THE easiest way to find new drugs may be to start with an old one. So say researchers who have transformed an antipsychotic drug into a potential treatment for prostate cancer. Their technique could offer low-budget academic labs and biotech start-ups a chance to compete with big pharma.
鈥淭he team began with a database of all approved drugs, then used software to model the best one to treat prostate cancer鈥�
One of the main goals of prostate cancer research is to find a drug that blocks the androgen receptor, a hormone receptor that plays a key role in the development of the disease. Existing drugs are unsatisfactory, so Ruben Abagyan, a computational structural biologist at the Scripps Research Institute in La Jolla, California, and his colleagues set out to find a better one.
Advertisement
The standard approach to this task would be to screen tens of thousands of small molecules in lab cultures looking for anti-androgenic activity. Once a few candidates are found, developers would begin the expensive process of testing how the compounds behave in the body. Most candidates fail, often because they are not easily taken up by the body or because they prove to be toxic.
In the hope of speeding up the process, Abagyan鈥檚 team began with a database of the chemical formulas of all approved drugs, then used three-dimensional molecular modelling software to predict which would fit into the critical region on the androgen receptor. The best fits came from three antipsychotic drugs called phenothiazines.
In lab tests, two of the three weakly blocked the androgen receptor. The researchers then tested 10 other molecules that differed slightly from the originals. One of these turned out to have strong anti-androgen activity and, as a bonus, it lacked the antipsychotic effect of its parent molecule (Proceedings of the National Academy of Sciences, ). Drug companies have now expressed interest in testing this and a related anti-androgenic compound the group found later, says Abagyan.
Phenothiazine antipsychotics are known to have some anti-androgenic side effects, so in retrospect the new use is not totally unexpected. However, the researchers were not aware of this at the time and chose the drugs solely on their shape. They say the strategy of looking for potentially useful side effects may actually be harder than screening all drugs using molecular modelling. 鈥淲e鈥檙e interested in molecular effects, and the side effects are in terms of symptoms,鈥� says Abagyan. 鈥淭he connection isn鈥檛 simple.鈥�
Abagyan鈥檚 approach may indeed offer a useful short cut to new drugs, says Michael Sundstrom, who heads the University of Oxford鈥檚 Structural Genomics Consortium. However, he cautions, just because an existing drug has been approved does not mean a related molecule will be just as safe. 鈥淓xperimental testing needs always to follow,鈥� he says. 鈥淏ut to save resources, and perhaps for smaller laboratories that do not have the resources for a major screening effort, it is definitely an area that will grow.鈥�
