The objective of this study was to evaluate the ability of the clinically available histone deacetylase (HDAC) inhibitor valproate to enhance the cytotoxicity of the Bcr-Abl inhibitor imatinib in imatinib-resistant cell lines.
The objective of this study was to evaluate the ability of the clinically available histone deacetylase (HDAC) inhibitor valproate to enhance the cytotoxicity of the Bcr-Abl inhibitor imatinib in imatinib-resistant cell lines.
Interactions between imatinib, and valproate have been examined in imatinib-sensitive and -resistant chronic myeloid leukemia (CML)cell lines (K562, KCL-22, CML-T1) and in bone marrow mononuclear cells (MNCs) derived from imatinib-resistant CML patients.
In imatinib-sensitive cell lines, cotreatment with imatinib 0.5 μM and valproate 5 μM for 48 hours potently enhanced imatinib-induced growth arrest and apoptosis. In resistant cell lines and in primary MNCs derived from imatinib-resistant patients, valproate restored sensitivity to the cytotoxic effects of imatinib. Coexposure of cells to valproate and imatinib was associated with repression of several genes involved in Bcr-Abl transformation. In particular, the combination valproate–imatinib downregulated the expression of Bcr-Abl and the antiapoptotic protein Bcl-2, which is particularly overexpressed in imatinib-resistant clones.
Data from this study suggested that administration of the clinically available HDAC inhibitor valproate may be a powerful strategy to enhance cytotoxic effects of imatinib in those patient resistant to imatinib or in which complete cytogenetic remission has been not reached. Cancer 2006. © 2006 American Cancer Society.
Chronic myeloid leukemia (CML) is a myeloproliferative disorder characterized by t(9;22) chromosomal translocation, which encodes for the chimeric protein Bcr-Abl. The constitutive active tyrosine kinase Bcr-Abl mediates a continuous and aberrant activation of several signal transduction pathways, including NF-kB, PI3K-Akt, STAT-5, MAPK-ERK, and others.1 These pathways promote cellular proliferation, survival, and resistance to apoptotic stimuli induced by conventional cytotoxic drugs. Recently, CML therapy has been completely revolutionized by the development of the Abl tyrosine kinase inhibitor imatinib, formerly STI-571.2, 3 In vitro studies have shown that imatinib causes growth arrest and potently induces apoptosis in Bcr-Abl–positive cell lines. Imatinib treatment of CML patients during the chronic phase of this disease determines 96% of hematologic responses and 76% of complete and major cytogenetic remissions (defined as a percentage of metaphases Ph-positive ranging from 0 to 35) after a follow-up of 18 months.4 These clinical outcomes are surprisingly higher than those achieved with traditional therapy (interferon plus cytarabine) and are associated with a better clinical tolerance. However, 24% of patients in complete hematologic remission are still positive for the t(9;22) rearrangement, when detected by cytogenetics. Moreover, in some rare cases, imatinib therapy may select and favor the emergence of drug-resistant clones.5–7 In blast crises, imatinib may also be an effective therapy, but patient responses are very short because of the development of imatinib-resistant clones.8 These clinical data suggest that additional therapies are required to enhance imatinib in the complete eradication of Bcr-Abl–positive clones.9 One possible approach involves the combination of imatinib with different agents such as chemotherapeutic agents, NF-kB inhibitors, HDAC inhibitors, proteosome inhibitors, geldanamycin, and others.10 HDAC inhibitors promote histone acetylation and subsequently chromatin relaxation and uncoiling.11, 12 These events facilitate transcription of different genes, especially those involved in cellular differentiation. Several reports have attributed to HDAC inhibitors the role of differentiating agents, and, in same leukemic cell lines, HDAC inhibitors also have been associated with the induction of apoptosis. HDAC inhibitors have been described as able to determinate terminal differentiation of Bcr-Abl–positive leukemic cell lines by a mechanism that involves the downmodulation of MAPK.13 Bcr-Abl inhibition with protein tyrosine phosphatase PTP1B overexpression or with imatinib determines terminal differentiation of K562 and reversion of the phenotype of Bcr-Abl–transformed Rat-1 fibroblasts.14 These data suggest that both Bcr-Abl inhibition and HDAC inhibition stimulate differentiation and reverse transformation of Philadelphia-positive cells. According to these considerations, HDAC inhibitors have been combined with imatinib to potently enhance imatinib antileukemic activity.15 SAHA,16 LAQ824,17 and MS-27518 enhance imatinib-induced growth arrest and apoptosis in both imatinib-sensitive and imatinib-resistant cell lines. Unfortunately, the majority of HDAC inhibitors are not readily available for routine clinical use. Recently, valproate, which is commonly administrated to treat chronic epilepsy, has been described as a potent HDAC inhibitor with antileukemic properties.19–22 In this work, we have evaluated the role of valproate as an imatinib enhancer in imatinib-sensitive and imatinib-resistant cell lines and in bone marrow samples derived from imatinib-resistant CML patients.
All reagents, unless specified, were from Sigma (Sigma-Aldrich, St. Louis, MO). Monoclonal antiphospho-tyrosine (PY-99; sc-7020), monoclonal antiBcl-2 (sc-509), polyclonal antiacetyl-Histone H3 (Lys 9/14; sc 8655), acetyl-Histone H4 (sc 8661-R), polyclonal antiHistone H3 (sc 10809) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Monoclonal antiPhospho-MAPK (M-8159) and antiβ-actin (A-5316) were from Sigma.
After we obtained informed consent, bone marrow samples from 4 imatinib-resistant patients were collected. These patients were in the chronic phase of CML as defined by published criteria.23 Three patients were interferon-α resistant and had upfront cytogenetic and hematologic resistance to imatinib as defined by published criteria.23 One patient never received therapy with interferon-α. All 4 patients were under imatinib therapy at the time of sample collection. As controls, bone marrow samples obtained from 2 normal donors and from 2 CML patients at diagnosis were collected. Both patients with newly diagnosed CML were sensitive to imatinib therapy.
K562, KCL-22, and CML-T1 are t(9;22)-positive cell lines established from an erythroid (K562) and lymphoid (KCL-22 and CML-T1) CML blast crisis. Cell lines are grown in RPMI supplemented with 10% fetal bovine serum and glutamine in a 5% CO2 saturated atmosphere. Resistant clones are a gift from Carlo Gambacorti-Passerini (National Cancer Institute, Milan, Italy) and have been established by continuous exposure to imatinib as described.24 No Bcr-Abl amplification nor Bcr-Abl point-mutations have been detected in this cell lines (data not shown).
To determinate proliferation, cells were plated at a concentration of 5 × 105 cells/mL. After 48 hours in 10% serum, cells were counted in a Burker hemocytometer chamber, with tripan blue exclusion of dead cells.
To evaluate the phosphorylation state of Bcr-Abl and its signal transduction, cells were incubated for 48 hours with the indicated drugs, washed with cold phosphate-buffered saline (PBS), and lysed with a buffer containing 20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM ethylenediamine tetraacetic acid (EDTA), 1 mM EGTA, 1% Triton X-100, 1 mM β-glycerolphosphate, and 1 mM orthovanadate and Protease Inhibitor Cocktail (Sigma, P8340). Protein concentration was determined using the Bio-Rad DC Protein Assay (Bio-Rad, Hercules, CA). Twenty μg of total protein were loaded onto 8% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and transferred to Hybond-ECL membranes (Amersham, now General Electric Health Care, Fairfield, CT). Filters were probed with appropriate antibodies and specific binding was detected by the Enhanced Chemiluminescence System (ECL system, Amersham). To evaluate the acetylation of histones, cells were treated for 12 hours with the indicated concentration of valproate, washed with ice-cold PBS and incubated on ice in 600 μL of buffer A (10 mM HEPES [pH 7.9], 10 mM potassium chloride (KCl), 0.1 mM EDTA, 0.5% NP-40, 1 mM dithiothreitol (DTT) and 1mM phenylmethylsulfonyl fluoride). After 10 minutes, nuclei were separated by centrifugation at 3000× g for 10 minutes. The supernatants (cytoplasmic extracts) were removed. Pellets were resuspended in 50 μL of buffer B (20 mM HEPES [pH 7.9], 400 mM KCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 2 μg each of leupeptin and aprotinin/mL) and incubated at 4 °C for 20 minutes with vigorous mixing. The nuclear lysates were clarified by high-speed centrifugation. Twenty μg of nuclear extract were separated on a 10% SDS-polyacrylamide gel electrophoresis. The experiments shown are representative of results obtained in at least 3 separate experiments.
An aliquot of bone marrow MNC cells, incubated with 1 μM imatinib, 5 μM valproate, and a combination of the 2 drugs for 18 hours, and the control sample were tested for their ability to give origin to colony growth in semisolid culture. The assay for multilineage colony-forming units (U) (CFU-Mix), erythroid bursts (BFU-E), and granulocyte–macrophage colony-forming U (CFU-GM) was carried out as described elsewhere.26 Briefly, 5× 104 MNCs were plated in 35-mm Petri dishes in 1-mL aliquots of Iscove-modified Dulbecco medium (Gibco Cell Culture Products, Invitrogen, Carlsbad, CA) containing: 30% FBS (Stem Cell Technologies, Vancouver, BC, Canada); 10−4 M 2-mercaptoethanol (Invitrogen); and 1.1% weight/volume (w/v) methylcellulose. Cultures were stimulated with IL-3 (10 ng/mL, Novartis, Basel, Switzerland), granulocyte–colony-stimulating factor (10 ng/mL, Amgen Inc.), granulocyte–macrophage-colony-stimulating factor (10 ng/mL, Sandoz, Holzkirchen, Germany) and erythropoietin (3 U/mL, Amgen Inc.). After incubation (37 °C, 5% CO2) for 14–18 days in a humidified atmosphere, progenitor cell growth was evaluated according to previously published criteria.26
After exposure to valproate, imatinib, and the combination of the 2 drugs, 1× 106 bone marrow mononuclear cells obtained from CML patients and cell lines were washed with PBS and resuspended in binding buffer containing Annexin V-fluorescein isothiocyanate and propidium iodide according to the manufacture's instructions (Bender MedSystems, Vienna, Austria). Annexin-V binding and propidium iodide positivity were assessed by flow cytometry on a Becton Dickinson FACScan (BD, Franklin Lakes, NJ), and analysis was performed with Cellquest (BD) software.
To prepare nuclear extract, 5× 106 cells were washed with ice-cold phosphate-buffered saline and incubated on ice in a shaker with 400 μL of cytosolic lysis buffer (10 mM HEPES [pH 7.9], 10 mM KCl, 0.1 mM EDTA, 0.5% NP-40, 1 mM DTT, and 1 mM phenylmethylsulfonyl fluoride). After 15 minutes, nuclei were separated by centrifugation at 5000 rpm for 10 minutes. The supernatants containing cytosolic extracts were removed. The pellets, containing nuclei, were resuspended in 50 μL of a second lysis buffer from TransAM™ ELISA kit (Active Motif, Rixensart, Belgium). After 30 minutes of incubation, nuclei have been clarified by high speed centrifugation.
Twenty μg of nuclear extract have been assayed for the DNA binding activity of NF-kB according to the protocol of the TRANS AM Kit. Basically the DNA binding motif of NF-kB (5′-GGGACTTTCC-3′) is coated to a 96 well plate. When nuclear extracts are added to the plate, nuclear transcriptionally active NF-kB binds to DNA determining the exposure of an epitope, which is recognized by a primary antibody directed against p65. HRP-conjugated secondary antibody provides sensitive colorimetric reaction, which is quantified by spectrophotometry. Adsorbances are measured at 450 nm with a reference wavelength of 655 nM.
At the end of the incubation period, total RNA was according to standard procedures. The reverse transcriptase step was adapted from the BIOMED one protocol.27–37 To evaluate the BCL-2 transcript amount, we used a real-time quantitative polymerase chain reaction (PCR) assay based on a specific set of primers and probe (Assays-on-Demand, Gene Expression Products) supplied by Applied Biosystems (Foster City, California). The values obtained were normalized using Abelson as control gene38 and the results were expressed using the ΔCt method as the efficiencies of both PCR reactions were determined and found to be equal. RQ-PCR reactions and fluorescence measurements were made using the iCycler (BIO-RAD).
K562, KCL-22, and CML-T1 are t(9;22)-positive cell lines established from erythroid (K562) and lymphoid (KCL-22 and CML-T1) CML blast crises. In K562 and KCL-22 cell lines, imatinib-resistant clones have been established after continuous exposure to low-dose imatinib, as previously described.24 As shown in Figure 1A, in imatinib-sensitive K562 cell line (K562-S), the exposure to a concentration of 0.5 μM imatinib for 48 hours reduces cell growth by about 60% with respect to untreated cells. Five μM valproate does not affect cellular proliferation, but when valproate is combined with imatinib, cell growth is completely blocked. Similar results have been obtained in imatinib-sensitive KCL-22 cell line (KCL-22-S) as shown in Figure 1B. In this cell line, the exposure to valproate alone reduces basal cell growth by about 40%. In imatinib-resistant K562 (K562-R), 0.5 μM imatinib does not affect proliferation, but when valproate is combined with imatinib, basal proliferation is reduced by about 60%. In imatinib-resistant KCL-22 (KCL-22-R), the exposure to the combination of imatinib and valproate reduces basal proliferation by about 50%. In KCL-22-R, treatment with valproate alone produces a 15% reduction in basal cell growth. Similar results have been observed in the imatinib-sensitive CML-T1 where combined imatinib–valproate causes a complete growth arrest (Fig. 1C).
To analyze whether valproate affects imatinib-induced apoptosis, cells were incubated with the indicated drugs for 24 hours. Apoptosis was evaluated by measuring AnnexinV-positive cells as explained in Materials and Methods. As shown in Table 1, valproate does not affect basal apoptosis in K562-S. After exposure to 0.5 μM imatinib, 20% of cells undergo apoptosis. When imatinib is combined with valproate, 28% of cells are apoptotic. In K562-R, imatinib does not induce significant apoptosis, but when it is combined with valproate, we observed a marked increased in apoptosis (18%). In KCL-22, although this cell line seems to be less sensitive to the proapoptotic stimulus of 0.5 μM imatinib, similar effects were observed. In particular, in K562-R, the combination imatinib–valproate duplicates the amount of apoptotic cells with respect to imatinib-treated cells. As a control for these experiments, proliferation and apoptosis were evaluated in the acute leukemic cell line HL60. In this cell line, 0.5 μM imatinib, 5 μM valproate, and a combination of the 2 drugs did not affect cell proliferation and apoptosis (data not shown).
|% Annexin V|
|Ctrl||2.8 ± 1||3.2 ± 2||3 ± 1||2 ± 1|
|STI-571||20 ± 3||4.2 ± 1||12 ± 3||5 ± 1|
|VA||4 ± 3||4 ± 2||3 ± 1||5 ± 2|
|STI ± VA||28 ± 2||18 ± 2||17 ± 2||12 ± 2|
An essential property of imatinib is the abrogation of CML multilineage colony-forming unit (CFU-Mix), granulocyte–macrophage colony-forming unit (GM-CFU), and erythroid burst-forming unit (BFU-e) colony formation. In imatinib-resistant patients, imatinib does not affect colony formation of bone marrow samples. Bone marrow mononuclear cells have been isolated from 4 imatinib-resistant patients to test the clonogenic potential in the presence of valproate, imatinib, and the combination of the 2 drugs. Imatinib alone does not affect growth of CFU-Mix, GM-CFU, or BFU-e colonies, but when it is combined with valproate, the number of colonies is sensibly reduced as reported in Figure 2A. In particular, imatinib alone produces a reduction of only about 15% of CFU-GM colony formation but has no effects on BFU-E and CFU-Mix. Valproate alone does not affect colony formation, but when it is combined with imatinib, CFU-GM is reduced by about 65%, BFU-E by 55%, and CFU-Mix by 65%. For controls, we evaluated effects of valproate and the combination with imatinib in imatinib-sensitive CML at initial diagnosis. Valproate alone did not affect colony formation and did not enhance imatinib-induced colony-formation inhibition (data not shown).
After 12 hours of incubation with imatinib, valproate, and the combination of the 2 drugs, mononuclear bone marrow cells from the 4 resistant patients were assayed for the expression of Annexin V by flow cytometry. One normal bone marrow sample and 1 case of CML at initial diagnosis were assayed as controls. As reported in Figure 2B, in 3 imatinib-resistant patients, the exposure to the combination imatinib–valproate is associated with a marked apoptosis induction. One imatinib-resistant patient was insensitive to the combination of the 2 drugs.
Prolonged culture of K562 cells with low-dose imatinib determines the onset of resistant clones, as previously described.24 In our resistant clones, Bcr-Abl gene was not amplified, and the tyrosine kinase domain was not mutated (data not shown). As reported in Figure 3A, the level of expression of Bcr-Abl protein in imatinib-sensitive and imatinib-resistant K562 is comparable, the state of phosphorylation of Bcr-Abl is similar, but K562-R shows a different set of tyrosine-phosphorylated proteins. In particular, K562-R expresses novel phospho-proteins between 60 and 100 Kd in weight. Moreover, resistant clone exhibits a different pattern of activation of MAPK, with an increased basal phosphorylation of the p44 subunit and a marked overexpression of the antiapoptotic protein Bcl-2. To determinate whether valproate interferes with Bcr-Abl signal transduction, both K562-S and K562-R were cultured for 48 hours in the presence of 0.5 μM imatinib, 5 μM valproate, and a combination of the 2 drugs. Cells were lysed, and Western immunoblotting was performed as described in Material and Methods. As shown in Figure 3B, the combination valproate–imatinib downmodulates the expression of Bcr-Abl protein in both K562-S and K562-R, as reported with other HDAC inhibitors.16, 17 To prove that valprote acts as an HDAC inhibitor, K562-S was treated for 12 hours with 5 μM and 1 mM valproate, and then nuclear extract was obtained as described. Western blot analysis of nuclear extracts revealed increased acetylation of Histone H3 and Histone H4 (Fig. 3C).
To assess whether valproate interferes with the expression of proteins involved in the control of differentiation and apoptosis, K562-S, K562-R, and primary CML cells were cultured for 12 hours in the presence of 5 μM and 1 mM valproate. Twenty μg of total lysate was assayed for expression of antiapoptotic protein Bcl-2 which mediates an antiapoptotic signal in CML27–29 and p21, which is involved in the control of G2-M transition. One mM valproate down-modulates Bcl-2 in K562 (both sensitive and resistant to imatinib) and in primary CML cells (Fig. 4A,B) and up-regulates p21 (Fig. 4A). Interestingly, in K562-R, 5 μM valproate determines an increase in p21 protein expression. In CML primary cells, valproate down-modulates Bcr-Abl protein expression as has been described with other HDAC inhibitors.16, 17 To further investigate the mechanism of Bcl-2 down-modulation, quantitative PCR was performed as described in Material and Methods. Figure 4C shows that valproate interfered directly with Bcl-2 gene transcription. K562 has also been cultured in the presence of valproate and MG-132, a proteosome inhibitor. Pretreatment with 1 μM MG-132 does not block valproate-induced Bcl-2 downmodulation, thus suggesting that valproate does not affect protein stability (data not shown).
Valproate has also been described as an inhibitor of the antiapoptotic transcriptional factor NF-kB.32 To verify whether valproate interferes with NF-kB DNA binding activiy, cells were treated for 24 hours in the presence of valproate at the indicated concentrations and then stimulated with 100 ng/mL of the prototypical NF-kB activator TNF-α for 15 minutes. As shown in Figure 5, valproate is associated with a modest, but significant, reduction of TNF-α-induced NF-kB activation both in K562-S and K562-R. Exposure to valproate alone does not cause activation of NF-kB as has been described with other HDAC inhibitors (data not shown).
In this work, we have demonstrated that the clinical available HDAC inhibitor valproate enhances imatinib-induced growth arrest and apoptosis both in imatinib-sensitive and imatinib-resistant cell lines and in primary cells derived from imatinib-resistant CML patients. Valproate is a short chain branched fatty acid that is extensively used to treat chronic seizures. Recently, it has been described as a potent inhibitor of HDAC at the therapeutic concentrations. HDAC inhibition is responsible for the acetylation of histones mediated by histone acetyl transferase (HATs). This event terminates in a differentiative and proapoptotic signal in many leukemic models. The importance of valproate is that it is already a clinically available HDAC inhibitor, whereas other HDAC inhibitors are associated with various toxic side effects. Initial clinical data on the effects of valproate in hematologic malignancies have already been published,30 suggesting that valproate could be easily combined with imatinib therapy. Several preclinical experiments have shown that HDAC inhibitors are potent enhancers of imatinib in blocking cellular proliferation and in activating proapoptotic pathways. Our data, using valproate with imatinib, are similar to data reported with other HDAC inhibitors. Despite the great number of these observations, the mechanism by which valproate, as with other HDAC inhibitors, enhances the antileukemic properties of imatinib is still a matter of debate. It has already been described that exposure to valproate interferes with expression of key proteins involved in control of cellular proliferation and differentiation, such as p21, PPARδ, MAPK, and proteins involved in the activation of catenin. This consideration is supported by recent experimental evidence, obtained in multiple myeloma cell lines, where gene expression profiling has shown that HDAC inhibitors determine repression of genes involved in oncogenic transformation and survival.31 In particular, HDAC inhibitors repress expression of IGF-1, IGF-1R, NF-kB, Ras, FLt3-L, Abl, and other genes that increase cellular proliferation and survival. These data revolutionize the common idea that HDAC inhibitors are associated with increased gene transcription, as reported in promyelocytic and core-binding factor leukemias. This novel interpretation suggests that the antileukemic properties of HDAC inhibitors may be related to impressive modification of the expression of genes involved in oncogenic signal transduction of cancer cells by rendering them more susceptible to apoptotic stimuli. In line with these observations, here we have demonstrated that valproate downmodulates the antiapoptotic protein Bcl-2. Bcl-2 is essential to the transforming potential of Bcr-Abl, as has been clearly demonstrated by Jaiswal and Cirinna.27, 28 Disruption of the expression of this protein may render cells more susceptible to apoptotic stimuli induced by imatinib exposure. Moreover, we have shown that Bcl-2 is expressed more in imatinib-resistant clones, which suggests that the combination valproate–imatinib may be of particular utility in resistant clones. Bcl-2 downmodulation appears to be regulated at the transcriptional level, because valproate dramatically reduces Bcl-2 mRNA levels. The combination valproate–imatinib downmodulates the expression of Bcr-Abl itself. This event may be extremely important in those cases in which imatinib resistance is due to Bcr-Abl overexpression or to the presence of a point mutation that reduces sensitivity to the drug. Bcr-Abl downmodulation may render the concentration of imatinib sufficient to completely inhibit Bcr-Abl kinase. An important point of discussion is related to the dosage of the drug. Our biologic assays have shown that low-dose valproate (5 μM) are sufficient to restore sensitivity to imatinib in imatinib-resistant clones, whereas biochemical downmodulation of Bcr-Abl and Bcl-2 is detectable only at 1 mM. The discrepancy between biologic and biochemical assays may in part be explained because of the higher sensitivity of resistant clones to valproate with respect to sensitive clones. In K562-R, but not in K562-S 5, μM valproate determines a modest, but detectable, acetylation of histone H3 and p21 overexpression. Anyhow, valproate may exhibit additional anticancer properties. Accumulating published evidence suggests that HDAC inhibitors activate the antiapoptotic transcription factor NF-κB.33–36 This event may limit HDAC-induced apoptosis. Inhibition of NF-kB nuclear translocation or transcriptional activity has been described as a powerful strategy to significantly increase apoptosis in nonsmall cell lung cancer cells after the addition of HDAC inhibitors.36 These data suggest that combined molecular targeting that inhibits both NF-kB and HDAC activities may provide a significant antineoplastic effect. Valproate has also been described as an inhibitor of NF-kB DNA-binding activity and transcriptional activity.32 Here we have shown that valproate is a modest inhibitor of TNF-α induced NF-kB activation in K562 or, at least, valproate treatment is not associated with NF-kB activation as has been described with Butyrate, Saha, and other HDAC inhibitors.33–36 This event may attribute to valproate a dual inhibitor role, acting against HDAC and NF-kB signal transduction pathway rendering cells more susceptible to apoptosis. In conclusion, we suggest that valproate should be combined with imatinib to completely eradicate residual Philadelphia-positive clones in those patient resistant to imatinib or in which complete cytogenetic responses have not been yet reached.