Carcinogenic Human Metabolites Are Revealed Using “Metabokiller”

Current approaches to identify carcinogens are expensive as well as time-consuming; however, artificial intelligence can accelerate the prescreening of numerous compounds. Some computational approaches have been developed to predict carcinogenicity, but these focus mainly on genotoxicity or mutagenicity alone, leading to suboptimal predictive performances, as this does not account for the mode of action of all carcinogens. To more accurately detect carcinogens, Mittal, Mohanty, Gautam, Arora, and colleagues developed the Metabokiller platform, which quantitatively assesses the electrophilicity of compounds as well as their potential effects on proliferation, oxidative stress, genomic instability, epigenome alterations, and antiapoptotic response. Comparative analyses of Metabokiller against nine other widely accepted carcinogenicity prediction tools demonstrated the superior ability of Metabokiller to capture both biochemical and functional properties of carcinogens, leading to more accurate classification. Moreover, when Metabokiller was used to predict carcinogenic cellular metabolites in the Human Metabolome Database, 632 metabolites were predicted to possess carcinogenic properties using a stringent cutoff, with specific enrichment in tyrosine and tryptophan metabolism pathways, nucleotide metabolism pathways, as well as metabolic pathways associated with drug metabolism, xenobiotics, and oxidative stress being observed. Experimental validation of two previously uncharacterized human metabolites, 4-nitrocatechol (4NC), involved in aminobenzoate degradation, and 3,4-dihydroxyphenylacetic acid (DP), a metabolic intermediate of the tyrosine metabolism pathway, was performed in Saccharomyces cerevisiae and indicated that both metabolites induce genomic instability and mutagenesis, with 4NC stimulating ROS levels. Additionally, both 4NC and DP supported malignant transformation of normal human epithelial cells in soft agar, supporting their function as carcinogens. In conclusion, this study describes the Metabokiller platform that allows for the comprehensive identification of potential carcinogens within the human metabolome.

Mittal A, Mohanty SK, Gautam V, Arora S, Saproo S, Gupta R, et al. Artificial intelligence uncovers carcinogenic human metabolites. Nat Chem Biol 2022 Aug 11 [Epub ahead of print].

Note:Research Watch is written by Cancer Discovery editorial staff. Readers are encouraged to consult the original articles for full details. For more Research Watch, visit Cancer Discovery online at http://cancerdiscovery.aacrjournals.org/CDNews.