Droplet microfluidic profiling of NK cell cytotoxicity with machine learning-enabled target-cell death analysis

Area of Science:

• Immunology
• Cell Biology
• Biotechnology

Background:

• Predicting Natural Killer (NK) cell immunotherapy efficacy is difficult due to cell variability and tumor microenvironment (TME) suppression.
• Current methods lack the resolution to capture single-cell functional dynamics at scale.

Purpose of the Study:

• To develop a high-throughput single-cell platform for quantifying NK cell cytotoxicity and understanding factors influencing their function.
• To compare the efficacy of expanded NK cells (exNK) versus peripheral blood NK cells (pbNK) under various conditions, including TME exposure.

Main Methods:

• A droplet microfluidic platform was developed, integrating machine learning for real-time target cell detection and classification.
• Single-cell cytotoxicity trajectories were mapped by analyzing effector-target interactions and NK cell morphology.
• The platform quantified NK cell cytotoxicity, serial killing, killing kinetics, and attachment outcomes.

Main Results:

• Expanded NK cells (exNK) demonstrated superior cytotoxic activity, serial killing, and faster killing dynamics compared to peripheral blood NK cells (pbNK).
• Ascites-conditioned peripheral blood NK cells (pbNK-asc) exhibited significantly reduced function across all measured cytotoxicity metrics.
• Expanded NK cells exposed to ascites TME (exNK-asc) maintained partial functionality, suggesting resilience conferred by ex vivo expansion.

Conclusions:

• The developed single-cell platform provides a scalable framework for assessing NK cell function and optimizing immunotherapies.
• Ex vivo expansion enhances NK cell potency and resilience against the suppressive tumor microenvironment.
• This technology offers valuable insights into NK-cancer cell interactions for improving off-the-shelf NK cell-based therapies.