Cancer Therapy: Preclinical Histone Deacetylase Inhibitors Induce a Very Broad, Pleiotropic Anticancer Drug Resistance Phenotype in Acute Myeloid Leukemia Cells by Modulation of Multiple ABC Transporter Genes

Purpose: Histone deacetylase inhibitors (HDACi) are being studied in clinical trials with the aim to induce cellular differentiation, growth arrest, and apoptosis of tumor cells. Recent reports suggest that the multidrug resistance-1 (MDR1) gene is regulated by epigenetic mechanisms. To investigate whether additional drug transporters are regulated by HDACi and how this affects cytotoxicity, acute myeloid leukemia (AML) cells were examined. Experimental Design: AML cells were cultured in the presence of phenylbutyrate, valproate, suberoylanilide hydroxamic acid, or trichostatin A and analyzed for drug transporter expression and function as well as sensitivity to anticancer drugs. Results: MDR1, breast cancer resistance protein (BCRP), and multidrug resistanceassociated proteins (MRP) 7 and 8 were induced in a doseand time-dependent manner as shown by semiquantitative PCR. The pattern of gene induction was cell line specific. Phenylbutyrate induced P-glycoprotein and BCRP expression and the efflux of drugs as determined with labeled substrates. KG-1a cells treated with phenylbutyrate developed resistance to daunorubicin, mitoxantrone, etoposide, vinblastine, paclitaxel, topotecan, gemcitabine, and 5-fluorouracil; as a result drug-induced apoptosis was impaired. Chromatin immunoprecipitation revealed the hyperacetylation of histone proteins in the promoter regions of MDR1, BCRP, and MRP8 on valproate treatment. Furthermore, an alternative MRP8 promoter was induced by HDACi treatment. Conclusions: Exposure of AML cells to HDACi induces a drug resistance phenotype broader than the “classic multidrug resistance,” which might negatively affect treatment effectiveness. Acute myeloid leukemia (AML) can be treated with nucleoside analogues such as cytosine arabinoside in combination with anthracylines (1). A major proportion of patients will, however, die from refractory or relapsing leukemia. Extrusion of chemotherapeutic agents by transporter proteins is a major obstacle for successful treatment of cancer. Overexpression of multidrug resistance-1 (MDR1), also known as ATP-binding cassette ABCB1, is an independent prognostic factor for treatment failure in AML (2). The encoded P-glycoprotein (P-gp) is a membranebound transporter that extrudes natural toxins across the plasma membrane (Supplementary Table S1; reviewed in refs. 3, 4). Similar patterns of resistance are conferred by breast cancer resistance protein (BCRP; also known as mitoxantrone resistance gene, MXR, or ABCG2) and multidrug resistance-associated proteins (MRP) 1 (ABCC1), 2 (ABCC2), 3 (ABCC3), and 7 (ABCC10; refs. 4–7). BCRP overexpression is associated with increased risk of relapse and a shorter disease-free survival in AML, especially in combination withMDR1 (8). The transport activity of P-gp, MRP1, MRP7, and BCRP is inhibited by relatively nontoxic compounds such as cyclosporine (9, 10). Other ABC transporters such as MRP4 (ABCC4), MRP5 (ABCC5), or MRP8 (ABCC11) render cells insensitive to nucleoside analogues (7). MRP8 confers resistance to antiviral drugs including 2′,3′-dideoxycytidine and fluoropyrimidines (7, 11, 12). Moreover, BCRP confers resistance to nucleoside analogues in addition to transport of natural toxins (13). In contrast, most ABC transporters are not involved in drug resistance (e.g., ABCG1). Authors' Affiliations: III. Medical Department, Technische Universität München, Munich, Germany; Internal Medicine I, Universitätsklinik Freiburg, Freiburg, Germany; and HELIOS Schlossbergklinik, Oberstaufen, Germany Received 8/4/08; revised 2/23/09; accepted 2/25/09; published OnlineFirst 5/19/09. Grant support: German José Carreras Leukemia Foundation Project R04/17 (T. Licht and U. Keller), R06/40f (M. Lübbert) and LaCaixa-DAAD grant 314, A/05/29785 (J. Duque-Afonso). The costs of publication of this article were defrayed in part by the payment of page charges. This articlemust therefore be herebymarked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). U. Keller and T. Licht contributed equally to this work. Requests for reprints: Ulrich Keller, III. Medical Department, Technische Universität München, Ismaninger Str. 15, 81675 Munich, Germany. Phone: 49-89-41407435; Fax: 49-89-41404879; E-mail: [email protected]. F 2009 American Association for Cancer Research. doi:10.1158/1078-0432.CCR-08-2048 3705 Clin Cancer Res 2009;15(11) June 1, 2009 www.aacrjournals.org Research. on April 14, 2017. © 2009 American Association for Cancer clincancerres.aacrjournals.org Downloaded from In clinical treatment, resistance to chemotherapy is frequently observed after initial response despite continued treatment. The MDR1 promoter is inducible by anticancer drugs (14), environmental stress (15), and epigenetic mechanisms such as DNA demethylation (16). Histone deacetylase inhibitors (HDACi) such as trichostatin A or valproate induce apoptosis or growth arrest or restore suppressed myeloid differentiation in AML (17). Several HDACi have been introduced into the treatment of leukemias. They were used alone or in combination with DNA demethylating agents or all-trans retinoic acid (18, 19). These studies have shown clinical remissions. HDACi can act synergistically with certain DNA-binding drugs in experimental models (20); thus, combination treatment is being investigated in clinical trials. Recent reports have, however, suggested that HDACi may modulate MDR1 expression. Leukocytes of patients and colon and renal cancer cells treated with the HDACi depsipeptide or trichostatin A, respectively, display increased MDR1 expression (21, 22). P-gp induction has also been observed in human and murine cells exposed to valproate (23). Treatment with the HDACi apicidine increases rhodamine-123 efflux and induces paclitaxel resistance in HeLa cells (24). Furthermore, depsipeptide combined with all-trans retinoic acid increases MDR1 levels in acute promyelocytic leukemia (25). Thereby, resistance to doxorubicin is induced, which is reversible by the P-gp inhibitor PSC833 (25). These reports argue for the possibility that combinations of HDACi and anticancer drugs might lead to reduced cytotoxic effects by inducing multidrug resistance-associated drug transport. In the current study, we show that HDACi induce in AML cells the expression of several ABC transporters. We find that exposure to HDACi can induce a pleiotropic anticancer drug resistance, which comprises nucleoside analogues in addition to the “classic multidrug resistance” phenotype. Materials and Methods Cells and culture conditions. Cell lines K562, CMK, HL-60, KG-1a, and SK-Hep-1, obtained from Deutsche Sammlung von Mikroorganismen und Zellkulturen, were cultured in RPMI 1640 (Invitrogen). The cancer lines KB-3-1 and KB-8-5 were kindly provided by Dr. Michael M. Gottesman (National Cancer Institute, NIH). KB cells were grown in DMEM (Invitrogen). KB-8-5 cells were selected with 10 ng/mL colchicine (Sigma-Aldrich) and transferred to DMEM without colchicine 1 day before experiments were done. Mononuclear cells were isolated from leukemic AML patients by Ficoll and cultured in Iscove's medium. Medium for cell lines contained 10% fetal bovine serum (PAA), whereas patient cells were grown in 20% fetal bovine serum. All media were supplemented with 2 mmol/L L-glutamine, 100 units/mL penicillin, and 100 μg/mL streptomycin. All cells were cultured at 37°C in a fully humidified atmosphere containing 5% CO2. Trichostatin A, valproate (Sigma-Aldrich), suberoylanilide hydroxamic acid (SAHA; vorinostat), and phenylbutyrate (all from Alexis Biochemicals) were used at the indicated concentrations. These concentrations did not show major toxicity in preliminary experiments. Semiquantitative reverse-transcription (rt) PCR for ABC transporters. Total RNA was extracted using the Miniprep Kit (Stratagene). cDNA was generated with SuperScript II reverse transcriptase (Invitrogen). Semiquantitative PCR was done with AmpliTaq Gold DNA polymerase (Applied Biosystems). Unpublished primers are summarized in Supplementary Table S2. The amount of cDNA was adjusted for equal loading as determined with a housekeeping gene, β-actin. Each PCR started with initial denaturing steps (96°C, 2 min and 92°C, 4 min). Amplification cycles consisted of 1 min at 94°C, 1 min at 56°C, and 2 min at 72°C. From preliminary experiments, cycle numbers were adjusted for each cell line so that the reactions were in the linear range. The reaction was finished by a 8 min incubation at 72°C. Amplified products were visualized by ethidium bromide on a 2.0% agarose gel. The band intensities were calculated with the image analysis software ImageJ. After substraction of background and normalization against the respective β-actin band intensities, the fold expression of MDR1, BCRP, MRP7, and MRP8 relative to untreated cells was calculated. Immunoblot analysis. Phenylbutyrate-treated or control cells were lysed with radioimmunoprecipitation assay buffer containing 5 mmol/L EGTA, 150mmol/LNaCl, 50mmol/L Tris, 1% Triton X-100, and protease inhibitors (complete protease inhibitor tablets; Roche Molecular Biochemicals) for 30 min on ice. Whole-cell lysates were separated by SDS-PAGE and subjected to immunoblot analysis with anti-P-gp (C-219; Signet Laboratories), anti-BCRP (clone BXP-21; Chemicon), or anti-β-Actin (Sigma-Aldrich). The antigen-antibody complex was detected with enhanced chemoluminescence kit (SuperSignal; Pierce) following manufacturer's protocol. Analysis of drug efflux by flow cytometry. After culturing in the presence or absence of phenylbutyrate or valproate, cells were harvested and transferred to DMEM/fetal bovine serum and 30 ng/mL BODIPYpaclitaxel (Molecular Probes). After 30 min at 37°C, cells were spun and resuspended in medium with or without 5 μmol/L cyclosporin A (Sigma-Aldrich) in addition to BODIPY-paclitaxel. One hour later, cells were spun, resuspended, and immediately analyzed by flow cytometry. To identify the amount of BODIPY-paclitaxel removed by drug efflux, histograms of cells in which P-gp was inhibited by cyclosporin A were overlaid. The distance between the two peaks thus indicates the amount of drug transport activity. Similarly, for analysis of BCRP transport activity, cells were loaded with 20 μmol/L mitoxantrone; fumitremorgin C (Alexis Biochemicals) at 10 μmol/L was used as an inhibitor. Accumulation of radiolabeled drugs. [H]daunorubicin, [H]paclitaxel, and [H]5-fluorouracil (5-FU) were purchased from Hartman Analytic. For drug transport experiments, they were mixed with the respective cold compounds to adjust final drug concentrations as indicated. Cells were washed with PBS and counted. Viable cells (1 × 10) were resuspended in medium containing radiolabeled drugs. After Translational Relevance Many cancers are initially responsive to chemotherapy but later develop cross-resistance to drugs that have not been administered before. The regulation of drug resistance genes is not completely understood. Histone deacetylase inhibitors (HDACi) are being investigated in clinical trials because these compounds can act synergistically with certain anticancer drugs in experimental models and may be useful in combination treatment. We find, however, that exposure of cancer cells to HDACi can induce several drug resistance-associated ABC transporters, leading to a broad-spectrum anticancer drug resistance, which comprises nucleoside analogues in addition to the “classic multidrug resistance” phenotype. The combined treatment may thus result in treatment failure. Furthermore, we find that the sequence of administration is important for the efficacy of drug treatment. Our data will help to design clinical protocols for a combination of chemotherapy with HDACi in which inactivation of anticancer drugs can be avoided. 3706 Clin Cancer Res 2009;15(11) June 1, 2009 www.aacrjournals.org Cancer Therapy: Preclinical Research. on April 14, 2017. © 2009 American Association for Cancer clincancerres.aacrjournals.org Downloaded from incubation at 37°C for indicated periods, ice-cold PBS was added to the samples. Samples were collected on ice, spun, and transferred to scintillation tubes. Opti-Fluor solution (Perkin-Elmer) was added, and incorporated activities were assessed with a liquid scintillation counter (Perkin-Elmer). Results are expressed as drug content per 1 × 10 cells. Analysis of apoptosis by Annexin V binding. Apoptosis induction was assayed by detection of phosphatidylserine on the plasma membrane as described earlier (26). In brief, cells were stained with FITC-labeled Annexin V (BD Pharmingen) and counterstained with propidium iodide. Numbers of apoptotic, that is, Annexin V-positive cells, were assessed by flow cytometry. Real-time PCR. For real-time PCR, cDNA was prepared from 2 μg RNA using the Omniscript cDNA synthesis kit (Qiagen). Real-time PCR was done using Platinum SYBR Green qPCR SuperMix (Invitrogen) on a ABI-Prism 7700 (Applied Biosystems). Data analyses were done by comparing Ct values with a control sample set as 1. The unpublished primers are included in Supplementary Table S2. Analysis of the MRP8 promoter region by rapid amplification of 5′-cDNA ends. CMK cells were incubated with or without 3 mmol/L valproate for 24 h. Total RNA was isolated. RNA (5 μg) was analyzed with the rapid amplification of 5′-cDNA ends (Invitrogen) according to the manufacturer's instructions. First-strand cDNA was synthesized Fig. 1. Induction of drug resistance genes in myeloid leukemia cells by HDACi. A, time course of ABC transporter gene expression in CMK cells exposed to SAHA. CMK cells were incubated for the indicated periods with SAHA at 2 or 8 μmol/L or in the absence of SAHA. Total RNA was isolated, and a semiquantitative rt-PCR was done with primers specific for human ABC transporters. β-Actin mRNA expression was used as a control for equal loading. cDNA from SK-Hep-1 hepatoma cells served as a positive control for MRP2 expression. Representative experiment. B, modulation of MDR1, BCRP, MRP7, and MRP8 mRNA expression by SAHA was assayed by rt-PCR. KG-1a and HL-60 cells were cultured for 20 h with or without 2 or 8 μmol/L SAHA. β-Actin expression was used as loading control. C, effects of various HDACi on gene expression was analyzed by rt-PCR. Cells were treated with 165 nmol/L trichostatin A, 3 mmol/L valproate, or 3 mmol/L phenylbutyrate for 24 h or, as a negative control, with medium (M) or the solvent DMSO (D). Representative experiment. Bottom, fold expression ofMDR1, BCRP, MRP7, and MRP8 after valproate or phenylbutyrate treatment relative to the expression in untreated cells. Data were calculated after quantification of band intensities with ImageJ software. Fold expression ± SD of three independent experiments. M, medium control. D, reversibility of gene induction. CMK cells were cultured for 24 h in the presence or absence of 3 mmol/L phenylbutyrate or 3 mmol/L valproate. Phenylbutyrateor valproate-treated and untreated cells were then harvested, transferred to fresh medium, and cultured for another 24 h with or without the respective HDACi. RNA was isolated and rt-PCR was done. M, medium; W, washout. 3707 Clin Cancer Res 2009;15(11) June 1, 2009 www.aacrjournals.org HDAC Inhibitors Induce Pleiotropic Drug Resistance Research. on April 14, 2017. © 2009 American Association for Cancer clincancerres.aacrjournals.org Downloaded from from total RNA using a gene-specific primer 1 (Supplementary Table S2) and SuperScript II reverse transcriptase. A homopolymeric tail was then added to the 3′-end of the cDNA using terminal deoxynucleotidyl transferase and dCTP. PCR was done using Taq DNA polymerase, a deoxyinosine-containing anchor primer, and a gene-specific primer. The following nested PCR used an Abridged Universal Amplification Primer in combination with further gene-specific primers (Supplementary Table S2). PCR products were separated by agarose gel electrophoresis, isolated with a gel extraction kit (Qiagen), and analyzed by automated sequencing (Supplementary Table S3). Chromatin immunoprecipitation. Chromatin immunoprecipitation was done as described elsewhere (27). Antibodies to acetylated histones H3, H4, and H3K9 were from Upstate. Rabbit IgG (Santa Cruz Biotechnology) served as a negative control. The protocol was slightly modified for background reduction by extension of a preclearing step to 3 h and extensive washing. Following reversal of histone-DNA crosslinks, DNA was isolated with the PCR Purification Kit (Qiagen). Fragments from promoter regions of MDR1, BCRP, and MRP8 were amplified by quantitative PCR using the SYBR Green Master Mix for LightCycler 480 (Roche). Specific primers are summarized in Supplementary Table S2. Fig. 2. HDACi increase the protein level and the transport activity of P-gp and BCRP. A, immunoblot analysis of P-gp and BCRP expression on phenylbutyrate treatment. CMK and KG-1a cells were treated with or without phenylbutyrate at the indicated concentrations for 24 h. β-Actin was used as a loading control. B, drug efflux on HDACi treatment. CMK and KG-1a cells were incubated for 24 h with or without 3 mmol/L phenylbutyrate or 3 mmol/L valproate, respectively. For assessment of drug transport, cells were first incubated in complete medium with 30 ng/mL BODIPY-paclitaxel for 30 min at 37°C and incubated for another 60 min at 37°C in 30 ng/mL BODIPY-paclitaxel in the presence or in the absence of 5 μmol/L cyclosporin A. BODIPY-paclitaxel fluorescence was determined by flow cytometry. The drug efflux activity is related to the distance between peaks in the presence (right peaks) and absence (left peaks) of the P-gp inhibitor cyclosporin A. Mean fluorescence relative to the mean fluorescence of untreated cells. Mean and SD of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001. C, experiments were done as in B. Mitoxantrone (20 μmol/L) was used with the inhibitor fumitremorgin C (10 μmol/L). Histograms of mitoxantrone in the presence and in the absence of fumitremorgin C are shown in separate panels because of short distance between the peaks. M, medium. 3708 Clin Cancer Res 2009;15(11) June 1, 2009 www.aacrjournals.org Cancer Therapy: Preclinical Research. on April 14, 2017. © 2009 American Association for Cancer clincancerres.aacrjournals.org Downloaded from Statistical analysis. The statistical functions of Excel and GraphPad Prism were used for all analyses.