- Gilead is tapping a California startup’s machine learning technology to find new drugs for NASH, a fatty liver disease that has garnered increased attention and investment from some of the industry’s largest companies.
- Through a three-year research collaboration, Gilead and Insitro will deploy the latter’s tech platform, which combines functional genomics, human genetics and machine learning, to better understand the different ways in which NASH, short for non-alcoholic steatohepatitis, progresses and regresses. The companies will then use that information to discover disease targets and predict patient responses to potential therapies.
- Deal terms hold that Gilead can advance up to five targets. Once those targets are identified, Gilead would be in charge of their chemistry and development. Insitro, however, carries an option to co-develop and co-commercialize them in the U.S. The smaller biotech gets $15 million upfront and could take home an additional $235 million in milestones, including a near-term $35 million payment based on operational achievements.
Gilead has been on a tear, trying to construct a deeper NASH pipeline as prospects dim for its lead candidate in the space.
Phase 3 data released in February showed the candidate, selonsertib, didn’t significantly outperform placebo in treating NASH patients who developed compensated cirrhosis because of severe liver scarring. While another trial evaluating the drug in less sick patients has yet to read out, Wall Street analysts aren’t holding their breath for a positive outcome.
With its leading position in NASH at risk, Gilead has turned to deals to create a kind of pipeline cushion.
The big biotech last week inked a research collaboration that will test a couple of its mid-stage NASH drugs with Novo Nordisk’s semaglutide, a GLP-1 agonist and the active ingredient in the Danish drugmaker’s diabetes medication Ozempic.
And in its latest move, Gilead is looking to find new therapies using Insitro’s tech.
Insitro CEO Daphne Koller explained to BioPharma Dive that her company’s platform works by analyzing human genetic data to create in vitro disease models. Functional genomics are applied to those models to determine the affects of different alleles or disease modifiers. Then comes the machine learning.
“The machine learning is used to relate what we see in a dish to what you would want to make predictions about, which is the human in vivo phenotypes, because ultimately that’s what we’re trying to modulate,” she said in an interview.
Koller noted that deal talks between Insitro and Gilead began about a year ago and were spurred by the fact that Gilead held considerable clinical data on NASH. Another benefit of NASH is that its various disease pathways have become relatively tractable compared to other therapeutic areas, according to Koller.
“You can’t necessarily differentiate stem cells into every possible lineage with high efficiency today,” she said, “but some of the recent papers just in the last year have really shown how one can create a NASH disease phenotype in an in vitro system.”
Insitro is backed by well-known life sciences investors including Arch Ventures, Third Rock Ventures and AZ16.