The waking dead: targeting zombie cells with AI

Researchers have designed an AI model that can identify novel drug-like compounds with anti-aging properties that target ‘zombie’ (senescent) cells.

It may be possible to ditch those wrinkles while simultaneously abandoning your 15-step skincare routine, according to new research published in Nature Aging. With the help of a novel AI platform, a team of scientists from the Massachusetts Institute of Technology (MIT) and the Broad institute of MIT and Harvard (all MA, USA) discovered three anti-aging drug candidates that selectively target zombie cells with more advantageous medicinal chemistry properties, and selectively comparable efficacy to known current senolytic compounds. These findings have the potential to accelerate the field of senolytic drug development and anti-aging medicine.

Cellular senescence refers to cell cycle arrest, in which a cell permanently stops dividing but does not die, AKA, a zombie cell. This can be caused by a variety of factors, for example, oxidative stress, DNA damage, inflammation, mitochondrial dysfunction and telomere erosion. As we age, these zombie cells can accumulate throughout the body. And like zombies, they also don’t just sit there doing nothing in their not-quite-alive but not-quite-dead state. They can wreak havoc.

Senescent cells contribute towards aging, through leaking senescence-associated secretory phenotypes, hindering tissue repair and regeneration, triggering inflammation and also through their pathological role in age-related diseases such as Alzheimer’s disease, diabetes and cancer.

As humans, we seem to be obsessed with combatting aging and turning back the clock. Money, resources and research have been poured into anti-aging research, to hinder, halt or reverse age-related diseases and improve health.

Death could really be the path to life: senolytics are a class of drugs that selectively induce death of senescent cells, alleviating the effects of age-related diseases and improving metabolic function. Although senolytic compounds have demonstrated hopeful preclinical and clinical results, their application has been limited by toxic side effects and lack of bioavailability.

Attempting to overcome these limitations, the researchers from Integrated Biosciences (CA, USA) screened 2,352 compounds to predict the senolytic activities of over 800,000 molecules utilizing a trained deep learning neural network. They identified three novel highly selective and powerful senolytic compounds from this database, and discovered important favorable chemical properties including high-oral bioavailability and favorable toxicity profiles following hemolysis and genotoxicity tests. Molecular docking stimulations also revealed that the three compounds all bound and inhibited Bcl-2, a cell apoptosis regulator and chemotherapy target.

The team then tested one of these senolytic compounds on 80-week-old aged mice (approximately representative of 80-year-old humans) and discovered that the treatment significantly decreased the number of zombie cells and the gene expression of senescence-associated genes in the kidneys of the aged mice.

“One of the most promising routes to treat age-related diseases is to identify therapeutic interventions that selectively remove these cells from the body similarly to how antibiotics kill bacteria without harming host cells. The compounds we discovered display high selectivity, as well as the favorable medicinal chemistry properties needed to yield a successful drug,” explained Satotaka Omori, Head of Aging Biology at Integrated Biosciences. “We believe that the compounds discovered using our platform will have improved prospects in clinical trials and will eventually help restore health to aging individuals.”