Whether cancer is maintained by a small number of stem cells or is composed of proliferating cells with approximate phenotypic equivalency is a central question in cancer biology1. In the stem cell hypothesis, relapse after treatment may occur by failure to eradicate cancer stem cells. Chronic myeloid leukaemia (CML) is quintessential to this hypothesis. CML is a myeloproliferative disorder that results from dysregulated tyrosine kinase activity of the fusion oncoprotein BCR–ABL2. During the chronic phase, this sole genetic abnormality (chromosomal translocation Ph+: t(9;22)(q34;q11)) at the stem cell level causes increased proliferation of myeloid cells without loss of their capacity to differentiate. Without treatment, most patients progress to the blast phase when additional oncogenic mutations result in a fatal acute leukaemia made of proliferating immature cells. Imatinib mesylate and other tyrosine kinase inhibitors (TKIs) that target the kinase activity of BCR–ABL have improved patient survival markedly. However, fewer than 10% of patients reach the stage of complete molecular response (CMR), defined as the point when BCR-ABL transcripts become undetectable in blood cells3. Failure to reach CMR results from the inability of TKIs to eradicate quiescent CML leukaemia stem cells (LSCs)2,3,4. Here we show that the residual CML LSC pool can be gradually purged by the glitazones, antidiabetic drugs that are agonists of peroxisome proliferator-activated receptor-γ (PPARγ). We found that activation of PPARγ by the glitazones decreases expression of STAT5 and its downstream targets HIF2α5 and CITED26, which are key guardians of the quiescence and stemness of CML LSCs. When pioglitazone was given temporarily to three CML patients in chronic residual disease in spite of continuous treatment with imatinib, all of them achieved sustained CMR, up to 4.7 years after withdrawal of pioglitazone. This suggests that clinically relevant cancer eradication may become a generally attainable goal by combination therapy that erodes the cancer stem cell pool.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Nguyen, L. V., Vanner, R., Dirks, P. & Eaves, C. J. Cancer stem cells: an evolving concept. Nature Rev. Cancer 12, 133–143 (2012)
Chomel, J. C. & Turhan, A. G. Chronic myeloid leukemia stem cells in the era of targeted therapies: resistance, persistence and long-term dormancy. Oncotarget 2, 713–727 (2011)
de Lavallade, H. et al. Imatinib for newly diagnosed patients with chronic myeloid leukemia: incidence of sustained responses in an intention-to-treat analysis. J. Clin. Oncol. 26, 3358–3363 (2008)
Corbin, A. S. et al. Human chronic myeloid leukemia stem cells are insensitive to imatinib despite inhibition of BCR-ABL activity. J. Clin. Invest. 121, 396–409 (2011)
Hu, C. J., Sataur, A., Wang, L., Chen, H. & Simon, M. C. The N-terminal transactivation domain confers target gene specificity of hypoxia-inducible factors HIF-1α and HIF-2α. Mol. Biol. Cell 18, 4528–4542 (2007)
Du, J. & Yang, Y. C. Cited2 in hematopoietic stem cell function. Curr. Opin. Hematol. 20, 301–307 (2013)
Graham, S. M. et al. Primitive, quiescent, Philadelphia-positive stem cells from patients with chronic myeloid leukemia are insensitive to STI571 in vitro. Blood 99, 319–325 (2002)
Jørgensen, H. G., Allan, E. K., Jordanides, N. E., Mountford, J. C. & Holyoake, T. L. Nilotinib exerts equipotent antiproliferative effects to imatinib and does not induce apoptosis in CD34+ CML cells. Blood 109, 4016–4019 (2007)
Copland, M. et al. Dasatinib (BMS-354825) targets an earlier progenitor population than imatinib in primary CML but does not eliminate the quiescent fraction. Blood 107, 4532–4539 (2006)
Luo, J., Solimini, N. L. & Elledge, S. J. Principles of cancer therapy: oncogene and non-oncogene addiction. Cell 136, 823–837 (2009)
Prost, S. et al. Human and simian immunodeficiency viruses deregulate early hematopoiesis through a Nef/PPARγ/STAT5 signaling pathway in macaques. J. Clin. Invest. 118, 1765–1775 (2008)
Berria, R. et al. Reduction in hematocrit and hemoglobin following pioglitazone treatment is not hemodilutional in Type II diabetes mellitus. Clin. Pharmacol. Ther. 82, 275–281 (2007)
Avagyan, S., Aguilo, F., Kamezaki, K. & Snoeck, H. W. Quantitative trait mapping reveals a regulatory axis involving peroxisome proliferator-activated receptors, PRDM16, transforming growth factor-β2 and FLT3 in hematopoiesis. Blood 118, 6078–6086 (2011)
Wang, Z., Li, G., Tse, W. & Bunting, K. D. Conditional deletion of STAT5 in adult mouse hematopoietic stem cells causes loss of quiescence and permits efficient nonablative stem cell replacement. Blood 113, 4856–4865 (2009)
Hoelbl, A. et al. Stat5 is indispensable for the maintenance of bcr/abl-positive leukaemia. EMBO Mol. Med. 2, 98–110 (2010)
Kieslinger, M. et al. Antiapoptotic activity of Stat5 required during terminal stages of myeloid differentiation. Genes Dev. 14, 232–244 (2000)
Fatrai, S., Wierenga, A. T., Daenen, S. M., Vellenga, E. & Schuringa, J. J. Identification of HIF2α as an important STAT5 target gene in human hematopoietic stem cells. Blood 117, 3320–3330 (2011)
Matsumoto, A. et al. CIS, a cytokine inducible SH2 protein, is a target of the JAK-STAT5 pathway and modulates STAT5 activation. Blood 89, 3148–3154 (1997)
Liu, S. et al. Targeting STAT5 in hematologic malignancies through inhibition of the bromodomain and extra-terminal (BET) bromodomain protein BRD2. Mol. Cancer Ther. 13, 1194–1205 (2014)
Wang, L., Giannoudis, A., Austin, G. & Clark, R. E. Peroxisome proliferator-activated receptor activation increases imatinib uptake and killing of chronic myeloid leukemia cells. Exp. Hematol. 40, 811–819 (2012)
Szanto, A. & Nagy, L. Retinoids potentiate peroxisome proliferator-activated receptor gamma action in differentiation, gene expression, and lipid metabolic processes in developing myeloid cells. Mol. Pharmacol. 67, 1935–1943 (2005)
Chen, Y., Haviernik, P., Bunting, K. D. & Yang, Y. C. Cited2 is required for normal hematopoiesis in the murine fetal liver. Blood 110, 2889–2898 (2007)
Du, J. et al. HIF-1α deletion partially rescues defects of hematopoietic stem cell quiescence caused by Cited2 deficiency. Blood 119, 2789–2798 (2012)
Koschmieder, S. & Schemionek, M. Mouse models as tools to understand and study BCR-ABL1 diseases. Am. J. Blood Res. 1, 65–75 (2011)
Rousselot, P. et al. Targeting STAT5 expression resulted in molecular response improvement in patients with chronic phase CML treated with imatinib. ASH Annual Meeting Abstracts (2012)
Davies, G. F., Juurlink, B. H. & Harkness, T. A. Troglitazone reverses the multiple drug resistance phenotype in cancer cells. Drug Des. Devel. Ther. 3, 79–88 (2009)
Yuan, H. et al. Activation of stress response gene SIRT1 by BCR-ABL promotes leukemogenesis. Blood 119, 1904–1914 (2012)
Pitulis, N., Papageorgiou, E., Tenta, R., Lembessis, P. & Koutsilieris, M. IL-6 and PPARγ signalling in human PC-3 prostate cancer cells. Anticancer Res. 29, 2331–2337 (2009)
Chen, Y., Hu, Y., Zhang, H., Peng, C. & Li, S. Loss of the Alox5 gene impairs leukemia stem cells and prevents chronic myeloid leukemia. Nature Genet. 41, 783–792 (2009)
Kominsky, D. J. et al. Abnormalities in glucose uptake and metabolism in imatinib-resistant human BCR-ABL-positive cells. Clin. Canc. Res. 15, 3442–3450 (2009)
Lu, D. & Carson, D. A. Repression of beta-catenin signaling by PPAR gamma ligands. Eur. J. Pharmacol. 636, 198–202 (2010)
Ito, K. et al. A PML–PPAR-δ pathway for fatty acid oxidation regulates hematopoietic stem cell maintenance. Nature Med. 18, 1350–1358 (2012)
Gabert, J. et al. Standardization and quality control studies of 'real-time' quantitative reverse transcriptase polymerase chain reaction of fusion gene transcripts for residual disease detection in leukemia — a Europe Against Cancer program. Leukemia 17, 2318–2357 (2003)
Nègre, D. et al. Characterization of novel safe lentiviral vectors derived from simian immunodeficiency virus (SIVmac251) that efficiently transduce mature human dendritic cells. Gene Ther. 7, 1613–1623 (2000)
Roth, O. et al. Imatinib assay by HPLC with photodiode-array UV detection in plasma from patients with chronic myeloid leukemia: comparison with LC-MS/MS. Clin. Chim acta. 411, 140–146 (2010)
Chou, T. C. Drug combination studies and their synergy quantification using the Chou-Talalay method. Cancer Res. 70, 440–446 (2010)
We thank C. Costa, V. Tran Chau, F. Goullieux, A. Krief, P. Raynal, C. Terré, S. Tabore and T. Andrieu for their experimental contributions. This work was supported by the Association Laurette Fugain, Paris, France, by the Association pour la Recherche sur le Cancer, Villejuif, France to S.P., P.R. and P.L. and by the Chaire industrielle de l’Agence Nationale pour la Recherche (ANR) to P.L.
The authors declare no competing financial interests.
Extended data figures and tables
Extended Data Figure 1 Clonogenicity assays in the presence of various PPARγ agonists and validation of STAT5B overexpression and anti-PPARγ, anti-STAT5 and anti-HIF2α siRNA.
a, Clonogenic capacities of BM CD34+ cells were assayed following pre-incubation for 2 days with culture medium alone (control) or supplemented with PPARγ agonists, PGJ2, troglitazone (Tro), ciglitazone (Cig), rosiglitazone (Ros), pioglitazone (Pio) or MCC-555 (MCC) (25 μM each) (samples from 4 donors in triplicate). The number of colonies scored is expressed as percentage of control (untreated) values with standard deviation (s.d.), *P < 0.05 using the nonparametric Wilcoxon rank test. b, Validation of anti-PPARγ siRNA used in Fig. 2b. CD34+ cells were transfected with irrelevant or PPARγ targeting siRNA (25 nM each). An anti-PPARγ shRNA was used as a positive control. PPARγ transcripts were normalized to GAPDH transcripts and expressed relative to the levels measured in untransfected cells. c, Western blot analysis with PPARγ, pan-STAT5, HIF2α and anti-actin antibodies (Ab). Validation of siRNA against PPARγ or STAT5 and lentivector expressing STAT5B (LvSTAT5B) were realized on CD34+ cells from human UCB. Validation of siRNA against HIF2α was realized on K562 cell line. Ctrl, scrambled siRNA; −, untreated. Quantification of western blot signal was realized with ImageJ software (http://rsb.info.nih.gov/ij/). Histograms show mean values with s.d., n = 3.
a, CFC assays with CD34+ CP-CML cells from patients at diagnosis. Imatinib and/or pioglitazone were added for 48h before CFC assays. Means of 29 patients with standard deviation (s.d.). b, CFC assays after lentivector-mediated expression of BCR–ABL or eGFP (negative control) in human cord blood CD34+ cells. Imatinib and/or pioglitazone were added for 48 h before CFC assays. Means of 3 individuals in triplicate with s.d. c, d, Limited dilution analysis (LDA) of CML LSCs by LTC-IC and frequency analysis. Plotted are means for CD34+ cells from 2 CP-CML patients, 16 replica each. Imatinib 1 μM, Rosi 10 μM. e, f, CFSE analysis of CD34+ cells (>96% Ph+) from CP-CML patients (for all experiments, imatinib 1 μM, dasatinib 0.146 μM, pioglitazone and rosiglitazone 10 μM, JQ1 1 μM. imat, imatinib; dasa, dasatinib; pio, pioglitazone; (P), undivided). To confirm the pivotal role played by STAT5 in the mechanism of action of pioglitazone in eroding the pool of TKI-resistant CML-LSCs, we investigated here the effect of the bromodomain inhibitor JQ1, which inhibits the transcriptional function of STAT5 by decreasing its activity through targeting the bromodomain-containing protein 2 (BRD2), a key cofactor of STAT5. Although this study with JQ1 is corroborative, one cannot completely exclude the possibility that these effects are coincidental, as targeting BRDs may cause a series of effects independent of STAT5.
Extended Data Figure 3 Pioglitazone slowly decreases STAT5 expression whereas imatinib rapidly inhibits STAT5 phosphorylation.
a, Differential kinetics of action of imatinib and pioglitazone. CD34+ CP-CML cells (patient 4) in liquid culture in serum-free medium without cytokines. b, Rate of apoptosis in CP-CML cell populations after 4 days of culture with imatinib and/or pioglitazone (n = 5; *P < 0.05). Solid bars (black for CD38− and grey for CD38+), percentage of recovery relative to input and normalized to untreated controls. Hatched bars, percentage of apoptosis, defined by the expression of annexin V. c, Flow cytometry analysis of permeabilized K562 cells with IgG against phosphorylated (Tyr694) STAT5. Untreated (black) and drug treated (red or blue). Control panel, no drug treatment but irrelevant IgG isotype control (grey peak). d, Western blot analysis with pan-STAT5 and anti-actin antibodies, showing a decrease of STAT5 by 3.5 fold ± 0.5 (s.d.) in lane 4 (n = 3). Lanes 1 and 2 for imatinib (15 and 30 min exposure, respectively); lanes 3 and 4 for Pio (72 and 96 h exposure, respectively). Ratio indicates ratio of STAT5 expression/β-actin expression relative to lane 1. Quantification of western blot signals (n = 3 for each condition) was realized with ImageJ software (http://rsb.info.nih.gov/ij/).
Extended Data Figure 4 Forced expression of STAT5 in CP-CML CD34+ cells increases the compartment of quiescent cells.
a, CFSE analysis of CP-CML CD34+ cells treated with pioglitazone after transduction with lentivectors (Lv) expressing eGFP or STAT5B, whose transcription is PPARγ-independent. Representative CP-CML patient 2 in triplicate (data for all patients are in Extended Data Fig. 2e). One coloured peak for each cell division number. P, colcemid arrested “parent-cells”. b, Distribution (%) of CD34+ cells in each division peak shown in Extended Data Fig. 3a. c, STAT5 mRNA expression analysis. d, Transduction efficiency of STAT5 lentivector. (5 replica with s.d.). e, Data for the 3 patients tested.
LTC-IC (LDA) showing differential toxicity of pioglitazone for CP-CML vs. normal CD34+ cells (n = 3, 16 replica for each).
Extended Data Figure 6 Erosion of undivided and imatinib-resistant CD34+ CP-CML cells is OCT1-independent.
a, Efficiency of LvOCT1 transduction (D8). b, OCT1 mRNA expression. Results are normalized to GAPDH mRNA levels and represented relative to mRNA expression for the “imatinib alone” condition. c, CFSE analysis and absolute cell count in the presence of imatinib, with or without OCT1 overexpression. Left scale (black), total cells showing CD34+ vs. CD34− cells (histograms). Right scale (red), undivided CD34+ cells (red dots) (representative for n = 3 CP-CML patients).
Extended Data Figure 7 The viability of undivided (P) and imatinib-resistant CD34+ CP-CML cells depend on HIF2α expression.
Representative CP-CML patient 8 in triplicate (data for all patients are in Extended Data Fig. 6e). a, CFSE analysis in presence of siRNA against STAT5 or HIF2α in CD34+-Ph+ cells treated or not with imatinib. One colored peak for each cell division number. P, colchemid arrested ‘parent-cells’. b, Distribution (%) of CD34+ cells in each division peak. c, HIF2α mRNA expression 72 h after siHIF2α transfection. d, STAT5 A and B mRNA expression 72 h after siSTAT5 transfection into human UCB CD34+ cells. Results are normalized to GAPDH mRNA levels (means of 5 experiments with s.d. for each gene assessed). e, Data for the 5 patients tested (*P < 0.05 relative to siCtrl; #P < 0.05 relative to Imatinib + siCtrl). f, CFSE analysis of cord blood CD34+ cells after transduction with lentivectors (Lv) expressing HIF2α or eGFP. One coloured peak for each cell division number. P, colchemid arrested ‘parent-cells’. g, Distribution (%) of CD34+ cells in each division peak (n = 5). h, Transduction efficiency of HIF2α Lv. i, HIF2α mRNA expression (means of 5 experiments with s.d.).
a, mRNA expression of target genes in CD34+ cells from UCB transduced or not by BCR–ABL expressing lentivector (Lv). BCR–ABL+ cells were cultured in serum-free medium without cytokines for 7 days with either imatinib alone (1 μM) or imatinib and pioglitazone (1 μM and 10 μM, respectively) (means of 5 experiments with s.d. for each gene assessed). Results are normalized to GAPDH mRNA levels and represented relative to mRNA expression for the ‘untreated’ condition. Overexpression of BCR–ABL in CD34+ cells from umbilical cord blood induced expression of STAT5 and HIF1α mRNAs by 2.7- and 2.8-fold, respectively (P = 0.043), while HIF2α and CITED2 mRNAs were increased by 12.5-fold and 9-fold (P = 0.043), respectively. In the presence of imatinib, STAT5 and HIF1α mRNAs were increased by 7.5- and 4.9-fold (P = 0.043), respectively, while HIF2α and CITED2 were increased by 19.3- and 22-fold (P = 0.043), respectively. Either pioglitazone or an siRNA against STAT5 (A and B) significantly reduced the levels of HIF2α and CITED2 mRNAs, while an siRNA against HIF2α significantly reduced CITED2 mRNA expression (>threefold each, P < 0.05). b, mRNA expression of target genes in Ba/F3 cell sub-lines independent of IL3 for viability after transduction with LvBCR–ABL or constitutively activated Stat5A1*6 (A*) or Stat5B1*6 (B*). Results are normalized to GAPDH mRNA levels and represented relative to mRNA expression for the original Ba/F3 cell line (means of 5 experiments with s.d. for each gene assessed). Forced expression of BCR-ABL increased the level of murine endogenous Stat5 (a and b) mRNAs by 2.7 fold (P = 0.043). When BCR-ABL or constitutively activated murine Stat5 1*6 (a or b) were overexpressed, murine endogenous Hif1α mRNA level was decreased by threefold (P = 0.043) and murine endogenous Hif-2α and Cited2 mRNAs increased by more than eightfold each (P = 0.043).
Extended Data Figure 9 The key regulator of HSC quiescence, CITED2, is overexpressed in TKI-resistant CD34+ cells from CP-CML patients.
a, mRNA expression of CITED2 and target genes thereof BMI1, HES1 and p57 after 9 days of culture with or without imatinib and pioglitazone. Results are normalized to GAPDH (n = 4). b, mRNA expression of endogenous murine Cited2 and its target genes Bmi1, Hes1 and p57 in Ba/F3 cell line with or without forced expression of BCR–ABL or constitutively active Stat5A 1*6 (A*) in the presence or not of siRNA against Cited2. Results are normalized to GAPDH (mean ± s.d. of 3 independent experiments in triplicate). Forced expression of a constitutively active form of murine Stat5 1*6 (A or B) in Ba/F3 cells, in and of itself, was sufficient to increase endogenous expression of murine Cited2 markedly (52-fold) as well as that of its target genes Bmi1 (2.5-fold), Hes1 (13-fold) and p57 (18-fold) c, Proliferation analysis by EdU incorporation assay of the Ba/F3 cell line that expresses or not constitutively active forms of Stat5 A 1*6 (A*) or B 1*6 (B*) in the presence or not of siRNA against Cited2 (representative result of 5 independent experiments).
About this article
Cite this article
Prost, S., Relouzat, F., Spentchian, M. et al. Erosion of the chronic myeloid leukaemia stem cell pool by PPARγ agonists. Nature 525, 380–383 (2015) doi:10.1038/nature15248
Overcoming imatinib resistance in chronic myelogenous leukemia cells using non-cytotoxic cell death modulators
European Journal of Medicinal Chemistry (2020)
Comparison of chronic myeloid leukemia stem cells and hematopoietic stem cells by global proteomic analysis
Biochemical and Biophysical Research Communications (2019)
Gain Fat—Lose Metastasis: Converting Invasive Breast Cancer Cells into Adipocytes Inhibits Cancer Metastasis
Cancer Cell (2019)
European Journal of Medicinal Chemistry (2019)
Identification of leukemia stem cell expression signatures through Monte Carlo feature selection strategy and support vector machine
Cancer Gene Therapy (2019)