Organ-on-a-chip, Organoids, but Drug Testing on Whole Organs? This Startup is Betting On It

Since 2022, the Food and Drug Administration has been on a journey to reduce the reliance on animal testing in drug development. The reasons are manifold, according to the agency: improving drug safety and speeding up the drug evaluation process, reducing animal experimentation, lowering research and development (R&D) costs, and ultimately, the cost of drugs. Separately, another widely held belief for reducing animal testing is tied to failures when the drug introduced in humans doesn’t perform as it did in preclinical animal studies. Then there’s the challenge of new complex drugs being developed in order to be delivered to targets directly within specific organs.

In late May, FDA announced a new proposed guidance that would reduce unnecessary animal testing for cancer drugs. The comment period will end later this month, after which a final guidance will be published.

The shift away from animal testing to the so-called NAMs( New Approach Methodologies) is widely seen as a boost to companies that have bet on alternatives for drug testing — they include everything from organoids and lab-grown human tissue-based models to organ-on-a-chip and insilico/computational models. For example, having the FDA validate their thesis is a big boost to companies like Vivodyne and Emulate, both of which are focused on human-based testing in drug development.

Here’s how the CEO of Emulate put it.

“The FDA said, ‘Let’s try to assist the pharma industry in general on how we can start to really chip away and reduce animals and then ideally eliminate them.’ And they picked areas like monoclonal antibodies where you do excessive non-numeric primary testing, and it doesn’t translate [to humans],” said Jim Corbett. “And so why continue the madness?

Whole-organ perfusion

One biopharma executive who is similarly buoyed by FDA’s moves says organoids or organ-on-a-chip company approaches may not be sufficient when it comes to developing drugs. Her company, Revalia Bio, is all about drug testing using donated human organs that cannot be used in transplants. She believes whole organ perfusion must be part of the vocabulary of human-based testing and the industry shouldn’t solely rely on organ-on-a-chip, organoid or other forms of human-based in-silico testing. To that end, the company has developed what it calls its Human Data Trials platform that integrates data from donor organs, organoids, and organ-on-chip systems into a comprehensive translational system.

“A human organ can be a Rosetta stone where these organs are ultimately going to feed the MPS systems (microphysiological systems) — the organoid, organ-on-a-chip-type systems — but we need to be able to grab data from across the spectrum,” said Jenna DiRito, co-founder and chief operating officer of Revalia, in a recent interview. “And so we partner with organoid, organ-on-a-chip-type companies. We’re in all these NAM consortiums. I think the goal for us is to make sure that the data is 1, complete and 2, integrated with each other.”

She argued that when drugs are tested on whole organs like Revalia offers, certain answers can be provided that cannot be gleaned from testing on organoids or organ-on-a-chip-type technologies. And vice versa is also true. So it depends on the question being asked and the research goals.

“So the high throughput drug selection toxicity [is] much better, much better in organ-on-a-chip when you don’t really know what you want to do,” DiRito said of drugmakers looking to test safety of a drug early on. “But delivery? That’s where we have a really strong suit in whole organs. So it’s really about blending all of those technologies together and just making sure we’re asking the right question at the right level for the right stage [of drug development].”

New Haven, Connecticut-based Revalia Bio’s Human Data Trial Platform has intrigued government agencies. Last December, the company won an ARPA-H (Advanced Research Projects Agency for Health) contract worth up to $26.7 million to “lead the effort to build advanced AI-driven models trained on Human Data Trials, living systems derived from donated human organs, organ-on-chip platforms, and multimodal datasets,” according to a Revalia news release.

“We have programs across liver and kidney where we’re creating models of ADME-tox (Absorption, Distribution, Metabolism, Elimination, Toxicity) to ultimately three years from now, or I guess two and a half years from now if we’re successful with the program, we should have a digital trial, a digital human data trial for ADME-toxin in liver and kidney,” she said about ARPA-H’s Catalyst program.

If the company founded in the early 2020s as a spinout from Yale, is successful, then this will lead to the creation of a digital twin model for testing ADME-tox in the liver and kidney. Digital twins have been talked about for years as a panacea for clinical trial recruitment challenges but working examples are still hard to come by in the real world.

Now, where does AI play a role in developing this human data stack? Revalia uses AI in two different ways.

“We have agentic models that help us screen all of the organ offers that are coming in every night and help allocate them to studies. They’ll really parse the data from all of the charts that are given to us. But then the other part of it too is we have so much data that comes off of all of these different organs. So unlike a mouse model where you maybe get a single tissue sample, we get hundreds of blood and urine and tissue samples and we can do multiomic analysis,” she explained. “We are using AI models to really parse down that data and help identify the top hits for receptors or really help recommend — we have all of these different models — which one is the best from kind of a multifaceted type of pipeline to move forward with.”

While the company, founded in 2021 by Yale scientists, currently offers whole organs for research, it will eventually provide hybrid offerings as well.

“We have the physical offering today,” DiRito, who is a Revalia co-founder, said. “We’re eventually hoping to have some type of hybrid offering where our customers can run these digitized trials and get an idea, get a confidence level of what their simulation would be. Will their drug work? Will it get to the place that it needs to go?”

The company calls three of the top 10 pharmaceutical companies its customers, though DiRito declined to name them. Revalia also works with medical device and biotech companies.

“The questions that they’re asking are really, they’re both about PKPD (pharmacokinetic, pharmacodynamic) and toxicity, but a lot of our programs are focused on unique methods of delivery within organs,” she explained.

Revalia Bio works in three main areas: chronic kidney disease, oncology in both liver and lung cancer, and finally in transplantation. DiRito declared that the company has been revenue-generating from day one, adding that the company has raised $20 million from venture capital firms like America’s Frontier Fund and Sierra Ventures, which led the company’s seed round.

Lab-grown tissue-based testing

While Revalia works on whole organs, Vivodyne has built its human-testing platform on lab-grown human organ tissue models. The idea for the biopharma world is that you can get preclinical data on human organs as opposed to on animals, thereby creating safer, more predictable drug candidates that won’t fail once the drug is tested on a person.

“We want to have better predictive models than the animal models. And the idea is that these human-based and human tissue-based models will allow that,” said Anthony Bahinski, chief biotechnology officer at Vivodyne. “It’s a high attrition rate in the clinic. Of 10 compounds that go in the clinic, nine fail, one actually makes it.”

The challenge with human tissue-based models is that they cannot scale rapidly to the level where it would be of interest to pharma and Bahinski contends that Vivodyne’s robotics-based system, called the HIVE system can reproduce results at scale.

“So it’s a fully roboticized system,” he explained. “It actually seeds our individual 3D vascularized tissue chips, cultivates those in situ, and then we can interrogate it with a confocal microscope or we can dissociate the tissues at the end of the experiment and do single-cell sequencing in any kind of omics or sample the effluence from the tissue. These are all perfusable. So then it’s fully automated. So it runs 24/7 without any human intervention once the reagents are placed into the hive.”

The company has licensed 20 different tissue models and, of them, about five to seven have been developed for the HIVE — a fully enclosed, big incubator with a temperature of 37 Centigrade and 5% carbon dioxide, similar to a bio-incubator, Bahinski explained. Where does AI play a role? Bahinski said that if the data from the 3D vascularized tissues were being manually analyzed, it would take an eternity.

“We utilize machine learning and AI to actually do the analysis, things like skeletonizing the vasculature and then looking at different parameters like ‘Are the vessels the same diameter? Are they changing? Are they branching? Are they longer?'” he explained. “We can do that by analyzing all of the 3D images. So you can imagine we do 3D confocal stacks, each tissue, and we can look at immunohistochemistry, label cells in each one and we can actually see which cells are changing.”

Separately, Vivodyne is using the AI analysis to develop its own internal models using algorithms that can predict what the patient response might be to the drug being tested. Pharma companies can collaborate with Vivodyne to co-develop drugs and the company, based in Brisbane, California, works with the top 10 pharma companies, Bahinski said. The company was founded in 2021 and has raised $40 million in a Series A funding round with Khosla Ventures as lead investor.

Organ-on-a Chip Testing

Organ-on-a-chip systems are those that contain engineered or natural miniature tissues that are grown inside microfluidic chips and are meant to replicate the 3D structure, mechanical forces, and physiological functions of human organs. But the main criticism of this technology historically has been that the throughput — the number of samples or compounds a lab can analyze within a time period — is very low. As a result, large pharma companies haven’t adopted the technology in scale.

Enter Boston-based Emulate, one of the foremost players in the organ-on-a-chip market, which was founded in 2013 out of the Wyss Institute for Biologically Inspired Engineering at Harvard University.

“Look, I think that criticism is fair,” said Jim Corbett, CEO of Emulate Bio, about the low throughput characterization. “And our first-generation technology was a lower-throughput product as an instrument. It could only hold 12 chips. We have 600 of these in the marketplace today.”

But the company has come far since then.

“Last year we launched a new product. We totally redesigned the chip, and I’ve increased the throughput by eight fold. So you can now run 96 chips at once for experiments with this,” he declared.

He noted that this iteration also removes some burden on scientists in terms how much microscopy they need to do.

“With my new platform, we’ve built in onboard microscopy. So all of that is done in the background for the scientists, and they read it in the morning or whenever they want,” Corbett explained.

He noted that Emulate has increased throughput, lowered the manual portion of the work scientists must do and lowered costs all in one fell swoop.

In Emulate’s liver and kidney organ-on-a-chip. the company uses primary human cells. In the intestine models, colon and duodenum, the company use primary organoids and in its brain model, it uses iPSC-derived cells – induced pluripotent stem cells.

“All of our organs have blood flowing through them, so we can recreate that environment. And also we can apply mechanical forces to it,” he said.

But is the company’s technology enough to jettison animal testing completely?

“Our customers have published on 30 different models that they’ve developed on our chips,” Corbett declared. “And so to give you a perspective, when you do a safety trial, and you’re doing animal studies, you’re looking at about 44, 45 organs when you’re doing the pathology analysis and so forth. So the science is evolving, the adoption by our customers is evolving. And again, it’s an open platform. So our customers build their own models on these.”

Corbett said that the company’s technology is aiming to address some unmet needs, particularly in brain diseases where you really can’t test in animals.

“We have launched a brain chip, which is an isogenic model with five cell types. Now a researcher can look at the complete neurovascular unit, the blood-brain barrier, and can test some of these large molecules, penetrate it or not depending on what their objective is,” he explained. “Those types of things can’t be done on animals today.”

Compared with Revalia and Vivodyne which didn’t name customers, Emulate has announced a few collaborations with pharma companies over the years — they include AstraZeneca, Roche, Takeda, Merck and Janssen.

While all three companies are in different stages of commercial growth and proving out their technologies, they are all united by the fact that over time, animal testing for drug development will be used less and less often given the path the FDA and the National Institutes of Health are on. And that’s a path that seems to be unaffected by which party is in charge or even the turnover of top ranking personnel at the FDA, like Commissioner Marty Makary.

This started in the Biden administration with the FDA Modernization Act 2.0 getting approved, Corbett said. “Remember, the two senators pushing that was bipartisan” he said in reference to Democratic Senator Cory Booker from New Jersey and Republican Senator Rand Paul from Kentucky.

“So when you have bipartisan approach on this thing, I think there’s continued momentum,” he said.

Photo: ClaudioVentrella, Getty Images

Correction: An earlier version of the story incorrectly stated the timeframe of Revalia Bio’s founding and the fundraising. The company was founded in the early 2020s and not early 2000s. They have a raised a seed round not a Series A.