Aled Edwards is the Chief Executive of the Structural Genomics Consortium, an open science charity that carries out early-stage drug discovery research in collaboration with industry, and places all the research into the public domain, without restriction. He is a Professor at the University of Toronto and an adjunct Professor at McGill University. This piece reflects the input of ~50 drug discovery experts in academia, industry and drug discovery non-profits.
For decades, academic institutions and science funders have championed the idea that screening libraries of old drugs with a variety of tests could uncover new uses for them. Because their safety is already proven in clinical trials for other uses, repurposing known drugs is expected to bring medicines to patients faster and more cost-effectively. This concept of throwing many tests at drug libraries without a hypothesis for how they might work in a disease is called hypothesis-free drug re-positioning.
The idea has a sound rationale, is attractive for many reasons, and has been implemented widely. However, after two decades of activity, the real-world evidence for its efficacy is less compelling. Indeed, we have been unable to document a single instance of a drug approved for clinical use where the idea for the clinical trial derived first from a lab-based screen of old drugs.
With dozens of these re-positioned drugs now in clinical trials for COVID-19, we are engaged in a very public test of this approach.
A solid case for drug re-positioning
For more than 50 years, astute clinicians and drug discoverers have made observations that have led to medicines designed for one disease finding uses in another. Celebrated examples of this form of re-positioning include minoxidil, sildenafil, and thalidomide.
There are also many examples of drugs re-positioned based on a scientific understanding of mechanism of action. In oncology, this repositioning strategy is now more and more common. The clinical use of imatinib for gastrointestinal stromal tumours (GIST) was one of the first examples. GIST tumours have driver mutations in the c-kit kinase, which was a known “off target” of imatinib. This led to the hypothesis that imatinib would have efficacy in GIST, and this proved correct when tested in patients.
Indeed, the list of re-positioned medicines is now long, and provided the germ for the idea that a systematic (hypothesis-free) screen of all known drugs in research labs might uncover more re-positioning opportunities. The hope was this strategy would be faster to yield clinical benefit, in that the “hits” were drugs already approved for use in humans, and this would short-cut the typical three to five years of hit to clinical candidate optimization.
How the “new uses for old drugs” idea grew
In the early 2000s, the NIH’s Neurodegenerative Drug Screening consortium launched what was among the first systematic re-positioning initiatives. Ahead of their time, they assembled ~1,000 FDA-approved compounds with the idea of identifying new uses for old drugs, and described the effort in a set of influential papers.
The experiments were designed well; compounds were sent to many investigators blinded for testing in their various models. When the assay results were unblinded, minocycline, an antibiotic, showed great promise in a number of assays for amyotrophic lateral sclerosis (ALS). It was a brilliant concept, a beautiful experiment, and a beautiful result.
The activity of minocycline in ALS models would never have been discovered had it not been for the ability to screen such a library. Understandably, this is the point where the idea took off, and hundreds of institutions and funding organizations helped put together and assay these “drug repurposing libraries”.
Fast forward to 2007, when the results from the clinical testing of minocycline in ALS patients were published, and that story did not end well. Not only did minocycline not prove effective in ALS, but it actually made patients worse. This lesser known part of the story had little impact on the repositioning horse, which had long since left the stable.
Twenty years and no success yet?
Given the enthusiasm for this strategy in combating COVID-19, and this historical background, we decided to informally examine the real-world clinical impact of the academic drug repositioning efforts. Specifically, after 20 years of effort in dozens and dozens of centers, a large (>50) group of us, including proponents of the idea, tried to think of instances in which an idea borne from a hypothesis-free drug screen led to a new use for an old drug — defined as approved for use in clinical practice.
We could not think of one.
If there are indeed few (or no) examples despite the considerable effort, there may be many possible explanations. Perhaps cell or animal assays do not reflect the human condition. Perhaps the pharmacology is mis-aligned and the drugs were optimized for a different target, dosing or tissue. Or perhaps it is the business model; there is little commercial incentive to support clinical trials in many of the indications.
Alternatively, there may be many successful repositioned drugs out there that we could not find because they did not go through formal regulatory approval, and part of the purpose of this piece is to canvas the community for success stories that did not emerge from our informal search.
However, if the successes are few or none, then hundreds and hundreds of millions of dollars have been, and are being spent supporting repositioning efforts whose explicit aim was to find drugs that treat patients, only to have none emerge. Perhaps the COVID-19 trials will change this.
“How often has a re-run of Gilligan’s Island won an Oscar?” — North Carolina proverb