Fax: (011) 33-05-56-93-88-83
Overproduction of BCR-ABL induces apoptosis in imatinib mesylate-resistant cell lines
Version of Record online: 22 NOV 2004
Copyright © 2004 American Cancer Society
Volume 103, Issue 1, pages 102–110, 1 January 2005
How to Cite
Desplat, V., Belloc, F., Lagarde, V., Boyer, C., Melo, J. V., Reiffers, J., Praloran, V. and Mahon, F.-X. (2005), Overproduction of BCR-ABL induces apoptosis in imatinib mesylate-resistant cell lines. Cancer, 103: 102–110. doi: 10.1002/cncr.20758
- Issue online: 17 DEC 2004
- Version of Record online: 22 NOV 2004
- Manuscript Revised: 20 SEP 2004
- Manuscript Accepted: 20 SEP 2004
- Manuscript Received: 12 MAR 2004
- Fondation de France
- Fondation pour la Recherche Médicale
- drug resistance;
- imatinib mesylate;
- BCR-ABL overexpression;
- chronic myeloid leukemia
Imatinib mesylate, a BCR-ABL tyrosine kinase inhibitor, induces apoptosis in chronic myeloid leukemia cells. Resistance to imatinib is currently the most important concern of this treatment. One of the main mechanisms of this resistance is overexpression of BCR-ABL.
In the current study, the authors investigated the correlation between BCR-ABL overexpression and apoptosis in BaF/BCR-ABL and LAMA84 cell lines resistant to imatinib suddenly deprived of the inhibitor, and compared with their sensitive counterpart.
Removal of imatinib from culture medium led to a decrease in Bcr-Abl protein expression by Day 5, which was sustained for ≥ 3 weeks of imatinib deprivation. Apoptosis was observed after 3 days of imatinib deprivation in resistant lines accompanied by caspase activation, loss of membrane asymmetry (annexin V staining), and alteration of mitochondrial potential (dihexyloxacarbocyanine iodide [DiOC6]). Transient activation of the STAT5/Bcl-xL pathway and Akt kinase activity preceded these responses.
Thus, imatinib removal led to apoptosis of BCR-ABL–overexpressing leukemic cells, a phenomenon that could be exploited to sensitize imatinib-resistant cells to the cytotoxic effect of other drugs. Cancer 2005. © 2004 American Cancer Society.
Chronic myeloid leukemia (CML) is characterized by the presence of the Philadelphia (Ph) chromosome that results from a t(9;22)(q34;q11) reciprocal translocation. The Ph chromosome harbors a BCR-ABL hybrid gene that encodes a fusion protein with deregulated tyrosine kinase activity, which is responsible for leukemogenesis in vitro and in vivo.1, 2 The process of BCR-ABL–induced in vitro transformation leads to growth factor independence and resistance to apoptosis. The BCR-ABL tyrosine kinase modulates several signaling pathways, activating proteins such as Ras, Akt, NFκ-B, and STAT5. Constitutive phosphorylation of the latter is essential for the transformation of hematopoietic cells by BCR-ABL and plays a dominant role in resistance to apoptosis through enhancement of expression of the antiapoptotic factor, Bcl-xL.3–10
Imatinib mesylate (previously known as STI571) is a BCR-ABL tyrosine kinase inhibitor that competes with adenosine triphosphate (ATP) for binding to the ABL kinase domain, thereby selectively suppressing the growth and inducing the apoptosis of BCR-ABL–positive cells.11–15 Imatinib-induced programmed cell death is accompanied by caspase activation, inhibition of the STAT5/Bcl-xL pathway, and down-regulation of PI3 kinase.10, 14–16 The impressive effects of imatinib in patients with CML have been demonstrated and confirmed in large clinical trials.17 However, imatinib resistance is a well recognized problem, particularly in patients with advanced-stage disease. To investigate this phenomenon, cell lines resistant to imatinib were generated in various laboratories,18–21 and these cells can escape the inhibitory effect of imatinib by diverse mechanisms, including overexpression of the BCR-ABL protein.
In the current study, we investigate the correlation between execution of apoptosis and dynamics of BCR-ABL expression in the presence and absence of imatinib in BaF/BCR-ABL and LAMA84 cell lines resistant to imatinib compared with their sensitive counterparts. We used 1 of 4 clones resistant to 1 μM of the inhibitor, BaF/BCR-ABL-r1, and 1 of 2 clones resistant to 2 μM, BaF/BCR-ABL-r2.18 We observed that imatinib removal from the culture of these cells induced a decrease in their BCR-ABL overexpression, a disturbance of the STAT5/Bcl-xL signaling pathway, and an increase in apoptosis.
MATERIALS AND METHODS
The imatinib-sensitive and resistant cell lines (e.g., BaF/BCR-ABL-s, BaF/BCR-ABL-r1, BaF/BCR-ABL-r2, and LAMA84-r) have been described previously.18 The sensitive line was grown in RPMI 1640 medium (Life Technologies, Cergy-Pontoise, France) supplemented with 10% fetal calf serum, antibiotics (penicillin, streptomycin), and L-glutamine (complete medium referred to as RF-10). As previously reported, resistance was defined by the capacity to survive indefinitely in the continuous presence of a given concentration of imatinib. In the current study, we used 1 of the 4 established BaF/BCR-ABL clones resistant to 1 μM, and 1 of 2 BaF/BCR-ABL clones resistant to 2 μM. LAMA84-r was resistant to 1 μM of imatinib.
ABL Flow Cytometric Analysis
Cells (5 × 105) were fixed in 1% paraformaldehyde/phosphate-buffered saline for 10 minutes, permeabilized with 0.3% saponin, and stained with an anti-ABL antibody (clone 24-11, Santa Cruz Biotechnology, Santa Cruz, CA) and then with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse immunoglobulin G (IgG; Becton Dickinson, San Jose, CA) as the secondary reagent. Next, the cells were incubated with 10 μg/mL propidium iodide (PI) for 30 minutes at 4 °C in the dark. Cells were then analyzed by flow cytometry using an XL cytometer (Beckman-Coulter, Fullerton, CA).
Annexin V Staining
Experiments were performed with FITC-annexin V (Coulter-Immunotech, Marseille, France) according to the manufacturer's instructions. Briefly, 5 × 105 cells were incubated with 1 μL of FITC-annexin V resuspended in 500 μL of 1 × binding buffer on ice for 10 minutes and then analyzed by flow cytometry. The FITC-annexin V–positive cells were considered to be apoptotic.
Mitochondrial Signs of Apoptosis
To assess mitochondrial membrane potential, 5 × 105 cells were incubated with 100 ng/mL dihexyloxacarbocyanine iodide, DiOC6(3), which produces a green fluorescence, for 20 minutes at 37 °C and then analyzed by flow cytometry. Cells exhibiting low DiOC6(3) fluorescence were considered to be apoptotic.
The level of caspase-3 activation was assessed by measurement of its capacity to cleave an Ac-Asp-Glu-Val-Asp-amino-4-methyl-coumarine (Ac-DEVD) substrate, as modified from a previously published method.14, 22 Aliquots of 7.5 × 104 cells were cultured in triplicate in RF-10 into 96-well plates in the presence or absence of imatinib. Triplicate wells with RF-10 only were used as background control. After 1 or 3 days, the plate was centrifuged at 1500 revolutions per minute for 5 minutes, the supernatant fluid was removed, and the cells were recovered with 50 μL of a buffer containing 10 mM HEPES, 5 mM dithiothreitol, 2 mM ethylenediaminetetraacetic acid (EDTA), 0.02% saponin, 1 mM phenylmethylsulfonyl fluoride, 10 μg/mL pepstatin A, 10 μg/mL leupeptin, and 72 mM fluorogenic Ac DEVD substrate (UBI, Euromedex, Souffelweyersheim, France). The plate was read on an automatic spectrofluorometer (Perkin-Elmer, Oak Brook, IL) using λexc = 380 nanometers (nm) and λem = 480 nm. After this reading, 200 μL of a 15-μg/mL PI solution was added to each well and a new reading was taken at λexc = 360 nm and λem = 600 nm to evaluate cell proliferation. The caspase-3 activity was calculated for 105 cells after taking into account the degree of cell proliferation.
Western Blot Analysis
Protein lysates were prepared according to the method of Kabarowski et al.23 with minor modifications. Protein concentrations were determined by the Bradford method (Bio-Rad, Marnes-La-Coquette, France). Approximately 100 μg of proteins was resolved on 12% or 6% polyacrylamide gels. After electrophoresis, proteins were transferred to a PVDF membrane by a semidry electrophoretic transfer. The blots were blocked in 5% milk/Tris-buffered saline (TBS) 1X (except for anti-phosphoSTAT5 antibody when blots were blocked in 5% bovine serum albumin/TBS1X-Tween) at room temperature with gentle shaking. They were then incubated with the respective antibody (1:1000 dilution) followed by an anti-rabbit or anti-mouse peroxidase-conjugated secondary IgG antibody. The reaction was revealed with an enhanced chemiluminescence detection (ECL) method (Western Lightning Chemiluminescence Reagent Plus; Perkin-Elmer). The following antibodies were used: anti-STAT5 (Santa Cruz Biotechnology), anti-phosphoSTAT5 (Cell Signaling Technology; Saint Quentin Yvelines, France), anti-ABL (clone 8E9; PharMingen, San Diego, CA), anti-Bcl-xL (PharMingen), and anti-actin (Sigma Chemical Company, St. Louis, MO).
Assessment of Akt Kinase Activity
Akt kinase activity was measured using the Akt kinase assay kit (Cell Signaling Technology) as specified by the provider. Briefly, 107 cells were suspended in lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM goat anti-mouse (GAM) ethylene glycol bis (aminoethylether tetraacetic (EGTA), 1% Triton X100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na3VO4, and 1 μg/mL leupeptin). An amount of cell extract corresponding to 2 × 106 cells was then immunoprecipitated with anti-Akt antibodies adsorbed on agarose beads. The beads were washed and incubated in the presence of 200 μM ATP and a recombinant glycogen synthase kinase (GSK)-3αβ as a substrate. After centrifugation to eliminate the agarose beads, the supernatant fluid underwent electrophoresis and Western blot analysis, using an anti-phospho GSK-3αβ (Ser21/9) as primary antibody. The antibody binding was detected by ECL.
BCR-ABL Expression Decreases on Imatinib-Deprived BaF/BCR-ABL–Resistant Cells
Analysis of BCR-ABL protein by Western blot analysis confirmed that the BaF/BCR-ABL–resistant cell lines overexpressed BCR-ABL compared with the sensitive line. The level of overexpression was higher in BaF/BCR-ABL-r2 than in BaF/BCR-ABL-r1 (Fig. 1A) as also demonstrated by flow cytometry analysis (Fig. 1B). After 5 days of culture in the absence of imatinib, BCR-ABL expression was reduced in the resistant cell lines. This reduced level was sustained for ≥ 3 weeks of imatinib deprivation (Fig. 1B) but remained higher than in the BaF/BCR-ABL–sensitive line (data not shown).
Sub-G1 Cells Express a Low Level of BCR-ABL
As shown in Figure 1B, ABL staining of the resistant cell lines exhibited a heterogeneous pattern, with a small population demonstrating low FITC fluorescence (Fig. 2A). To investigate the nature of this subpopulation, we performed double staining with FITC-ABL and PI to eliminate the apoptotic sub-G1 cells. After exclusion of these by gating only on the cycling cells, the small population with low ABL content disappeared (Fig. 2B), suggesting that it was constituted of apoptotic cells.
Imatinib Removal Induces Apoptosis in Resistant Cell Lines
Apoptosis induces loss of membrane asymmetry resulting in phosphatidyl serine (PS) exposure and alterations in mitochondrial membrane potential (Δψm). Measurement of this phenomenon by flow cytometric analysis of FITC-annexin V–stained cells showed that an average of 20% of both BaF/BCR-ABL-r1 and BaF/BCR-ABL -r2 cells exhibited signs of apoptosis after 3 days of imatinib withdrawal, and this proportion increased to 39% and 48% for BaF/BCR-ABL-r1 and BaF/BCR-ABL -r2, respectively, by Day 5 (Fig. 3A). A previous report showed that withdrawal of imatinib from LAMA84-r cultures reduced Bcr-Abl protein expression.24 Similarly to BaF/BCR-ABL-r, we showed that LAMA84-r exhibited 17% and 41% of annexin V-positive cells, respectively, after 3 and 5 days of imatinib removal (Fig. 3B).
Similar results were obtained when apoptosis was assessed by flow cytometry using the Δψm-dependent probe DiOC6(3), which detects alterations in mitochondrial membrane potential. A 3-day period of imatinib deprivation led to a decrease in the incorporation of the DiOC6(3) probe, particularly in the BaF/BCR-ABL-r2 line, indicating an increase in apoptosis in the imatinib-resistant cells (Fig. 4).
Caspase-3 activation in BaF/BCR-ABL-resistant cell lines was assessed by measurement of DEVDase activity after 1 or 3 days of culture without imatinib. Compared with the control (resistant cell lines cultured in the presence of imatinib), resistant cell lines exhibited a decrease in caspase-3 activity after 1 day of imatinib deprivation (−30% and −40% for r1 and r2, respectively; Fig. 5). By contrast, this activity was increased above control values on Day 3 (+44% and +58% for r1 and r2, respectively; Fig. 5).
Imatinib Removal Induces a Transient Activation of STAT5/Bcl-xL and PI3 Kinase/Akt Pathways
To determine whether imatinib removal plays a role in BCR-ABL–induced apoptotic signaling pathways, we analyzed the level of STAT5 phosphorylation and Bcl-xL expression in resistant BaF/BCR-ABL and LAMA84 cell lines by immunoblotting.
After 1 day of imatinib deprivation, STAT5 was hyperphosphorylated and the level of Bcl-xL was increased as compared with the patterns found in resistant lines cultured in the presence of the drug (Fig. 6). After 3 days of imatinib removal, STAT5 phosphorylation returned to the basal level (Fig. 6). The amount of STAT5 protein remained unmodified in the three resistant cell lines. Because the level of Bcl-xL was maintained relatively high after 3 days in all cell lines, other mechanisms must be suggested to allow the apoptotic events described earlier.
Imatinib-sensitive cell lines treated with increasing doses of imatinib during 2 hours showed a decrease in STAT5 hyperphosphorylation.25 To determine whether an increase in STAT5 phosphorylation level occurred rapidly, resistant cell lines were deprived of imatinib during 0.5, 2, 4, 6, 12, and 24 hours. Hyperphosphorylation of STAT5 appeared after 0.5 minutes of imatinib deprivation and was stable until 6 hours. After 12 hours, it decreased but remained more elevated than in controls. Figure 7 showed the results found with BaF/BCR-ABL-r2 (the most representative of the three resistant lines studied). The amount of STAT5 protein remained unmodified in the three resistant cell lines (data not shown).
Another mechanism responsible for resistance to apoptosis in BCR-ABL–expressing cells is the activation of the PI3 kinase/Akt pathway. After 1 day of imatinib withdrawal, we observed an increase in Akt activity in BaF/BCR-ABL-r2 compared with Time zero (Fig. 8). This activation was abolished after 3 days of imatinib removal and appeared to be at the same level as the control (BaF/BCR-ABL-r2 with imatinib).
Therefore, we conclude that imatinib deprivation-induced apoptosis is related to activation of the STAT5/Bcl-xL and PI3 Kinase/Akt pathways in our resistant lines.
Our previous study showed that the BaF/BCR-ABL imatinib-resistant cell lines established in our laboratory overexpressed BCR-ABL compared with their sensitive counterpart and that the degree of overexpression correlates directly with the concentration of imatinib to which each clone is resistant, being higher in BaF/BCR-ABL-r2 than in BaF/BCR-ABL-r1. In the current report, we demonstrate that such overexpression can be modulated by the presence or absence of imatinib in the culture medium. Indeed, imatinib removal induced a decrease in BCR-ABL expression in resistant cell lines, which, nevertheless, remained higher than in BaF/BCR-ABL-s. Imatinib removal also induced an increase in cell death that is due to an apoptotic process as demonstrated by exposure of PS groups on the plasma membrane, a decrease in DiOC(6) incorporation, and caspase-3 activation by Days 3–5 of imatinib deprivation. Furthermore, our results also indicate that resistant cell lines spontaneously exhibit a slight degree of apoptosis as detected by PI staining and caspase-3 activity.
It has been reported that BCR-ABL was localized in the cytoplasm of leukemic cells and imatinib stimulates its nuclear entry.26 Therefore, it would be of interest to determine the subcellular distribution of Bcr-Abl in our experimental system. Using immunofluorescence, we found that the localization of Bcr-Abl in resistant cell lines was cytoplasmic and that the fluorescence intensity was higher in resistant cells than in sensitive cells. Unfortunately, as imatinib removal induces apoptosis in resistant cells provoking cellular alterations, we could not interpret the results found with these cells growing without the inhibitor.
BCR-ABL confers resistance to apoptosis in leukemic cells.16, 27, 28 Protection from programmed cell death may be mediated, in part, through STAT5 up-regulation of the antiapoptotic molecule Bcl-xL9 and, in part, by phosphorylation and inactivation of the proapoptotic molecule BAD by Akt kinase.29, 30 Several reports have demonstrated that imatinib induces apoptosis in BCR-ABL–positive cells12, 14, 15 by removing the oncogenic signal that activates the antiapoptotic response, such as that triggered by STAT5/Bcl-xL. Indeed, blockade of the BCR-ABL tyrosine kinase induces dephosphorylation of STAT5, thereby decreasing Bcl-xL expression.10, 31 The other signaling pathway affected is PI3 kinase,14, 15 which interacts with STAT5 to induce leukemogenesis.32 Our data show for what to our knowledge is the first time that imatinib withdrawal disturbed STAT5/Bcl-xL and PI3 kinase/Akt signaling pathways and led to unbalanced proapoptotic and antiapoptotic signaling in BaF/BCR-ABL-r and LAMA84-r cell lines. Indeed, after 24 hours of imatinib removal, we observed hyperphosphorylation of STAT5, leading to an increase in the level of Bcl-xL, elevated Akt kinase activity, and inhibition of caspase activation. This might explain why no signs of apoptosis were detected at the membrane or mitochondrial levels at that time. In fact, STAT5 hyperphosphorylation occurred rapidly, as soon as 30 minutes after imatinib deprivation, demonstrating that Bcr-Abl kinase was dependent on the drug level. However, after 3 days of imatinib deprivation, when STAT5 phosphorylation and Akt kinase activity decreased and most likely returned to basal levels, executioner caspases were activated, leading to the membrane and mitochondrial features of apoptosis. Gesbert and Griffin9 suggest that STAT5 and PI3 kinase can synergize to enhance Bcl-xL expression. In the current study, STAT5 and PI3 kinase activation were related to the increased expression of Bcl-xL, suggesting that these pathways induced an antiapoptotic response in imatinib-deprived resistant cells. In contrast, their inhibition did not result in the decrease of Bcl-xL expression, suggesting that another mechanism most likely was involved in the apoptotic process.
The results of the current study confirm that imatinib deprivation in leukemic cells whose mechanism of resistance is BCR-ABL overexpression can lead to apoptosis, possibly via STAT5/Bcl-xL and PI3 kinase/Akt kinase pathways. This result, detected in three resistant cell lines, can be considered as a general phenomenon found in imatinib-resistant cell lines.
The overexpression of BCR-ABL, one of the main mechanisms of resistance developed by imatinib-resistant cell lines, also was detected in a few patients with CML who are resistant to imatinib therapy.33–35 Although overexpression of BCR-ABL can confer resistance to imatinib, increased expression of the protein was toxic to cells when imatinib is withdrawn abruptly as we demonstrated. This phenomenon could be exploited in the treatment of imatinib-resistant patients for whom sudden withdrawal of imatinib could sensitize the leukemic cells to an apoptosis-inducing drug. The evidence that an excess of uninhibited Bcr-Abl kinase is deleterious to the leukemic clone is provided in the clinical setting by the observed negative selection of cells with BCR-ABL gene amplification upon withdrawal of the drug when the patient loses response as described by Gorre et al.33 Indeed, in 1 patient resistant to imatinib therapy, BCR-ABL gene amplification was no longer detectable 4 weeks after discontinuation of imatinib. This patient received an alternative treatment for her leukemia. This phenomenon reflects the removal of selective pressure followed by the loss of amplified copies of the BCR-ABL gene. It has been report that reexposure of imatinib in the LAMA84-r cell line grown without the compound during several weeks induced an increase in cell death. A complete restoration of sensitivity after 4 months without imatinib was detected by the same level of BCR-ABL expression in deprived LAMA84-r and the parental LAMA84-s.24 These data and our results demonstrated that it may be preferable to stop treatment with imatinib and to use it as an alternative treatment for these patients.
The authors thank Dr. Elizabeth Buchdunger (Novartis, Basel, Switzerland) for the generous gift of imatinib mesylate.