Advanced Cell Mapping Visualizations for Single Cell Functional Proteomics Enabling Patient Stratification

Abstract

Highly multiplexed single-cell functional proteomics has emerged as one of the next-generation toolkits for a deeper understanding of functional heterogeneity in cell. Different from the conventional population-based bulk and single-cell RNA-Seq assays, the microchip-based proteomics at the single-cell resolution enables a unique identification of highly polyfunctional cell subsets that co-secrete many proteins from live single cells and most importantly correlate with patient response to a therapy. The 32-plex IsoCode chip technology has defined a polyfunctional strength index (PSI) of pre-infusion anti-CD19 chimeric antigen receptor (CAR)-T products, that is significantly associated with patient response to the CAR-T cell therapy. To complement the clinical relevance of the PSI, a comprehensive visualization toolkit of 3D uniform manifold approximation and projection (UMAP) and t-distributed stochastic neighbor embedding (t-SNE) in a proteomic analysis pipeline is developed, providing more advanced analytical algorithms for more intuitive data visualizations. The UMAP and t-SNE of anti-CD19 CAR-T products reveal distinct cytokine profiles between nonresponders and responders and demonstrate a marked upregulation of antitumor-associated cytokine signatures in CAR-T cells from responding patients. Using this powerful while user-friendly analytical tool, the multi-dimensional single-cell data can be dissected from complex immune responses and uncover underlying mechanisms, which can promote correlative biomarker discovery, improved bioprocessing, and personalized treatment development.

Keywords: IsoCode chip; data visualization; single-cell proteomics; t-distributed stochastic neighbor embedding; uniform manifold approximation and projection.