Multi-Dimensional Resistance Phenotype Allows Subpopulation of Quiescent Chronic Myeloid Leukemia Stem Cells to Universally Escape From Therapeutic Attack

Abstract 203

Chronic myeloid leukemia (CML) is well responsive to various therapeutic strategies, including conventional chemotherapy, tyrosine kinase inhibitors (TKI), but also allogeneic stem cell transplantation combined with donor lymphocyte infusions (DLI). Although initial complete responses are frequently achieved, recurrence of the disease is a common phenomenon occurring sometimes years after the therapeutic intervention(s). Although acquired resistance is sometimes observed, the majority of relapses can be controlled with re-introduction of therapeutic pressure. This indicates that an initially undetectable subpopulation of leukemic precursor cells is capable of escaping the therapy and may give rise to a relapse of the disease after discontinuation of the therapeutic pressure. Since the different therapeutic strategies are based on different anti-tumor effector mechanisms, different escape variants of leukemic precursor cells may reside. To test this hypothesis, we investigated in detail the phenotype and apoptosis gene expression fingerprint of the subpopulations of CML precursor cells residing after conventional chemotherapy (Ara-C, daunorubicin, and camptothecin), TKI treatment (imatinib, dasatinib), and immunological interventions with high affinity antigen-specific cytotoxic donor T cells or NK cells. CD34+ CML precursor cells were isolated from patients and labeled with the green or red flurorescent dyes PKH67 or 26 to allow visualization of individual cell divisions. Quantitative flowcytometric analysis applying counterstaining with antibodies for different cell surface molecules allows a detailed analysis of the phenotype of the precursor cells capable of escaping the different treatment modalities. Although as expected the majority of CML precursor cells and their malignant progeny were efficiently targeted, a small subpopulation of non-dividing CD34 bright cells, comprising on average 0.05–0.2% of the initial CD34+ population, resided after the different strategies. In the absence of therapeutic pressure these cells were capable of producing malignant progeny, illustrating their proliferative capacity. In contrast to the quiescent population persisting during therapeutic pressure, the malignant progeny developing from this population after discontinuation of the treatment was again normally responsive to re-introduction of the therapeutic intervention. Next, we performed cross-resistance analysis by combining the different strategies in sequential order. We demonstrated that despite the different anti-tumor effector mechanisms, the population of quiescent leukemic precursor cells residing after each individual therapeutic intervention shows cross-resistance to the other treatment modalities. To investigate the escape mechanisms underlying this general resistance phenotype, we performed detailed analysis of the cell surface expression of molecules involved in the interaction with immune effector cells like T and NK cells. In addition, we isolated the population of quiescent CML precursor cells residing after therapeutic intervention and compared the apoptosis gene expression fingerprint of these cells with the profile of total CD34+ CML precursor cells, their proliferating malignant offspring and CD34+ precursor cells isolated from peripheral blood of healthy stem cell donors after G-CSF induced stem cell mobilization. This fingerprint was made using a quantitative array PCR containing probes for 86 apoptosis-related genes (SABiosciences). The phenotype analysis showed inferior expression of HLA-class I and several adhesion molecules crucial for the formation of a high avidity interaction with immune effector cells, including CD54, CD58 and CD49d. Moreover, the genetic profiling revealed significant downregulation (4–182 fold) of crucial players of the different main apoptosis pathways, including caspase-3 and -8, Fas, TNFR, and different association molecules like FADD, TRADD, and TRAFs, and a 4.8–6.2 fold upregulation of the most important inhibitor of the death receptor pathway c-FLIP (CFLAR). Interestingly, we observed a 55 fold downregulation of the TKI-targeted fusion gene partner Abl1. These data show that a subpopulation of CML precursor cells with potential self-renewal capacity harbors a multi-dimensional resistance phenotype allowing them to escape from different therapeutic strategies.