Comparative Evaluation of STEAP1 Targeting Chimeric Antigen Receptors with Different Costimulatory Domains and Spacers.

ABSTRACT

We have developed a chimeric antigen receptor (CAR) against the six-transmembrane epithelial antigen of prostate-1 (STEAP1), which is expressed in prostate cancer, Ewing sarcoma, and other malignancies. In the present study, we investigated the effect of substituting costimulatory domains and spacers in this STEAP1 CAR. We cloned four CAR constructs with either CD28 or 4-1BB costimulatory domains, combined with a CD8a-spacer (sp) or a mutated IgG-spacer. The CAR T-cells were evaluated in short- and long-term in vitro T-cell assays, measuring cytokine production, tumor cell killing, and CAR T-cell expansion and phenotype. A xenograft mouse model of prostate cancer was used for in vivo comparison. All four CAR constructs conferred CD4+ and CD8+ T cells with STEAP1-specific functionality. A CD8sp_41BBz construct and an IgGsp_CD28z construct were selected for a more extensive comparison. The IgGsp_CD28z CAR gave stronger cytokine responses and killing in overnight caspase assays. However, the 41BB-containing CAR mediated more killing (IncuCyte) over one week. Upon six repeated stimulations, the CD8sp_41BBz CAR T cells showed superior expansion and lower expression of exhaustion markers (PD1, LAG3, TIGIT, TIM3, and CD25). In vivo, both the CAR T variants had comparable anti-tumor activity, but persisting CAR T-cells in tumors were only detected for the 41BBz variant. In conclusion, the CD8sp_41BBz STEAP1 CAR T cells had superior expansion and survival in vitro and in vivo, compared to the IgGsp_CD28z counterpart, and a less exhausted phenotype upon repeated antigen exposure. Such persistence may be important for clinical efficacy.