Clinical trials in-a-dish for cardiovascular medicine

Abstract

Cardiovascular diseases persist as a global health challenge that requires methodological innovation for effective drug development. Conventional pipelines relying on animal models suffer from high failure rates due to significant interspecies variation between humans and animal models. In response, the recently enacted Food and Drug Administration Modernization Act 2.0 encourages alternative approaches including induced pluripotent stem cells (iPSCs). Human iPSCs provide a patient-specific, precise, and screenable platform for drug testing, paving the way for cardiovascular precision medicine. This review discusses milestones in iPSC differentiation and their applications from disease modelling to drug discovery in cardiovascular medicine. It then explores challenges and emerging opportunities for the implementation of 'clinical trials in-a-dish'. Concluding, this review proposes a framework for future clinical trial design with strategic incorporations of iPSC technology, microphysiological systems, clinical pan-omics, and artificial intelligence to improve success rates and advance cardiovascular healthcare.

Keywords: Artificial intelligence; Cardiovascular diseases; Clinical genomics; Clinical trials; Microphysiological systems; Organoids; Personalized medicine; iPSC.

Graphical Abstract

The landscape of iPSC technology is swiftly shifting from 2D models to more sophisticated 3D structures that recapitulate cardiac biology and diseases with unprecedented precision. This transformation presents exciting opportunities to conduct clinical trials ‘in-a-dish'. By harnessing patient-specific iPSC derivatives like cardiac organoids, trialists may be able to effectively filter out toxic drugs, monitor drug behaviours, and select trial participants who respond to drug candidates. Furthermore, the integration of cutting-edge synergistic technologies, such as artificial intelligence, microphysiological systems, and clinical pan-omics, is instrumental in facilitating the implementation of clinical trials ‘in-a-dish’, ultimately leading to improved success rates.

Figure 1.

The technological leap of human-induced pluripotent stem cell models from 2D to 3D for modelling healthy and diseased heart. Human-iPSCs can be reprogrammed from easily accessible biopsies like blood, skin, or urine, collected from individuals with or without cardiovascular diseases. There is a growing number of protocols tailored to differentiating iPSCs towards specific cardiovascular cell types, including cardiomyocyte, endothelial cell, fibroblast, smooth muscle cell, pericyte, and a range of inflammatory cells. Encouragingly, significant progress has been made in enhancing the maturity of iPSC-derived cells. A recent notable breakthrough is the development of induced pluripotent stem cell-derived 3D models, including engineered heart tissue, assembloids, and cardiac organoids, which hold immense potential for faithfully replicating human heart morphology and function ‘in-a-dish’. Continual refinement of these iPSC-based models promises to deepen our understanding of human heart development and a broad spectrum of cardiovascular diseases, such as congenital heart diseases, cardiovascular diseases with or without genetic drivers, and cardiotoxicity. LVNC, left ventricular non-compaction; DCM, dilated cardiomyopathy; HCM, hypertrophic cardiomyopathy; MI, myocardial infarction; CM, cardiomyocyte; EC, endothelial cell; SMC, smooth muscle cell; EHT, engineered heart tissue; OFT, outflow tract; RV, right ventricle; LV, left ventricle; AVC, atrioventricular canal; CVDs, cardiovascular diseases. Created with BioRender.com

Figure 2.

Artificial intelligence accelerates ‘clinical trials in-a-dish’. Artificial intelligence plays a pivotal role in modern drug discovery, facilitating virtual screening and drug design. Artificial intelligence-driven models, like the graph neural network, facilitate virtual screening where molecular property prediction models analyse compounds within vast chemical libraries. Moreover, generative artificial intelligence methods have the capability to design novel compounds from scratch. Computational models also predict drug-like properties, expediting hit identification and evaluation in drug discovery pipelines. In iPSC-based clinical trials, artificial intelligence enhances various stages, from reprogramming and differentiation to assay analysis. Convolutional neural networks automate quality control by discerning successfully reprogrammed or differentiated cells in iPSC culture. Moreover, artificial intelligence aids morphological analysis in iPSC assays and organoid models, detecting drug effects and facilitating drug discovery. Lastly, artificial intelligence facilitates the analysis of functional measurements to personalize drug evaluation based on patient genetics, thereby improving the efficacy of iPSC clinical trials. Created with BioRender.com

Figure 3.

Integrative induced pluripotent stem cell-based trial: from dish to bedside. The integration of iPSCs with synergistic technologies holds promise to improve clinical trials across various phases. In Phase 1, artificial intelligence assists in designing drug-like molecules, while large-scale human-iPSC-cardiomyocyte cohorts aid in evaluating cardiotoxicity. iPSC-cardiac organoids cultured in a dish serve as a reliable platform for pharmacokinetics/pharmacodynamics studies, and multiple organ chips allow the measurement of absorption, distribution, metabolism, and excretion features. Moving to Phase 2, patient-derived iPSC-cardiac organoids help stratify trial participants into non-responders and responders. Furthermore, clinical pan-omics techniques enable the discovery of molecular markers for drug responsiveness, guiding trialists to recruit responders for Phase 3 trials thus improving success rates. Additionally, iPSC technology facilitates the implementation of clinical trials for ultra-rare diseases through collecting patient-specific samples, followed by iPSC reprogramming, biobanking, cardiac organoid differentiation, and drug testing. Created with BioRender.com