A new study from researchers at UCSF and collaborating institutions, published in Nature, introduces a new framework for designing drugs that can bind to two sites on the same disease-causing protein, a discovery that ultimately could lead to more effective and safer therapies for patients.

The work, led by co-first authors Jack W. Stevenson, PhD, and Kevin Lou, PhD, builds on earlier research from the lab of UCSF’s Kevan Shokat, PhD, including the development of “bitopic” inhibitors for mTOR, a protein involved in cancer. The new study extends that concept to other cancer targets, including ABL1 and EGFR, and offers guidelines for designing these two-site molecules more systematically.

Here, Lou, a Fellow at the Quantitative Biosciences Institute in the UCSF School of Pharmacy, explains how targeting two sites at once could make drugs more potent and selective.

Your research uses a different approach to designing drugs. What are you trying to accomplish?

We’re trying to bring another layer of precision to precision medicine.

Traditionally, medicinal chemists design molecules that fit into a specific site on a protein, almost like a lock and key. The problem is that similar proteins can sometimes have similar-looking binding sites, so a drug can end up affecting other proteins that we don’t want it to. That can contribute to side effects and narrow the drug’s therapeutic window, limiting the dose that can be given safely and effectively.

The concept is pretty simple: What if, instead of designing a molecule that grabs onto a protein at just one site, we designed one that grabs onto it at two sites?

Why might binding to two sites be better than binding to just one?

Think about rock climbing. If you’re climbing a wall and you’re holding on with one hand, you have one point of contact. But if you can grab the wall with both hands, you have a stronger grip.

That’s essentially what we’re trying to do with these molecules. We take two molecules that each recognize a different site on the same protein and link them together. The resulting molecule can grab onto the protein at both sites — like using two hands instead of one.

That can give us a molecule that holds onto its target more tightly and, importantly, is more selective for that target.

Why does that matter for cancer drugs?

Cancer cells are very good at developing resistance. If a drug grabs onto a protein at just one site, a small change at that site can sometimes prevent the drug from working.

Going back to the rock-climbing analogy, if you have two hands on the wall and one hand loses its grip, you still have the other hand. In the same way, a change at one binding site may not be enough to make a two-site drug stop working.

We also want these drugs to act on the target we care about while leaving other proteins alone. By requiring the molecule to recognize two sites on the same target, we can potentially make it much more selective.

What did you learn from this study?

Bitopic inhibitors — molecules that bind to two sites on the same target — aren’t new. What we wanted to do was understand the principles that govern how well they work and develop a more generalizable approach that could help researchers design and build them.

We tested those principles with two cancer targets, ABL1 and EGFR, and found that this approach can produce molecules that are potent and selective and can work inside cells. We hope these guidelines can serve as a blueprint for researchers designing drugs against other targets.

Where could this research go next?

There are other proteins where molecules are already known to bind at two distinct sites, and we’re excited to explore those. But an even more interesting possibility is to look for these second binding sites earlier in the drug-discovery process.

If we can find new ways to identify those sites, we may be able to use this two-handed approach against many more drug targets.

Ultimately, we hope that the enhanced potency and selectivity of bitopic inhibitors can translate into more effective and safer therapies for patients.