Feature: Shortening the odds in drug design

The practice of seeking and refining chemical compounds to improve the human condition has been around since Adam discovered the benefits of apple juice and snake oil. The same ideals now fuel a multibillion-dollar pharmaceutical industry, underpinned by a staggering array of biological and chemical research.

It is surprising, then, that over the last few decades pharmaceutical companies have been struggling a little to find enough ‘new’ drugs to meet the ever- increasing complexity of human disease and, of course, their projected bottom line.

A number of major technological and scientific advances over the last couple of decades, such as combinatorial chemistry, high-throughput screening and the sequencing of the human genome, have injected fresh ideas and techniques into drug development research.

One of the newest kids on the block – fragment-based drug design – is further improving the chances of finding novel lead compounds. This approach relies on identifying low-molecular weight compounds that bind to a target protein. ‘Hits’ can then be modified or linked to other compounds to increase potency.

One of those employing this relatively new approach is Associate Professor Martin Scanlon, who leads a research group in structure-based drug design at the Monash Institute of Pharmaceutical Sciences (MIPS) in Melbourne.

At this year’s Lorne Conference on Protein Structure and Function, Scanlon discussed his group’s use of fragment- based drug design to search for novel antimicrobial compounds and a way to combat the increasing problem of bacterial resistance.

“Our primary interest is in characterising the interaction of proteins with small molecules that selectively bind to therapeutically interesting protein targets,” says Scanlon.

“We chose to implement the fragment-based drug design approach in this work as an alternative to high-throughput screens (HTS), which requires a very large number of molecules to look for the one that presents itself in exactly the right orientation to bind optimally to the target protein, even then the chances of a hit are actually incredibly remote.”

By “optimally”, Scanlon means those small molecules that are both potent at their target and good drug-like candidates for therapeutics in terms of their physicochemical properties, such as solubility, lipophilicity and polarity.

---PB---

“Back of the envelope calculations of just how many molecules would fulfil the standard criteria of being drug-like produce numbers like 10⁶⁰. People argue about this wildly, but even if they are out by an order of magnitude or two, it is still a very very big number.

Since your typical HTS library might have 10⁶ compounds in it, the chances of finding the molecule that optimally binds to your protein is therefore miniscule, virtually nothing.”

This doesn’t mean that nothing will bind in an HTS screen. Obviously, molecules do bind, and there will be many non-optimal hits. Those candidates then have to go through a modification process to meet the desired drug-like criteria.

“The problem with all of this, particularly in an academic environment, is that you need a very large library, lots of time and lots and lots of money, and all that with a very small chance of success. Not exactly a recipe for attracting grant funding!

“So we decided that rather than looking at drug-like molecules, we would look at fragments of such molecules that are much smaller than in a typical drug screen library, about the same size as each of the different epitopes that bind to your protein of interest.”

This means that the compound ‘hits’ almost always bind with only a low affinity or potency. However, because they are smaller and generally less complex than molecules you find on an HTS library for example, the chance of them binding optimally to the target protein is actually much higher.

“The positive hits in fragment-based screening can therefore generate some very efficient binding partners, and then you have to work at making them more potent. The big advantage is that it is much easier to make a molecule more complex and potent whilst maintaining the binding efficiency than the other way around,” says Scanlon.

The end result with using smaller molecules is that far fewer compounds are needed in the initial screen to get a good coverage of the chemical space, and even with a library of around 1100-1200 molecules, you get much higher hit rates than people get typically in HTS.

Read part II, Making drug resistance futile.