快猫短视频

IT WAS a meeting that might transform medicine. In 1998, biochemists Craig Crews and Raymond Deshaies starting chatting at a conference at the Semiahmoo Resort on the picturesque coast of Washington state. 鈥淎fter a few beers, we started doing some wishful thinking,鈥� says Crews. 鈥淎nd we came up with something new.鈥� Their idea was to create a new kind of drug 鈥� one that could target and destroy proteins by hijacking cells鈥� own machinery.

鈥淎fter a few beers we came up with something new. If it works, it will be revolutionary鈥�

Deshaies started writing a research proposal almost as soon as he got back to his lab at the California Institute of Technology in Pasadena. Now, more than seven years on, he and Crews 鈥� of Yale University 鈥� are getting tantalisingly close to achieving their goal. 鈥淚f this works,鈥� says Deshaies, 鈥渋t will be completely revolutionary.鈥�

Almost all drugs work by interfering with proteins. Some bind to and block the active site on an enzyme, others prevent signalling molecules from binding to a receptor. They all consist of small molecules that can easily get into the bloodstream and then into cells.

But the active site of a protein is a tiny target, so finding small molecules that bind to it and it alone is not easy. And cancers, viruses and bacteria often mutate and become resistant to drugs. Worse still, a huge number of proteins do not work in a way that is easy to inhibit 鈥� many interact via domains far larger than an active site, which drugs are too small to interfere with. 鈥淭here are so many proteins out there that don鈥檛 have enzymatic activities and aren鈥檛 traditionally druggable,鈥� says Crews.

The bottom line is that there would be huge advantages if, instead of trying to block the activity of a problem protein, we could just get rid of it. This can be done in animals by resorting to genetic engineering, for instance, but what researchers have long sought is a less drastic way of achieving it in people.

Many are now excited about one promising method, called RNA interference (RNAi). Adding the right bits of RNA to a cell can trigger a built-in mechanism that blocks the production of a protein. The problem with RNAi is delivery. The snippets of RNA are still a lot larger than conventional drugs and getting them inside cells is a challenge that has yet to be overcome. 鈥淚鈥檓 sceptical about their ability to translate it into a therapy,鈥� says Deshaies.

What he and Crews dreamed up back in 1998 was a way to create small-molecule drugs that are easily delivered yet offer the same therapeutic potential as RNAi. In fact, in some ways their method is even more powerful. RNAi prevents all forms of a protein being made. Crews and Deshaies will, if they succeed, be able to destroy one form of a protein while leaving others untouched.

As with so many brilliant ideas, evolution got there first. The human papillomavirus (HPV) makes cells grow madly 鈥� occasionally causing cervical cancer 鈥� partly by destroying a key regulatory protein. Well over a decade ago, Peter Howley of Harvard Medical School showed that HPV destroys it by hijacking the cellular machinery that chews up proteins.

This dirty work is done by cellular recycling factories called proteasomes. These complexes cut up proteins 鈥� which are chains of hundreds of amino acids 鈥� into small peptides just eight or so amino acids long. The peptides are then broken down into individual amino acids, the building blocks for new proteins.

The brains of the recycling operation are enzymes called E3s. There are several hundred different E3 enzymes, each of which targets a specific protein or class of proteins. Once activated, E3s wander about the cell attaching 鈥渞ecycle me鈥� labels to their target protein in the form of a chain of ubiquitin molecules (see Diagram). The proteasomes then dismantle any appropriately labelled proteins.

While the ubiquitin-proteasome system (UPS) does clear away damaged or misfolded proteins, it is much, much more than a waste-disposal system. Most of the proteins it breaks down are in perfect working order. What the cell gets in return for this apparent wastefulness is a way of regulating processes with exquisite precision. Many proteins play a vital role under certain conditions or at particular stages of development but become a hindrance when the cell鈥檚 activities change 鈥� at which point the UPS disposes of them.

HPV hijacks the UPS by producing a protein called E6, which binds both to the regulatory protein that the virus knocks out and to an E3 enzyme. This tricks the E3 into adding the ubiquitin 鈥渞ecycle me鈥� label to the regulatory protein, condemning it to destruction.

If a virus could hijack the UPS and get it to degrade a protein it would not normally touch, Howley realised, maybe we could too. 鈥淎t that point,鈥� he recalls, 鈥渨e became interested in targeted degradation.鈥�

鈥淭argeted degradation has immense potential for treating disease鈥�

Initially, Howley鈥檚 team tried to alter the virus鈥檚 E6 to make it target other proteins. That worked, but the altered E6s were unstable. In 1997 Pengbo Zhou, then a postdoc in Howley鈥檚 lab, tried a slightly different tack. He knew that in one class of E3 enzymes, a subunit called an F-box binds to the E3鈥檚 target protein. So he tinkered with the DNA coding for the F-box to make the F-box bind to a different target protein. When the modified gene is added to cells, E3s with the new F-box subunit label the desired target protein for destruction by the UPS.

Now at Cornell University in New York, Zhou has successfully used this system to target over half a dozen proteins for degradation. He has also shown that by adjusting the amount of a modified F-box that a cell produces, he can fine-tune the level of the target protein rather than destroying it all. This is an advantage over RNAi, which tends to be all or nothing.

Even more significantly, Zhou can target specific forms of a protein. Almost all proteins have extra bits, such as sugars, tacked onto them. These modifications can have profound impacts on where the proteins go and what they do. Some cancer cells produce modified proteins that promote growth, for instance.

Zhou has created a modified F-box that leaves a protein with an added phosphate group untouched, but triggers the destruction of the unmodified form. Another of his F-boxes targets a form of a protein that floats free inside cells while leaving the membrane-bound version intact.

For conditions where one form of a protein causes disease while other forms are vital for normal functioning, being able to target only the disease-causing form would be a huge advantage. Again, this cannot be done with RNAi, since it blocks protein production long before different forms develop.

Zhou鈥檚 work proves that targeted protein degradation has immense potential for treating a huge range of diseases. But his approach relies on delivering the DNA coding for a modified F-box to a cell. 鈥淚t has all the problems of a gene therapy approach,鈥� says Howley. So while Zhou鈥檚 method is a powerful tool for studying proteins, it is unlikely to be used for treating people any time soon.

Around the time that Zhou constructed his first modified F-box, Crews and Deshaies met at Semiahmoo. At the time, they were unaware of his work. But like Zhou and Howley, they were inspired by HPV to try to hack into the UPS. Their idea, though, was to create a drug of two halves: take one small molecule that binds to an E3, join it to another small molecule that binds to a target protein and the result, they reasoned, would be a molecule that labels the target protein for destruction, yet should still be small enough to slip through the cell membrane. A drug, in other words, that could be swallowed like a painkiller and provide the power of RNAi.

For their initial proteolysis-targeting chimeric molecule, or protac, Crews and Deshaies chose ovalicin 鈥� which binds tightly and specifically to a protein called MetAP-2 鈥� as the target-protein-binding part. The E3-binding part proved more problematic. They could not find a small molecule to do the job, so they ended up using a peptide, a tiny protein 10 amino acids long. This meant the protac could not pass through the cell membrane after all, but tests on cells that had been broken open showed that it did trigger the destruction of MetAP-2 by the UPS. The concept was sound.

With their next protac, Crews and Deshaies decided to target the receptor that responds to the male hormone androgen. Androgen encourages the growth of some prostate cancers, so one treatment strategy is to inhibit its receptor. In cancer patients the receptor frequently mutates and becomes resistant to drugs, so it is a good target for a protac. Crews and Deshaies attached the 10-amino-acid peptide to dihydroxytestosterone, a small molecule that binds to the androgen receptor. This time, they injected the protac into cells. Within an hour, almost all the androgen receptors were destroyed.

But microinjection into cells is not going to cut it as a drug-delivery method. So Crews and Deshaies tinkered with the peptide part to enable the protac to get through cell membranes on its own. It worked 鈥� but it still isn鈥檛 enough. 鈥淚t鈥檚 a great idea,鈥� says Neal Rosen of the Memorial Sloan-Kettering Cancer Center in New York. 鈥淭he problem is they don鈥檛 have a drug.鈥� For starters, drugs containing peptides are hard to manufacture. More importantly, any stray peptides in the bloodstream are rapidly destroyed. 鈥淭he body is very good at getting rid of unnatural molecules,鈥� points out chemist Kevan Shokat of the University of California, San Francisco.

But in recent months Crews has finally come up a small molecule that can replace the peptide as the E3-binding part. He did this by modifying a chemical called nutlin-3 so it can be joined to the target-protein-binding half of a protac. If this latest generation of protacs works, Crews and Deshaies will be well on their way to producing a real drug that could be tested in animals and humans. 鈥淭hey are, I鈥檇 say, reasonably close,鈥� says Shokat. Even if nutlin does not work, it should be possible to find other small molecules that bind to E3s.

A host of new drugs

There are still other many obstacles to overcome, though. 鈥淭he biggest challenge is reducing the size of the molecules,鈥� says Shokat. Because protacs have two parts with distinct functions, they are always going to be relatively large. And the bigger the molecule, the less likely it is to work as a drug.

Then there is safety. 鈥淭he potential toxicity issue should be rigorously addressed,鈥� says Zhou. Because they hijack E3 enzymes, he says, protacs could allow proteins to survive when they should be destroyed. But this might not be too serious: there is an anti-cancer drug called Velcade that blocks the entire UPS, by inhibiting proteasomes, yet it is far less toxic than expected, given its lack of specificity.

The opposite effect could also be a problem. Depending on how specifically they bind to their target proteins, protacs might mark for destruction proteins other than the intended targets. Here too there is an encouraging precedent, in the form of an experimental drug called 17-AAG. It causes a large number of proteins to be destroyed by the UPS because it disrupts protein folding. 鈥淲e thought this drug had to be toxic, given what it was doing,鈥� recalls Rosen. In fact, it is surprisingly non-toxic.

There is still a long way to go before anyone pops the first protac. But if they do turn out to be safe and effective, protacs could completely change the way we think about drug design. At the moment, finding a new drug usually involves screening tens of thousands of molecules to see if any happen to affect a protein鈥檚 activity, or working out the structure of the protein鈥檚 active site and trying to design a drug that blocks that site. Neither is easy.

To create a protac, by contrast, the first step is to find a small molecule that binds somewhere 鈥� anywhere 鈥� on a protein. 鈥淲e are inhibiting proteins just by destroying them, independent of their activity,鈥� says Deshaies. It should not matter at all what part of a target protein a protac binds to. It鈥檚 an important advantage, says Shokat, especially since many families of proteins have very similar active sites.

Admittedly, protacs will never be as easy to design as the bits of RNA used to trigger RNAi, where all you need is part of the gene sequence coding for the protein you want to destroy. And the tricky bit will be joining the protein-binding part to the E3-binding part. But with lots of candidates to choose from, this should always be possible. Crews points out that the first step in drug discovery is often to find molecules that bind to a target protein. 鈥淭his is exactly what pharmaceutical companies have been doing for decades,鈥� he says. An untold number of such molecules have been found but shelved because they do not inhibit a protein鈥檚 activity. The companies are sitting on a gold mine of molecules that could form half of a protac 鈥� if they can be persuaded to share. 鈥淭hese are their lemons,鈥� Crews says. 鈥淚鈥檓 offering to make lemonade.鈥�