Human iPSC-derived preclinical models to identify toxicity of tumor-specific T cells with clinical potential

Abstract

The balance between safety and efficacy of Tcell therapies remains challenging and Tcell mediated toxicities have occurred. The stringent selection of tumor-specific targets and careful selection of tumor-specific Tcells using Tcell toxicity screenings are essential. Invitro screening options against vital organs or specialized cell subsets would be preferably included in preclinical pipelines, but options remain limited. Here, we set up preclinicalmodels with human induced pluripotent stem cell(hiPSC)-derived cardiomyocytes, epicardial cells, and kidney organoids to investigate toxicity risks of tumor-specific Tcells more thoroughly. CD8+T cells reactive against PRAME, HA-1H, CD20, or WT1, currently used or planned to be used in phase I/II clinical studies, were included. Using these hiPSC-derived preclinical models, we demonstrated that WT1-specific Tcells caused on-target toxicity that correlated with target gene expression. Multiple measures of Tcell reactivity demonstrated this toxicity on the level of Tcells and hiPSC-derived target cells. In addition, phenotypic analysis illustrated interaction and crosstalk between infiltrated Tcells and kidney organoids. In summary, we demonstrated the benefit of hiPSC-derived models in determining toxicity risks of tumor-specific Tcells. Furthermore, our data emphasizes the additional value of other measures of Tcell reactivity on top of the commonly used cytokine levels.