Article pubs.acs.org/molecularpharmaceutics
Multidrug Efflux Pumps Attenuate the Effect of MGMT Inhibitors Karl-Heinz Tomaszowski,† Ralf Schirrmacher,‡ and Bernd Kaina*,† †
Department of Toxicology, University Medical Center, Obere Zahlbacher Strasse 67, D-55131 Mainz, Germany Montreal Neurological Hospital and Institute, 3801 University Street, Montréal, Quebec H3A 2B4, Canada
‡
ABSTRACT: Various mechanisms of drug resistance attenuate the effectiveness of cancer therapeutics, including drug transport and DNA repair. The DNA repair protein O6methylguanine-DNA methyltransferase (MGMT) is a key factor determining the resistance against alkylating anticancer drugs inducing the genotoxic DNA lesions O6-methylguanine and O6-chloroethylguanine, and MGMT inactivation or depletion renders cells more susceptible to treatment with methylating and chloroethylating agents. Highly specific and efficient inhibitors of the repair protein MGMT were designed, including O6-benzylguanine (O6BG) and O6-(4-bromothenyl)guanine (O 6 BTG) that are nontoxic on their own. Unfortunately, these inhibitors do not select between MGMT in normal and cancer cells, causing nontarget effects in the healthy tissue. Therefore, a targeting strategy for MGMT inhibitors is required. Here, we used O6BG and O6BTG conjugated to β-D-glucose (O6BG-Glu and O6BTG-Glu, respectively) in order to selectively inhibit MGMT in tumors, harnessing their high demand for glucose. Both glucose conjugates efficiently inhibited MGMT in several cancer cell lines, but with different extents of sensitization to DNA alkylating agents, with lomustine being more effective than temozolomide. We further show that the glucose conjugates are subject to ATP-binding cassette (ABC) transporter mediated efflux, involving P-glycoprotein, MRP1, and BCRP, which impacts the efficiency of MGMT inhibition. Surprisingly, also O6BG and O6BTG were subject to an active transport out of the cell. We also show that pharmacological inhibition of efflux transporters increases the induction of cell death following treatment with these MGMT inhibitors and temozolomide. We conclude that strategies of attenuating the efflux by ABC transporters are required for achieving successful MGMT targeting. KEYWORDS: MGMT, inhibitors, ABC transporter, drug targeting, DNA repair
■
INTRODUCTION Therapy of most cancers includes a combination of surgery, radiotherapy, and systemic chemotherapy. Alkylating agents, such as temozolomide (TMZ), dacarbazine, fotemustine, and lomustine (CCNU), are important anticancer drugs for brain tumors, melanomas, lymphomas, and sarcomas.1 The major cytotoxic properties of these chemotherapeutic agents are mediated by O6-alkylguanine adducts, introduced into the DNA upon treatment.2 Methylating anticancer drugs such as TMZ generate O6-methylguanine (O6MeG), a highly mutagenic, carcinogenic, clastogenic, and cytotoxic DNA lesion.3 The cytotoxic effect of O6MeG is cell cycle-dependent and requires processing through mismatch repair (MMR).4 If not repaired, O6MeG mispairs with thymine during DNA replication, leading to G:C to A:T transition mutations.5,6 The O6MeG:T mispair is also recognized by the MMR proteins MutS homologue 2 and 6 (MSH2 and MSH6). These proteins initiate the excision of the misincorporated thymine, but due to the mispairing properties of O6MeG, thymine is reinserted opposite O6MeG, resulting in futile cycles of MMR creating long stretches of single-stranded DNA gaps.7 The replication of DNA containing these structures generates DNA double-strand breaks (DSBs), which trigger cell death by activating the apoptosis pathway.8,9 © 2015 American Chemical Society
In contrast, the cytotoxic effect of chloroethylnitrosoureas such as lomustine occurs in an MMR-independent manner by generating O6-chloroethylguanine (O6ClEtG).3 If not repaired, intramolecular rearrangement of O6ClEtG produces the intermediate N1-O6-ethanoguanine, ultimately leading to the formation of an N1-guanine-N3-cytosine interstrand cross-link (ICL).10 These cross-links block DNA polymerases, and, in contrast to O6MeG, formation of DSBs occurs in the first round of replication.11 The DNA repair protein O6-methylguanine-DNA methyltransferase (MGMT) protects against these lesions by repairing O6-alkyl adducts in a stoichiometric manner, transferring the alkyl group to an internal cysteine residue resulting in restoring of guanine and inactivation and proteasomal degradation of MGMT.12 MGMT plays a critical role in protecting normal tissues against the adverse effects of O 6 -alkylguanine adducts.13−15 In tumors, MGMT is the primary cause of resistance against O6-alkylating anticancer drugs.2,16 Therefore, Received: Revised: Accepted: Published: 3924
May 4, 2015 August 28, 2015 September 17, 2015 September 17, 2015 DOI: 10.1021/acs.molpharmaceut.5b00341 Mol. Pharmaceutics 2015, 12, 3924−3934
Article
Molecular Pharmaceutics
MGMT inhibitors, which needs to be tested in preclinical/ clinical trials.
pharmacological inactivation of MGMT renders cancer cells more susceptible to treatment with SN1-alkylating agents. The MGMT inhibitors O6-benzylguanine (O6BG) and O6-(4bromothenyl)guanine (O6BTG, lomeguatrib) efficiently inactivate MGMT in tumors without being toxic. Clinically applied, however, MGMT inhibitors have thus far failed to produce the expected therapeutic results, because they inhibit MGMT in both normal and tumor tissue, thus leading to a general sensitization, which in turn requires a dose reduction of the alkylating anticancer drug. Since the conventional MGMT inhibitors increase the bone marrow toxicity of alkylating agents, a targeting strategy is desired in order to improve chemotherapy effects by selective inhibition of MGMT in tumor tissue.17 So far, strategies for a selective MGMT inactivation were based on local, intracranial administration18 or chemical modifications of the commercially available MGMT inhibitor O6BG. Thus, approaches have exploited the overexpression of folate receptors on cancer cells by folate-O6BG conjugate or targeted the hypoxic microenvironment in solid tumors by an analogue of O6BG which is converted to a potent MGMT inhibitor under oxygen-deficient conditions.19,20 Unfortunately, despite initial encouraging findings, no successful clinical trials have been completed to date. One reason might be drug resistance mechanisms, including multidrug resistance (MDR), which were shown to limit the cytotoxic effect of many anticancer drugs. Most commonly, MDR in tumors is mediated by ATP-binding cassette (ABC) transporters, which eliminate drugs by active efflux.21 There are 49 of these transporter proteins that form a superfamily of membrane proteins that transfer structurally diverse molecules across biological membranes in an ATPdependent manner.22 Due to their role as a detoxification system, ABC transporters are overexpressed in a broad range of human cancers. The transporters P-glycoprotein (P-gp), multidrug resistance protein 1 (MRP1), and breast cancer resistance protein (BCRP) have been well characterized as to their contribution to therapy resistance and reduced patient survival.23−26 Therefore, several approaches have been investigated to reverse efflux-mediated MDR, including pharmacological inhibition of ABC transporters, targeted downregulation of MDR genes, or the design of anticancer drugs to evade efflux by ABC transporters.27 Previously, we synthesized compounds to achieve an active targeting of cancer cells by coupling the MGMT inhibitors O6BG or O6BTG to β-D-glucose. Two of these glucose conjugates, referred to as 2-amino-6-(benzyloxy)-9-(octyl-β-Dglycosyl)purine (O6BG-Glu) and 2-amino-6-(4-bromotiophen2-yl-methoxy)-9-(octyl-β-D-glycosyl)purine (O6BTG-Glu), efficiently inhibited MGMT repair activity.28 Here, we analyzed the toxicity of glucose-conjugated MGMT inhibitors in combination with alkylating agents on cancer cells. Maximal sensitization provoked by the glucose conjugates was strongly dependent on the alkylating agent, the MGMT activity, and the influence of multidrug transporters. For the first time, this work demonstrates an active efflux out of the cell of the nonconjugated and glucose-conjugated MGMT inhibitors by the ABC transporters P-glycoprotein, MRP1, and BCRP. Consequently, we posit that maximal sensitization of cancer cells by MGMT inhibitors is dependent on efflux mechanisms that control the intracellular concentration of the therapeutics. We conclude that inhibition of ABC transporters in cancer cells should be taken into account as part of a strategy of targeting
■
EXPERIMENTAL SECTION 2.1. Cell Culture and Drug Treatment. Cell lines were grown at 37 °C under a humidified atmosphere with 5% CO2. T98G and A375 were purchased from American Type Culture Collection (ATCC) and HT29 from Cell Lines Services (Heidelberg, Germany). The human glioblastoma cell line GBP61 was a kind donation from Prof. Hopf (Institute of Neurosurgery, Mainz, Germany) and the melanoma cell line D05 from C. W. Schmidt (Queensland Institute of Medical Research, Queensland, Australia). The cell lines T98G, GBP61, and A375 were maintained in Dulbecco’s modified Eagle medium (DMEM) but HT29 and D05 in RPMI 1640 medium. Media contained 10% fetal calf serum (FCS), 10 U/mL penicillin and 10 mg/mL streptomycin. The MGMT inhibitor O6-benzylguanine (O6BG) was purchased from Sigma-Aldrich (Munich, Germany), and O 6 -(4-bromothenyl)guanine (O6BTG) was kindly provided from Dr. Geoff Margison (Paterson Institute for Cancer Research, Manchester, U.K.). The synthesis of the glucose-conjugated MGMT inhibitors has been previously described.29 The conjugate 2-amino-6-(benzyloxy)-9-(octyl-β-D-glycosyl)purine (O6BG-Glu) was synthesized on large scale by us. The second glucose-conjugated MGMT inhibitor 2-amino-6-(4-bromotiophen-2-yl-methoxy)-9-(octylβ-D-glycosyl)purine (O6BTG-Glu) was synthesized by and purchased from Haoyuan Chemexpress Co., Limited (Shanghai, China). All MGMT inhibitors were dissolved in dimethyl sulfoxide (DMSO) and stored at −20 °C. Unless otherwise noted, MGMT inhibitors were added 4 h before the treatment with alkylating agents. Temozolomide (TMZ) (ScheringPlough, Kenilworth, NJ, USA) was dissolved in DMSO and then diluted with sterile water to a final concentration of 35 mM. Aliquots were stored at −80 °C. A solution of 10 mM lomustine [1-(2-chloroethyl)-3-cyclohexyl- L -nitrosourea; CCNU] (Sigma, Munich, Germany) was prepared by dissolving in ethanol and stored at −20 °C. For the treatment of exponentially growing cells, the alkylating agents were added to the medium. The ABC transporter inhibitors verapamil, MK571, and Ko143 were purchased from Sigma-Aldrich (Munich, Germany), dissolved in DMSO, and stored at −20 °C. Unless otherwise noted, ABC transporter inhibitors (10 μM verapamil, 10 μM MK571, and 5 μM Ko143) were added to cells at least 15 min before treatment with an MGMT inhibitor. 2.2. MGMT Activity Assay. Determination of MGMT activity was performed as previously described.30 Briefly, after treatment, cells were washed with ice cold phosphate buffered saline (PBS), harvested, and resuspended in a sonification buffer [20 mM Tris-C1, pH 8.5, 1 mM EDTA, 1 mM 2mercaptoethanol, 5% glycerol, and protease inhibitors (0.1 mM PMSF and protease inhibitor cocktail (Roche, Mannheim, Germanny)]. Cells were sonicated on ice, and protein concentration was determined by the method of Bradford et al.31 For the transfer of tritium-labeled methyl groups ([3H]methyl) on MGMT, a certain amount of protein extract was incubated with tritium-labeled thymus DNA for 90 min at 37 °C. Afterward proteins were precipitated with 13% trichloroacetic acid (TCA), and DNA was hydrolyzed by boiling the samples for 45 min at 95 °C. Proteins were washed 3 times with 5% TCA and solubilized in 0.1 M sodium hydroxide (NaOH). Finally, radioactivity was determined by liquid scintillation 3925
DOI: 10.1021/acs.molpharmaceut.5b00341 Mol. Pharmaceutics 2015, 12, 3924−3934
Article
Molecular Pharmaceutics counting (Canberra Packard, Dreieich, Germany) and MGMT activity was expressed as fmol of [3H]-methyl transferred from radioactively labeled DNA to MGMT protein per mg of total protein extract. Each assay included a determination with HeLa MR extract, which does not express MGMT (background control). 2.3. Determination of Apoptosis. Induction of apoptosis was determined by measuring the subdiploid DNA content (Sub-G1). After specified times, adherent and detached cells were collected, washed with PBS, and fixed in ice cold 70% ethanol for 30 min. After RNAase (30 μg/μL) digestion, DNA was stained with propidium iodide (PI) (16.7 μg/mL) in PBS. Flow cytometric analysis was performed using a FACSCalibur (Becton Dickinson, Heidelberg, Germany). For each sample, 10,000 events were acquired and the proportion of apoptotic cells was calculated with WinMDI (Joseph Trotter, http://facs. scripps.edu/software.html) 2.4. Colony Formation Assay. This method was performed as previously described.32 Briefly, cells were seeded at appropriate cell numbers in 60 mm cell culture dishes to yield at least about 50 colonies after treatment with TMZ or CCNU. Cells were allowed to attach for 6 h and then exposed to MGMT inhibitors and specified concentrations of alkylating agents. After incubation for 15 days, the colonies were fixed (acetic acid:methanol:H2O 1:1:8) and stained (1.25% Giemsa, 0.125% crystal violet). Colonies containing 50−100 cells were counted, and the surviving fraction was calculated. 2.5. Preparation of Cell Extract and Western Blotting. Samples were separated by SDS−PAGE and blotted onto a PVDF membrane (Millipore, Billerica, MA, USA), based on the method of Renart et al.33 Proteins were visualized using the Odyssey system (LI-COR Biosciences). The following antibodies were used: anti-MGMT, anti MRP1, anti-P-gp, antiBCRP (Millipore, Billerica, MA, USA), anti-HSP90, and antiERK2 (Santa Cruz Biotechnology, Heidelberg, Germany). 2.6. Fluorescent Dye Accumulation Assay. Quantification of fluorescent dye accumulation was performed based on the method of Hollo et al.34 Exponentially growing cells were harvested and resuspended in transport buffer [0.952 mM CaCl2, 5.36 mM KCl, 0.441 mM KH2PO4, 0.812 mM MgSO4, 136.7 mM NaCl, 0.385 mM Na2HPO4, 25 mM glucose, 10 mM HEPES, 5% FCS, pH = 7.4]. Afterward, 2.5 × 105 cells were incubated with either 30 μM rhodamine 123 (Rho 123), 20 μM 5(6)-carboxy-2′,7′-dichlorofluorescein diacetate (CFDA), or 0.5 μM pheophorbide A (Pheo A) (Sigma-Aldrich, Munich, Germany) in the absence or presence of a MGMT inhibitor (100 μM) or ABC transporter inhibitor (100 μM verapamil (for P-gp), 40 μM MK571 (for MRP1), and 50 μM Ko143 (for BCRP)) as positive controls. Accumulation of fluorescent dyes occurred for 30 min at 37 °C with continual shaking at 800 revolutions per minute (rpm). Cells were then washed with ice cold PBS and stained with PI (20 μg/mL) in PBS, to exclude dead cells. Flow cytometric analysis was performed using a FACSCanto (Becton Dickinson, Heidelberg, Germany). For each sample 10,000 events were acquired and the median fluorescence intensity (MFI) was analyzed with BD FACSDiva v6 (Becton Dickinson, Heidelberg, Germany). Accumulation of fluorescent dyes was calculated with the following equation: relative accumulation =
where MFIdye corresponds to the MFI after loading cells with only fluorescent dye and MFIdye+inh refers to the MFI in the presence of MGMT or ABC transporter inhibitors. The values MFIAF and MFIAF-inh represent the autofluorescence of cells or respective inhibitor. Statistics are based on calculated induction factors related to cells treated with fluorescence dye only. 2.7. Fluorescent Dye Efflux Assay. Quantification of the active efflux of fluorescent dyes was performed based on the method of Homolya et al.35 Exponentially growing cells were harvested and resuspended in transport buffer. For the accumulation phase, 2.5 × 105 cells were loaded with either 30 μM Rho 123, 20 μM CFDA, or 0.5 μM Pheo A for 30 min at 37 °C with continual shaking at 800 rpm. Cells were then washed with ice cold PBS and reincubated in transport buffer, including either ABC transporter inhibitors or 100 μM of one of the MGMT inhibitor. The efflux phase was performed for 20 min (for CFDA) or 120 min (Rho 123, Pheo A) at 37 °C with shaking at 800 rpm. Cells were then washed with ice cold PBS and stained with PI (20 μg/mL) in PBS, to exclude dead cells. Flow cytometric analysis was performed using a FACSCanto. For each sample 10000 events were acquired and the median fluorescence intensity (MFI) was analyzed with BD FACSDiva v6. The difference of the fluorescent dye retained in cells before and after the efflux phase reflects the transport activity of certain ABC transporters. To calculate the reversal mediated by MGMT inhibitors or ABC transporter inhibitors, the following equation was used: efflux inh =
MFIeff+inh − MFIAF‐inh − (MFIeff − MFIAF) MFI 0‐value − MFIAF − (MFIeff − MFIAF)
where MFI0‑value corresponds to the MFI after loading cells with fluorescent dyes (accumulation phase) and MFIeff+inh or MFIeff refers to the MFI after efflux phase in the presence or absence of MGMT or ABC transporter inhibitors, respectively. The values MFIAF and MFIAF‑inh represent the autofluorescence of cells or respective inhibitor. The effects of MGMT inhibitors on ABC transporters were related to the respective transporter inhibitor (100 μM verapamil (for P-gp), 40 μM MK571 (for MRP1), and 50 μM Ko143 (for BCRP)). Calculations were based on the assumption that the ABC transporter inhibitors blocked the complete efflux (100%) of the respective fluorescent dye. Statistics are based on calculated induction factors related to cells treated with fluorescence dye only. 2.8. Statistics. All of the in vitro experiments were performed at least three times. Statistical analysis was performed using GraphPad Prism version 6 (GraphPad Software, La Jolla, CA, USA). Results were expressed as mean ± standard deviation (SD). Comparison between samples was performed using the unpaired t test or the oneway ANOVA analysis. A p-value of
