Background: Relapse after allogeneic transplant represents the most common cause of treatment failure in patients with acute lymphoblastic leukemia, despite significant graftversus-leukemia effects.Objective: We wished to test the hypothesis that the gene expression profile of leukemia cells surviving in a setting of significant alloreactivity post-transplant differs from the gene expression profile of leukemia cells surviving in a setting of reduced alloreactivity.Methods: 10e4 GFP-labeled, C57/BL6-derived murine pre-B ALL cells carrying the human bcr/abl and INK4A/ARF mutations were incorporated into C3.SW hematopoetic cell grafts transplanted into normal C57/BL6 mice. These strains are MHC I- and II- matched, but differ at multiple minor histocompatibility antigens. The experimental group received cells from donors primed against recipients to produce a high graft-versus-host (GVH) disease environment. Control mice received naïve donor marrow. At 14–21 days post-transplant, leukemia cells were purified from marrow using immunomagnetic selection and, in a second experiment, flow sorting for enhanced purification. Affymetrix whole genome expression profiles were generated and analyzed using dCHIP software. Expression levels in a univariate analysis with ³ 2-fold changes over controls at p <0.0001 were considered significant. Expression of selected genes was further studied by flow cytometry and real-time PCR in leukemia cells cultured with or without interferon gamma, and in leukemia cells cocultured with B6-primed C3.SW bone marrow.Results: A small number of genes were consistently expressed at higher levels in ALL cells surviving in a high-GVH environment across both transplant experiments. These included interferon gamma (IFNg) dependent genes (interferon-inducible GTPase, +205.7 fold; interferon-induced transmembrane protein 3, +3.15 fold), hematopoetic growth factors (macrophage activation factor 2, +53.3 fold increase) and tumor suppressors (HRAS-like suppressor, +7.3-fold increase). Of particular physiologic interest was a 7.2 fold increase in Sca-1 (Ly6.A) expression. In the presence of IFNg in vitro, 91 ± 1.7 % of leukemia cells express Sca-1, compared with 52 ± 1.8 % of leukemia cells cultured without IFNg (p <0.0001, n=3). Comparably, there was a 1.3 fold increase in relative mRNA expression (p=0.0004, n=2) in these cells. Leukemia cultured with alloreactive (B6 primed) C3.SW bone marrow showed a 1.6 fold increase in Sca-1 surface expression (p =0.031) and a 2.0 fold increase in relative mRNA expression (p <0.0001) when compared with control leukemia cells. There was also a 1.2 fold increase in relative Sca-1 mRNA expression in leukemia cells cultured with alloreactive vs. naïve bone marrow (p = 0.0072).Conclusions: These data provide compelling evidence that leukemia cells change their gene expression profile in response to the immunologic pressure of the post-transplant environment. In vitro validation studies specifically demonstrate that Sca-1 is upregulated in the presence of IFNg and alloreactivity. Though the roles of Sca-1, as a hematopoetic lineage molecule and as a marker of activated lymphocytes, have traditionally been described separately, recent evidence suggests that Sca-1 is capable of coordinating multiple signals via lipid rafts. Upregulation of Sca-1 may therefore represent a powerful adaptive mechanism by which leukemia cells simultaneously promote their own hematopoetic colony formation and actively engage the host immune response. Further study of this novel mechanism has important implications for targeted leukemia therapy and immune modulation post-transplant.