It has been nearly impossible to get a good look at Rommie Amaro's favorite protein in action. Called p53, the protein sounds the alarm to kill cells with DNA damage and prevent them from becoming cancerous—one reason why it has been called the "guardian of the genome." But it is big and floppy, a molecular shapeshifter that is hard to follow with standard imaging tools. So Amaro, a computational biologist at the University of California (UC), San Diego, turned to supercomputers. She plugged in new x-ray snapshots of p53 fragments and beefed up her program to make a movie of the quivering activity of each of the protein's 1.6 million atoms over a full microsecond, an eternity on the atomic scale that required about a month of supercomputer time. She watched as four copies of p53 linked up and wrapped themselves around a DNA strand, an essential dance the protein performs before it sends off messages for cellular self-destruction.
Amaro wasn't just interested in the behavior of healthy p53: She wanted to understand the effects of mutations that the gene for p53 is prone to. In dozens of simulations, she and her colleagues tracked how common p53 mutations further destabilize the already floppy protein, distorting it and preventing it from binding to DNA. Some simulations also revealed something else: a fingerhold for a potential drug. Once in a while, a small cleft forms in the mutated protein's core. When Amaro added virtual drug molecules into her models, the compounds lodged in that cleft, stabilizing p53 just enough to allow it to resume its normal functions.
For Amaro and a few other researchers, those computer simulations are an inspiration. "A long-standing dream of cancer biology is to find small molecule drug compounds to restore the activity of p53," Amaro says. "We're very excited about this."
The beauty of these [drugs] is that they are broadly applicable.
Of the nearly 1.7 million people diagnosed with cancer each year in the United States alone, about half have mutated versions of p53—a sign of how important the normal protein is in preventing the disease. It is one of the most intensely studied proteins in science, and a highly sought-after target for drugmakers. But of the dozens of p53 drugs in development, the vast majority simply try to boost levels of healthy p53. And despite decades of effort, none has made it to market.
Amaro's work illustrates how a handful of academic labs and small companies are making progress with a fresh approach to targeting p53: rescuing it when it's sick. They're finding drugs that bind to and prop up copies of mutated p53, restoring its shape and ability to carry out its job. One such drug has already passed an early stage safety trial in humans, and a more advanced clinical trial is now underway in Europe. Other would-be medicines are nearing human tests. If any succeed in the clinic, they could dramatically change the landscape for cancer treatment—and for other diseases that involve misfolded proteins, perhaps even Alzheimer's.
It won't be easy. Restoring normal function to a mutated protein is more difficult than simply blocking a protein, the strategy used by most medical therapies, says Klas Wiman, a tumor cell biologist at the Karolinska Institute in Stockholm. As a result, large drug companies have shied away from the rescue approach and progress has been slow, he says. "It's a little out of the mainstream for big pharma."
The payoff could be big, however. Not only could the strategy treat many kinds of cancer, but just a handful of drugs might be enough, particularly when coupled with chemotherapy drugs that induce the tumor cell damage to which p53 responds. P53 mutations tend to be clustered in the core of the protein, where it binds to DNA, and they have similar effects on its shape. Cell assays and animal studies suggest that drugs that restore p53's activity work with not just one mutant form of the protein, but many, says Alan Fersht, a chemist at the University of Cambridge in the United Kingdom. "The beauty of these things is that they are broadly applicable."
An understanding of p53's seemingly magical powers to suppress tumors didn't emerge until well after the protein's discovery in 1979. Early on it was thought to be an oncogene capable of turning a cell cancerous under some circumstances. Only a decade later was it confirmed to bind to DNA and turn on the expression of other genes aimed at healing cell damage. If that damage is deemed too extensive by other cellular actors that interact with p53, it triggers p53 to launch the call for the cell to commit suicide.
The protein is now known to interact with and control dozens of different genes and proteins, and it helps regulate the cycle of molecular events by which cells grow and reproduce. Because of its outsize importance, its presence in cells is tightly controlled. Another protein, MDM2, latches onto p53 molecules and destroys them, keeping their numbers in check.
But this control mechanism can fail in multiple ways. For starters, when p53 itself is mutated, MDM2 cannot attack it. As a result, the malfunctioning protein builds up in cells unchecked and keeps the remaining healthy p53 from doing its job. Without the genome's guardian on patrol, precancerous cells survive and reproduce. This gives them the opportunity to build up the additional mutations they need to become fully malignant.
Most attempts to target p53 in the cancer fight involve trying to boost its level. One popular approach has been to prevent MDM2, and a relative called MDMX, from reducing p53 levels. The hope has been that doing so might allow some nonmutated p53 to stick around longer and kill damaged cells. "Everyone is working on these [drugs] like crazy," says David Lane, a cancer biologist at the Agency for Science, Technology and Research in Singapore, and a co-discoverer of p53.
A host of cell culture and animal data suggests they should work. However, Lane adds, "the clinical trials have not been as successful as we hoped." For example, an MDM2-targeting drug shrank deep-tissue fat cell tumors in just one out of 20 patients in a phase I safety trial, published in 2012. Blocking MDM2 comes with other potential downsides, Fersht says—not least that the approach can only work if there is still healthy p53 left inside cells. Also, because p53 is involved in so many cellular processes, boosting its level too much can have side effects. In the 2012 trial, for example, eight people had serious side effects, such as a sharp decrease in immune cells called neutrophils, and 14 people had comparatively mild side effects like nausea. Yet many MDM2 blockers remain in human trials, and if they pass muster, they are likely to be the first p53 drugs on the market.
Still, Fersht, Lane, Amaro, and others want to target the heart of the problem: mutant p53. In the early 1990s, in vitro tests on cancer cells by Lane and his colleagues hinted that some compounds could restore mutant p53's normal function. But the drugs weren't always doing what investigators thought. It turned out that one compound, called CP-31398, was indeed triggering cell death—but not by restoring p53. It killed cells by gumming up their DNA.
Later candidates have done a better job. In 1998, for example, Wiman's team screened a library of 2000 compounds from the U.S. National Cancer Institute and found two that appeared to restore mutant p53's ability to kill cancer cells. One, known as MIRA-1, turned out to kill more than cancer cells: It was toxic in mice. But the other, called PRIMA-1, proved more promising. A later study showed that PRIMA-1 breaks down into another compound, abbreviated MQ, and 3 years ago Amaro and her colleagues reported computer modeling results that suggested MQ was binding to the inside of the pocket that forms within a mutated p53 core. Her results also showed that the compound props the protein back into shape, rescuing its function.
Wiman and his colleagues have since come up with a more active version of PRIMA-1, and the Karolinska Institute has spun out a biotech startup called Aprea AB to commercialize the drug. Now called APR-246, in 2012 it made it through the first round of safety trials on patients with a type of blood cancer that shows high rates of mutations in p53. It is now in a phase II clinical trial in women with ovarian cancer, which almost always has p53 mutations. The trial, at centers across Europe, will likely be completed within 2 years, says Lars Abrahmsen, Aprea AB's chief scientific officer in Stockholm.
Side effects are a concern because APR-246 binds to the amino acid cysteine and irrevocably changes it—and cysteines are abundant on numerous other proteins besides p53. But so far, Abrahmsen says, APR-246 has been tolerated well in clinical trials, even in relatively high doses. Wiman suspects that this is because the drug's shape makes it interact primarily with the cysteines in the core of mutated p53. But he and his colleagues are now working to confirm this.
Meanwhile, Fersht's group is making headway with other cysteine-binding compounds. And Amaro's computational studies suggest that the right molecule could lodge temporarily in the mutated p53's cleft, staying in place just long enough for the guardian protein to do its job. Yet unlike APR-246, it would eventually fall off and so would avoid making permanent changes to other proteins in the body, reducing the risk of side effects. In fact, a team led by Amaro, UC Irvine biochemist Peter Kaiser, and UC Irvine computer scientist Rick Lathrop recently used computer modeling to screen more than 1 million different compounds for binding in the cleft. The team found several hundred that might do so and restore the protein to its functional shape; more than 30 did the job when tested in cell culture, although they don't know whether they attached to p53 in the pocket. Several have now been licensed by a biotech startup in San Diego called Actavalon, founded in 2013 by Kaiser, Amaro, and others.
Another biotech startup, called Z53 Therapeutics, in Cleveland, Ohio, is taking aim not at the cleft in mutated p53, but at mutations that knock out a key site for binding zinc ions in p53's core. Without zinc, the protein loses the shape it needs to bind DNA. Drugs designed to shuttle zinc into cells can help restore the activity of the most common of these mutant p53s, lab studies suggest.
At this point, it's unclear which—if any—p53 stabilization strategy will pan out, Lane says. But success could leave a lasting mark on cancer care, by helping hundreds of thousands of patients every year, far more than other genetically targeted therapies. The protein-rescuing strategy could also pave the way for similar medications aimed at restoring other mutant proteins. The strategy is already helping fight cystic fibrosis, where a newly approved drug known as Orkambi helps stabilize the proper shape of proteins that balance the flow of ions in cells lining the lung's airways. And the same approach may eventually help treat patients with other misfolded protein disorders, such as Alzheimer's and Parkinson's.
But perhaps the most far-reaching goal of this approach would be to prevent tumors from ever arising in the first place. Wiman notes that current blood screening techniques can already reveal whether a person is shedding cancer-linked proteins into their blood stream, even before they show signs of having a full-blown tumor.
Someday it may be possible to give people with such warning signs drugs that rescue p53, getting their cellular guardians to snap to attention and wipe out the cancer before it ever gets started. "In the long run," Lane says, "it's a very attractive idea."