T Cell Exhaustion Contributes to Immune Evasion in Chronic Lymphocytic Leukaemia

Abstract 1773

T cell exhaustion is a state of T cell dysfunction in response to chronic antigen exposure, marked by impaired effector function and the continued expression of inhibitory receptors such as Programmed Death 1 (PD-1) (Wherry EJ. Nat Immunol. 2011 Jun;12(6):492–9.). Because tumour growth in chronic lymphocytic leukaemia (CLL) occurs over a long period of time, we hypothesized that the continued exposure of T cells to a CLL-derived antigen could also lead to a state of T cell exhaustion. We therefore investigated whether T cell exhaustion is induced in CLL by using the Eμ-TCL1 transgenic (tcl1^tg) tumour transfer mouse model for this disease (Hofbauer JP, et al. Leukemia. 2011 Sep;25(9):1452-8) and by analyzing primary samples from CLL patients. We found that the number of PD-1⁺ T cells was increased in both CD4⁺ and CD8⁺ populations and in all lymphoid compartments examined of the Eμ-TCL1 transgenic (tcl1^tg) tumour recipient mice, but not in recipient mice receiving wildtype (WT) splenocytes showing that leukemic mice have an increased number of T cells displaying an exhausted phenotype that is induced by the presence of CLL cells. We next assessed the expression of the ligands for PD-1 on the surface of murine CLL cells. Peripheral CLL tumour cells showed only a modest increase in PD-L1 expression as compared to WT B cells. However, lymph node and spleen residing tumour cells showed a marked increase in PD-L1 expression, which suggests a microenvironment-induced upregulation of PD-L1 on tumour cells, e.g. by their close contact to accessory cells.

To validate our results on primary human CLL samples, we collected peripheral blood from 89 unselected CLL patients and 18 healthy donors and observed an increase in surface expression of PD-1 on the CD4⁺ and CD8⁺ T cell populations. While the percentage of PD-1⁺ CD4⁺ T cells in chemonaive patients was comparable to healthy donors, chemotherapy drastically increased the number of PD-1-expressing CD4⁺ T cells (63.81% ±19.75 vs 35.70% ±19.22; p<.001). In contrast, treatment apparently had little impact on the PD-1⁺ CD8⁺ population, indicating that PD1 induction on CD8⁺ T cells might be a more general CLL specific phenomenon. PD-1 expression in chemonaive patients did not correlate with the prognostic markers CD38, Zap-70 and IgVH mutations, however, it significantly correlated with Rai stage (CD4⁺: p<.001; CD8⁺: p=.003; spearman correlation). Our immune phenotyping analyses also included markers to distinguish between naïve, memory, and effector T cell subsets. We observed that PD-1 expressing T cells belong primarily to the memory compartment, characterized by the absence of CD45RA expression. PD-1 expression levels in T cells was independent from chronic human cytomegalovirus (HCMV) infection, indicating that increased PD-1⁺ T cell numbers in CLL do not simply reflect a higher incidence of HCMV infections, which has been described in this patient population.

As we speculated that PD1/PDL1 pathway might be exploited by CLL to evade a T cell dependent cytotoxic attack, we next wanted to know whether blocking this pathway by recombinant PD1 (rPD1) or PD-L1 specific Fab fragments in vivo would lead to reinvigoration of CLL specific T cells reflected in CLL clearance. We found that inhibition of PD1/PD-L1 binding in vivo in the tcl1^tg tumour transfer model led to a decrease in CLL load in the peripheral blood and in all analysed organs (lymph node, spleen, bone marrow, liver, lung) while in vitro treatment with rPD-1 or PD-L1 Fab fragment did not affect the viability of mouse tumour cells. Our results imply that in tumour bearing mice, a vast number of tumour-specific exhausted CTLs exist that can be immediately reactivated by PD1/PD-L1 blockage leading to cytolysis of the tumour, thereby representing a promising target for therapeutic intervention in CLL patients.