In chemical genetics, a new strategy could speed drug discovery

HTS allows researchers to test thousands of small drug-like molecules at once for a specific biological activity, such as inhibiting the cell movements that allow cancer cells to spread in the body. Screening for potential new drug compounds in complex systems differs from the traditional drug discovery approach, which begins with one particular protein of interest and tries to find inhibitors for that specific target.

By starting with many small molecules and screening for a certain biological effect, researchers rapidly pinpoint drug candidates. But that’s where chemical genetics hits a wall. Once researchers find an inhibitor, or potential drug, they want to know how it works, by identifying the protein on which the inhibitor acts. But how?

“This is a significant problem for the entire field,” Peterson says. “What particular protein is the small molecule hitting to get that biological effect?” In this case, Peterson and his colleagues identified an inhibitor called pirl1, which, when added to cell extracts, blocks the assembly of actin fibers used for cell motility.

The methods scientists typically use to find protein targets only work in certain cases. For instance, if an inhibitor binds tightly to its target, then researchers could use that to pull both the inhibitor and its target protein back out of the cell extract. Pirl1, however, like many small molecule inhibitors, binds to its target only weakly.

Or, if the researchers knew most of the proteins involved in cell motility, they could add back each pure protein individually to the mix until actin fiber assembly was restored. But this method is time-consuming and only works if the protein players in a biological system are already known.

“This forced us to think outside the box. We didn’t have a lightening strike out of the blue, but we were inspired by classical genetics,” Peterson says.

In classical genetics, a technique called genetic suppression screening is used to identify an unknown gene that causes a certain biological effect. Imagine you have a taste for a particular ice cream flavor, but you can’t remember the name. So you decide to taste all 32 flavors at the ice cream shop, one at a time, until you find the flavor that gives you that delicious hit. Genetic suppression works something like this.

Peterson and his colleagues adapted this approach, creating a technique called biochemical suppression. The team tested batches of separated cell extract proteins to see if they could add back the one protein being inhibited, and therefore restore actin fiber assembly. Think of starting with triple ice cream cones–instead of 32 single cones–to speed the process of identification. If a triple cone has the tasty hit, you can then test each of the three flavors in that cone individually to see which one is the key.

Similarly, when a batch of proteins restored the activity, that batch was separated into even smaller batches, until finally, the researchers narrowed the search down to two possible target protein complexes, Cdc42/RhoGDI and Arp2/3. In the test tube, pirl1 did not directly inhibit Arp2/3, but it did inhibit Cdc42/RhoGDI. It turns out that Arp2/3 acts downstream of Cdc42/RhoGDI in the pathway that assembles the actin fibers. This shows an advantage of the biochemical suppression method–discovering other proteins in the same biological pathway.

Also, biochemical suppression allows researchers to identify the target even if it is composed of a more than one protein. Going back to ice cream, imagine that the perfect flavor is a mix of vanilla AND chocolate. If you tested all possible flavors individually you would never get a “hit.” In fact with pirl1, the target turned out to be a complex of two proteins, Cdc42 and RhoGDI. These advantages make this technique a powerful way to identify multiple components of a complex biological system, such as the spread of cancer cells.

“In these small molecule inhibitor screens, we are letting the biology tell us which proteins are going to ’druggable’,” Peterson notes. Unlike traditional drug development where a target is studied intensely, or validated, before the search for inhibitors of that specific target begins, phenotypic screens with multiple potential targets are faster out of the starting gates. But before now, they had faced a roadblock in discovering the protein target’s identity.

“This biochemical suppression technique is a new tool that makes it feasible to do these screens because now you have a method to identify the target at the end,” Peterson says.

Testing pirl1, Peterson and his collaborators placed the inhibitor on monkey kidney cells, where it stifled cell membrane ruffling, a process related to cell motility. Next, they hope to determine whether pirl1, or chemicals with similar structure and activity, will work similarly on other cell types and in animal models of cancer.