On the hunt for monomeric degraders

Lori Friedman was sitting in her office at Genentech when her colleague Kyung Song walked in with a Western blot gel and a chemical biology conundrum. “I can’t find where the protein is going. It’s not in the membrane fraction. It’s not in the nuclear or cytosolic fraction. I don’t know where it is,” Song said. Most small-molecule drugs act by blocking a protein’s activity, but Song was dissecting cells to see whether her team’s lead molecules instead changed where the protein clustered in the cell. These gels opened a more interesting possibility.

Friedman, who is now chief scientific officer of Oric Pharmaceuticals, realized her team had accidentally stumbled onto a set of degraders. “This is fantastic. It’s disappearing,” she recalls telling Song. So began the hunt to figure out how Genentech’s inavolisib and related anticancer compounds lead to the breakdown of the PI3Kα enzyme, which adds phosphate groups to lipids.

A class of small molecules called targeted protein degraders is already all the rage in the pharma industry. Large pharmaceutical firms and biotechs are exploring various approaches to destroy, rather than just inhibit, proteins of interest. But drug developers are focusing mostly on big, bulky, heterobifunctional molecules that bind to a protein with one arm and the cell’s trash disposal machinery with the other. These large molecules can be tricky to work with, sometimes refusing to dissolve in solution or to cross cellular membranes, for example.

Genentech’s research, just published in Cancer Discovery, is a reminder that smaller molecules called monomeric degraders can achieve similar effects through different medicinal chemistry paths.

“This is pretty intriguing work,” says Georg Winter, a chemical biologist at the CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences. “Monomeric degraders are not new. But they might be more common than we had perceived.”

Cancer researchers have long been intrigued by phosphatidylinositol 3-kinases (PI3K), a family of proteins that plays a role in cell growth, proliferation, and survival. But attempts to develop drugs that block individual members have been complicated by complex downstream biology, including compensatory signaling that kicks in when one pathway is shut down. Compounds that can discriminate between related PI3K proteins are also rare.

Genentech, for example, put a drug candidate called taselisib all the way through Phase 3 clinical trials but abandoned it because of its modest clinical benefit and considerable side effects—likely a result of its activity on more than one PI3K family member.

In 2019, Novartis secured US Food and Drug Administration approval for alpelisib, the first drug to inhibit just the PI3Kα isoform of the protein. Genentech is close behind, with a taselisib-related PI3Kα inhibitor called inavolisib in Phase 3 studies.

But while working with taselisib and inavolisib, Friedman and her colleagues realized that these drugs’ mechanism of action might go beyond simply blocking the protein’s activity. Perhaps, she thought, these small molecules were preferentially pushing just the mutant PI3Kα out of the cell membrane. After Song’s gels showed that the drugs depleted PI3Kα in the membrane, the cytoplasm, and the nucleus, the researchers revised that hypothesis.

The mechanism they unraveled sheds new light on PI3Kα biology. Cancer-causing mutations destabilize PI3Kα, speeding up its journey to the garbage bin. Taselisib and inavolisib further rattle mutant PI3Kα, making it even more vulnerable to the cell’s protein-disposal machinery. These molecules therefore have a dual mechanism of action: they block PI3Kα signaling from both the wild-type and mutant protein, and induce the degradation of the mutant protein.

“I think it’s fantastic,” says Bart Vanhaesebroeck, a PI3K trailblazer at University College London. He has tracked Friedman and colleagues’ work since they first started presenting it at conferences in 2017. “They have opened up a completely new field.”

With new mechanisms of action come new therapeutic opportunities.

Because inavolisib degrades mutant PI3Kα while sparing wild-type protein levels, it could offer fewer side effects and more lasting activity in people with PI3Kα-mutated cancers. This improved safety and efficacy profile “provides us with an opportunity to go into previously inaccessible combinations,” says Genentech’s Anwesha Dey, director and principal scientist of discovery oncology, who now leads the research side of this program.

Genentech also found that this inavolisib-induced degradation relies on the activity of HER2 or related receptors at the cell membrane. This means the drug might work well in HER2-positive PI3Kα-mutated breast cancers, a type of cancer that has not yet been effectively treated with PI3Kα inhibitors.

In Genentech’s hands, Novartis’s alpelisib was not able to break down PI3Kα. But Novartis reported at the 2018 American Association for Cancer Research meeting that alpelisib induces PI3Kα degradation in some tumor models. It is also testing the drug in HER2-positive cancers.

Inavolisib is far from the first monomeric degrader. Selective estrogen receptor degraders and selective androgen receptor degraders are decades-old target destabilizers and still a keen focus of research. Thalidomide, first approved by the FDA in 1998, found new life in 2013 as a molecular glue when researchers discovered that it and its derivatives work by sticking various targets to the E3 ubiquitin ligase component of the cell’s waste-management system.

But for Steve Staben, who was Genentech’s research team leader for the discovery of inavolisib, the PI3Kα work underscores the need to keep an open medicinal chemistry mind. After stumbling onto inavolisib’s unexpected capabilities, he reviewed the literature and found 39 underappreciated monomeric degraders.

Monomeric degraders are smaller and easier to work with than heterobifunctional degraders, his analysis showed.

“The goal of that perspective was as a call to action for drug discovery scientists, that we should be looking for this [degrader] mechanism more frequently,” says Staben, who is now chief scientific officer at Lycia Therapeutics.

There are always trade-offs. With monomeric degraders, the specifics of degradation may not be widely transferable to other protein targets. How often do mutations destabilize a protein enough that a small molecule can then push it over the edge and into the trash can, for example? How do you screen for this activity?

These questions provide all the more reason to dig in, Staben says. “The more we can understand different types of mechanisms that promote degradation, the closer we’ll get to understanding which targets will be susceptible,” he says.

And new search strategies are emerging. Last year, CeMM’s Winter showcased one approach for identifying nonobvious glue degraders. By screening many anticancer compounds in wild-type and in E3 ligase–deficient cells in parallel, his team set out to find small molecules that only have activity when the trash-disposal system is working. This led to the identification of small molecule degraders of the protein cyclin K, he reported in Nature Chemical Biology.

Winter cofounded the biotech firm Proxygen to develop this screening platform to discover glue degraders.

“The jury is out as to how often we will really see monomeric degraders. But a lot of labs and companies now take the bet that this will probably happen more often than not,” Winter says.

Drug hunters can also keep taking an opportunistic approach to discovering simpler degraders by more conscientiously tracking small molecules’ effects on protein levels.

For Friedman, inavolisib changed her approach to drug discovery. She instantly started having Western blots done for every single target she worked on to see if they were being degraded.

That mentality has caught on at Genentech as well. “There are enough examples of this out there that I think we should always keep an eye on this,” says Gina Xiaojing Wang, director and senior principal scientist of discovery chemistry at Genentech.

These mechanism-of-action analyses are easier said than done. Friedman and her colleagues were lucky to be working with cells that overexpressed mutant PI3Kα, or they would have missed the mutant-dependent signal. Look too soon or too late, and the protein levels may appear normal. And figuring out the underlying biology of degradation can take years.

“It starts easy, but it gets harder the closer you look,” Winter cautions.

But this is part of the fun of the protein degradation puzzle. “Proteins have lives, and you have to understand who they interact with and what happens when you interfere with any part of a protein’s life cycle,” Friedman says. “This is a huge area of basic biology that will only help drug discovery.”

Asher Mullard is a freelance writer based in Ottawa, Ontario.