|Home | About | Journals | Submit | Contact Us | Français|
Imatinib mesylate (IM) induces remission in chronic myelogenous leukemia (CML) patients but does not eliminate leukemia stem cells (LSC), which remain a potential source of relapse. Here we investigated the ability of HDAC inhibitors (HDACi) to target CML stem cells. Treatment with HDACi combined with IM effectively induced apoptosis in quiescent CML progenitors resistant to elimination by IM alone, and eliminated CML stem cells capable of engrafting immunodeficient mice. In vivo administration of HDACi with IM markedly diminished LSC in a transgenic mouse model of CML. The interaction of IM and HDACi inhibited genes regulating hematopoietic stem cell maintenance and survival. HDACi treatment represents an effective strategy to target LSC in CML patients receiving tyrosine kinase inhibitors.
Chronic myelogenous leukemia (CML) is a lethal hematological malignancy resulting from transformation of a primitive hematopoietic cell by the BCR-ABL oncogene (Sawyers, 1999). Leukemia-initiating cells or leukemia stem cells (LSC) in CML share several properties with normal hematopoietic stem cells (HSC), including ability to regenerate multilineage hematopoiesis and quiescence (Holyoake et al., 1999; Wang et al., 1998). Progeny of transformed stem cells have a proliferative advantage over normal hematopoietic cells, allowing the Philadelphia (Ph)-positive clone to displace residual normal hematopoiesis. Without treatment CML progresses from a chronic phase (CP) to an accelerated phase (AP) and terminal blast crisis (BC). Deregulated tyrosine kinase activity of the BCR-ABL protein plays an important role in CML pathogenesis. Treatment with BCR-ABL tyrosine kinase inhibitors (TKI) reverses the proliferative advantage of CML progenitors, inducing remission and allowing regrowth of normal hematopoietic cells. The BCR-ABL kinase inhibitor imatinib mesylate (IM, Gleevec) has emerged as the first-line treatment for CML patients (Druker et al., 2001; O’Brien et al., 2003). Most CP CML patients achieve complete cytogenetic response (CCR) with IM treatment, and demonstrate major reductions in BCR-ABL transcript levels as assessed by real-time quantitative RT-PCR (Q-PCR) (Hughes et al., 2003). However there is evidence that primitive leukemia stem and progenitor cells are retained in patients achieving remission with IM treatment (Bhatia et al., 2003). Disease recurrence is usually seen following cessation of drug treatment, even in CML patients who are BCR-ABL negative by Q-PCR (Cortes et al., 2004; Rousselot et al., 2007). These observations suggest that “cure” of CML remains elusive following treatment with TKI alone.
The mechanisms underlying persistence of LSC in IM-treated CML patients are not well understood. BCR-ABL kinase domain mutations associated with IM resistance may be seen in some CML patients in CCR, but are not consistently found (Chu et al., 2005). Although reduced drug uptake or increased efflux together with high levels of BCR-ABL expression in primitive progenitors could theoretically contribute to IM resistance, previous studies have found adequate drug levels and effective inhibition of BCR-ABL activity in CML progenitors following IM treatment (Chu et al., 2004; Copland et al., 2006; Jordanides et al., 2006). Our studies show that IM effectively inhibits proliferation of CML primitive progenitors but only modestly increases progenitor cell apoptosis (Graham et al., 2002; Holtz et al., 2002). Growth factor (GF) or other microenvironmental signals may preserve viability of CML cells despite BCR-ABL kinase inhibition by IM (Chu et al., 2004). Importantly IM-induced apoptosis is restricted to dividing CML progenitors, whereas non-dividing CML progenitors are especially insensitive to IM-induced apoptosis (Holtz et al., 2005; Jorgensen et al., 2005). The relative insensitivity of non-dividing CML progenitors may contribute to the persistence of BCR-ABL+ progenitors in patients achieving remission on IM therapy. Similar results have been obtained with more potent BCR-ABL TKI including dasatinib, nilotinib and bosutinib (Copland et al., 2006; Jorgensen et al., 2007; Konig et al., 2008a; Konig et al., 2008b). These results suggest that BCR-ABL independent mechanisms contribute to survival of primitive CML cells after TKI treatment, and indicate the need to identify additional strategies to eliminate CML LSC.
Histone deacetylase inhibitors (HDACi) are a class of agents that have shown promise as a therapy for several cancers (Marks et al., 2004). HDACi can modulate gene expression through increased histone lysine acetylation. Anti-cancer effects may also be related to modulation of the acetylation status of non-histone proteins (Bolden et al., 2006). In contrast to most other pro-apoptotic agents that preferentially target dividing cells, HDACi have been shown to induce apoptosis in non-proliferating cancer cell lines, which may have important implications for elimination of quiescent primitive LSC (Burgess et al., 2004). Treatment with the hydroxamic acid analog pan-HDACi SAHA, LAQ824 (LAQ) or LBH589 (LBH), alone and in combination with TKI has been reported to induce apoptosis in CML cell lines and BC CML cells (Fiskus et al., 2006a; Fiskus et al., 2006b; Nimmanapalli et al., 2003a; Nimmanapalli et al., 2003b). However, BC CML cells may originate from a more mature progenitor population rather than a stem cell, and differ markedly in behavior and therapeutic response from CP CML cells (Calabretta and Perrotti, 2004; Jamieson et al., 2004). The effect of HDACi on primitive LSC from CP CML patients has not been characterized. It is not known whether HDACi are capable of inducing apoptosis in quiescent stem and progenitor cells that resist elimination by BCR-ABL TKI. Here we investigated the effects of the HDACi, LAQ and LBH, alone and in combination with IM, on primitive CML leukemia stem and progenitor cells.
We investigated the effects of the HDACi, LBH and LAQ, alone and in combination with IM, on CP CML leukemia stem and progenitor cells. We confirmed that exposure to HDACi enhanced levels of acetylated histones in CML CD34+ cells by Western blotting of chromatin extracts. Acetylated histone H3 and H4 levels were significantly increased in cells treated with LBH (Fig 1A) or LAQ (Fig S1A–C). Histone acetylation was not altered after treatment with IM alone. Treatment with HDACi has been reported to reduce BCR-ABL expression in cell lines and BC CML cells.(Fiskus et al., 2006a; Nimmanapalli et al., 2003a) We observed moderate reduction in BCR-ABL protein levels in CP CML CD34+ cells after treatment with LBH (Fig 1B). Treatment with IM resulted in reduced tyrosine phosphorylation of the BCR-ABL substrate CrkL and reduced overall tyrosine phosphorylation in CML CD34+ cells, confirming effective inhibition of BCR-ABL kinase activity (Fig 1B). LBH treatment also resulted in modest reduction in tyrosine phosphorylation in CML CD34+ cells compared with controls, and in further reduction in tyrosine phosphorylation in combination with IM (Figure 1B). The cell cycle regulatory genes p21 and p27 are known targets of HDACi treatment. In our studies, p21 and p27 levels were observed to be increased in CML CD34+ cells after LBH treatment and further increased with the combination of LBH and IM (Fig 1B).
To determine the effect of LBH on apoptosis and proliferation of CML progenitor cells, CML CD34+CD38− and CD34+CD38+ cells were labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE), cultured for 96 hours with LBH, IM or the combination, and then labeled with Annexin-PE and analyzed by flow cytometry (Fig S2A). Assessment of CFSE fluorescence allowed analysis of cell division since fluorescence is reduced by half in successive cell generations. We observed significantly increased apoptosis of CML CD34+CD38− primitive and CD34+CD38+ committed progenitors following treatment with the combination of IM and LBH, but not with IM or LBH alone (Fig 2A). Treatment with 50nM LBH with or without IM resulted in modest increase in apoptosis of normal primitive progenitors (Fig 2B). The IM and LBH combination induced significantly more apoptosis in CML compared with cord blood (CB) progenitors [CD34+CD38− cells IM+LBH (25nM) p<0.05, IM+LBH (50nM) p<0.001]. Similar results were observed for LAQ (Fig S2B–E). Reduced viability of CML CD34+CD38− and CD34+CD38+ cells treated with the combination of IM and LBH compared with IM or LBH alone was also observed after trypan blue labeling (Fig S2P), on evaluation of cell morphology (Fig S2Q), and by Caspase 3 labeling (Fig S2R).
Previous studies have shown that primitive quiescent CML CD34+ cells are especially resistant to IM-induced apoptosis (Graham et al., 2002; Holtz et al., 2005; Holyoake et al., 1999). We observed that treatment with LBH or LAQ combined with IM resulted in increased apoptosis of undivided CML CD34+ cells, with CFSE fluorescence equivalent to the parent generation, compared to IM alone (Fig 2C, Fig S2F). Although LBH and LAQ treatment also increased apoptosis in undivided normal progenitor cells compared with IM alone (Fig 2D, Fig S2G), the increase in apoptosis was significantly less than in undivided CML progenitor cells [CD34+CD38− cells IM+LBH (25nM) p<0.01, IM+LBH (50nM) p<0.01]. In conclusion, the combination of IM and LBH effectively induces apoptosis in CML primitive and committed progenitors, and to a significantly greater extent than in normal progenitors.
Treatment of CML CD34+CD38− primitive and CD34+CD38+ committed progenitors with IM or the combination of IM and LBH resulted in significant inhibition of proliferation as measured by reduction in CFSE fluorescence, whereas only modest reduction in proliferation of CML CD34+CD38− cells was seen with LBH589 alone (Fig 2E). Treatment with LBH alone did not inhibit normal progenitor proliferation, whereas the combination of LBH and IM also resulted in modest inhibition of CB progenitor proliferation. Significantly greater proliferation inhibition of CML compared with CB CD34+CD38− cells was seen for IM and IM combined with LBH 50nM (p<0.05) (Fig 2F). Similar results were seen with LAQ (Fig S2H–K).
The effect of LAQ on CML and normal committed progenitor frequency was assessed in methylcellulose progenitor assays (Fig S2L–O). Significant reduction of CML CFC frequency was seen with IM and at the highest LAQ concentration (100nM, p<0.001). LAQ combined with IM resulted in enhanced inhibition of CML CFC numbers compared with LAQ alone or IM alone (p<0.05). High concentration of LAQ also resulted in reduction in normal CFC growth which was not enhanced by combination of IM with LAQ.
The immunodeficient non-obese diabetic/severe combined immune deficiency (NOD.CB17-Prkdcscid or NOD/SCID) mouse model is widely used to assay primitive human hematopoietic stem cells with in vivo engraftment capacity (SCID-repopulating cells or SRC) (Larochelle et al., 1996). NOD/SCID interleukin-2 receptor-γ chain (IL2Rγ) deficient (NOD.Cg-Prkdcscid Il2rgtm1Wjl/Sz or NSG) mice support superior engraftment of human hematopoietic cells compared with NOD/SCID mice (Shultz et al., 2005). We determined that larger numbers of CML CD34+ cells (1×106–2×106) were required to establish engraftment in NSG mice compared to CD34+ cells from normal bone marrow (BM) (1×105–8×105). Interestingly CML LSC demonstrated enhanced short-term compared with long-term engraftment and increased myeloid skewing compared with normal HSC (Fig 3A, B).
We tested the effect of IM (1μM), LBH (50nM), or LBH in combination with IM (1μM) on CML and normal cells capable of engraftment in NSG mice at 16 weeks. We observed reduced engraftment of CML CD34+ cells treated with IM alone (p<0.05) and LBH alone (p<0.001) compared to untreated controls. However treatment with the combination of IM and LBH resulted in further inhibition of engraftment of CML CD34+ cells compared with no treatment (p <0.001), IM alone (p<0.05), and LBH alone (p<0.05) and indeed abrogated CML cell engraftment in mice (Figure 3C, D). Engraftment of both myeloid and lymphoid cells was eliminated (not shown). Human CFC were also eliminated from the marrow of mice receiving cells treated with LBH combined with IM (p<0.01, p<0.05 and p<0.01 compared with no treatment, IM alone and LBH alone respectively) (Fig 3E). FISH analysis showed that 86% of human cells from controls engrafted in mice were BCR-ABL+ confirming that engraftment is truly CML in origin. Sufficient numbers of cells were not available to perform FISH analysis after combination treatment. Q-PCR analysis confirmed that BCR-ABL+ LSC contributed to engraftment of human cells. BCR-ABL and BCR signal were not detectable in mice receiving cells treated with the combination, confirming abrogation of engraftment of human CML cells (Fig 3F). CB CD34+ cells showed significantly reduced engraftment following treatment with IM alone, LBH alone or the combination of IM and LBH compared with no treatment (p<0.001 each). Treatment with LBH and IM combination did not result in significantly enhanced inhibition of engraftment compared to IM alone (p= 0.37) or LBH alone (p=0.43) (Fig 3G, H). Significantly less inhibition of normal cell engraftment was seen following LBH and IM treatment compared to CML cell engraftment which was completely eliminated (p=0.01). Similar results were obtained when evaluating the effect of LAQ treatment on engraftment of CML and normal cells in NSG mice at 6 weeks (Fig S3A–E). These results show that HDACi in combination with IM effectively target primitive CML cells capable of multilineage engraftment that resist elimination following treatment with IM alone.
The low levels of long-term engraftment of CML LSC in the xenogeneic transplant model limit its use for in vivo drug treatment studies. We therefore utilized a transgenic Scl-tTa-BCR-ABL mouse model to investigate the effect of HDACi treatment on CML stem cells in vivo (Huettner et al., 2003). Withdrawal of tetracycline results in reversible induction of BCR-ABL expression and induction of a CML-like myeloproliferative disorder characterized by neutrophilic leukocytosis and splenomegaly. Transplantation of leukemic BM cells to wild-type recipients consistently resulted in development of a myeloproliferative disorder 3–4 weeks after transplantation. The transplantation approach allows generation of a large cohort of mice with similar time of onset of leukemia, providing a robust and consistent CML model suitable for preclinical studies of therapeutic interventions against CML stem cells in the setting of the in vivo host microenvironment.
Scl-tTa-BCR-ABL mice were crossed with transgenic GFP expressing mice to facilitate identification of donor cells. BM cells were obtained from Scl-tTa-BCR-ABL/GFP mice 4 weeks after induction of BCR-ABL expression by tetracycline withdrawal. GFP expressing cells were selected using flow cytometry and transplanted into wild-type FVB/N recipient mice irradiated at 900cGy. Recipient mice developed leukocytosis and neutrophilia 4 weeks after transplantation. Mice were then treated for 4 weeks with IM (200mg/kg body weight daily by gavage), LBH (30 mg/kg body weight intraperitoneally 3 days per week on Monday, Wednesday and Friday), LBH in combination with IM, or vehicle alone (controls) (Fig 5A). Mice treated with the IM and LBH combination demonstrated significantly increased apoptosis of Lin-Sca-1+Kit+ (LSK) stem cells compared with vehicle or either drug alone (p<0.005, Fig 4A, B). In vivo EdU labeling studies showed reduction in the percentage of LSK cells in S-phase in mice treated with the combination of IM and LBH and to a lesser extent with IM (Fig 4C). White blood cell (WBC), neutrophils and GFP+ cells (representing BCR-ABL expressing cells) were elevated in controls but were reduced to normal levels in mice receiving the different treatments (Fig 4D–H). LBH combined with IM resulted in greater reduction in WBC, neutrophils and GFP+ cells than single agent LBH or IM. Control mice became moribund and 3 of 10 mice died between 3 to 4 weeks after start of treatment (7 to 8 weeks after transplantation), whereas the mice receiving the different treatments remained well. Mice were euthanized at 4 weeks and BM and spleen cells were analyzed by flow cytometry. A profound reduction of GFP+ WBC (Fig 5B), myeloid cells (Gr-1+Mac-1+, Fig 5C), granulocyte-macrophage progenitors (GMP, Lin-Kit+Sca-1-CD34+FcγRII/IIIhi, Fig 5D), common myeloid progenitors (CMP, Lin-Kit+ Sca-1-CD34+FcγRII/IIIlo, Fig 5E), and stem cells (LSK cells, Fig 5F, G) was observed in the BM of mice treated with IM combined with LBH compared with IM or LBH alone. LBH as a single agent also reduced GFP+ WBC and progenitors but to a significantly lesser extent than the combination. A marked reduction of total GFP+ WBC (Fig 5H), immature myeloid cells (Gr-1+Mac-1+) (Fig 5I), GMP (Fig 5J), CMP (Fig 5K) and LSK cells (Fig 5L, M) was also seen in spleens of mice treated with IM in combination with LBH. In contrast to effects on BM cells, LBH as a single agent was ineffective in reducing numbers of GFP+ progenitors in the spleen. Treatment of wild type FVBN mice with IM, LBH or the combination for 4 weeks did not cause clinical symptoms or affect survival. There was a trend towards less weight gain in treated mice compared with controls although not statistically significant (Fig S4A). Treatment with LBH or LBH combined with IM caused moderate suppression of WBC and red blood cell (RBC) counts in peripheral blood and suppression of GMP, MEP and LSK cells in BM from wild-type mice (Fig S4B–G). Therefore LBH with or without IM results in moderate in vivo toxicity to normal hematopoietic cells, although considerably less than the toxicity to leukemia cells.
We conducted survival studies on mice after 4 weeks exposure to treatment (Fig 6A). Control mice that survived the 4 weeks treatment period all died within 90 days after discontinuing treatment (Fig 6B). Mice treated with the combination of IM and LBH showed markedly improved survival, and showed normal WBC counts at 8 weeks after discontinuation of treatment with only small number of residual GFP+ WBC (Fig 6C, D). To accurately quantify the loss of CML stem cells, BM cells from leukemic mice treated with IM, LBH or the combination for 4 weeks were transplanted in limiting dilutions to secondary mice (Fig 6E). Treatment with the combination of LBH and IM resulted in abrogation of engraftment of GFP+ leukemia cells in secondary mice measured at 8 and 16 weeks after transplantation (Fig 6F–H). In contrast LBH alone or IM alone did not significantly affect levels of GFP+ cells in secondary recipients (Fig 6F) or the frequency of LSC quantified on limiting dilution analysis (Fig 6G, H). These results clearly demonstrate that the LBH and IM combination is capable of markedly reducing the number of LSC capable of engraftment in secondary recipients and causing disease relapse.
To further investigate potential mechanisms underlying CML LSC targeting we conducted gene expression analysis of CML CD34+CD38− cells exposed to IM, LBH and the combination of IM and LBH. We focused our analysis on genes whose expression was significantly affected by the combination of IM and HDACi, compared with IM and HDACi alone, which we refer to as the “interaction” of IM and LBH (Fig 7A). The gene expression programs represented by these genes were identified using Gene Set Enrichment Analysis (GSEA). IM and LBH combination resulted in increased inhibition of HDACi-downregulated gene sets compared to LBH alone, suggesting that co-treatment with IM could potentiate HDACi-induced gene expression alterations. We observed reduced expression of gene sets related to the HSC state; the HOX, MYC and WNT related pathways; cell cycle regulation; protein translation through the mTOR-EIF4 pathway; cellular stress response and cell survival; and increased expression of G-protein coupled receptor genes (Fig 7B and Table 1A). We also observed reduced expression of genes with transcription-factors (TF) binding motifs for E2F which has important roles in regulating cell cycle progression and apoptosis (Chen et al., 2009); Ying-Yang 1, which directs histone acetylation changes at promoters (Gordon et al., 2006); GABP/NRF2 and NRF1, which regulate stress response (Mathers et al., 2004); and MYC-MAX, which regulate HSC proliferation and maintenance (Laurenti et al., 2008) (Table 1B). These results suggest possible molecular mechanisms that may contribute to the effects of the HDACi and IM combination (Fig 7C).
Several leukemias including CML are propagated by small populations of LSC, eradication of which may be required to achieve long-term remission and cure. Populations of cancer stem cells have also been identified in several solid tumors (Clarke et al., 2006). Treatment with BCR-ABL TKI including IM, nilotinib and dasatinib reverses the proliferative advantage of CML progenitors resulting in remission induction but does not eliminate LSC, which continue to be detectable in patients despite prolonged IM treatment. Persistent LSC represent a reservoir of disease and potential source of relapse. CML patients have a high likelihood of disease relapse on discontinuation of treatment and it is currently recommended that patients should be treated with IM indefinitely to prevent relapse. Concerns regarding the lack of disease cure with IM alone, the potential risk of side effects with long-term treatment, issues of non-compliance, and the high financial burden associated with drug treatment provide a strong impetus to develop strategies to target residual LSC.
In the current study we found that treatment with HDACi in combination with IM was significantly more effective in inducing apoptosis in CP CML progenitor cells compared with IM alone. The combination of HDACi and IM induced apoptosis in quiescent CML progenitor cells that are highly resistant to elimination following treatment with IM alone. HDACi in combination with IM also inhibited CML progenitor proliferation and CML committed progenitor growth in colony assays. In contrast to reports that HDACi could induce apoptosis in CML cell lines and BC CML cells (Fiskus et al., 2006a; Fiskus et al., 2006b; Nimmanapalli et al., 2003a; Nimmanapalli et al., 2003b), we found that HDACi by themselves had minimal effect on apoptosis of CP CML cells. CML progenitor proliferation and CFC frequency were reduced only at the highest HDACi concentrations. Therefore targeting of primitive CP CML cells by HDACi is potentiated by concomitant inhibition of BCR-ABL tyrosine kinase activity. Importantly treatment with the combination of HDACi and IM markedly depleted CML LSC capable of long-term, multilineage engraftment in immunodeficient mice. Furthermore administration of the HDACi and IM in combination resulted in profound depletion of LSC with secondary repopulating capacity in the in vivo setting in a transgenic BCR-ABL mouse model of CML, and prevented leukemia relapse after discontinuation of treatment. Therefore our data consistently show that the combination of IM with an HDAC inhibitor effectively targets primitive CML LSC and is superior to either agent alone. On the other hand this combination may be less suitable for treating patients harboring IM resistance-conferring BCR-ABL mutations, since the mutant kinase would be inhibited poorly by IM, thus eliminating its required contribution to the effect of the combination.
The effects of HDACi are complex, and involve multiple genes and pathways. HDACi-induced reduction in BCR-ABL protein expression and kinase activity could potentiate the effects of IM on CML LSC. We also observed that combination with IM enhanced HDACi-induced gene expression changes, compared with HDACi alone. Although the mechanisms underlying this effect are still unknown, novel nuclear functions of cytoplasmic tyrosine kinases related to histone and transcription factor modulation have been recognized. Potential mechanisms involved in CML LSC inhibition by the IM and HDACi combination, compared with IM or HDACi alone, suggested by gene expression analyses included reduced expression of genes related to the primitive HSC state such as downregulation of HOX, MYC and WNT related genes; reduced expression of E2F regulated genes, which may play an important role in protecting non-proliferating cells from stress-induced apoptosis (Moon et al., 2005); and increased G-protein coupled receptor expression, possibly influencing microenvironmental interactions of CML progenitors (Fig 7C). These observations although preliminary, indicate promising avenues for further investigation of molecular mechanisms underlying the effects of the HDACi and IM combination.
Our studies indicate significant activity of HDACi against normal hematopoietic progenitors with increased levels of apoptosis, inhibition of proliferation, and inhibition of CFC and SRC growth. In vivo administration of LBH was associated with moderate inhibition of normal blood cell counts and BM stem and progenitor populations and reduced weight gain. These inhibitory effects on normal progenitors are consistent with clinical observations of thrombocytopenia and myelosuppression in clinical trials of LBH (Bruserud et al., 2007). On the other hand the HDACi and IM combination resulted in significantly less apoptosis of normal compared with CML progenitors, and significantly less inhibition of normal SRC compared with CML SRC. In addition, in vivo administration of the combination resulted in significantly less inhibition of normal blood cell counts, and BM stem and progenitor populations, compared with near complete elimination of CML LSC. These observations suggest a therapeutic window for HDACi and IM effects on CML LSC compared with normal stem cells that could be exploited clinically. However, the toxicity of LBH to normal progenitors indicates a need for continued exploration of mechanisms underlying activity of the LBH and IM combination on CML LSC to aid development of more selective, non-toxic approaches for targeting LSC in future.
There is considerable interest in devising improved approaches to target CML LSC. The therapeutic application of these studies will be towards using a combination of LBH and IM to achieve elimination of residual LSC in IM-treated and responsive CML patients. We have recently shown that a farnesyltransferase inhibitor BMS-214662 can selectively kill quiescent primitive CML progenitor cells (Copland et al., 2008). However this agent is not being developed for clinical testing in CML. In contrast, several HDACi are currently in clinical trials for hematological malignancies as well as solid tumors (Glaser, 2007). The role of HDACi in targeting cancer stem cells has not been previously described. Based on our observation that LAQ and LBH combined with IM can eliminate CML LSC, we have developed and initiated a clinical trial to determine the safety and tolerability of LBH in combination with IM in CML patients in cytogenetic remission with evidence of residual BCR-ABL+ cells. The ultimate measure of success for these studies in achieving elimination of residual LSC will be the ability of patients to maintain long-term remission after discontinuation of IM treatment. Historically, studies in CML have greatly enhanced our understanding of chromosomal translocations and oncogenes in cancer biology and have led the way in successful application of targeted therapies. It remains to be determined whether the current studies of targeting of LSC in CML using HDACi will have broader application to targeting of primitive, quiescent cancer-initiating cells in other leukemias and solid tumors.
CB samples were kindly provided by StemCyte (Arcadia, CA). Normal BM samples were obtained from donors at City of Hope National Medical Center (COHNMC). Mononuclear cells were isolated using Ficol separation. CD34+ cells were then isolated using a positive magnetic bead selection protocol (StemCell Technologies, Vancouver, BC, Canada). CML samples were obtained from patients in CP who had not received prior IM treatment from the COHNMC and Glasgow University. BM samples were processed as described above. Leukopheresis samples were processed for CD34+ cell selection using CliniMACS (Miltenyi Biotech, Germany). CD34+CD38− and CD34+CD38+ cells were obtained by flow cytometry sorting. All patients and healthy donors signed an informed consent form. Sample acquisition was approved by the Institutional Review Boards at the COHNMC, in accordance with an assurance filed with and approved by the Department of Health and Human Services, and the North Glasgow University Hospital Division of NHS Greater Glasgow and Clyde, and met all requirements of the Declaration of Helsinki.
Cells were exposed to IM, LBH and LAQ (Novartis) during culture in Stemspan serum-free medium (StemCell Technologies), supplemented with low concentrations of growth factors (GF) similar to those present in long-term BM culture stroma-conditioned medium [granulocyte-macrophage colony-stimulating factor (GM-CSF) 200 pg/mL, leukemia inhibitory factor (LIF) 50 pg/mL, granulocyte colony-stimulating factor (G-CSF) 1 ng/mL, stem cell factor (SCF) 200 pg/mL, macrophage-inflammatory protein-1α (MIP-1α) 200 pg/mL, and interleukin-6 (IL-6) 1 ng/mL] at 37°C with 5% CO2 and high humidity (Bhatia et al., 1995).
Cells were labeled with 1.25μM CFSE (Molecular Probes, Eugene, OR) for 10 minutes at 37°C, incubated overnight in Stemspan serum-free medium supplemented with low concentrations of growth factors as described above to release unbound CFSE, labeled with CD34-APC and CD38-PE, and flow cytometry (MoFlo; Cytomation, Fort Collins, CO) sorted for CD34+CD38− and CD34+CD38+ cells with narrow and uniform CFSE labeling. Cells were then cultured as described above for 96 hours with or without IM, LBH or LAQ, labeled with Annexin V-PE (BD-PharMingen, San Diego, CA) and analyzed by flow cytometry (FACSCalibur; BD) for Annexin V and CFSE fluorescence. The level of CFSE in the parent cells was determined using aliquots of cells fixed in 4% paraformaldehyde directly after cell sorting. ModFit software (Verity, Topsham, ME) was used to assess the different cell generations and proliferation index of each sample. The percentage of apoptotic cells in total, divided and undivided cells was determined. The effect of drug treatment on cell viability was also assessed by trypan blue staining, assessment of cell morphology and Caspase 3 labeling as described in the Supplemental methods.
CML CD34+ cells (1–2×106 cells/mouse) or CB CD34+ cells (1×105 cells/mouse) were cultured for 96 hours in the absence of drug (control), or with addition of IM (1μM) alone, LBH (50nM) alone, or IM (1μM) in combination with LBH or LAQ (50nM) in medium with low concentrations of GF. Cells were then harvested, washed and transplanted via tail vein injection into sublethally irradiated (300 cGy) 8 weeks old NOD.Cg-Prkdcscid IL2rgtm1Wjl/SzJ mice (NSG mice, The Jackson Laboratory, Bar Harbor, ME). Mice were euthanized after 6 or 16 weeks and marrow contents of femurs, spleen cells and blood cells were obtained at necropsy. To assess human cell engraftment, cells were labeled with anti-human CD45 antibody and analyzed by flow cytometry. Specific human cell subsets were detected by staining with antibodies to human CD34, CD33, CD11b, and CD19. Human CD45+ cells were selected by immunomagnetic column selection. To assess human CFC CD45+ selected cells were placed in methylcellulose progenitor culture with human specific cytokines. To assess engraftment of malignant BCR-ABL expressing cells, CD45+ selected cells obtained were evaluated for the BCR-ABL translocation by interphase FISH and for BCR-ABL mRNA levels by Q-PCR. Mouse care and experimental procedures were performed in accordance with established institutional guidance and approved protocols from the Institutional Animal Care and Use Committee of Beckman Research Institute at COHNMC.
Inducible, transgenic Scl-tTa-BCR-ABL mice in the FVB/N background (Huettner et al., 2003) were crossed with transgenic GFP expressing mice (FVB.Cg-Tg (ACTB-EGFP) B5Nagy/J, Jackson Lab). BM cells were obtained from Scl-tTa-BCR-ABL/GFP mice 4 weeks after induction of BCR-ABL expression by tetracycline withdrawal and a pure population of GFP expressing cells was selected using flow cytometry and transplanted by tail vein injection (106 cells/mouse) into wild-type FVB/N recipient mice irradiated at 900cGy. Blood samples were obtained 4 weeks after transplantation to confirm development of neutrophilic leukocytosis. Mice were treated with IM (200mg/kg daily by gavage for 28 days), LBH (30 mg/kg body weight intraperitoneally three times a week for 28 days), LBH in combination with IM, or with vehicle alone (control). After 4 weeks of treatment, animals were euthanized and the total marrow content of femurs and tibiae and spleen cells were obtained. The number of total nucleated cells, GFP-expressing cells, and GFP+ myeloid, progenitor and stem cell populations were measured by flow cytometry as described in the methods. The effect of drug administration on apoptosis and cycling of stem cells in vivo was evaluated as described in the Supplemental methods. BM cells from a subset of treated mice were pooled and 2×106, 1×106, 5×105 cells/mouse (8 mice/dose/condition) mixed with 2×105 BM cells/mouse from wild-type FVBN mice were transplanted into wild-type FVB/N recipient mice irradiated at 900cGy. Engraftment was monitored by drawing peripheral blood (PB) every 4 weeks. The percentage of GFP+ cells in PB was analyzed by flow cytometry. The fraction of mice showing evidence of engraftment at 16 weeks after secondary transplantation was determined and the frequency of LSC was calculated using Poisson statistics. Another subset of mice was followed after discontinuation of treatment and survival and PB counts were monitored for 90 days. Mouse care and experimental procedures were performed under pathogen-free conditions in accordance with established institutional guidance and approved protocols from the Institutional Animal Care and Use Committee of Beckman Research Institute at COHNMC.
To measure BCR-ABL mRNA in cells engrafted in mice total RNA was extracted using the RNeasy Mini Kit (Qiagen, Valencia, CA), and first strand cDNA was synthesized using Superscript III first strand kit and Q-PCR analysis performed using primer and probe sequences for BCR-ABL (B3A2) as described elsewhere (Branford et al., 1999). BCR levels were measured as internal controls. The amount of BCR-ABL mRNA per unit input RNA was calculated based on the standard curves.
Cells were cultured with or without IM, LBH and LAQ for 24 hours. For analysis of histones, cells were lysed in buffer containing 0.5% Triton X 100; 2mM phenylmethylsulfonyl fluoride (PMSF), and 0.02% sodium azide, chromatin isolated using centrifugation, and histones extracted overnight in 0.2N HCl. For evaluation of other proteins, cells were lysed in buffer containing 0.5% Nonidet P-40, 0.5% sodium deoxycholate, 1mM PMSF, 50mM NaF, 1 mM Na3VO4, and a protease inhibitor cocktail (all from Sigma Diagnostics). Proteins were resolved on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred to nitrocellulose membrane. Membranes were sequentially reprobed with primary and secondary antibodies. Primary antibodies included anti-Histone-H3 and anti-Histone-H4 rabbit polyclonal and anti-acetyl-Histone-H3 (Lys9/Lys14) rabbit polyclonal antibody (9715 and 9677) (Cell Signaling Technology, Danvers, MA), anti-acetyl H4 (Millipore, Billerica, MA), anti-P-CrkL (Cell Signaling Technology), anti-phosphotyrosine (Upstate), anti-P27 (c-19) and anti-P21(C-19) (Santa Cruz Biotechnology, Santa Cruz, CA), anti-actin mouse monoclonal antibody (AC-15) (Sigma-Aldrich Corp., St. Louis, MO) and anti-ABL (Ab-3) (Oncogene Science, Cambridge, MA). Horseradish peroxidase-conjugated secondary antibodies were from Jackson ImmunoResearch Laboratories (Westgrove, PA). Antibody detection was performed using the Superfemto kit (Pierce Biotechnology, Rockford, IL). Protein levels were determined by densitometry using Image-Quant software (Amersham Pharmacia Biotech, Piscataway, NJ).
CML CD34+CD38− cells selected using flow cytometry sorting were treated with IM (1μM), LBH (50nM) and the combination of IM and LBH or cultured without exposure to drugs (controls) for 24 hours (n=3 each). Total RNA from 5000 cells was extracted using the RNeasy kit (Qiagen), amplified and labeled using GeneChip® Two-Cycle Target Labeling and Control Reagents (Affymetrix, Santa Clara, CA), and hybridized to Affymetrix GeneChip® Human Genome U133 Plus 2.0 Arrays. Microarray data analyses were performed using R (version 2.9) with genomic analysis packages from Bioconductor (version 2.4). Expression data were normalized using the robust multiarray average (RMA) algorithm, with background adjustment, quantile normalization and median polish summarization. Probesets with low expression levels or low variability across samples were filtered. For genes with multiple probesets, the gene level expression was set to be the median of the probesets. Linear regression was used to model the gene expression with the consideration of 2×2 factorial design and matched samples. Differentially expressed genes were identified by calculating empirical Bayes moderated t-statistic, and p-values were adjusted by FDR using the “LIMMA” package. We focused our attention on the “interaction” between IM and LBH to identify genes for which the extra effect of the combination of IM and LBH is significant; that is the additional effect of the combination drug that cannot be explained by the additive effect of IM and LBH treatment [Interaction = (Combination − LBH) − (IM − Control)]. Genes significantly altered by interaction between IM and LBH (p<0.01, fold change=3) were selected. Gene Set Enrichment Analysis (GSEA) was performed using GSEA software version 2.04 [http://www.broadinstitute.org/gsea/] to detect enrichment of predetermined gene sets using t-scores from all 13812 genes for 1263 gene sets in C2 (curated gene sets) category from the Molecular Signature Database (MsigDB) (Subramanian et al., 2005). Gene sets representing common functional categories were categorized and grouped. We also analyzed enrichment of gene sets with common TF binding sites (586 sets) from MsigDB.
Data from independent experiments were reported as the mean ± SEM. Student t test analysis was performed to determine statistical significance.
Chronic myelogenous leukemia (CML) results from transformation of a hematopoietic stem cell by the BCR-ABL gene. The BCR-ABL tyrosine kinase inhibitor imatinib mesylate (IM) is effective in inducing remissions and improving survival in CML patients but does not eliminate leukemia stem cells (LSC). Patients need continued treatment to prevent disease relapse and strategies to eliminate residual LSC are required. Our studies indicate that treatment with the histone deacetylase inhibitors (HDACi) combined with IM is effective in inducing apoptosis in CML LSC that resist elimination by IM alone. Several HDACi are in clinical development, and our studies support clinical trials of HDACi in combination with tyrosine kinase inhibitors to eliminate LSC in CML patients.
This work was supported by NIH grants R01 CA95684 and R01 HL77847, and a Translational Research Grant from the Leukemia and Lymphoma Society to Ravi Bhatia; the General Clinical Research Center Grant #5M01 RR00043; and NIH grant CA34196 to the Jackson Laboratory. We acknowledge the excellent technical support of the COHNMC Analytical Cytometry, Functional Genomics and Bioinformatics cores, and the Animal Resources Center. We thank Allen Lin and UK Hematologists for assistance with obtaining samples, StemCyte for their generous gift of CB samples, Dr Elisabeth Buchdunger and Peter Atadja from Novartis Pharmaceuticals for supplying IM, LAQ and LBH and Jeff Tauer for assistance with NSG mouse studies. Ravi Bhatia is a Novartis advisory board member. Tessa Holyoake is a Novartis advisory board member and has received research funding from Novartis.
Accession number. Microarray data has been deposited in the Gene Expression Omnibus database (Accession number GSE20876).
Author contributions;Bin Zhang: Designed and performed research, analyzed data, wrote manuscript
Adam Strauss: Designed and performed research, analyzed data, wrote manuscript
Su Chu: Performed experiments, analyzed data, wrote manuscript
Min Li: Analyzed data, wrote manuscript
Yinwei Ho: Performed experiments, reviewed manuscript
Keh-Dong Shiang: Analyzed data, wrote manuscript
David Snyder: Provided material, interpreted data, reviewed manuscript:
Claudia Huettner: Provided material, interpreted data, reviewed manuscript:
Leonard Shultz: Provided material, interpreted data, reviewed manuscript:
Tessa Holyoake: Provided material, interpreted data, reviewed manuscript:
Ravi Bhatia: Designed study, analyzed data, wrote manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.