A Jeff Bezos-backed startup just inked a deal with $84 billion biotech Gilead to make drugs for a troubling and increasingly common disease
- Pharmaceutical giant Gilead is teaming up with Silicon Valley startup Insitro to develop better drugs for a quietly emerging disease called NASH.
- Short for non-alcoholic steatohepatitis, NASH can lead to liver failure and is estimated to affect some 16 million Americans.
- Insitro is led by the former chief computing officer of Google life-extension spinoff Calico. It has backing from top-notch tech investors like Jeff Bezos and life science firms like Verily.
- The three-year deal could be worth as much as $250 million for Insitro.
- Visit BusinessInsider.com for more stories.
The world's leading drug companies are racing to capitalize on the $35 billion market for a quietly-emerging disease that's believed to affect 16 million Americans.
Called non-alcoholic steatohepatitis, or NASH, the condition can lead to liver failure. It's caused by a buildup of liver fat and is often found in people with diabetes and obesity.
Several potential treatments for the disease have fallen short when tried in people, leading some pharma companies to abandon the space. Not Gilead, the California biotech with a market value of about $84 billion. Instead of giving up on its existing work in NASH, the company is beefing up its efforts in the area.
On Tuesday, Gilead unveiled a partnership with a Silicon Valley startup called Insitro. Insitro is backed by Jeff Bezos and led by Daphne Koller, the former chief computing officer of Google life-extension spinoff Calico.
How Insitro can help Gilead find new drugs
Insitro puts data-computing experts and biologists in the same room to try and boost the speed and efficiency of traditionally tedious drug development work.
In its partnership with Gilead, Insitro will plug nearly a decade's worth of the drug company's clinical trial data in NASH into its system. The two companies will work together for three years, according to the terms of the agreement, and may pursue as many as five new drug target areas.
"We've been trying to tackle this problem for a while," Mani Subramanian, Gilead's senior vice president for liver diseases, told Business Insider. Gilead currently has three NASH drug candidates in advanced stages of research and began pursuing NASH in 2011.
"NASH is a unique challenge. No one's succeeded there." he said. "Why not jump into an area where no one's succeeded?"
Subramanian said that Gilead had been in talks with Insitro since the company launched last year. Gilead considered partnerships with other startups as well, but Insitro stood out as the clear leader in the space, he said.
As part of the agreement, Insitro will get an upfront payment of $15 million, with additional near-term payments of up to $35 million, based on whether the company meets certain operational milestones. In addition, Insitro will be eligible for up to $200 million for meeting other milestones for each of the five targets they pursue.
At its most basic, the partnership will involve comparing big biological datasets from patients with and without various forms of NASH to learn what differentiates them. Researchers will zero-in on hotspots in the data where the disease appears to take a turn, and try to stop or reverse the disease's course with a drug.
For example, "you could imagine taking skin cells from healthy people and sick people and then using machine learning to see what distinguishes the sick cells from the healthy cells," Koller told Business Insider.
"Once you have that information, you can start to look for interventions," she said.
An 'allergic reaction' to AI hype
For years, the drug development process has relied heavily on data from cells and mice, which often fails to translate into real results in people. As a result, it takes roughly a decade and about $2.6 billion to create each new drug.
Recently, machine learning and artificial intelligence have been heralded as one potential solution, leading to the creation of dozens of new drug development startups that claim to use AI in their approach. Koller believes some of the excitement in the area has merit.
"I generally have an allergic reaction to a lot of the hype about AI," Koller said, "But when it comes to interpreting and extracting certain kinds of data, honestly it's exceeded my expectations."
In the field of voice recognition, for example, approaches combining machine learning with large datasets have led to developments like Amazon's Alexa. Like with voice recognition, the field of drug development benefits from the ability to take advantage of enormous sets of data from places like clinical trials.
But traditionally in the life science arena, that data has been separated into two domains: computational science and biology. Unifying them, Koller believes, will allow Insitro to build the scientific datasets that are needed to create better drugs for the first time.
Insitro launched just last year, with $100 million in Series A financing from backers like Andreessen Horowitz and Foresite Capital. More recently, the company has received backing from Jeff Bezos' personal investment fund, Bezos Expeditions, and Alexandria Venture Investments, Two Sigma Ventures, and Verily.
"We're not just looking at the same datasets and trying to squeeze whatever is in there," Koller said. "If you need hundreds of thousands of samples, we can generate those."