ABSTRACT: Chimeric antigen receptor T cells (CAR-Ts) are transformative cellular therapies; however, current forms of CAR-T face multiple major hurdles. These include antigen loss, tumor microenvironment suppression, trafficking, proliferation, toxicity, and persistence. An efficient way to overcome these challenges is to engineer thousands of different CAR-T variants, and systematically select the best ones. To achieve this, we develop CLASH, a versatile platform that enables high-efficiency massively parallel CAR-T engineering. In CLASH, pooled AAVs mediate simultaneous gene editing and precise CAR knock-in via massively parallel homology-directed repair (HDR). Such library-scale HDR events produce a pool of stably integrated CAR-T variants each with targeted immune gene editing. In vitro or in vivo CLASH experiments using human CD8 and CD4 T cells subject the CAR-T pools to long-term co-culture or animal models of cancer, enabling unbiased selection of favorable CAR-T variants with enhanced cancer killing and persistence. Next-generation sequencing of genomic integrations deconvolves the time-course dynamics of knock-in CAR-T variants and identify winning CAR-Ts that survived specific selections. Validations of convergent top genes JADE1, PELI1, EHMT1, NLRP10, KDM4E, TET2 and PRDM1 show that their perturbations in CAR-Ts modulate T cell characteristics such as proliferation, exhaustion and memory phenotypes. A PRDM1 exon3-skip mutant CAR-T exhibits increased proliferation capacity, central memory cell phenotype and longevity, resulting in high-performance in vivo therapeutic efficacy in adoptive cell therapy models across multiple cancer models including solid tumor. Time-course transcriptomics and immune profiling and molecular interrogations demonstrate that PRDM1 exon3-skip CAR-Ts have multiple rewired gene expression patterns and immunological programs. Interrogations of the epigenetic mechanism via histone array, co-immunoprecipitation and genome-wide chromatin binding by Cut-and-Run reveal that deletion of exon-3 leads to disruption of histone H4 binding to the PR domain, thereby altering the regulation of critical downstream immune factors including SELL, CCR7, STAT1, STAT6, CDCA7 and IL7R. The versatility of CLASH is broadly applicable to the engineering of various desired CAR features in a wide range of cell therapies.
