Skip to main content

New Uses for Old Medications

Even drugs whose development was stalled or canceled might show promise for illnesses they were never meant to treat

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


Despite decades of research, diseases of the brain have proven especially difficult to treat. Consider Alzheimer’s disease—every single clinical trial of an Alzheimer’s drug to date has failed. In January Pfizer announced that it had ended research on drugs for the disease, as well as for Parkinson’s. Autism has been similarly frustrating; despite the attention it has received, we still have no effective treatments. Then there’s schizophrenia. It hasn’t seen a breakthrough for more than 60 years, since the discovery of chlorpromazine (better known as Thorazine), which happened largely by chance.

Still, as a neuroscientist who has studied schizophrenia for decades, I am optimistic. The story of chlorpromazine, an antihistamine that was repurposed for schizophrenia, offers a powerful lesson. We now have the ability to find other drugs that could be repurposed to treat brain diseases, thanks to new technologies. Medicines already on our shelves may hold untapped promise for treating brain diseases—if only pharmaceutical companies can be prompted to share their data with scientists.

Thorazine has its origins in the late 1940s, when surgeons prescribed an antihistamine called largactil as a way to relax anxious patients about to go into surgery. This led doctors to try the drug on people suffering from pathological anxiety and agitated psychotic patients. With a few modifications, in 1952, largactil was reborn as the antipsychotic drug Thorazine, ushering in an entire generation of drugs to treat a variety of psychiatric disorders, from schizophrenia and bipolar disorder to severe depression and anxiety.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


Thorazine was a critical breakthrough, but it wasn’t a silver-bullet cure. This stems from the fact that schizophrenia is likely not a single disorder but rather a collection of many. Various approaches are needed to attack the diverse underlying mechanisms of disease—and that starts with drug repurposing.

Because an existing drug has already passed the U.S. Food and Drug Administration’s safety requirements to prove it is nontoxic to humans, successfully repurposing it—that is modifying it to treat another ailment—can take less than half the estimated 13 years and one-third of the average $3 billion cost than developing a single drug from scratch. The several-thousand FDA-approved drugs in existence thus represent a vast resource that can potentially be repurposed to target any number of conditions.

But what’s missing from this database is information on the thousands of drugs that are not FDA approved, such as those stalled in clinical trials or discontinued by pharmaceutical companies. When these companies abandon development of a drug, information on that drug gets locked away in company coffers. Trillions of dollars’ worth of research—some of which has come from taxpayers—is lost.

Scientists need unfettered access to this information and we need it now. Starting in 2012, the National Institutes of Health (under its National Center for Advancing Translational Sciences, or NCATS), along with the U.K.’s Medical Research Council, have been striking deals with major pharmaceutical companies to take abandoned drugs from their pipelines and release information about them publicly. Other independent initiatives to create similar databases of approved and failed drugs are also under way.

To date, large-scale computational efforts for analyzing such data have not been very effective in drug repurposing efforts largely because there haven’t been enough data points to work from. But if this influx of information on abandoned drugs could be funneled into a single, centralized resource along with existing data on approved drugs—and combined with the explosion in genetic knowledge related to the underlying disease mechanisms—it would be a revelation. Researchers could at last effectively employ the latest tools in bioinformatics, data science and machine learning to uncover common molecular themes between diseases, or between diseases and potential drugs.

Ultimately, the key to success is access, and therein lies our greatest hurdle. Many pharmaceutical companies are still reticent to reveal anything that might jeopardize intellectual property. Even academic scientists may hesitate to share too broadly with competing labs. To remedy this, entities like the FDA must develop incentives for sharing data, such as creating a legal framework that safeguards privacy and commercial interests. These incentives could then open the floodgates for easy-to-use, open platforms for efficiently sharing and mining data, expanding and augmenting the efforts of NCATS.

The breadth and scale of this type of investigation would not have been possible five years ago. But now is a pivotal moment in neuroscience, and we have never been closer to real breakthroughs.

Currently in my lab we are testing whether certain cancer drugs—either clinically approved or in trials—that target and restore some of the same biological processes that are also disrupted in schizophrenia have the same restorative properties in the brain cells of schizophrenia patients.

This is only one promising example, and proof of concept that a systematic and strategic approach to drug repurposing could actually move the needle. There’s no time to waste. We now have the data-mining capabilities to deploy a legion of virtual researchers in search of these eureka moments. What we need is a full-throated buy-in from pharmaceutical companies and academic scientists alike—and access to the life-saving data they hold.