AI-Designed Drug Shows Potential to Slow Biological Ageing

Last year, a clinical trial conducted by Insilico Medicine, a company focused on accelerating drug discovery through artificial intelligence, suggested that one of its drug candidates may assist in treating patients afflicted with a chronic lung disease. The company asserts that data from the same trial reveals a more compelling prospect: that the drug may also decelerate the ageing process. The molecular structure of the drug, known as rentosertib, was generated with the assistance of AI. Results from the new study, published Monday indicate that the drug reduced biological markers of age as assessed by six “ageing clocks,” a distinct type of AI technology aimed at predicting an individual’s morbidity and mortality. This elaborate clinical trial represents a significant advancement in the ongoing endeavour to enhance health care through the application of AI methodologies that also support widely used chatbots and image generation tools. Insilico represents merely one of numerous start-ups, technology behemoths, and academic laboratories striving to expedite drug discovery and refine various medical tasks through the application of these methodologies.

The recent rise of ageing clocks is facilitating a shift in focus from short-term treatments to the domain of longevity research. These systems provide estimates regarding the rate at which an individual’s body is ageing, as well as the relative ageing rates of specific organs in comparison to the overall body. However, there remains an ongoing debate among scientists regarding the extent of useful information that these so-called clocks can actually provide. While Insilico’s clinical trial shows the promise of several AI techniques, the company has not yet tested its anti-ageing effects in healthy patients. “This drug looks encouraging,” said Eric Topol. “But we do not yet have a definitive trial to make the final judgement.” Insilico commenced its exploration of the emerging wave of AI technologies over a decade ago. Established in 2014 by Alex Zhavoronkov, a biotechnologist of Latvian-Canadian descent, the company was one of the pioneering initiatives aimed at optimising drug discovery through the application of neural networks. Insilico employed techniques akin to those utilised in training AI systems such as ChatGPT to construct a model that analyses an extensive dataset.

This dataset encompasses health records from thousands of medical patients, blood tests that detail the microscopic proteins produced within their bodies, and a vast array of academic papers that explore the effects of these proteins. Through this system, the company aims to pinpoint specific proteins associated with illness and disease. Insilico subsequently developed a second system that examines extensive datasets detailing the physical structure of proteins and the manner in which these minute biological entities interact with other molecules. “It is like scanning a lock and generating a key that fits the lock,” Dr Zhavoronkov said in an interview. This is how he and his company developed rentosertib, which is intended to address a condition known as idiopathic pulmonary fibrosis, or IPF. Sometimes referred to as “the Alzheimer’s of the lungs,” IPF is a chronic condition characterised by the thickening and scarring of pulmonary tissue, which diminishes its capacity to transfer oxygen into the bloodstream. The condition has the potential to result in mortality, potentially within a span of a few years. Last year, through its clinical trial, Insilico demonstrated that its drug candidate could substantially enhance the air capacity within the lungs of patients suffering from IPF.

Zhavoronkov stated that his company also developed the drug with the aim of broadly prolonging a patient’s life span. Distinct from the assessments related to IPF, the trial employed six ageing clocks to evaluate the age markers in patients prior to and following treatment with the drug. Across the 43 patients who participated in the trial, the clocks – which essentially attempt to estimate a person’s age based on assessments of the functioning of their cells, tissues, and organs – demonstrated notable decreases in the predicted age. Each clock depends on a distinct interpretation of those indicators, resulting in varying forecasts. During the trial, all six clocks indicated a decrease in predicted age following a 12-week period of drug administration by the patients. He acknowledged, however, that the study was far from conclusive, noting that the sample size was limited and that ageing clocks do not consistently provide reliable results. Most notably, Insilico’s drug has not yet undergone testing in healthy patients.