Improving CAR-T efficacy against solid tumors by expanding lymph node reservoirs of “stem-like” CAR-T cells

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT Immunotherapy using CAR-T cells has induced dramatic responses in hematological malignancies, but extending this success to more common epithelial cancers, which cause the greatest mortality, has proved more challenging. In a phase 1 trial we conducted, CAR-T cells targeting the tumor-associated antigen ROR1 induced responses in patients with chronic lymphocytic leukemia but became terminally exhausted in breast and lung cancer patients, indicating that strategies to preserve CAR-T function are needed to achieve efficacy in solid tumors. Recent work has demonstrated that a subset of PD-1+Tcf1+ “precursor” exhausted cells (Tpex) is critical for mediating successful tumor responses to immunotherapy. Tpex retain a stem-like capacity for self-renewal, proliferation, and the ability to differentiate into effector-like cells. Interestingly, Tpex preferentially co-localize with antigen-presenting cells (APCs) and reside in tumor-draining lymph nodes (tdLN), where they may be protected from factors in tumors that drive terminal exhaustion, such as more chronic antigen stimulation, mTOR activation, and hypoxia. Unlike conventional T cells, CAR-T cells are activated by intact antigen expressed on tumor cells, not by peptide/MHC complexes expressed on both tumor cells and APCs. This design restricts CAR- T activation exclusively to tumors, where preservation of stem-like qualities may be impaired relative to tdLNs. To test this, we developed a Kras/p53 (KP) autochthonous model of ROR1+ lung cancer that recapitulates the dysfunction of ROR1 CAR-T cells we observed in patients. Using this model, we found ROR1 CAR-T cells did not accumulate in tdLNs and showed faster attrition of Tpex and terminal exhaustion in tumors compared to tumor-specific T cells (TCR-T) that were cultured identically prior to infusion. Tcf1+ CAR-T cells showed greater proliferative capacity, stem-like self-renewal, and tumor control compared to Tcf1- CAR-T cells, suggesting that strategies to preserve Tcf1+ CAR-Ts could significantly improve efficacy. Interestingly, CAR-T cells in tdLNs formed a small reservoir of “pre-exhausted” PD-1-Tcf1+ central memory cells that we could increase by vaccination and that may be able to generate more polyfunctional effectors than Tpex, which are epigenetically committed to producing hypofunctional effectors. We hypothesize that insufficient activation of CAR-T cells by APCs in tdLNs impairs maintenance of stem-like Tpex reservoirs, resulting in less durable tumor control, and that using vaccination to increase reserves of Tcf1+ CAR-T cells in LNs will overcome this barrier. In this proposal, we will use the KP model to determine: 1) whether CAR-T are impaired in maintaining a stem-like Tcf1+ reservoir relative to TCR-T cells and whether this is due to insufficient activation by APCs; and 2) whether activating CAR- T cells with endogenous APCs in LNs via a novel vaccine platform will improve numbers of Tcf1+ CAR-T cells, tumor control, and response to PD-L1 blockade. Our results will reveal whether CAR-T exhaustion in solid tumors is driven in part by impaired Tpex formation and will evaluate a novel vaccine platform to overcome these defects that can be translated to the clinic to enhance the efficacy of CAR-T cells in patients with solid tumors.