Insilico Medicine Brings First AI-Designed Drug to Clinical Trials

Image of a pill on a background of zeros and ones to indicate how artificial intelligence and deep learning can help predict drug-drug interactions
Source: Yuuji/Getty Images

Insilico Medicine, an artificial intelligence (AI) company based in New York and Hong Kong, has brought the first AI-designed and discovered drug candidate to clinical trials in less than three years.

The drug candidate is designed to target idiopathic pulmonary fibrosis, a rare lung disease with a poor prognosis affecting 13-20 people per 100,000 around the world. There is currently a lack of treatments available for this lung disease.

The first-in-human trial is microdosing healthy volunteers with the small molecule drug candidate, currently known as ISM001-055, and is taking place in Australia.

Insilico Medicine has developed a machine learning model called Generative Tensorial Reinforcement Learning (GENTRL) that uses neural networks to search for new drug targets.

“Deep generative models are machine learning techniques that use neural networks to produce new data objects. These techniques can generate objects with certain properties, such as activity against a given target, that make them well suited for the discovery of drug candidates,” explained Alex Zhavoronkov, Founder and CEO of Insilico Medicine, and co-authors in a 2019 Nature Biotechnology paper describing the method.

According to the company, this is the first time a drug target has been developed entirely using AI from target discovery to drug development to human trials.

“We are very pleased to see Insilico Medicine’s first antifibrotic drug candidate entering into the clinic… to our knowledge the drug candidate is the first ever AI-discovered novel molecule based on an AI-discovered novel target,” said Feng Ren, CSO of Insilico Medicine, in a press statement.

The company has developed a platform called Pharma.AI that encompasses target identification, production of new molecules and prediction of clinical trial outcomes. This platform was used to develop ISM001-055 to the preclinical stage in less than 18 months for $2.6M. In February this year, the company announced it had discovered a preclinical candidate drug and 9 months later it has started its first clinical trial.

The timing and costs of this project are impressive given that standard preclinical drug development costs are over $430M and the timing can take between 3-6 years to complete.

Zhavoronkov emphasizes that another factor the company has overcome is that very few companies actually manage to take a prospective drug all the way to clinical trials.

“There are very few examples of a pharmaceutical company discovering a new target for a broad range of diseases, designing a novel molecule, and initiating human clinical trials. To my knowledge, nobody has achieved this with AI to-date,” he explained.

“The failure rates in preclinical target discovery are very high and even after the targets are validated in animal models, over half of Phase 2 clinical trials fail primarily due to the choice of target.”

Of course, it has yet to be seen how successful their candidate will be in later trials, but if it is effective this process has the potential to revolutionize the drug development space and allow the design of more targeted drugs in areas of unmet medical need.

This site uses Akismet to reduce spam. Learn how your comment data is processed.