GabyOver the past year, artificial intelligence has driven major advances. It has moved well beyond creating images and text—researchers are now using machine learning to hunt for drugs that target the human biological clock.
AI reduces the time needed to search for anti-aging drugs
The more data a model has, the more accurate its predictions. AI has powered chess-playing programs, self-driving cars, and on-demand TV recommendations. This time, researchers used a specific algorithm to search for new senolytic drugs.
Senolytics are drugs that can slow aging and help prevent age-related diseases. They eliminate senescent cells—cells that are damaged and no longer divide but that release substances that cause inflammation.
Developing senolytics is expensive and slow. Noting that, Vanessa Smer-Barreto, a research fellow at the Scottish Institute of Genetics and Molecular Medicine, turned to AI.
She says generating your own biological data can be very costly and time-consuming, and researchers often have limited funds. Smer-Barreto says they used training data from existing studies and trained an AI model to speed up the search.
Using the algorithm, the team identified three promising candidate anti-aging drugs. They trained the model on examples of known senolytics and non-senolytics so it could distinguish between them, then used it to predict previously unknown molecules that might fight cellular aging.
Just 5 minutes—and the AI returned candidates
About 80 senolytics are known, but only two have been tested in humans. With huge costs involved, it can still take 10–20 years for anti-aging drugs to reach the market.
The Smer-Barreto team reviewed a wide range of studies but filtered the results. They limited their training set to 58 compounds, excluding any with ambiguous results.
They screened 4,340 molecules through the model, which returned results in just five minutes. The model flagged 21 top-scoring molecules it predicted were likely senolytics. Without AI, this screening would have taken weeks and cost far more.
The candidates were then tested on two types of cells: healthy and senescent. Of the 21 molecules, three killed senescent cells while sparing normal cells. These potential senolytics now face further testing to see how they interact with the body.
One step closer to anti-aging drugs
Although the research produced promising results, Science Focus describes it as just the beginning. Smer-Barreto says the next step is to collaborate with clinicians to test the identified drugs on samples of healthy human lung tissue.
With upcoming tests, the team hopes to see whether these drugs can combat aging in damaged organ tissues. Smer-Barreto says large doses probably won’t be necessary, especially early on. The drugs will first be tested on tissue models, then administered locally or in microdoses if needed.
Smer-Barreto says the drugs still have many stages to clear. If they ever reach the market, they will have to pass a series of safety tests first.
Although this data-driven method was used to search for anti-aging drugs, the same approach can be applied elsewhere. Similar models could help discover drugs for other diseases, such as cancer.
