Beyond viability: Post treatment single cell and spatial transcriptomic profiling in mixtures of PRISM barcoded cancer cell lines

PRISM is a highly multiplexed, genomic barcode-based cell line viability technology of nearly 1000 solid tumor and hematopoietic cell lines. The unique cell line barcodes enable the cell lines to be multiplexed into pools of dozens to hundreds of different cell lines per well or dish, which greatly reduces the time and costs required to profile the viability effects of compounds or biologics across 900+ cell lines. Leveraging underlying genomic characterization of these cell lines enables the discovery of biomarkers of sensitivity and resistance. In addition to enabling viability screening in 900+ cell lines at scale, the PRISM barcoding technology also enables us to profile post-perturbation RNA transcript and protein changes in mixtures of cell lines by adapting commercially available single cell sequencing (10x Genomics Flex v2) and spatial transcriptomic assays (Element Biosciences AVITI24). Expanding the phenotypic readout capabilities of the PRISM technology beyond viability to include gene and protein changes will provide deeper insight into drug mechanisms of action, resistance, off-target effects, and heterogeneity of response. Here we present our work developing multiomic phenotypic profiling assays in barcoded PRISM cell line pools. We leverage 10x Genomics Flex v2 (for genomewide RNA single cell sequencing and PRISM barcode identification) in conjunction with the Flex-compatible Proteintech Human Discovery antibody panel (to measure abundances of 300+ proteins) in response to perturbations. In a pilot of 100 cell lines spanning 20+ lineages, we reliably detect 99 cell lines, recovering on average ~7000 unique genes per cell. In parallel, we are exploring the potential of Element Biosciences AVITI24 for spatial transcriptomic readouts in pools of PRISM cell lines. In a pilot of 25 cell lines spanning 15 lineages, we detect each of the 25 cell lines and although we recover fewer counts per cell, we recover morphology features from 6 cell paint stains (membrane, nucleus, actin, mitochondria, golgi, and ER). In both assays, we treated with a pan-RAS inhibitor and observed a concordant gene signature in differentially expressed genes. The 10x Flex v2 assay has the benefit of depth (detecting more transcripts per cell), while the Element imaging workflow enables culturing cells directly on the flowcell and provides spatial and morphological features. Thus, the two assays are highly complementary. We identify commonalities as well as distinct insights obtained by the two technological approaches and identify cell line specific responses.