The use of existing drugs in new therapeutic areas has a rich history, with Viagra's serendipitous pivot from anti-angina flop to erectile dysfunction blockbuster being the standout example. Bioinformatics is now taking the guesswork out of repositioning, slashing the time and money needed to reach the clinic.
Writing in the journal Cancer Discovery, Stanford University researchers describe how they tapped into thousands of gene-expression profiles to find a new candidate for small cell lung cancer (SCLC). The researchers developed a computerized drug discovery tool to anaylze the profiles and then took a closer look at any patterns that emerged. Having found potentially druggable targets among the patterns, the algorithm identified which class of drugs could act against them.
In the case of SCLC, the algorithm singled out tricyclic antidepressants. A Phase II trial of desipramine--sold by Sanofi ($SNY) as an antidepressant--is now underway. "We are cutting down the decade or more and the $1 billion it can typically take to translate a laboratory finding into a successful drug treatment to about one to two years and spending about $100,000," Big Data guru and Stanford University associate professor Atul Butte said. It took Butte and his colleagues 20 months to find a candidate and advance it to the clinic.
The wealth of safety data from earlier trials and approved drugs helped speed the process. Tricyclic antidepressants were largely replaced by new drugs with fewer side effects. The risks are well known though, and the poor prognosis for SCLC patients--who have a 5-year survival of 5%--means they are more tolerant of side effects. NuMedii is commercializing the computational approac, and raised $3.5 million earlier this year.
- read the journal's release
- here's the abstract
- check Stanford's post
- and its accompanying blog