The potent ability of allogeneic hematopoietic stem cell transplantation (Allo-HSCT) and donor leukocyte infusion (DLI) to cure leukemia remains the most striking example of the ability of the immune system to recognize and destroy tumors. Unfortunately, both allo-HSCT and DLI are often complicated by graft-versus-host disease (GVHD). In addition, durable responses to conventional DLI for acute leukemias have been disappointing. A better understanding of the mechanisms of alloreactive T cell-mediated anti-leukemia activity will be important for separating the GVL effect from GVHD. Adoptive transfer of selected subsets of T cells specific for miHA- or leukemia associated antigens might offer the chance to maximize GVL while minimizing GVHD. Using mouse models of human GVHD directed against miHAs, we recently demonstrated that antigen-experienced CD44loCD62LhiCD8+ T cells contain T memory stem cells that have greater ability than naïve T cells and mature memory T cells to proliferate and generate alloreactive effector cells and all memory T cell subsets (

Nature Medicine, 2005, 11:1299
). Using the same mouse model, we have now found that although B6/SJL mice receiving donor CD44hiCD8+ T cells (mature memory cells) primed against B6 mouse-derived myeloid leukemia C1498 cells do not develop clinical GVHD, most will die from C1498 leukemia by day 45 following injection of C1498 cells. Adoptive transfer of CD44loCD8+ T cells primed against C1498 leukemic cells caused clinical GVHD, but the majority of recipients (75%) survived long term free of C1498 leukemia. Surprisingly, the GVL effect of donor CD44loCD8+ T cells primed against C1498 leukemia cells was significantly inhibited when C1498 leukemia cell-primed CD44hiCD8+ T cells and CD44loCD8+ T cells were co-injected into B6/SJL mice, with only 25% of the mice surviving without leukemia. These results suggest that while the GVL effect is clearly mediated by antigen experienced CD44loCD8+ T cells, CD44hiCD8+ T memory cells primed against tumor cells are not only functionally defective in eliminating leukemia cells but are also potent inhibitors of alloreactive T cell-mediated GVL activity. We found that host-reactive effector memory CD8+ T cells produced 10-fold higher IL-10 than unstimulated naïve T cells and T memory stem cells, while CD8+ T memory stem cells expressed upregulated IL-10 receptors. These findings suggest that the inhibitory effect of mature memory T cells on alloreactive T cell-mediated GVL effect may be associated with increased production of IL-10 by mature memory cells and/or enhanced susceptibility of T memory stem cells to IL-10 secreted by mature memory cells. In addition, host dendritic cell activation of donor CD8+ naïve T cells progressively induced the generation of memory stem cells (CD44loCD62LhiSca-1hi), central memory cells (CD44hiCD62Lhi) and effector memory cells (CD44hiCD62Llo). CD8+ T memory stem cells displayed a TCR V-beta repertoire similar to that of unstimulated naive T cells. In contrast, both central memory and effector memory T cells showed a skewed TCR V-beta repertoire. Thus, selective elimination of suppressive CD44hiCD8+ T cells may represent an approach to augmenting GVL activity while preserving a diverse TCR V-beta repertoire.

Disclosure: No relevant conflicts of interest to declare.

Author notes

*

Corresponding author

Sign in via your Institution