A promising strategy for tumor therapy is the adoptive transfer of tumor specific T cells which are endowed with chimeric antigen receptors (CAR). First generation CARs are constructed by single chain antibodies and as signal domain the ζ chain of the CD3 complex. However, clinical trials are disappointing as adoptive transferred T-cells showed only modest persistence in patients resulting in limited clinical activity. We there for hypothesized that CAR expressing T-cells in comparison to unmodified T-cells display signaling defects when stimulated via their CARs.
Cytomegalovirus(CMV)pp65 MHC I restricted CD8+ T-cells were generated, isolated by tetramer selection and modified with first generation CAR targeting CD19 and purified based on their receptor expression to more than > 95% purity. T-cell receptor (TCR) and CAR expression were quantified by Quantibright beads. Effector function of both T-cell populations were analyzed for specific lysis, cytokine production (IFN-g, TNF-a) and proliferation (CSFE) upon target cell stimulation. Phosphorylation of Erk, Jnk, p38 and PLC-γ was measured and analyzed with CBA Flexsets from BD. All statistical analyses have been performed using the statistical software package R. Signal peak intensities have been compared using the nonparametric wilcoxon rank sum test.
CMV-specific MHC-I restricted TCR as well as the CARs are expressed at same density levels and T-cells show equally lysis of targets either in the time of lysis onset as well the maximal lysis. In contrast, cytokine production (IFN, TNF-a) as well as antigen driven proliferation was reduced in CAR expressing T-cells when compared to CMV-specific CD8+ T-cells upon target exposure. PLC-γ was phosphorylated within minutes after target contact by CMV-specific CD8+ T-cells whereas CAR transduced CMV-specific CD8+ T-cells showed no significant phosphorylation of PLC-γ to target cell exposure. T-cell activated via CAR's demonstrated a statistically significant reduction of maximal phosphorylation in comparison to CMV-specific T-cells for ERK, for JNK and for p38. To exclude that CAR modification of CMV-specific CD8+ T-cells may impair signaling, CAR-CMV-specific CD8+ T-cells were exposed to CMVpp65 expressing targets. Killing, cytokine production and signal intensity were restored in comparison to parental CMV-specific CD8+ T-cells.
CAR expressing T-cells show functionally signs of split anergy by efficient target elimination but fails to produce significant levels of cytokines and do not proliferate in response to target stimulation. Split anergy is not due to reduced expression of the CAR's but due to a complete lack of phosphorylation of PLC-γ as well as reduced phosphorylation of MAP-kinases ERK, p38 and JNK. These results potentially explain why primary CAR expressing T-cells fail to show significant clinical efficacy. Analysis of adequate phosphorylation, as proposed here, may be a powerful tool to identify the most promising second generation CARs for clinical studies.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.