Lattice light-sheet microscopy Multi-Dimensional Analyses (LaMDA) of T-cell receptor dynamics predict T-cell signaling states

Abstract

Abstract Cell surface receptors dynamically change in response to cell signaling, but a direct linkage between receptor dynamics and cell state has not been established. Here we report the development of lattice light-sheet microscopy multi-dimensional analyses (LaMDA), a pipeline that combines high spatiotemporal-resolution four-dimensional lattice light-sheet microscopy, machine learning, and diffusion maps to analyze T-cell receptor (TCR) dynamics and predict T-cell signaling states without the need for complex biochemical measurements. LaMDA images thousands of TCR microclusters on the surface of live primary cells to collect high-dimensional dynamic data for machine learning, which extracts key dynamic features to build predictive diffusion maps. LaMDA spatiotemporally reveals global changes of TCRs across the 3D cell surface, accurately differentiates stimulated cells from unstimulated cells, precisely predicts attenuated T-cell signaling after CD4 and CD28 receptor blockades, and reliably discriminates between structurally similar TCR ligands. We anticipate broad usage of this approach for other receptors and cells, as well as for guiding the design and development of future immunotherapies for cancer, infection, and autoimmunity.