ABSTRACT
A potential cause of cancer relapse is pretreatment chemoresistant subpopulations. Identifying targetable features of subpopulations that are poorly primed for therapy-induced cell death may improve cancer therapy. Here, we develop and validate real-time BH3 profiling, a live and functional single-cell measurement of pretreatment apoptotic sensitivity that occurs upstream of apoptotic protease activation. On the same single cells, we perform cyclic immunofluorescence, which enables multiplexed immunofluorescence of more than 30 proteins on the same cell. Using cultured cells and rapid ex vivo cultures of colon cancer patient-derived xenograft (PDX) models, we identify Bak as a univariate correlate of apoptotic priming, find that poorly primed subpopulations can correspond to specific stages of the cell cycle, and, in some PDX models, identify increased expression of Bcl-XL, Mcl-1, or Her2 in subpopulations that are poorly primed for apoptosis. Last, we generate and validate mathematical models of single-cell priming that describe how targetable proteins contribute to apoptotic priming.