0
Facebook
Volume 87 Issue 45 | p. 11 | News of The Week
Issue Date: November 9, 2009

Old Drugs, New Tricks

Polypharmacology: Ligand-based approach finds new targets of known drugs
Department: Science & Technology
Keywords: Drug Interactions, Polypharmacology, Targets, Side Effects
[+]Enlarge
WEAVING A WEB
Ligand-based association increases the number of identified targets for known drugs.
Credit: Nature
8745notw1_figure3cxd_lg
 
WEAVING A WEB
Ligand-based association increases the number of identified targets for known drugs.
Credit: Nature

A new computational method to identify previously unknown targets for approved and investigational drugs could help researchers find new uses for those drugs and better understand their side effects. The method was devised by a team led by Bryan L. Roth of the University of North Carolina, Chapel Hill, and Brian K. Shoichet of the University of California, San Francisco (Nature, DOI: 10.1038/nature08506).

"Drugs are special molecules; there aren't that many of them," Shoichet says. "Let's make the best use of the ones we already have and find new indications for established drugs."

Many drugs are more than one-trick ponies, binding to multiple targets. Sometimes those interactions lead to new uses for well-established drugs. Other times, they cause harmful side effects. In either case, knowing about those interactions allows better use of drugs. The challenge is finding them.

In the new method, Shoichet and coworkers define drug receptors not by structure or sequence but by the ligands that bind to them. This approach differs from structure-based approaches, which use a receptor's crystal structure as a starting point. "For us the target is just a bunch of ligands," Shoichet says.

Using a modified version of an algorithm used to search gene-sequence databases, they screen compounds against a database of targets, asking how much the compounds look like the ligands. "Essentially, the algorithm looks for shared functional groups and shared bond separations between them and weights those patterns by their occurrences among the set of ligands and what one would expect to find just at random," he says.

The team compared 3,665 approved or investigational drugs with hundreds of targets, which are defined by their ligands. The researchers predicted thousands of unanticipated interactions and experimentally tested 30 of them. They confirmed 23 of those 30 interactions.

In one case, the team found that Rescriptor, which inhibits the enzyme HIV reverse transcriptase, also inhibits the histamine H4 receptor. "There's no evolutionary relationship between this viral enzyme and the human histamine receptor," Shoichet says. "Nevertheless, Rescriptor binds to histamine H4 at physiologically relevant concentrations. This may have a role in some of the painful side effects the drug has."

In another example, the antidepressant Prozac, whose primary target is the serotonin transporter, binds the β1 adrenergic receptor, a G-protein-coupled receptor (GPCR) that usually binds such compounds as epinephrine and norepinephrine. "These transporters have no sequence or structural similarities to GPCRs. Nevertheless, here is Prozac hitting this GPCR," Shoichet says.

"New methods are needed to predict the range of biological targets a drug may bind to," says Andrew L. Hopkins, a longtime researcher in the pharmaceutical industry and now at the University of Dundee, in Scotland. "Validation of the new method is an important step forward in building a toolbox for chemists to design drugs that act on multiple targets."

"As a scientist, I find the combination of the simplicity of the method and its applicability fascinating," says Irilenia (Irene) Nobeli of the Institute for Structural & Molecular Biology at the University of London. "As a member of the public, I want easy access to the predictions about the drugs I might be taking and the effects they might have on my health."

 
Chemical & Engineering News
ISSN 0009-2347
Copyright © American Chemical Society