Groundbreaking Discovery: Artificial Intelligence Unveils Potent Anti-Aging Compound!

Finding new drugs – “drug discovery” – is a costly and time-consuming task. But a type of artificial intelligence called machine learning can greatly speed up this process.

Vanessa Smear-Barreto, a research fellow at the Institute of Genetics and Molecular Medicine at the University of Edinburgh, and her colleagues recently used this technique to find three promising candidates for senolytic drugs – drugs that slow down aging and prevent age-related diseases.

Senolytics work by killing aging cells. These are “living” (metabolically active) cells, but they can no longer reproduce, hence their nickname: zombie cells.

These cells suffer from DNA damage—skin cells damaged by sunlight, for example—so stopping reproduction stops the damage from spreading.

But aging cells are not always good. They secrete a cocktail of inflammatory proteins that can spread to nearby cells. Over the course of our lives, our cells are exposed to many attacks, from ultraviolet rays to exposure to chemicals, and so they accumulate.

A large number of senescent cells is associated with a number of diseases, including type 2 diabetes, coronavirus, pulmonary fibrosis, osteoporosis and cancer.

Studies in mice have shown that killing senescent cells with senogenic drugs can alleviate these diseases. These drugs can kill zombie cells while keeping healthy cells alive.

Currently, about 80 anti-aging drugs are known, but only two of them have been tested in humans: a combination of dasatinib and quercetin. It would be great to find more geriatric drugs that could be used for various diseases, but it will take 10 to 20 years and billions of dollars to enter the market.

Researchers at the University of Edinburgh and the Spanish National Research Council IBBTEC-CSIC in Santander, Spain wanted to see if we could train machine learning models to identify new drug candidates for treating aging.

To do this, they provided the AI ​​models with examples of both a known aging state and a non-aging state. Models have learned to distinguish between them and can be used to predict whether previously unseen molecules can cause aging.

When solving a machine learning problem, the data is usually tested against a set of different models first, because some of them tend to perform better than others.

To determine the best performing model, at the beginning of the process, a small portion of the available training data is separated and hidden from the model until the training process is completed.

This test data is then used to determine how many errors the model allows.

The researchers identified the best model and developed it to make predictions. The AI ​​model identified 21 highest-scoring molecules that it believes are highly likely to age. If we were to test 4,340 of the original molecules in the lab, it would take at least several weeks of hard work and only £50,000 to buy the compounds, not including the cost of experimental equipment and setup.

These drug candidates were then tested on two types of cells: healthy and aging. The results showed that among the 21 compounds, three (priproleucine, oleandrin and genexetin) were able to destroy senescent cells while keeping the majority of normal cells alive. These new anti-aging drugs were then further tested to learn more about how they work in the body.

The most detailed biological experiments showed that of the three drugs, oleandrin was more effective than the most famous anti-aging drug of its kind.

The potential ramifications of this interdisciplinary approach involving data scientists, chemists, and biologists are enormous. Given enough high-quality data, AI models can accelerate the amazing work that chemists and biologists are doing to find cures for diseases, especially those that are unknown.

Having confirmed their effectiveness in removing senescent cells, the researchers are now testing three aging candidates in human lung tissue.

Source: Science Alert