Under the first aim of this project we have generated a precursor B cell leukemia line derived from mice with transgenic expression of E2aPBX1, a recurring translocation present in approximately 5% of pediatric ALL (Bijl et al, Genes and Development, 2005). The cell line was confirmed to be immunogeneic as vaccination with irradiated leukemia cells protects against subsequent challenge with E2aPBX1 but not other tumors. Using monoclonal antibodies to deplete cell subsets, we demonstrated that protection in immunized mice requires both CD4 and CD8 T cells and is impaired following NK cell depletion. Under aim 2 we have performed bone marrow transplantation experiments to assess GVL in this model. Allogeneic transplantation followed by E2aPBX1 leukemia challenge and subsequent transfer of primed allogeneic T cells results in cure of leukemia in all mice. However, the mice develop weight loss and histologic changes consistent with GVHD that results in late mortality. Interestingly, priming T cell donors with recipient (and leukemia) strain non-malignant B cells did not cure the mice indicating that both minor antigens and leukemia-associated antigens are responsible for cure in this model. We next sought to separate the anti-leukemic GVL effect from GVHD by selecting for T cells subsets. Neither CD4 nor CD8 T cells from primed donors alone were sufficient to cure all of the mice. Using flow sorting based on expression of CD44 and CD62L (L-selectin) we have demonstrated that central memory phenotype T cells (CD44+/CD62L+) can cure leukemia without the induction of GVHD whereas nave T cells (CD44-/CD62L+) induce rapidly lethal GVHD. In summary, we have demonstrated that ALL can be effectively targeted by a T cell response in vivo but that this response requires vaccination of the donor T Cell inocula. Second, allogeneic antigens contribute to the cure following T cell infusion but results in GVHD. Finally, sorted populations of T cells from primed donors can mediate selective graft versus leukemia responses. Using this model we have begun studying the early progression of the leukemia in bone marrow and the impact of this progression on T cells. We have identified that a surprisingly large percentage of T cells in leukemia-infiltrated compartments express high levels of the negative regulator of T cell function, programmed death 1 (PD-1) receptor. Addition studies have shown that the percentage of PD-1+ T cells correlates with the extent of leukemic involvement and that PD-1+ T cells also express other markers of a senescent phenotype such as T cell immunoglobulin and mucin domain 3 (Tim-3). Using an E2aPBX1 cell line expressing ovalbumin and ovalbumin-specific T cells we have demonstrated the induction of PD1 occurs only when T cells are able to recognize antigens on the ALL. Interestingly, careful assessment T cells during early leukemia progression have shown that the induction of PD1 occurs early (by day 5 after injection of leukemia) whereas acquisition of other T cells senescent markers such as Tim-3 and Lymphocyte Activation Gene 3 (LAG3) occur later suggesting that these markers may be more functionally relevant in terms of antileukemic potential. Indeed, T cells from irradiated tumor cell primed mice also express PD1 but not Tim-3 or LAG3 and mediate an antileukemic effect. Finally, preliminary a data from human bone marrow samples leukemia samples from patients with ALL (obtained from Dr. Alan Wayne) have shown expression of PD1, Tim-3 and LAG3 on a subset of T cells. In summary, this data suggests that for pediatric ALL blocking Tim-3 or LAG3 may be more effective than targeting PD1. Another area of investigation in the laboratory is focused on understanding the impact of the response to minor histocompatibility antigens expressed in normal tissues (graft versus host) on the anti-tumor immune response following allogeneic HSCT. Under project Project ZIA BC 011320, we have demonstrated and published that even mild GVHD can significantly impair responses to vaccines targeting tumors (Capitini et al, Blood, 2009) and that this attenuation of vaccine responses results from both diminished proliferation and increased apoptosis (manuscript accepted, Journal of Immunology). The tumor antigenic complex used in these studies is the HY system in which a solid tumor that naturally expresses Y chromosome-derived antigens was injected into female mice receiving female allogeneic bone marrow and T cells (where the mouse strain-specific minor allogeneic antigens are distinct from the HY-derived tumor antigens). We have now assessing HY tumor responses in male mice of the same strain from which the tumor is derived. Importantly, in this model, the tumor antigens completely overlap with tumor antigens, a clinically relevant scenario in which tumor-specific antigens may be weak or absent. The use of CD4 and CD8 HY specific T cell receptor transgenic donors allows careful tracking of T cells targeting shared antigens. We have shown that, while HY-specific T cells mediate mild GVHD and expand to a much larger extent in males than in female hosts (despite HY vaccination), these T cells with specificity to both normal tissues and a solid subcutaneous tumor are less potent at inducing tumor regression. We have now generated and HY expressing pre-B cell ALL line. Interestingly, using female recipients of male bone marrow (where HY expression is restricted to the hematopoietic compartment) we have shown that the impairment of T cell responses against malignancy occurs only when the target minor antigen is coexpressed in the same non-malignant tissue compartment as the tumor since HY specific T cells reject solid tumors but not leukemia in this model. Using candidate molecule (assessment of PD1, Tim3, LAG3) and non-candidate (global gene expression) approaches to assess T cells from different compartments (bone marrow vs secondary lymphoid tissues) we are exploring what distinguishes T cells with anti-tumor potential from those unable to reject tumors. Aim 3 is ongoing and involves the extension of work conducted under aim 2 to clinically relevant leukemia targets. We have established that E2aPBX1 overexpresses the Wilm's Tumor 1 gene, also overexpressed on approximately 70-80% of human leukemias and validated as a target in patients. In order to obtain large numbers of WT-1 specific T cells, we have generated mice that express a T cell receptor specific for the dominant class I epitope derived from WT-1 (called Db126). T cells from these mice show robust expansion to Db126 peptide in vitro resulting T cells with the capability of targeting peptide pulsed targets in vitro. While E2aPBX1 cells can also be targeted, the results have been variable. We are in the process of optimizing the use of these T cells in vivo against E2aPBX1 and other hematologic malignancies that overexpress WT-1. An open protocol in the Pediatric Oncology Branch (Dr. Alan Wayne, Principal Investigator) is utilizing dendritic cell vaccination targeting WT-1 peptides as a therapeutic intervention in patients relapsing following allogeneic HSCT. Although this trial involves peptide-pulsed DC vaccination and is, thus, restricted to HLA-A2+ individuals, we are currently evaluating RNA electroporation of DCs with full-length WT1 RNA as a means to generate a non-HLA-A2 restricted platform. We have obtained RNA transcription vectors from Duke University where clinical trials are ongoing using RNA-electroporated DCs targeted antigens other than WT1. These vectors will be used to produce ?clinical-grade? WT1 RNA in the Duke facility. A proposal to use this platform as a post-transplant pre-emptive therapy in children with high risk ALL and AML has been presented to the Pediatric Blood and Marrow Transplant Consortium.