Central Memory T Cells That Express High Level Of CD44 Mediate Strong Anti-Tumor Response Without Gvhd

Several groups have demonstrated directly and indirectly that central memory T cells (T_CM) have decreased ability to induce graft-versus-host disease (GVHD) compared with naïve T cells (T_N). We have previously demonstrated that T_CM expressing high level of CD44 (CD44^high T_CM) do not cause GVHD. In this study, we further defined the role of CD4⁺ and CD8⁺ subsets of this population in GVHD using the same GVHD model as we used previously (C57BL/6→BALB/c). While all CD4⁺ T_N recipients developed lethal GVHD and died within 28 days after transplantation, none of the recipients of CD4⁺CD44^high T_CM developed GVHD and all survived more than 100 days after transplantation (P<0.01). Similarly, 36% of CD8⁺ T_N recipients developed lethal GVHD but none of CD8⁺CD44^high T_CM recipients developed GVHD (P<0.05). One of the unique features of effector memory T cells (T_EM) is that these cells are able to facilitate stem cell engraftment and mediate some degree of anti-tumor response without causing GVHD. To test the ability of T_CM to deplete host radioresistant T cells, BALB/c recipients were lethally irradiated (8.5 Gy) and transplanted with 1x10⁷ T-cell-depleted (TCD) bone marrow cells together with graded numbers of T cell subsets from C57BL/6. Host T cells persist in all mice receiving only TCD bone marrow cells. Addition of 1x10⁵ of either T_N or CD44^high T_CM to the graft led to 100% depletion of all host radioresistant T cells. In contrast, none of 1x10⁵ T_EM recipients and only 50% of 5x10⁵ T_EM recipients depleted all host radioresistant T cells. These data indicate that T_N and CD44^high T_CM have equivalent abilities to deplete host radioresistant T cells and both of them are at least 5 times more potent than T_EM. To test the ability of CD44^high T_CM to mediate anti-tumor effects, host-type leukemia/lymphoma cells (BCL1) cells were added to the stem cell graft. All recipients transplanted with TCD bone marrow together with BCL1 cells developed tumor and died within 28 days after transplantation. Addition of 1x10⁴ T_N and CD44^high T_CM protected 42% and 36% of the recipients from developing tumor respectively. In contrast, only 7% of TEM recipients were protected (P<0.05) at the same dose level. Similarly, addition of 1x10⁵ T_N and CD44^high T_CM protected 79% and 80% of the recipients from developing tumor respectively. However, only 25% of 1x10⁵ T_EM recipients were protected (P<0.01). These data indicate that T_N and CD44^high T_CM have equal anti-tumor activity and both of them are at least 10 times more potent than T_EM. Similar results were obtained in CD8 subset of CD44^high T_CM. A critical question is why CD44^high T_CM are able to kill host radioresistant T cells and tumor cells so efficiently without inducing GVHD. Even though CD44^high T_CM were able to respond to alloantigens, bioluminescence imaging analyses suggest that CD44^high T_CM also mediate an abortive alloresponses upon transfer into allogeneic recipients. Using an in vivo functional assay, we further demonstrated that alloreactive CD44^high T_CM lost the reactivity over time, suggesting a mechanism by which CD44^high T_CM were able to eliminate host resistance without causing GVHD. In conclusion, these data demonstrate that selective transplantation of CD44^high T_CM is able to efficiently separate anti-tumor effects from GVHD.