Skip to Main Content

NEW HAVEN, Conn. — The idea was born over beer in a hotel bar.

It would spark a revolutionary approach to treating cancers, attract tens of millions in investments, and set rival companies on a scientific race. At the center of it all? A tiny garbage truck.


It’s called the proteasome, and its job is to chew up and get rid of your cells’ obsolete and broken proteins. When biochemists Craig Crews and Raymond Deshaies discussed it over brews at a conference in 1998, the proteasome had been described but not domesticated.

They knew that it naturally destroys proteins the cell no longer needs. But could they hijack that machinery to treat diseases?

Crews has built a career on that question. A 51-year-old professor at Yale, he has spent the last 20 years fiddling with cellular garbage trucks. “I was always a tinkerer,” he said — but “tinkering” doesn’t exactly capture the complexity of his work.


He has already taken one molecule given up on by a pharma giant, and with help from colleagues, turned it into a successful cancer drug. Now, using the proteasome, he aims to revisit hordes of other drug discovery projects that have stalled or failed. And he’s founded a company, called Arvinas, built around the premise that harnessing the proteasome will open the door to new treatments for diseases that are currently considered “undruggable.”

He isn’t the only one. “Every big pharmaceutical company in the world is thinking about this area,” said Andrew Phillips, the chief scientific officer of C4 Therapeutics, a biotech startup in Cambridge, Mass., that’s working on the same question. “There is a lot of excitement.”

Stumbling on a garbage truck

Crews was a new professor at Yale, sniffing around for projects, when he read about a strange molecule produced by certain soil-dwelling bacteria.

Scientists at the pharmaceutical company Bristol-Myers Squibb had found that it was startlingly good at killing melanoma cells. But they couldn’t figure out why. In 1992, they broke their secretive habits, publishing a detailed report on everything they knew about the substance.

“This had a particular appeal to me, because it had such potency, but there was this mystery about it,” said Crews. “They didn’t know how it worked. All they knew was that it would kill tumor cells. And that to me is this wonderful invitation. I was curious.”

Even as a kid, he’d been drawn to problems that seemed unsolvable. His father was a NASA engineer who researched lightweight materials for aircraft wings. He would bring home valves and levers that his lab could no longer use. Crews kept them in a box in the garage along with old car parts, as raw materials for robot building. It went with his loose-leaf notebook of inventions — including a design for a perpetual motion machine.

He’d had glimpses inside his father’s lab, and there was never a moment when he didn’t want to be a scientist. “NASA in the ’60s and ’70s was the most exciting place to be a researcher, and I would argue that the biotech arena today is the equivalent,” he said.

Unable to access the mysterious molecule described by Bristol-Myers Squibb, Crews had his team make their own version in the lab.

Then, they took a string of minuscule plastic beads, each about the size of a grain of sand, coated them with the molecule they were interested in — and went fishing. They poured the contents of a cell over the molecule-slathered beads. Only one thing stuck: a cylinder of proteins called the proteasome.

It turned out that the mystery molecule blocked the proteasome from doing its job, namely disposing of junk proteins in the cell. “By gumming it, by blocking the action of the proteasome, you get buildup of toxic proteins that should have been removed,” explained Crews.

This affected all cells. But it affected cancer cells more than healthy cells, because cancer cells proliferate uncontrollably, constantly making and discarding proteins. So with the proteasome jammed, toxic levels of old proteins built up quicker in the cancer cells. And then the cells died.

“It’s coming in from a totally different direction that no one else has come in from before.”

Derek Lowe, drug discovery researcher

To Crews, that sounded like more than basic science. It sounded like a drug.

So he and Deshaies started a company called Proteolix. With a few tweaks, that molecule was turned into a drug for multiple myeloma, approved by the FDA in 2012.

And if money is any measure, that drug — called Kyprolis — was a success. Proteolix was sold to Onyx Pharmaceuticals in a deal worth $850 million in 2009, and then Onyx was sold to Amgen for around $10 billion in 2013.

A seek-and-destroy mission

Stopping the garbage truck from working was one promising path to treatment.

Crews, however, had already been wondering about the opposite. He wanted to hijack the proteasome so it gobbled up the proteins that help cause disease.

With genomic analysis, there has been an explosion in our knowledge of the proteins that are involved in diseases. They do their damaging work by docking to various structures in the cell, causing chemical reactions.

Many, many drugs — what Crews calls “traditional” drugs — try to stop these rogue proteins by blocking up their docking ports.

That approach has saved a lot of lives. But it can also have some serious problems. The molecules that block the docking sites fall off after a while, so you need to keep flooding the cell with drug molecules to keep the docking sites plugged. And if the drug is toxic, as many cancer drugs are, that means patients need to have a whole lot of toxic molecules floating around in their cells.

What Crews and Deshaies first discussed in 1998 was a different tack completely. They wanted to tag the nasty proteins that cause disease, the way a city puts pink ribbons around trees that need cutting down.

“You can imagine a small molecule, a drug, that works under this new paradigm, will truly be one that can seek and destroy rogue, disease-causing proteins.”

Craig Crews, Biochemist at Yale

The proteasome would notice the tags, and do what it does best: “taking the protein that’s been tagged, unwinding it, threading it in, and chewing it up,” said Crews.

The niftiest part of the idea: The chain of molecules that tags harmful proteins, and feeds them to the cellular garbage disposal, does not get chewed up. Instead, it’s freed to go off and hunt for more proteins.

“You can imagine a small molecule, a drug, that works under this new paradigm, will truly be one that can seek and destroy rogue, disease-causing proteins, rather than simply binding and falling off, binding and falling off,” said Crews. “So you don’t need as much drug. It gets the job done, first time around.”

It also could mean that disease-causing proteins without active docking sites — such as those involved in Alzheimer’s — could potentially be “druggable” for the first time.

They called this chain of molecules PROTACs.

A race to develop drugs

Even getting these large molecule-chains across the greasy, slippery cell membrane was a challenge; at first, a postdoctoral scientist in Deshaies’ lab at the California Institute of Technology was microinjecting them inside cells by hand.

But in 2009, Crews decided to use smaller molecules as one link in the chain, rather than larger sequences of amino acids called peptides. And his lab has continued to tweak these tiny molecules so that they are better at nabbing harmful proteins and flagging down the garbage truck.

“As far as tagging the protein itself and dragging it off to the shredder, no one has ever tried to treat disease like that,” said Derek Lowe, a longtime drug discovery researcher for pharmaceutical companies. “It’s coming in from a totally different direction that no one else has come in from before.”

Lowe still isn’t sure if it will work in humans, but he said “it really has a lot of promise.”

That promise was reinforced in 2010, when a Japanese team published a paper showing that an old drug, in use since 1957 but never fully understood, worked by tagging proteins so that they’d be destroyed by the proteasome. That drug was thalidomide, and it was widely prescribed to pregnant women to prevent nausea, until it was found to cause birth defects.

Crews found the news exciting: It proved the approach could work. In 2012, he founded Arvinas, to try to capitalize on the promise of the proteasome. It’s now racing against competitor C4, which was unveiled earlier this year, to develop possible cancer treatments. C4 calls its molecular chains “degronomids,” but the idea is the same.

Both companies emerged from basic science labs, and both have big-name partnerships: Arvinas with Merck and Genentech, and C4 with Roche.

Crews hopes that within a year, Arvinas will be testing the idea he hashed out over beer two decades ago. All he needs are patients willing to have their cellular garbage trucks hijacked.