PhD Student The University of Hong Kong, Hong Kong
Abstract: Chimeric antigen receptor macrophages (CAR-M) have demonstrated efficacy in clinical HER2+ solid tumors, yet the development of macrophage-optimized CAR architectures remains underexplored. A critical challenge lies in identifying design principles for CAR structures that maximize macrophage-specific functions such as tumor infiltration, persistence, and microenvironment remodelling. To address this, we developed a high-throughput platform to systematically evaluate synthetic macrophage-specific engulfment receptors (SMSERs) incorporating diverse domains: macrophage specific signalling peptides, hinge regions, transmembrane (TM) segments, and intracellular domains (ICDs).
Using Golden Gate modular cloning, we generated three CAR libraries: single ICD library containing 491 constructs, dual ICD library containing 2,205 variants and triple ICD library containing 1,145 CARs. An in vivo screening platform was established using intra hepatically implanted HepG2 tumors to select CAR-M variants with enhanced tissue infiltration and persistence. Single-cell RNA sequencing analysis of tumor-infiltrating CAR-M further mapped receptor designs to functional phenotypes like cytokine secretion and phagocytosis.
To extrapolate design rules, neural networks analyzed screening data to predict optimal CAR configurations. This integrated approach—combining empirical testing with machine learning—reveals structural principles governing CAR-M performance.
Our strategy provides a framework for rational design of next-generation CAR-M therapies, enabling personalized receptor architectures tailored to specific immune cell functionalities. This advances the translation of CAR-M from preclinical validation to clinical application in solid tumors.
Funding Source: This research is supported by Health@InnoHK, the Innovation and Technology Commission of the Government of the HKSAR.
