“Drug repurposing doesn’t work. We’ve all tried it and it doesn’t work.”
This was a comment from a distinguished speaker at the Computational Chemistry, Gordon Research Conference (GRC) in July, 2022. For the uninitiated, the GRC is a unique forum where researchers from all over the world come together in an informal setting to discuss their unpublished work to facilitate the exchange of ideas.
My interest in drug repurposing began in graduate school, where I used structure based virtual screening methods to predict a combination of drugs to target the malaria parasite, Plasmodium Falciparum. Almost 10 years later, I am still obsessed with repurposing and had the privilege of presenting a poster at the July 2022 GRC showcasing the use of Cyclica’s PolypharmDB database to identify drug-centric repurposing candidates. In this project, three out of seven predicted repurposing candidates tested in vitro were the most potent molecules ever tested against a gain-of-function (GOF) Src variant associated with an ultra-rare disease.
While drug repurposing doesn’t often pique the interest of the drug discovery community in the same way as de novo drug design, it’s just one of many ways that we can use PolypharmDB, within Cyclica’s MatchMaker™ platform, an all-against-all database of ~17k molecules screened against 18K proteins. In addition to identifying non-obvious biological targets of known drugs (drug-centric repurposing), we are using PolypharmDB to identify polypharmacological drug opportunities with the hope of de-risking clinical candidates and to identify known clinical molecules that are active at low data targets. Our hope is that this will help to address unmet medical needs and improve patient outcomes. Currently, the PolypharmDB portfolio has validated repurposing candidates for immunology, inflammation, cancer, and rare disease targets.
A recent study has classified repurposing efforts into three approaches:
To date, most repurposing examples fall into the first two categories. The majority of successfully repurposed drugs have been identified using a disease-centric approach, where drugs are repurposed for different types of a similar disease ‘family’, such as cancer. Target-centric repurposing is successful when the target protein is relevant to multiple indications, the most famous example being Viagra.
Drug-centric repurposing identifies a completely novel target for a drug that has not yet been identified, a much more difficult proposition than target-centric or disease-centric, likely due to insufficient approaches to predict a drug’s protein binding profile, which is where Cyclica’s science has proven to be particularly successful.
Cyclica’s AI-enabled drug discovery platform, MatchMaker™, makes it possible to systematically and rapidly explore the predicted polypharmacology (“proteome binding profile”) of a library of small molecules across the proteome. Cyclica developed PolypharmDB in 2018, a database of computed proteome binding profiles for clinical candidates and approved drugs. PolypharmDB has expanded to include the computed binding profiles for preclinical as well as clinical candidates and approved drugs, which opens up the possibility of using PolypharmDB as a “pre-positioning” tool. Correct prediction of the polypharmacological profile of a preclinical candidate could help to mitigate the failure of clinical candidates, as it could help inform the disease model selection to support the intended clinical trial design, i.e. “drug-centric prepositioning”.
As the opening quote indicates, there are some well known challenges associated with drug repurposing. One challenge is identifying the intellectual property (IP) position since these molecules are already patented (both composition of matter and method of use). However, this can be viewed as an opportunity, albeit mostly to large pharma companies, by creating a user designed and owned PolypharmDB containing a proprietary library of compounds associated with non-obvious protein(s) and therefore non-obvious disease opportunities. For companies that don’t own a large library of proprietary compounds, there is a way to use drug-centric repurposing to gain IP. The repurposing candidate could be used as a “drug-like” starting point from which to develop novel, patentable drugs; as opposed to de novo identification. This does create a challenge of designing drugs that are novel enough to be patentable, yet similar enough to maintain the desired polypharmacological profile. In my opinion, this particular issue is one that can be overcome, and something that I look forward to discussing further with my peers at this year's Gordon Conference.
With our repositioning strategy and platforms, Cyclica is capable of identifying drug-centric repurposing opportunities that accurately predict the polypharmacological profiles of drugs in various stages of development. Hopefully this helps to bring therapies to patients faster.
Check out another blog post on drug repurposing here.