An alternative approach to drug screening searches for small fragments that snugly fit into a protein drug target. 

For more than two decades, high throughput screening (HTS) has been the main technique used to begin the process of developing new drugs. Often described as the “needle in the haystack” approach, HTS involves screening about a million compounds against a disease target — usually proteins — that are found in the body. It’s like a matchmaking game: the goal is to find a drug molecule that can bind with the target protein, to either switch off or turn up its action. The fit is never perfect at first, but it’s a starting point in the years-long drug discovery process. 

In recent years, an alternative approach to finding drug candidates has emerged, known as fragment-based drug design.  In HTS, you’re taking a million or so different complex-shaped compounds and hoping that just a few might be a good fit with the protein. But with the fragment-based approach, you’re starting with a smaller pool of much simpler compounds about half the size of HTS molecules — mere fragments. 

A Game of Tetris

Picture it as the popular video game Tetris. Geometric blocks are rapidly falling from the sky, as the player has to quickly manipulate how they fit into a base pattern. With HTS, the falling blocks are large and complicated, and finding a fit is challenging. But with fragment-based design, the dropping shapes are smaller and simpler, making it easier to match them up with the base pattern. “Because the molecules we’re working with are quite simple, the chances that you’ll have an efficient and snug fit into the protein is much higher than for HTS,” says David Hepworth, Head of Medicinal Chemistry in the Inflammation and Immunology research unit at Pfizer in Cambridge, Mass., who has been using the fragment-based approach.

The drawback to starting simpler, says Hepworth, is that the initial fragment has “a low affinity” for the protein. In other words, even if that one piece fits nicely, the number of potential places it can bind to the protein is limited  because it has a relatively small surface area. “With fragment-based screening, it generally provides a really nice fit in one part of the protein. This provides an excellent starting point to build a more complex structure.”

Throughout the optimization effort drug designers use a technique called X-ray crystallography understand the shape of their target protein and how the shape of the drug molecule can be adjusted to achieve an optimal fit. 

As a drug designer, the simplicity and minimalism to this approach appeals to Hepworth. “I like the idea of starting with something small and building it up as efficiently as possible.” A smaller drug also means that it is more likely to have all the other properties required of a successful drug molecule, such as, the ability to be absorbed in the intestines and be less likely to cause liver damage, he adds. 

The Case of IRAK4

Starting small is proving to be a viable approach. Fragment–based screening helped create a new potential treatment for auoimmune diseases such as rheumatoid arthritis (RA) — by targeting an enzyme that had eluded researchers for over a decades. This investigational compound is known as a kinase inhibitor, which may block IRAK4, an enzyme involved in the inflammatory process, potentially helping to reduce the joint swelling and damage associated with the disease. 

Finding a suitable candidate compound had its share of challenges. IRAK4 is a type of kinase in a family of aboutone of about  500 highly related kinase enzymes in the body. “You want to find something that will inhibit this kinase but not the rest of the family, to avoid potential of the compound to cause side effects” says Hepworth. IRAK4 also has a relatively small binding site, which presents some challenges. “Scientists across the industry have been trying to find IRAK4 inhibitors for several years,” he says. “Because it appeared to be difficult, we decided to go about it a different way.” 

In 2009, they did their initial fragment-based screen with a pool of about 3,000 compounds. While they got hundreds of tiny compounds that bound to IRAK4, their winning compound stood out from the pack because of its unorthodox structure. “A medicinal chemist can usually pick out a kinase inhibitor based on it chemical structure, but this one structure  looked different. That attracted the team,” he says. 

Once they had their drug lead, a team of Pfizer drug designers spent about three years “elaborating” from the initial structure, adding parts to increase its binding affinity to other sites on the protein and to optimize the other properties required of the molecule. Throughout the process, they used X-ray crystallography to fine-tune the structure of the drug-lead to match the shape of the IRAK4 binding site.

Hepworth believes this custom-built approach made the difference to help them narrow down the search for the potential new rheumatoid arthritis medicine. “This molecule has a different structure, which may be why we have evidence that it’s selective,” he says. “It’s quite a nice story how an alternative approach to medicinal chemistry is helping to enable the development of potential new medicines.”