Abstract
Adoptive immunotherapy using T cells armed with chimeric antigen receptors (CAR) has proven extremely effective against CD19+ B-lineage acute lymphoblastic leukemia (ALL) with remission rates as high as 70-90% in recent clinical trials in relapsed/refractory patients. CD8 T cells are typically thought of as the primary antitumor effector cells in an adoptively transferred product due to their potent cytolytic capabilities, whereas CD4 T cells are thought to primarily provide "help" to enhance CD8 T cell activity via cytokine production. However, CARs are synthetic constructs that likely alter the functionality of T cells. Further, CAR T cells are endowed with 2 specificities, one through the CAR and one through the endogenous T cell receptor (TCR). To evaluate the biology of CAR T cells, we sought to evaluate the impact of CAR expression on the functionality of T cells and to study the impact of TCR on CAR T cell activity using both human CAR T cells in a xenograft model and murine CAR T cells in a syngeneic murine model. A human or murine second-generation anti-CD19 scfv/CD28/CD3ζ CAR was transduced into human or mouse CD8 (CAR8) and CD4 (CAR4) T cells, respectively, and tested against pre-B ALL cell lines in human murine xenografts or syngeneic models. Surprisingly, human CAR4 cells alone had equivalent ability to eradicate the Nalm6 ALL in vivo as CAR8 cells. Although CAR8 cells more rapidly cleared leukemia (2 vs 4 days), relapse eventually occurred. In contrast, CAR4 cells eradicated leukemia more slowly, but persisted longer and prevented relapse. In mice receiving CAR4+CAR8 cell products, only CAR4 cells were detectable at day 55. Next we utilized our syngeneic murine system to study CAR4 and CAR8 activity in an immunocompetent system using TCR transgenic T cells with known TCR specificity against the male histocompatibility antigen, HY. As expected, when stimulated through the TCR, CAR8 cells produced cytokines and degranulated, as manifested by CD107a expression, whereas CAR4 cells only demonstrated a cytokine production response. However, murine CAR4 cells activated through the CAR receptor develop "CD8-like" cytolytic features characterized by degranulation and were able to completely eradicate leukemia in vivo when administered without CAR8 cells, similar to the human CAR4 cells. We then evaluated the effect of TCR antigen in vivo on CAR T cell function by comparing activity in male (HY+) vs female (HY-) recipients. CAR4 cells were curative in both male and female recipients. However, CAR8 treatment has no anti-tumor activity in male recipients indicating that CAR8 cells are very susceptible to the negative effects of TCR signaling. TCR antigen availability in vivo greatly increased the number of infused CAR4 cells (P=0.003, male vs female recipients), while they significantly decreased the number of CAR8 cells (P=0.0023, male vs female recipients) after 1 week. Interestingly, we found that there was significant down-regulation of CAR expression on both CAR4 and CAR8 cells in male recipients (HY+). PD-1 and Tim3 were expressed at low levels on CAR4 cells and CAR8 cells in female recipients. However, there was significantly higher PD-1 (P=0.0079) and TIM3 (p=0.014) expression on CAR8 cells in male vs female recipients, whereas there was no difference in PD-1 or Tim3 on CAR4 cells recovered from male vs female recipients. These data indicate that CAR4 cells are more functional and become less exhausted than CAR8 cells in the presence of TCR antigen. Finally, we evaluated the long-term persistence of CAR4 and CAR8 cells in our syngeneic model in which there is no xenogeneic reactivity allowing for long-term monitoring. At day 60 and day 80, CAR4 cells were detectable in bone marrow and spleen, whereas CAR8 cells had completely disappeared. Despite the long-term persistence of CAR4 cells in recipients with or without TCR antigen, CAR4 cells are not completely immune to effects of TCR activation. In male recipients after 60 days, CAR4 cells also start to express high PD-1 levels and have decreased CAR T cell counts. These findings have important implications for CAR therapy, particularly in the allogeneic setting where the presence of TCR antigen may be more likely to be present in the recipient. In addition, these results suggest that CAR4 cells alone are cytolytic, equally potent to CAR8 cells at eradicating leukemia in vivo, and may be superior due to better persistence and reduced susceptibility to exhaustion.
Disclosures
No relevant conflicts of interest to declare.
Structure of the Signal Transduction Domain in Second-Generation CAR Regulates the Input Efficiency of CAR Signals
