Organs-on-Chips: a new paradigm for safety assessment of drug-induced thrombosis

Blood hypercoagulability and thrombosis have been observed in patients during clinical trials of candidate drugs, yet these safety risks are seldom identified during preclinical testing, leading to increased mortality and morbidity, and increased attrition rates in the clinic. Current preclinical models — standard cell cultures, flow chambers, and animal models — are often ill-equipped to predict thrombosis in the clinic. In vitro models are typically assembled without critical biomechanical forces, such as shear stress and mechanical strain, or relevant cytoarchitecture, such as interactions between different tissue types, which are essential to physiological function. In addition, animal models not only are expensive and costly but also possess inherent cross-species biological differences that are difficult, if not impossible, to reconcile for accurate human predictions. As a preclinical platform with a potentially higher predictive value, organs-on-chips are fluidic systems that reproduce organ-level function via cellular components of human origin, tissue–tissue interfaces, and dynamic mechanical forces. Compared with other current preclinical models, organs-on-chips combine the advantages of tunability and ease of biochemical, histological, and image analysis, while bypassing difficulties in cross-species translation. In this review, we delineate the limitations of current preclinical models, which are often unable to predict drug-induced thrombosis, and report some recent advancements in Organs-on-Chips technology that represent a promising alternative for modeling tissue-specific thrombotic events and derisking next-generation drug discovery.