals used in testing while still keeping the advantages of long-term, systemic data. It has developed a technique it calls mosaic screening, which enables it to evaluate multiple gene therapies in one animal by introducing its drug compounds into a single cell. By changing the DNA of that cell, which still interacts with other systems in the body, you can get a good idea of what the drug’s chronic effects might be.
That reduces costs enough to allow Gordian to use animals that are a better match for humans in research, like horses, which have similar diseases that come with old age rather than mice. The $170 million company is already using its technique to develop gene therapies for age-related diseases including osteoarthritis and fatty liver disease.
“The core clinical challenge that literally every biotech faces is that animals are not human, and organoids are not either, and cells are not,” Gordian CEO Francisco LePort told Forbes. “You can't know what's going to work in a human until you really try it in a human.”
More robust organoids could help. San Francisco-based Vivodyne, which in May raised a $40 million series A, is creating more complex organoids that are bigger than the microscopic versions and work more like organs in the body do. Some are even able to circulate blood. CEO Andrei Georgescu said this will improve scientists’ ability to predict how well drugs work much earlier in the development process. Right now, he said, “we do not have enough observations to be able to easily correlate that. And that is really the limiting factor.”
Scientists are also looking to artificial intelligence to help model the ways drugs interact with the body. The FDA Modernization 2.0 Act, as well as the guidance documents from the FDA and NIH, all refer to AI as a potential animal testing alternative. But while AI has accelerated our understanding of the basic building blocks of the human body, scientists are still a long way away from being able to reliably simulate how a drug might interact with human patients.
“At best, we probably understand 10% to 15% of the fundamentals of biology,” said Najat Khan, chief R&D officer at pharmaceutical company Recursion, which is working on its own AI models for drug discovery and has several cancer drugs currently in human trials.
Cambridge, Massachusetts-based Parallel Bio, an early-stage biotech company with $30 million in funding, is using AI to model the human immune system. To do so, it’s testing out immune responses to potential medicines on organoid lymph nodes derived from patients from a wide variety of demographics. It then uses its AI tools to better predict how safe and effective a treatment might be for its pharmaceutical customers, which currently include vaccine developer Centivax. The company’s ultimate goal is “to pull away drug discovery from animal models into more relevant models,” CEO Robert DiFazio told Forbes.
He admitted that while technology may never replace animal testing completely, the collective alignment towards that goal is energizing.
“This is one of the few things right now that is securely bipartisan,” he said. “I think the general public understands that this is one of the reasons why drugs are so expensive and why it's so hard to make medicine.”