AI for Drug Repurposing- An Interview with Ian Roberts, Healx
The Digital DNA series explores the role of large-scale genetic testing in science, industry and society. We aim to understand both the benefits and risks of this emerging technology and see what the future may hold.
In the finale of our Digital DNA series we talked to Dr Ian Roberts, the Chief Technology Officer of Healx. He describes how Healx uses artificial intelligence (AI) to repurpose already-approved drugs as treatments for rare diseases.
FLG: Can You Introduce Healx?
IR: Healx are an AI-powered biotech, and we accelerate the discovery and development of treatments for rare diseases.
Our mission is to advance 100 treatments towards the clinic by 2025. We are doing this through the use of big data and artificial intelligence, which will enable us to repurpose existing drugs and combinations to make the drug discovery process quicker, cheaper and more successful. A repurposed drug is when we take a drug which is already approved for use in one disease and find a new use for that drug in another disease.
Healx was founded in 2014 by our CEO Dr Tim Guilliams and Dr David Brown, the chair of Healx and co-inventor of Viagra. Healx aims to improve on the current drug discovery process, with the idea that by re-inventing the way that drugs are discovered we can bring treatments to rare diseases that are often overlooked in the traditional pharma model.
FLG: Why is the Current Pharmaceutical Model Ineffective for Rare Disease Patients?
IR: The current pharma model is a target-based one; the idea is that you start with the disease, identify a target, then find small molecules to modulate that target and bring therapeutic benefit.
However, this requires you to have selected both the right disease, and the right target to modulate it. In many cases these hypotheses are often incorrect. That’s why a significant proportion of drug candidates fail – they either fail to engage, or act on the wrong target for that particular disease.
Now if you imagine that there are over 7,000 rare diseases with a plethora of potential targets, that hypothesis-driven approach is not going to work. In contrast, Healx uses a hypothesis-free approach to find treatments for rare diseases using big data and algorithms.
FLG: Can you Describe how Healnet Works?
IR: Healnet is our artificial intelligence drug discovery platform which uses big data and a suite of AI algorithms, combined with scientific and patient group expertise, to identify potential drug candidates for repurposing. In contrast to the traditional pharma hypothesis-led approach, we are hypothesis free. We take all the big data that represents biological, clinical, pharmacological knowledge and assemble it into a knowledge graph to form a comprehensive representation of rare disease therapeutic space. This knowledge graph then serves as input to our drug repurposing predictive algorithms.
When our algorithms find a match between an existing approved drug and a new disease, we then work backwards to see whether the drug could in fact bring therapeutic benefit. We often find the mechanism by which the repurposed drug is working in the new disease is through a different or secondary mechanism to the disease for which it was originally designed.
We then pass these findings onto our resident pharmacologists who look at the evidence that supports the link between the repurposed candidate and the new disease. The drug candidate will then move onto pre-clinical testing, and, if successful, into clinical trials. This entire process can be achieved within 24 months, which is significantly quicker and cheaper than developing a new chemical entity.
We have already carried out this process successfully. FRAXA Research Foundation approached us a couple years ago to collaborate on a drug repurposing programme for fragile X syndrome. We are now about to enter Phase 2a clinical trials with repurposed drugs identified through the Healnet platform.
FLG: What are the Advantages of Drug Repurposement?
IR: A repurposed agent has already been in use in humans for a long period of time; so we understand its safety, its benefits and its side-effects. Because of that information we can seek regulatory approval much more quickly with a repurposed agent, meaning that treatments can be developed more quickly and cost-effectively.
If you think about the traditional drug discovery model, the path from discovery through to clinical trials can take 12-14 years for a new chemical entity. However, for a repurposed drug we are getting all the way through discovery to clinical in around 24 months. For monotherapies we can sidestep the need to do Phase 1 studies and often enter clinical trials at Phase 2a. We anticipate that following the start of a clinical trial the treatment will be available to patients within 3-5 years.
FLG: How Do you Work with Pharmaceutical and Biotechnology Companies?
IR: Pharma companies are interested in identifying new diseases in which to apply their drugs, which our Healnet platform can help with.
Similarly, we may identify a promising connection between a drug and a disease, but in an area where we either don’t have the capacity or internal experience to lead it ourselves. In that case we would make that information available to a pharma or a biotech or form a partnership to co-develop the treatment
FLG: Anything Else?
IR: One of Healx’s key values is patient benefit. It’s the reason Tim and David set up the company in the first place. Healx works extremely closely with patient groups and we are currently in the process of establishing a new programme called the Rare Treatment Accelerator. This programme offers select patient groups the opportunity to partner with us to develop new treatments for their rare disease of interest. Any patient groups interested in getting involved should look at our website and apply when the call opens.
Catch up on the rest of the Digital DNA series here.