Microfluidics-based single-cell characterization of innovative cellular immunotherapies

Chimeric Antigen Receptor (CAR)-T cells have fundamentally transformed the landscape of cancer treatment, demonstrating remarkable clinical efficacy in hematological malignancies. Despite achieving unprecedented clinical success, major challenges hamper their broader clinical implementation, including complex mechanisms of action, characterization methodologies, and manufacturing scalability. CAR-T cell products exhibit substantial phenotypic and functional heterogeneity, which current approaches struggle to capture at the single-cell level. As this heterogeneity critically impacts therapeutic efficacy, there is a pressing need for novel technologies capable of accurately profiling these complex functional dynamics over time. To tackle this challenge, this thesis leverages a novel microfluidic analytical platform combining live-cell fluorescence imaging with an AI-driven analytical pipeline for the simultaneous profiling of thousands of miniaturized tumor-immune co-cultures by integrating cytotoxicity, phenotypic, and kinetic readouts at single-cell resolution. To mitigate the risk of on-target off-tumor toxicity, adapter CAR-T platforms such as the UniCAR-T system have emerged as a powerful alternative for enhanced controllability and targeting flexibility. However, their single-cell dynamics and serial killing capabilities remain unexplored. This work investigates the single-cell functional landscape of UniCAR-T cells targeting CD33+ acute myeloid leukemia, uncovering distinct functional subsets coexisting within the same UniCAR-T cell population, including a highly potent subset of serial killers. These findings underscore a degree of single-cell functional heterogeneity not detectable by conventional bulk assays. In addition, the present work advances the development of a novel microfluidics-based cytokine release assay with the potential to integrate multifunctional readouts on cytotoxicity, phenotype, and cytokine secretion on individual immune cells. Overall, the present study offers a novel data-driven framework for the single-cell profiling of immune cell therapies while providing an in-depth characterization of adapter UniCAR-T cells, with the ultimate goal of supporting the design and development of next-generation immune cell therapies.