dermatology unit for one of their businesses about 10 years ago, and met with Narain, at the time the head of cutaneous oncology and therapeutics research at the University of Miami Miller School of Medicine. The investors ended up licensing an experimental cancer drug based on Narain’s work in Miami, and formed a company called Cytotech Labs around it in 2006.
Narain had advised Cytotech for two years when Carl Berg decided it was time to “ramp up” the company. He convinced Narain to join full-time, and the Miami scientist relocated to Boston where the R&D operations of the newly-named Berg started to take shape.
The company has grown significantly from there. It’s filled out its executive team, staffed up to nearly 200 employees, and added marquee names to its group of advisors, like Schadt and Eric Nestler, the chairman of Mount Sinai’s neuroscience department. (Even actor Stanley Tucci is listed as a “senior advisor of patient advocacy.”) Berg has effectively turned itself into a drug discovery and biomarker production shop equipped with various Big Data tools. Narain says the idea at Berg is to “flip” the model of drug discovery by starting out with detailed knowledge of biology, understanding the many pathways in cells and how they change when disease strikes. Then the company uses sophisticated computer tools to figure out how those pathways should be changed—and what type of drug could do. That’s in contrast with the old traditional drug discovery approach of forming a hypothesis about a target, screening for compounds that can hit it, and going from there.
Say Berg is looking at cancer, for example. The company gathers biological samples—blood, tumor tissue, or urine—from a diverse group of people who have cancer, while also getting samples of their healthy tissue (or from other healthy donors). Berg creates cell lines from those tissue samples, and subjects them to various different environments—like the low oxygen, high glucose habitats cancer likes to live in—to mimic the certain disease state the person is suffering from.
Once those cell lines are established, Berg identifies the genes, proteins, metabolites, and lipids that are in them, and generates trillions of data points that come from both the diseased and healthy cells, Narain says.
“We’re the only company that’s integrating all of those ‘omics’ internally, “ Narain says. “The rest of the industry is kind of concentrated on the genomics. I feel that you’ve got to go much deeper than the genomics.”
Berg then puts all of that data into a computer system, which creates what looks like an “airline map,” consisting of big hubs and various lines going in and out of them. The big hubs on that map, say, the New York or Atlanta, represent the key proteins that are causing the biggest difference between health and disease—either because they are in short supply, or because they need to be silenced. Once it has pinpointed these proteins, Berg can engineer replacement versions of the missing ones, or treatments using technology like RNA interference to target the disease-causing ones. Meanwhile, the lines going in and out of each hub, like planes flying to and from the airport, become potential biomarkers for a disease, Narain says.
Berg says that this method cuts out the entire process of screening thousands of chemicals to see which ones might work as drugs, which can take several years and hundreds of millions of dollars. As a result, Narain says, the company can go from “zero to target” in 18 to 24 months, and to a drug in half the time it usually takes.
“I was very excited about a company that actually started with systems biology and this more holistic integration of biology and computation as their driving force. I think it’s very unusual, you’d be hard pressed to find another company like it,” Schadt says. “It’s still a bit of a fantasy to think we can generate all of the types of data needed, but they’re definitely generating high impact data that’s exactly the type that needs to be generated.”
But regardless of the secret sauce it’s whipped up, Berg is facing its fair share of skeptics. Berg is far from the first to try to use computer-generated models to discover drugs—and none of those methods as of yet have really caused a sea change in drug development. Merrimack Pharmaceuticals (NASDAQ: [[ticker:MACK]]), for instance, combines a high-density protein array—a way to capture how proteins interact with one another—with interactive computer models that researchers can use to figure out what target to attack, or what type of drug they should add, as part of a supposedly capital-efficient way to design drugs. It doesn’t have any FDA approved drugs yet, though the company does have a few candidates in mid- and late-stage studies. Boston’s Nimbus Discovery has a different method, but is early on in its journey—its drug candidates are in preclinical testing.
The VC I spoke with likens Berg’s approach to
