Abstract 3875: <i>De novo</i> spatial proteomic profiling of immune synapses using machine learning-guided microscoop

Abstract The spatial proteome of the immune synapse (IS) between a target cell and a lymphocyte is fundamentally important to understand the mechanism of cell-mediated immunity for both immuno-oncology and therapeutic applications. In this research, we used Microscoop, a fully-automatic microscope system integrated with a machine learning-based algorithm, to best determine ISs for proteomic mapping. We used Raji B cells as antigen-presenting cells (APCs) and induced the formation of ISs by incubating with Jurkat T cells. Multiple IS images were applied to train our algorithm using convolution neural network-based deep learning. A sequential process including fluorescence imaging, deep learning-enabled pattern generation, and photochemical labeling was implemented to achieve spatial-specific biotinylation of the IS proteins. Moreover, Microscoop is capable to repeat the process automatically on thousands of fields of view to label sufficient immune-synaptic proteins with biotin-tag for further enrichment allowing protein identification with mass spectrometry. We have successfully labeled and isolated proteins from spatially reorganized interfaces between T cells and APCs. Following MS-based proteomic analysis, several hundreds of proteins were identified, including the proteins known to be specifically associated with T-cell receptor (TCR) activation such as LCK, one of the major factors involved in TCR signaling at ISs. More interestingly, we identified several proteins novel for ISs, including proteins involved in phosphatidylinositol signaling. Our data showcases the capability of subcellular de novo spatial proteomics of the Microscoop technology, revealing novel factors responsible for initiating the immune response of a lymphocyte and shedding light on immune checkpoint signaling and tumor immunotherapy. Citation Format: You-Pi Liu, Weng Man Chong, Harry Huang, Yi-De Chen, Chia-Wen Chung, Hsiao-Jen Chang, Chih-Wei Chang, Jung-Chi Liao. De novo spatial proteomic profiling of immune synapses using machine learning-guided microscoop [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3875.