- Record: found
- Abstract: found
- Conference Proceedings: found
Biosimulation-based target deconvolution of cancer metabolism
Abstract
Cancer cells exhibit mutations in metabolic hub genes and dysregulated metabolic phenotypes towards biomass production, consistent with enhanced cell replication and avoidance of growth suppression. Here we explored the hypothesis that the reprogramming of cellular metabolism initiates and/or sustains the proliferation and replicative immortality of cancer cells through the accumulation of oncometabolites (i.e., metabolites with pro-oncogenic functions). We specifically leveraged biological knowledge, multi-omics data, and artificial intelligence (AI) for the construction and parametrization of a compartmentalized genome-scale (~104 species and reactions) mechanistic model of cellular metabolism. The model consisted of coupled ordinary differential equations (ODEs) whose solution provided the temporal dynamics of the included molecular species. To navigate the huge combinatorial space underlying the modulation of the chemical reactions within the model, we developed a deep reinforcement learning-based algorithm capable of identifying biological targets by simulating how the combined modulation of these targets could reproduce a given phenotype (in other words, solving the inverse problem of target deconvolution). We considered common patterns of altered oncometabolites level in personalized phenotypes from the Pan-Cancer Analysis of Whole Genomes (PCAWG) database for about 1,000 patients across 25 cancer types. We predicted enzyme combinations targeting mitochondrial bioenergetics that are largely consistent with the literature and could potentially be of therapeutic value. Notably, already marketed drugs exist for many of the involved targets. Overall, our AI-assisted modeling and simulation platform provides a simple and explainable framework to study combination therapy and drug repurposing in cancer.
Content
Author and article information
References
Wishart David. Metabolomics and the Multi-Omics View of Cancer. Metabolites. Vol. 12(2)2022. MDPI AG. [Cross Ref]
Hanahan Douglas, Weinberg Robert A.. Hallmarks of Cancer: The Next Generation. Cell. Vol. 144(5):646–674. 2011. Elsevier BV. [Cross Ref]
DiNuzzo Mauro. How artificial intelligence enables modeling and simulation of biological networks to accelerate drug discovery. Frontiers in Drug Discovery. Vol. 2:2022. Frontiers Media SA. [Cross Ref]
Reznik Ed, Luna Augustin, Aksoy Bülent Arman, Liu Eric Minwei, La Konnor, Ostrovnaya Irina, Creighton Chad J., Hakimi A. Ari, Sander Chris. A Landscape of Metabolic Variation across Tumor Types. Cell Systems. Vol. 6(3)2018. Elsevier BV. [Cross Ref]
Lee GaRyoung, Lee Sang Mi, Lee Sungyoung, Jeong Chang Wook, Song Hyojin, Lee Sang Yup, Yun Hongseok, Koh Youngil, Kim Hyun Uk. Prediction of metabolites associated with somatic mutations in cancers by using genome-scale metabolic models and mutation data. Genome Biology. Vol. 25(1)2024. Springer Science and Business Media LLC. [Cross Ref]
Sainero-Alcolado Lourdes, Liaño-Pons Judit, Ruiz-Pérez María Victoria, Arsenian-Henriksson Marie. Targeting mitochondrial metabolism for precision medicine in cancer. Cell Death & Differentiation. Vol. 29(7):1304–1317. 2022. Springer Science and Business Media LLC. [Cross Ref]