Background
Chimeric antigen receptor (CAR) T cell therapies have been used to effectively treat several hematological malignancies, resulting in six FDA approvals. However, even within these approved therapies, significant variation in response exists between patients. Interestingly, similar variation in CAR T cell efficacy is also observed in the laboratory setting between CAR T cells generated from different healthy peripheral blood mononuclear cell (PBMC) donors. Although it is unclear what causes variation in either setting, one underlying source of variation includes differences in T cell function between different donors and patients. The goal of this study is to determine the functional attributes of T cells that are critical for effective CAR T cell antitumor response in vivo.
Methods
We generated B7-H3 CAR T cells expressing a CD28 costimulatory domain and a CD3 signaling domain from a range of healthy PBMC donors. We next evaluated each donor’s CAR T cell performance in functional assays measuring 1) pre-antigen expansion, 2) antigen-induced expansion and persistence (repeat stimulation assay), 3) cytokine secretion (Multiplex ELISA), and 4) killing efficiency (MTS cytotoxicity assay). Antitumor efficacy of each donor’s CAR T cells was also evaluated in an intraperitoneal LM7 osteosarcoma model. A scoring system was developed to assess the performance of each donor’s CAR T cells in every functional assay, and scores were correlated with in vivo antitumor efficacy to determine critical T cell functional attributes of effective CAR T cell donors.
Results
Significant variation in functional assay performance was observed between CAR T cells generated from different donors. Further, we also observed significant variation in the survival of mice treated with CAR T cells from different donors. Scoring the donors based on their performance in functional assays led to a range of scores for each assay which could then be correlated with in vivo efficacy. Although data is still being generated for additional donors, to date, a significant positive correlation has been established between CAR T cell persistence scores and in vivo survival.
Conclusions
We have established that significant variation exists between CAR T cells generated from different healthy donors in both in vitro functional assay performance and in vivo antitumor efficacy. Further, we have demonstrated that specific functional attributes, namely persistence, are correlated with better antitumor efficacy in vivo. We are continuing to test additional healthy donors to further evaluate variation between donors and extending these studies to CARs with different antigen specificity and to patient samples.