A Vector-Space Representation of Cytoskeletal Drug Mechanisms for Intracellular Doppler Spectroscopy

Biodynamic imaging is sensitive to intracellular transport and has been successfully used to profile drug effects in 3-D cell culture. We report the use of biodynamic imaging to provide phenotypic profiles of cytoskeletal drugs that have a wide range of mechanisms of action (MoA). These profiles serve as “fingerprints” of the drug MoA and can be queried by machine learning clustering algorithms. In this study, 7 cytoskeletal drugs are used, including cytochalasins and latrunculin that inhibit the polymerization of actin, jasplakinolide that enhances actin polymerization, colchicine and nocodazole that inhibit microtubule polymerization, and taxanes that stabilize microtubules. In contrast, blebbistatin is a molecular motor inhibitor that inhibits ATPase activity. Biodynamic imaging is sensitive to subtle changes in intracellular motion, and drugs that affect the cytoskeleton are particularly strong inducers of biodynamic fingerprints. Biodynamic imaging uses short-coherence digital holography to profile cytoskeletal drug responses of 3D cultured tumor spheroids. Drug responses and their relations are examined in a high-dimensional space spanned by vectors of biomarkers and related with their mechanism of actions and targeted cytoskeletal components.