Immunorecognition of leukemic stem cells by NK cells : the role of HDAC inhibitors in NKG2D ligand-mediated anti-tumor responses in acute myeloid leukemia

Abstract

The diagnosis of acute myeloid leukemia (AML) is associated to a poor long-term outcome due to frequent relapse despite intensive chemotherapy, radiation and hematopoietic stem cell transplantation (HSCT) as well as continuous advances in treatment modalities. Relapses might be caused by leukemic stem cells (LSC). According to a recently emerging concept, LSC display many features of normal hematopoietic stem cells (HSC) like quiescence and self renewal capacity and therefore are poorly accessible for conventional therapies which primarily reach the rapidly proliferating cells. Additionally, LSC are apparently able to escape from immunorecognition and thereby sustain the disease. NK cells, as the main innate immune effectors against tumor cells, are able to recognize and kill malignant cells when triggered by cell surface expression of a multitude of activating ligands. The best-described receptor-ligand pair in humans is NKG2D and its ligands, ULBP and MICA/B. Furthermore, NCR is an important family of activating receptors on NK cells, whose ligands are not yet known. The regulation of NK cells is completed by several inhibitory receptors (KIR) specific for different HLA class I molecules on potential target cells. While preceding work in our lab was describing the interaction between NK cells and leukemic blasts of AML, there is no information available on the recognition of LSC by NK cells. In this study we aimed to elucidate the interaction of NK cells with LSC of AML. The cell surface expression of ligands for activating and inhibitory NK cell receptors on LSC was in focus of these studies. Moreover, we applied a pharmacological approach to treat the patient-derived primary AML leukemic cells and examined the consequences for cell surface expression of NK cell-specific ligands. By employing hematopoietic colony forming assays, cytotoxicity assays as well as in vivo NOD/SCID xenotransplantation we aimed to functionally assess the implications of the upregulation of activating ligands for NK cell immunorecognition of LSC. In initial experiments, we demonstrated that activating ligands for the NKG2D receptor and NCR receptors on NK cells are absent or only weakly expressed on the surface of patient derived AML blasts. This expression could be increased by pharmacological means applying bryostatin-1, a modulator of PKC activity. Upregulation of cell surface expression of NKG2D ligands on AML blasts led to increased immunorecognition by NK cells in cytotoxicity assays. Subsequently, we demonstrated that similarly to total blasts, LSC of AML as judged by the phenotype CD45dimCD34+CD38-, did not express ULBP and MICA/B on their surface. To pharmacologically increase their expression, we employed the HDAC inhibitor valproic acid (VA), a drug acting through epigenetic modification of gene expression and having long-term records in different clinical applications. This treatment with VA proved to be of importance for the immunorecognition by NK cells. In the functional assays we employed NK cells selected for the KIR-HLA class I mismatch in order to circumvent inhibitory signals inactivating the NK cells. Serial replating colony forming unit (CFU) assays with LSC after treatment with VA and after coincubation with KIR-HLA mismatched NK cells demonstrated an efficient reduction in colony formation capacity upon this synergistic treatment. The cytotoxicity assays with VA-treated LSC as targets and KIR-HLA mismatched NK cells as effectors revealed interindividual differences among patient samples, reflecting a complex regulation of NK cell activation and immunorecognition. Altogether, a direct interaction of NK cells and LSC could be demonstrated in vitro. In the in vivo setting, by transplantation of AML cells intrafemurally into NOD/SCID mice with consecutive treatment of VA and HLA-mismatched NK cells, we were able to achieve a stable engraftment of human AML in the mouse bone marrow. However, the combined treatment with VA and NK cells was not influencing the content of malignant cells as compared to untreated mice. The ongoing studies aim at optimization of AML treatment with NK cell-based immunotherapy in the preclinical NOD/SCID transplantation model. Taken together, these results showed the potential of VA as an applicable anti-neoplastic drug to enhance immunorecognition of LSC of AML by NK cells, mediated by increased cell surface expression of activating ligands. The functional consequences of an enhanced immunorecognition by NK cells in abolishing the colony forming capacity of patient derived LSC are promising beneficial effects for innovative AML treatments in future.