The strength of investor interest in artificial intelligence was proved again today with the $56m series B financing round secured by Healx, the UK technology company specialising in treatments for rare diseases.
The money will be used to kickstart development of 40 new rare disease projects and speed progress of Healx’s 10 existing programmes, one of which is in the clinic. Such is the appetite for one of the industry’s hottest areas that the latest funding injection came only 15 months after the group’s $10m series A round.
Tim Guilliams, Healx's chief executive, told Vantage that rather than exploring multiple rare diseases the company would focus on four disease clusters and had already identified neurology and oncology as two areas.
Healx, which this year celebrated its fifth birthday, has departed from the usual business model of AI companies working in drug discovery. Instead of looking for new products, the company is using its technology to find cures for rare diseases among existing, mainly off-patent, drugs.
An advantage is that the drugs already have a proven safety track record, potentially speeding development and reducing costs. As for the IP, Kate Hilyard, Healx's chief operating officer, said the use of the therapies in new indications and combinations could ensure a solid patent position, though method-of-use patents are widely seen as unreliable.
Digging for data victory
Key to Healx’s efforts is mining public and proprietary data from numerous sources including scientific papers, clinical trials and patient groups to create what the company calls a knowledge graph of rare disease.
The knowledge graph is then examined by the company’s algorithms and pharmacology experts to identify the most promising treatments for a given disease. Mr Guilliams said this had led to a hypothesis-free approach to drug discovery, where neither targets nor a particular rare disease were preselected for study, with decisions made by the data. “By doing this we take out the bias you can see with normal drug development,” he said.
The company says this data-driven approach has significantly cut development times, with targets typically moving into the clinic within 24 months. It also means that the group can rapidly scale drug development, hence the ambitious-looking target of 40 new programmes.
Healx says it has validated this approach with its own Fragile X programme, which was only identified two years ago and is set to begin phase IIa trials early next year, using an adaptive clinical trial design with multiple combinations of drugs. Some of the other targets from its 10 programmes are also set to enter the clinic next year.
As to the future of some of these pipeline products, Healx is adopting a flexible attitude. “There are some programmes we will be taking all the way through to the market, and we will patent-protect them, but there will be others we will partner,” Ms Hilyard said.
Healx in not alone in trying to repurpose old drugs for new treatments. However, AI-driven drug discovery, while fashionable, is still very much in its infancy. As such, the real proof of the validity of the company’s approach will be the first Fragile X data readouts due in 2021, an event that is bound to be closely watched.