SCAN-ACT: Adoptive T Cell Therapy Target Discovery Through Single-Cell Transcriptomics

Summary The FDA approval of T cell receptor-engineered T cells (TCR-T) for synovial sarcoma demonstrates the potential for adoptive T cell therapies (ACTs) in solid tumors. However, the paucity of tumor-specific targets without expression in normal tissues remains a major bottleneck, especially in rare cancer subtypes. Here, we present a comprehensive computational pipeline called SCAN-ACT that leverages single cell RNA sequencing and multi-omics data from tumor and normal tissues to nominate and prioritize targets for both chimeric antigen receptor (CAR)- and TCR-T cells. For surface membrane targets, SCAN-ACT proposes target pairs for bispecific Boolean logic-gated CAR T cells. For peptide-MHC targets, SCAN-ACT proposes intracellular peptides bound to a diverse set of human leukocyte antigens. We applied the SCAN-ACT pipeline to soft tissue sarcoma (STS), analyzing 986,749 single cells to identify and prioritize 395 monospecific CAR-T targets, 14,192 bispecific CAR-T targets, and 5,020 peptide-MHC targets for TCR-T cells. Selected targets were validated experimentally by protein expression and for peptide-MHC binding. Proposed targets and target pairs reflected the mesenchymal, neuronal, and hematopoietic ontogeny of STS. This work provides a robust data repository along with a web-based and user-friendly set of analysis tools to accelerate ACT development for solid tumors.