Oct 26, 2026 · 57 min
Human liver models challenge animal testing in drug development
Bioprinting 3D Human Tissue To Discover And Test New Treatments With Viscient Bio's Keith Murphy
More predictive human tissue models could reshape toxicity screening, disease modeling, and the financing and regulatory path for new medicines.
- 1Keith Murphy traces a path from Amgen and Organovo to scaffold-free bioprinted liver tissues for drug discovery.
- 2High-density, multicellular human liver models aim to expose clinically relevant toxicity and disease phenotypes animal studies may miss.
- 3Viscient’s next phase depends on financing, regulatory adoption, and advancing a MASH candidate toward clinical development.
Don't miss
Murphy explains how Organovo’s liver disease models led to discovery of an FXR agonist that Eli Lilly later acquired.
The brief
Keith Murphy recounts moving from MIT and Amgen into biotechnology entrepreneurship, founding Organovo before building new companies around human tissue models.
Organovo’s strategy shifted from engineered blood vessels toward disease modeling and therapeutic discovery, eventually producing an FXR agonist acquired by Eli Lilly.
Murphy explains how scaffold-free, high-density liver tissues combine multiple cell types and organized architecture in scalable formats for drug testing.
The central argument is regulatory as much as scientific: complex human-cell liver screening could improve phase-one decisions while reducing reliance on animal studies.
Murphy closes with Viscient’s clinical priorities, including a phase-two-ready MASH candidate, new liver-disease programs, and the venture funding needed to advance them.
What was said on this episode
22 statements · 15 positive · 6 negative · 1 neutral
Scaffold-free tissues can reduce biological changes induced by tissue-culture plastic.
“if we could make a tissue that actually didn't interact primarily with the plastic, but that all the cells interacted together apart from the plastic, that seemed really powerful to me as a way to get away from the changes to biology that tissue culture plastic can induce.”
Listen at 12:56
Organovo could produce ten bioprinted blood vessels weekly by 2010.
“by 2010, we made, you know, 10 blood vessels a week. You know, we could make 10 blood vessels a week if we wanted to.”
Listen at 16:30
No one has solved small-diameter blood-vessel grafting below the knee.
“no one has solved this issue.”
Listen at 17:43
Disease modeling offers the greatest patient impact and investor return.
“I think that gives you the biggest win for investors as well, right? We're going to have the biggest impact on patients and the biggest win for investors.”
Listen at 24:19
Bioprinted models best represent final tissue biology among their models.
“the bioprinted model is the most complex and the best representation of the final tissue in the biology”
Listen at 29:09
Their technology produced tissue forty cell layers thick.
“we were easily readily making 40 cell air thick tissue.”
Listen at 30:50
Their engineered blood vessels withstood 1,200 mmHg after 21 days.
“these blood vessels would grow. When we made blood vessels, for example, the extracellular matrix grew so that they could withstand, you know. 1200 millimeters of mercury by 21 days.”
Listen at 33:01
A bioprinted liver model contains one million cells.
“The liver. Bio-printed liver, it's three millimeters across, half a millimeter thick on a transwell surface. So that's one million cells.”
Listen at 33:37
Their liver model retains MASH phenotypes from patient cells.
“taking patient cells who have MASH and putting them into that liver model, we could retain the phenotype by handling it the right way and get MASH phenotype in that model and see deposited collagen.”
Listen at 34:22
Their MASH models matched clinical compounds’ antifibrotic effects.
“We matched all the known compounds that have been in the clinic in terms of whether or not they reduced fibrosis in the human. They matched that in our models.”
Listen at 34:38
Vizient synthesized 500 compounds and tested about 200 in spheroid tissues.
“for one of our drugs at Vizient, we synthesized 500 compounds and we probably tested 200 of those in SAR in the actual spheroid tissues”
Listen at 35:48
Primary cells best represent disease biology in their models.
“we focus on using primary cells because we think that best represents the disease”
Listen at 36:41
Chemically induced animal MASH fibrosis is not equivalent to human diet-induced disease.
“And that is not the same.”
Listen at 38:56
Maintaining patient-derived disease phenotypes is easier than inducing disease in animals.
“I think it's an easier challenge to broach. And we've done it successfully.”
Listen at 39:22
Animal models fail to reveal all compounds that are toxic in humans.
“they just don't show everything is toxic that ends up being toxic.”
Listen at 40:13
Animal-model misses can cost $50–$200 million before human liver toxicity is discovered.
“you're spending – if you make a mistake and pick a drug that has rat talk – that looks clean in a rat and other species but ends up being liver toxic in a human, you spend $50 to $200 million before you figure that out.”
Listen at 40:22
Three-day microscopy may predict liver toxicity observed at 21 days.
“the images at three days are going to tell us if that thing's going to end up with liver toxicity showing at 21 days”
Listen at 41:30
The FDA should use their model or a comparable human liver-toxicity model.
“I don't even think there's a case anymore for the FDA to not be using our model or a model like ours.”
Listen at 42:33
Drugs toxic after seven days in their model should not advance.
“if you look toxic at seven days in our model. You probably shouldn't take that drug forward”
Listen at 43:39
Compounds flagged by their model should be excluded from lead selection.
“we just say, hey, two of these have red flags. Don't take those.”
Listen at 43:57
Primary sclerosing cholangitis is an attractive next disease focus.
“primary sclerosing cholangitis is really attractive”
Listen at 54:37
VivoSim’s intestine model can predict drug-induced diarrhea.
“we repurposed the intestine then into VivoSim as a way to predict. diarrhea for drugs.”
Listen at 55:10
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.