Improved molecular methods enabling multiplexed in vivo discovery of genes influencing tissue accumulation

Abstract

Cell therapy is a transformative approach to medicine, yet engineering cell-based therapeutics remains immensely challenging due to the complex behavioral requirements of different cells across diverse disease contexts. A key limitation is the identification of genetic modifications that improve a cell’s ability to carry out beneficial cellular behaviors in disease-specific applications. One such cellular behavior which enables the active delivery of therapeutic cells to their intended targets, cellular homing, is a universal requirement for systemically administered cells. The ability to rapidly identify tissue-specific homing genes would have broad applications for cell therapies, but remains immensely challenging. To address this, we developed two novel molecular methods to improve pooled in vivo screening pipelines, High-N Barcoding (HNB) and Selective Molecular Barcode Enrichment (SMBE). HNB and SMBE facilitate genetic library creation and barcode processing, respectively, to improve construction, resolution, sensitivity, reproducibility, throughput, and cost parameters of in vivo pooled screens. Using these, we carried pooled in vivo gain of function screens using a rationally designed “homing factor” ORF library to discover genes that enhance solid tumor accumulation phenotypes in CD4 T cells. We identified CXCR3 as the top enrichment in murine subcutaneous and metastatic B16F10 melanomas, and a subcutaneous CT26 colorectal tumor. Validation experiments by flow cytometry and “oHNB”, a variation of our HNB method, confirmed that CXCR3 transduction was sufficient at enhancing tumor accumulation, despite endogenous expression in activated T cells. Characterization experiments demonstrated that the CXCR3 enrichment phenotype is conferred independently of TCR-antigen matching, and suggests that enhanced proliferation contributes a minor secondary role. Despite its well characterized necessity in tumor homing, we present the first demonstration of sufficiency of forced CXCR3 overexpression enhancing tumor accumulation. Lastly, we screen a ~1500 member “surface-ome” ORF library in stem cell derived muscle progenitor cells, and identify several significantly enriched ORFs in a muscle injury model. To our knowledge, these are the largest in vivo pooled native ORF screens carried out to date. Taken together, the tools and findings presented in my thesis broaden the in vivo functional genomics tool kit, and advance our understanding of engineering strategies to improve tissue specific homing.