The average small molecule drug development program starts with a mechanism, an intended outcome such as inhibition, and then screening of as many molecules as possible from the libraries. Sometimes it is possible to make educated guesses as to what types of molecule are more likely to be useful, but often screening must be very broad and with little direction. In principle, low cost computation makes it possible to dramatically reduce the cost of discovery of useful molecules given a specific target mechanism. This shift from physical to in silico screening has been underway for a while, for example at Insilico Medicine, but is still a work in progress.
Small molecule medicine has its limitations, and in the future it seems likely that much of its present portfolio will be overtaken by gene therapies that can act precisely on target mechanisms: greater efficacy, far fewer side-effects, no expensive initial screening needed. Extending the life span of small molecule development into that era will require a dramatic reduction in its costs. At the end of the day the whole industry revolves around how much time and effort is needed to produce the prospect of a given benefit to patients.
The materials here are an example of the present state of the art when it comes to the use of in silico initial screening of small molecules. It needs something like the senolytics field to spur the development of better infrastructure for small molecule development. Small molecules can work well here, as demonstrated by animal data for dasatinib and quercetin, among others; there are many clear mechanistic targets for the clearance of senescent cells; there is a very large market, meaning the entire elderly population impacted by the burden of senescent cells in aged tissues; and the field is still young enough for an end to end development program to run for years and nonetheless find sizable profit at the end of it, if successful.
Senolytics are compounds that selectively induce apoptosis, or programmed cell death, in senescent cells that are no longer dividing. A hallmark of aging, senescent cells have been implicated in a broad spectrum of age-related diseases and conditions including cancer, diabetes, cardiovascular disease, and Alzheimer's disease. In their new study, researchers trained deep neural networks on experimentally generated data to predict the senolytic activity of any molecule. Using this model, they discovered three highly selective and potent senolytic compounds from a chemical space of over 800,000 molecules.
All three displayed chemical properties suggestive of high oral bioavailability and were found to have favorable toxicity profiles in hemolysis and genotoxicity tests. Structural and biochemical analyses indicate that all three compounds bind Bcl-2, a protein that regulates apoptosis and is also a chemotherapy target. Experiments testing one of the compounds in 80-week-old mice, roughly corresponding to 80-year-old humans, found that it cleared senescent cells and reduced expression of senescence-associated genes in the kidneys.
The accumulation of senescent cells is associated with aging, inflammation, and cellular dysfunction. Senolytic drugs can alleviate age-related comorbidities by selectively killing senescent cells. Here we screened 2,352 compounds for senolytic activity in a model of etoposide-induced senescence and trained graph neural networks to predict the senolytic activities of more than 800,000 molecules.
Our approach enriched for structurally diverse compounds with senolytic activity; of these, three drug-like compounds selectively target senescent cells across different senescence models, with more favorable medicinal chemistry properties than, and selectivity comparable to, those of a known senolytic, ABT-737. Molecular docking simulations of compound binding to several senolytic protein targets, combined with time-resolved fluorescence energy transfer experiments, indicate that these compounds act in part by inhibiting Bcl-2, a regulator of cellular apoptosis. We tested one compound, BRD-K56819078, in aged mice and found that it significantly decreased senescent cell burden and mRNA expression of senescence-associated genes in the kidneys. Our findings underscore the promise of leveraging deep learning to discover senotherapeutics.