Background Chimeric Antigen Receptor (CAR) T cell therapy is an adoptive cell therapy in which cells are genetically modified to express a receptor that binds and kills tumor cells via a specific target antigen. This approach has demonstrated clinical success against hematologic B cell cancers; however, the therapy has failed against solid tumors due to several issues, such as the inability to identify tumor-specific antigens leading to on-target, off-tumor toxicities, and an immunosuppressive tumor microenvironment.

Methods To address these limitations, we are developing universal CAR T cells that target antigens on tumors and immunosuppressive cells via small molecule adaptors. Instead of directly binding to a target antigen, our universal CAR, SNAP-CAR, contains a mutated O⁶-alkylguanine-DNA alkyl transferase, SNAPtag. It is co-administered with one or more heterobifunctional small molecule adaptors containing a benzyl guanine (BG) motif. This adaptor forms a covalent bond with the SNAPtag receptor, arming it to recognize a cell surface antigen. We generated small molecule adaptor variants targeting surface receptors such as carbonic anhydrase IX (CAIX), folate receptor alpha (FOLR1), and beta (FOLR2) with different linker lengths. In vitro, we assessed CAIX and folate receptors targeting adaptors, by co-culturing SNAP-CAR T cells with a dose titration of CAIX and folate targeting adaptors and antigen-positive or negative cell lines or M2 polarized macrophages. We measured T cell activation and tumor cell lysis via flow cytometry post a 24-hour co-incubation. To address the on-target, off-tumor specificity issue for CAIX, we designed an approach to specifically activate the adaptor in the presence of tumor-microenvironment (TME) stimuli and tested it here by developing a chemically caged BG that only reacts with our SNAPtag receptor in the presence of reactive oxygen species (ROS) characteristic of the TME.

Results For 24-hour and 48-hour co-culture experiments, we observed T cell activation and tumor cell lysis. We demonstrated highly potent activity with low nanomolar concentrations of the adaptors on both tumor and tumor suppressor cell antigens. We observed FOLR2 expression on M2-polarized macrophages and demonstrated adaptor-mediated SNAP-CAR T cell proliferation and activation at 72 hours. We observed ROS-mediated decaging of the protective group on the caged BG and SNAP-CAR labeling in the presence of ROS.

Conclusions CAIX and FOLR1/FOLR2-targeting adaptors lyse antigen-positive cells and demonstrate robust SNAP-CAR T cell activation in an adaptor-dose-dependent manner. The ROS de-caging of the BG tag is highly efficient suggesting promise for this as a building block for a conditionally controlled SNAP-CAR adaptor.