In a major step forward for cancer drug discovery, researchers have demonstrated how computational simulations can unravel the complex role of water molecules in drug binding, potentially saving years of trial and error in the lab.
The study focuses on B-cell lymphoma 6 (BCL6), a protein implicated in several cancers. Using advanced computational techniques, the team showed how subtle changes in drug molecules can disrupt or stabilise water networks in the protein’s binding site, dramatically affecting drug potency.
This method could be used by researchers worldwide to actively guide the design of new drugs, improving both their potency and their selectivity. In the longer term, this could increase the availability of effective cancer medications that are kinder on the body.
The study was led by researchers at The Institute of Cancer Research, London, and the findings were published in the Journal of Chemical Information and Modeling. The Medical Research Council provided funding in the form of a grant, which was partly supported by AstraZeneca.
The importance of water molecules
Water molecules, often overlooked in drug design, play a critical role in how drugs interact with their targets. In protein binding sites, these molecules can form intricate hydrogen-bonded networks that influence the orientation, stability and effectiveness of drug compounds. Displacing a single water molecule can either enhance or weaken a drug’s binding affinity – an effect that it is difficult to predict experimentally.
“Water networks are like invisible scaffolding,” said first author Daniella Hares, a PhD student in the Division of Cancer Therapeutics at The Institute of Cancer Research (ICR). “They hold everything together, and if you remove one piece, the whole structure can shift. Our simulations help us see how that scaffolding behaves when we tweak a drug molecule.”
To explore this, the researchers used two powerful computational methods – Grand Canonical Monte Carlo (GCMC) simulations and alchemical free energy calculations – that model how water molecules behave and how changes in drug structure affect binding energy.
They focused on four BCL6 inhibitors developed in earlier ICR-led research, each of which was designed to grow into a water-filled subpocket of the protein. These compounds sequentially displaced up to three water molecules, resulting in a 50-fold increase in potency. The researchers wanted to understand why.
Achieving a more nuanced picture
Using GCMC, the team simulated how water molecules occupied the subpocket in the presence of each compound. Impressively, the simulations reproduced 94 per cent of the water sites observed in crystal structures, even when starting from different protein conformations. This suggests that GCMC could be a reliable tool early in drug development, before experimental data are available.
The first compound, known as compound 1, formed a stable network of five water molecules. When a small ethylamine group was added to create compound 2, one water molecule was displaced. Surprisingly, this only led to a modest two-fold increase in potency. Looking at the simulations, the team realised that although the new group formed additional interactions with the protein, it also destabilised the remaining water network, negating the benefits.
Compound 3 introduced a larger pyrimidine ring, displacing a second water molecule. This time, the potency jumped more than 10-fold. The simulations showed that the pyrimidine not only replaced the lost water interactions but also stabilised the remaining network by forming new hydrogen bonds. This stabilisation contributed significantly to the potency gain.
Finally, compound 4 added a second methyl group, displacing a third water molecule. Despite predictions that this would be unfavourable, the compound showed a further two-fold increase in potency. The simulations suggested that while the water network was destabilised, the methyl group helped prearrange the molecule into the ideal protein-binding conformation – offsetting the loss.
The scientists analysed each transformation using a combination of GCMC and alchemical calculations to dissect the contributions from water displacement and new protein interactions. The cycles showed excellent consistency, indicating reliable and converged simulations.
The study also compared GCMC with faster solvent analysis methods, such as SZMAP and 3D-RISM. The team found that GCMC provided a more nuanced picture, with the other techniques often failing to capture the cooperative effects between water molecules.
Promoting wider adoption
Looking ahead, the researchers plan to apply their methods to other drug targets with complex water networks. They believe that integrating GCMC into standard drug design workflows could lead to faster, more efficient development of cancer therapies.
The team also hopes the work will encourage more researchers to use GCMC in their pharmaceutical research projects. Despite its power, the method remains underused – due in part to both lack of awareness and limited availability in commercial software. However, the computational cost is very manageable, with GCMC simulations running overnight and alchemical calculations completing in a few days.
To support broader use, the authors have made their simulation scripts and data publicly available on GitHub.
“The implications for drug discovery are significant”
Senior author Professor Swen Hoelder, Group Leader of the Medicinal Chemistry 4 Group (including Analytical Chemistry) at the ICR, said:
“This approach accounts for how water molecules interact with each other, not just with the protein or drug. That’s crucial when dealing with complex networks.
“The implications for drug discovery are significant. Traditionally, optimising a drug to interact with water networks requires multiple rounds of synthesis and testing – a process that can take years. By using GCMC and alchemical calculations, we can predict which modifications are likely to succeed before entering the lab.”
First author Hares said: “Our work has demonstrated that managing water molecule interactions is very much a balancing act. You gain some interactions, but you lose others. By quantifying that trade-off, our methods could save enormous time and resources in drug discovery. Instead of guessing, we can design smarter from the start.
“Most excitingly, because water molecules are so important to how proteins function, this approach could be useful for any drug that targets a protein, meaning it could transform the treatment landscape across cancer types.”
Image credit: cromaconceptovisual from Pixabay