Bryan L. Roth, M.D., Ph.D., a member of the Brain & Behavior Research Foundation Scientific Council and three-time NARSAD Grantee, is among leaders of a team that has successfully tested an automated process to develop new medications, including those with promise to be effective in treating brain and behavior disorders. The team’s process is described in a paper appearing in the December 13 issue of the journal Nature.
The idea of designing new drugs to hit specific biological targets—for instance, a particular kind of receptor on the surface of nerve cells in the brain—is called rational drug design. While it has yielded only partly satisfying results to date, the concept holds great appeal to pharmaceutical developers, particularly as scientists learn more and more about human biology, including the brain.
Dr. Roth and colleagues set out to find new compounds that would “hit” multiple specific targets in the brain (‘therapeutic targets’) that they designated in advance. This is important because many medications, even if they target a single protein, are known to affect multiple proteins, sometimes causing unwanted side-effects or toxicity. The goal is to hit multiple therapeutic targets and minimize interactions with unintended targets that cause side-effects.
As a point of departure, the team used an existing drug called Donepezil sometimes prescribed to enhance cognition in Alzheimer’s patients. Applying a database containing a wealth of information about relationships between chemical structures and their known biological activity, Roth’s team succeeded in identifying several new drug candidates that proved in mouse tests to be able to hit dopamine receptor targets that Donepezil does not, and which are thought to have good potential to provide enhanced therapeutic benefits. According to Dr. Roth, one of the compounds the team identified showed efficacy in a mouse model of attention deficit-hyperactivity disorder (ADHD). “In principle, our approach is applicable to all drug-target classes,” he said, “limited only by the availability of structure-activity data to create useful models” for accurate prediction.
Read the complete study from Nature