Eradication of Chronic Myeloid Leukemia Stem Cells: A Novel Mathematical Model Predicts No Therapeutic Benefit of Adding G-CSF to Imatinib

Michor F (2009) Eradication of Chronic Myeloid Leukemia Stem Cells: A Novel Mathematical Model Predicts No Therapeutic Benefit of Adding G-CSF to Imatinib. PLoS Comput Biol 5(9): e1000503. doi:10.1371/journal.pcbi.1000503 Eradication of Chronic Myeloid Leukemia Stem Cells: A Novel Mathematical Model Predicts No Therapeutic Benefit of Adding G-CSF to Imatinib Jasmine Foo 0 Mark W. Drummond 0 Bayard Clarkson 0 Tessa Holyoake 0 Franziska Michor 0 Christophe Fraser, Imperial College London, United Kingdom 0 1 Computational Biology Program, Memorial Sloan-Kettering Cancer Center , New York , New York, United States of America, 2 Section of Experimental Haematology, Medical Faculty, University of Glasgow , Glasgow , United Kingdom , 3 Molecular Pharmacology and Chemistry Program, Memorial Sloan-Kettering Cancer Center , New York, New York , United States of America Imatinib mesylate induces complete cytogenetic responses in patients with chronic myeloid leukemia (CML), yet many patients have detectable BCR-ABL transcripts in peripheral blood even after prolonged therapy. Bone marrow studies have shown that this residual disease resides within the stem cell compartment. Quiescence of leukemic stem cells has been suggested as a mechanism conferring insensitivity to imatinib, and exposure to the Granulocyte-Colony Stimulating Factor (G-CSF), together with imatinib, has led to a significant reduction in leukemic stem cells in vitro. In this paper, we design a novel mathematical model of stem cell quiescence to investigate the treatment response to imatinib and G-CSF. We find that the addition of G-CSF to an imatinib treatment protocol leads to observable effects only if the majority of leukemic stem cells are quiescent; otherwise it does not modulate the leukemic cell burden. The latter scenario is in agreement with clinical findings in a pilot study administering imatinib continuously or intermittently, with or without G-CSF (GIMI trial). Furthermore, our model predicts that the addition of G-CSF leads to a higher risk of resistance since it increases the production of cycling leukemic stem cells. Although the pilot study did not include enough patients to draw any conclusion with statistical significance, there were more cases of progression in the experimental arms as compared to continuous imatinib. Our results suggest that the additional use of G-CSF may be detrimental to patients in the clinic. - The existence of cancer stem cells, a rare subpopulation of cancer cells responsible for tumor initiation and maintenance, was first postulated in the 1960s [1]. In leukemia in particular, increasing evidence suggests that leukemic stem cells are the only cells within the tumor capable of propagating the disease [26]. Leukemic stem cells share many properties such as self-renewal, pluripotency, and quiescence with tissue stem cells, and they appear to remain untouched by both conventional chemotherapy and targeted drugs [6]. Since the repopulating capabilities of leukemic stem cells necessitate their eradication for a cure of the disease, the development of new therapeutic approaches targeting leukemic stem cells would have a profound impact on cancer management. Chronic myeloid leukemia (CML) is associated with the Philadelphia chromosome, which results from a reciprocal translocation between chromosomes 9 and 22 generating the BCR-ABL fusion oncogene [7,8]. The ABL kinase inhibitor imatinib greatly improves outcome in CML patients [9]; however, some evidence suggests that it cannot eradicate the disease since it preferentially targets progenitor cells while being incapable of depleting leukemic stem cells [10]. Surviving leukemic stem cells are a potential source of relapse, as demonstrated by the dynamics of leukemic cells in patients who discontinue imatinib after a prolonged administration of therapy [1117]. Several mechanisms of leukemic stem cell insensitivity to ABL kinase inhibitors have been suggested [18]; those mechanisms include quiescence of leukemic stem cells, drug export from the cytoplasm, overexpression of BCR-ABL in stem cells as compared to differentiated cells, and a lack of immune responses against leukemic stem cells. Under normal circumstances, a fraction of hematopoietic stem cells is quiescent [1924]. Cycling and quiescent hematopoietic stem cells display major functional differences, mostly reflected in their homing and mobilization abilities [2529]. Quiescent stem cells are mobilized by Granulocyte-Colony Stimulating Factor (GCSF) and other agents and show preferential homing to the bone marrow as compared to dividing hematopoietic stem cells [29]. In contrast, CML stem cells are constitutively present in the circulation, but also contain a subpopulation of quiescent cells [30]. This quiescent subpopulation has been shown to be insensitive to imatinib therapy in vitro and might therefore represent a therapeutically relevant subpopulation of cancer cells [31]. However, the extent of quiescence of CML stem cells and the in vivo response of such cells to imatinib have not yet been established. In this paper, we present a novel mathematical model to investigate the response of cycling and quiescent leukemic stem cells to treatment with imatinib alone or imatinib combined with G-CSF. We use clinical data of a Phase II trial administering Imatinib mesylate (Gleevec) is currently the standard treatment for chronic myeloid leukemia (CML) and elicits a large reduction in leukemic cell burden in most patients. However, strong evidence suggests that imatinib does not cure the disease; approximately 20% of patients relapse within three years, and discontinuation of imatinib therapy often leads to a rebound of the leukemic cell burden. Laboratory studies have suggested that there exists a subpopulation of quiescent leukemia cells (i.e., cells that do not divide) that may be insensitive to imatinib treatment. It has been postulated that the disease outcome may be improved by administering imatinib in conjunction with the Granulocyte-Colony Stimulating Factor (G-CSF), a growth factor which wakes up the quiescent stem cells and sensitizes them to imatinib. In this study, we design a novel mathematical model of stem cell quiescence to investigate the treatment response to imatinib and G-CSF. We find that adding G-CSF to an imatinib treatment protocol leads to observable effects only if the majority of leukemic stem cells are quiescent. Our model also predicts that adding G-CSF leads to a higher risk of resistance, since it increases the number of leukemic stem cell divisions and thus the probability of acquiring a resistance mutation. imatinib and G-CSF and study the effects of various treatment strategies on the leukemic stem cell pool [32]. This study is part of a growing literature of theoretical approaches to CML [16,3339]. Mathematical Modeling of the CML Cell Hierarchy, Turnover, and Response to Imatinib Consider a differentiation hierarchy of hematop (...truncated)