Precision-cut tumor tissue slices, a novel tool to study the tumor microenvironment interactions with chimeric antigen receptor (CAR) T cells

Abstract

Up until present day, chimeric antigen receptor (CAR)-T cell therapy has only been approved for hematological malignancies, as CAR-T cells do not show comparable efficacy in solid tumors. Therefore, understanding the features of the tumor microenvironment (TME), is key to improve efficacy of adoptive cell therapies (ACTs) against solid tumors. In this context, robust workflows, which dissect the complex interactions between CAR-T cells and the TME are still lacking. To address this need, we have established an ex vivo workflow co-culturing tissue slices from patient tumor resections with CAR-T cells. The workflow is composed of assessing several complementary attributes, such as cytokine release via flow cytometry, quantification of cell infiltration into the tumor and assessment of the regions of the tissue slice the CAR-T cell infiltrate into by using the MACSima™ imaging cyclic staining technology. Using this workflow it is possible to observe the behavior of CAR-T cells within the tumor and its TME, their infiltration into distinct tumor compartments, as well as to dissect the underlying molecular mechanisms that drive T cell migration, thanks to MACSima™ multiplexing technology and its ability to image several markers at the same time. Assessment of ovarian carcinoma tissue slices revealed substantial release of specific cytokines and increased infiltration of T cells in the tumor areas when CAR-T cells were added to the tissue slices as compared to non-engineered T cells. The establishment of this novel approach will enable researchers to better characterize the interaction between CAR-T cells and the TME. Tissue slices present an intrinsic heterogeneity, which is indeed an advantage compared to other in vitro models but can turn itself into complex results interpretation. Therefore, we recommend that any conclusion derived from this assay should be verified with complementary models.