Integration of artificial intelligence and high-content screening enabled identification of drugs for long-term treatment of cerebral cavernous malformation disease

Résumé fourni par la source

Abstract Background Adults and children with cerebral cavernous malformations (CCMs) are at risk of experiencing lifelong complications such as hemorrhagic strokes, neurological deficits, and epileptic seizures. These complications can severely reduce quality of life. At present, there is no safe or effective therapeutic option for the long-term treatment of CCMs. Methods Using advanced artificial intelligence (AI) and machine learning models, powered by the Benevolent Platform™, we aimed to identify therapeutic drug targets for CCM pathology (e.g., CCM1, CCM2, CCM3). An AI integrative approach utilized various data types from biomedical entities, including diseases, genes, tissues, and biological mechanisms, together with CCM transcriptomic experimental data. High-throughput drug screening of AI-selected FDA-approved medications, analyses of mitochondrial morphology, and studies on pharmacokinetics, pharmacodynamics, and toxicology were conducted in CCM animal models to identify drugs that could potentially be repurposed for the long-term treatment of CCM disease. Results AI predicted the AMPK (AMP-activated protein kinase) and mTOR (mammalian target of rapamycin) pathways as potential therapeutic targets that contribute to CCM pathology. High-content screening validation revealed that the FDA-approved drug metformin, which acts as an AMPK agonist and mTOR inhibitor, can reverse changes in cell-cell junction organization and increase KLF4 expression, a marker for CCM, in human CCM endothelial cells in cultured assays. In addition, pharmacodynamic markers of metformin were observed in CCM mouse models ( Slco1c1-iCreERT2;Krit1 fl/fl ;Pten fl/wt and Slco1c1-iCreERT2;Pdcd10 fl/fl ) including reduced S6 kinase or ribosomal protein phosphorylation, a marker of decrease mTOR signaling, and increased AMPK phosphorylation, a marker of AMPK activation, that corresponded to reduced lesion burden. Pharmacokinetic and toxicological studies in CCM animal models showed that that metformin penetrates the brain and long-term administration has a favorable safety profile. We also demonstrated that brain endothelial cells in chronic CCM mouse models exhibit increased levels of the inflammatory marker VCAM-1, which is associated with altered mitochondrial phenotypes, as observed by immunofluorescence, MITO-tagging, and electron microscopy analysis. Additionally, we discovered that metformin and a potent AMPK activator, PF-06409577, can reverse mitochondrial phenotypic changes in brain endothelial cells and reduce the elevation of VCAM-1 expression associated with chronic CCM disease. Therefore, metformin can provide cytoprotection and may reverse the CCM endothelial phenotype by activating AMPK. Conclusions Predictions using AI technology and high-throughput drug screening, combined with pharmacokinetic, pharmacodynamic, and toxicological studies in CCM animal models, identified metformin as a promising drug candidate for repurposing for the long-term treatment of CCM disease. We propose that metformin enhances metabolic adaptation to brain vascular malformations by activating AMPK, which helps reverse mitochondrial fragmentation in brain endothelial cells. Graphical abstract

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.