Altered Methylation of the Human MDR1 Promoter Is Associated with Acquired Multidrug Resistance1

One of the most important forms of drug resistance in acute myeboid leukemia is the multidrug resistance (MDR) phenotype, which is characterized by the expression of the MDRJ gene product, P-glycoprotein. Although a number of factors affect MDR1 gene expression, the genetic events that “switch on” the human MDRJ gene in tumor cells that were previously P-glycoprotein negative have remained elusive. Here, we report evidence that the methybation status of the human MDR1 promoter may serve as a basis for this “switch.” Based on Southern analysis using methylationsensitive and methylation-insensitive restriction enzymes, a tight correlation was found between MDR phenotype and demethylation of the 5’ region of the MDRJ gene in a human T cell leukemia cell line. Similar results were obtained from the analysis of P-glycoprotein-positive and P-glycoproteinnegative samples of chronic lymphocytic leukemia. Treatment of the cell lines with the demethybating agent 5’azadeoxycytidine altered the methybation pattern of the MDR1 promoter in P-glycoprotein.negative cells to resemble that of P-gbycoprotein-positive cells and activated the promoter such that MDR1 mRNA was now detectable. Treatment also resulted in an increased resistance to epirubicin and decreased daunomycin accumulation, both of which were reversible by verapamil, a characteristic of the classical MDR phenotype in cells expressing P-gbycoprotein. These results suggest that the MDR phenotype may be Received I 1/13/96; revised 4/25/97; accepted 7/3/97. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with I 8 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by the Anti-Cancer Council of Victoria and Department of Veterans’ Affairs (Canberra, ACT. Australia) and the Sir Edward Dunlop Medical Research Foundation (Melbourne, Victoria, Australia). 2 To whom requests for reprints should be addressed, at Division of Haematology and Medical Oncology. Peter MacCallum Cancer Institute, East Melbourne 3000, Victoria, Australia. Phone: (613) 9656 1190; Fax: (613) 9656 1408; E-mail: [email protected]. acquired as a result of changes in methylation of the MDR1 promoter. INTRODUCTION Drug resistance remains a major obstacle in the successful treatment of human tumors by cytotoxic agents. Of the several forms of drug resistance described, the most extensively studied is classical MDR,3 which is characterized by the overexpression of the MDRJ gene product, Pgp; partial reversal of resistance by several modulators, including verapamil and cyclosporin A; and cross-resistance to a variety of naturally occurring cytotoxic agents ( 1 ). This plasma membrane glycoprotein functions as an energy-dependent drug efflux transporter, resulting in bower intracellular levels of a wide variety of chemotherapeutic agents that are substrates for Pgp (1). Despite the difficulties in successfully modulating this phenotype clinically (2), the expression of Pgp is of prognostic significance in various hematological malignancies, including acute myeboid leukemia. In a series of unrelated studies, Pgp expression represented an adverse prognostic factor that independently predicted for survival, response rate, and duration of response (reviewed in Ref. 2). In human cancers in which the corresponding normal tissue expresses Pgp, the resulting tumors also appear to express this protein. These include carcinomas of the colon, liver, pancreas, and kidney (1). However, in drug-sensitive tumors, the acquisition of MDR during the course of chemotherapy is thought to be due to the selection of resistant mutants in the tumor cell population by drugs known to be substrates for Pgp. This model fits with the mutation-selection hypothesis for drug resistance in cancer (3), in which genetic changes in a small number of drug-sensitive cells ultimately lead to the development of resistance via a process of selection. This scenario is thought to be particularly relevant in hematological malignancies. Although the regulation of Pgp expression has been studied in several cell types and much has been learned about factors regulating its expression (4-8), the genetic events that switch on Pgp expression in Pgp-negative cells remain poorly defined. No mutations in the sequence of the MDR 1 promoter have been identified to account for the activation of the MDR 1 promoter in cells that express Pgp. Although point mutations have been reported in a number of osteosarcomas at nucleotides 103 and 107 downstream of the MDR1 transcription initiation site, they were found in untreated tumors, and their significance remains to be determined (9). Attempts have been made to identify transcription factors that may be altered or present in higher or lower levels in Pgp-positive cells compared to Pgp-negative cells. One report 3 The abbreviations used are: MDR, multidrug resistance: Pgp. P-glycoprotein: CLL. chronic lymphocytic leukemia: azadC, 5 ‘-azadeoxycytidine; RT-PCR, reverse transcription-PCR. Research. on July 14, 2017. © 1997 American Association for Cancer clincancerres.aacrjournals.org Downloaded from 2026 DNA Methylation and Drug Resistance .1 Unpublished observations. suggested that the SP1 transcription factor was overexpressed in Pgp-positive cells, but this finding has not been confirmed in other cell lines ( 10). Ogura et a!. ( I I ) have also identified three factors that bind to various regions of the MDR I promoter and affect transcription in resistant but not sensitive cells. It is not known whether these factors can “switch on” an inactive MDR I promoter. Furthermore, extensive analysis of promoter function with the chloramphenicol acetyltransferase reporter gene including regions extending 4.7 kb upstream of the transcription initiation site did not reveal a cis-acting element responsible for the cell type-specific expression of Pgp (6). On the basis of cell phenotypes and the sequence data, one would expect the transfected MDR1 promoter to be inactive in drug-sensitive cells, paralleling the endogenous promoter. However a number of groups (6, 12) including ours,4 have demonstrated that the MDR 1 promoter was active in chloramphenicol acetyltransferase reporter gene assays in both drug-sensitive and drugresistant cells. This was in contrast to the endogenous promoter that was only active in the drug-resistant cells. In view of these findings and the finding that DNA methylation is often involved in regulating the tissue-specific expression ofgenes (13, 14), we reasoned that the MDR 1 promoter may be differentially methylated in drug-resistant compared to drug-sensitive cells. To elucidate the genetic events leading to the up-regulation of the MDRI gene in drug-sensitive tumor cells, we studied an i,l vitro model of acquired MDR. Our results indicate that DNA methylation may play a role in regulating the activity of the MDR I promoter. More specifically, Pgp expression correlates inversely with DNA methylation in the MDRI promoter region. Demethylation by treatment with azadC converted Pgp-negative cells to Pgp-positive cells with the typical MDR phenotype. These findings were confirmed in a clinical example of MDR. MATERIALS AND METHODS Tissue Samples and Cell Lines. Three CLL samples were used, from which cells were isolated from 10 ml of heparinized blood by Ficoll density gradient centrifugation. The cells were washed twice in PBS and analyzed on a flow cytometer for CD5/CD19, and they were analyzed for Pgp expression using the MRKI6 monocbonal antibody as described previously ( I 5). All of the samples contained more than 95% B-CLL cells. Aliquots were frozen at -70#{176}C. The CCRF-CEM line is a human T-cell leukemia cell line (16) that does not contain any detectable level of MDR1 mRNA, even after 40 cycles of PCR, nor can Pgp be detected by flow cytometry using MRK16. The CEM/A7 cell line was isolated from the CCRF-CEM cell line by selection for resistance to doxorubicin (17) and is currently maintained in 0.07 p.g/ml doxorubicin. The CEM/A7R cell line was isolated by culturing the CEMJA7 line in the absence of doxorubicin for over S years (18, 19). Both CEM/A7R and CEM/A7 behave like classical MDR cell lines, with easily detectable MDR1 mRNA and Pgp by Northern analysis and flow cytometry respectively, but the CEM/A7R line expresses less Pgp compared to CEM/A7. The hybrid cell lines used in this study (2Gb, 4G9, and 2H6) were derived by the fusion of the CEM-CCRF and CEM/A7 parental lines as described previously (20). DNA Isolation and Southern Analysis. Genomic DNA was isolated from cell lines and clinical samples using the Progenome DNA isolation kit (Progen Industries, Darra, Queensland, Australia). To control for complete digestion, DNA samples were added to 0.5 p.g of pGEM7Zf(+) plasmid (Promega), the digestion pattern of which is known. After digestion, DNA was fractionated on 1 .5% agarose and transferred to Hybond-N+ nylon (Amersham Corp.) with 0.4 M NaOH for 3-4 h. Membranes were prehybridized for 1 h in S X salinesodium phosphate-EDTA (0.9 M NaCl, 50 m i sodium phospate, and S msi EDTA, pH 7.7), 50% formamide, 1% SDS, 200 p.g/ml herring sperm DNA, and S X Denhardt’s solution (0.5% BSA, 0.5% Ficoll, and 0.5% polyvinyl pyrollidone) at 42#{176}C.Hybridization was carried out in the same buffer at 42#{176}C for 16 h with either the 979-bp Psi’! or the 402-bp MspI MDRI promoter fragments (see Fig. 1A), which were labeled with 32P-dCTP by random priming. Filters were washed twice in 1 X SSC (0. 15 M NaC1, 15 mM sodium citrate), 0.5% SDS for 10 mm at room temperature and twice with 0.2 X SSC, 0. 1% SDS for 15 mm at 65#{176}C. mRNA Isolation and RT-PCR cDNA Titration Analy515. Following lysis of cells in a solution containing guanidine isothiocyanate. poly(A)+ RNA were isolated using oligo(dT) cellulose (AMRAD-Pharmacia, Boronia, Victoria, Australia). First-strand cDNA was synthesized using poly(A)+ RNA isobated from 2 x l0 cells with random hexamers in 20 p.1 of reaction mixture. All PCR reactions were carried out with Taq pobymerase (Life Technologies, Inc.) in 50 p.1 of a solution containing a number of template serial 1 :2 cDNA dilutions (0.6125-10 ng), 25 pmol of each primer, 200 p.M dNTPs, 2 msi MgC12. and supplied buffer. Reactions were carried out using an Omnigene thermocycler (Hybaid, Teddington, Middlesex, United Kingdom) as follows: 94#{176}C for 1 mm followed by 34 cycles each at 94#{176}Cfor I mm, 58#{176}Cfor 1 mm, 72#{176}Cfor 1 mm, and terminated with a final extension at 72#{176}C for 5 mm. PCR products were fractionated on 2.5% agarose and were quantified using the Gel Doc 1000 system (Bio-Rad). At 34 cycles, amplification was in the exponential phase and histone 3.3 levels were equivalent in all samples (data not shown). The primers for MDR1 were as follows: sense, CCCATCATFGCAATAGCAGG (2596-2615); and antisense, GTFCAAACTTCTGCTCCTGA (2733-2752), described previously by Noonan et a!. (21), resulting in a PCR productof 157 bp. Histone 3.3 was used as the endogenous internal control to ensure mRNA integrity and to normalize PCR amplification. The primers were as foblows: sense, CCACTGAACTI’CTGATFCGC (282-301 ); and antisense, GCGTGCTAGCTGGATGTCTF (476-496), resulting in a product of214 bp. Both MDR1 and histone 3.3 primers span an intron. azadC Treatment of Cells: Growth and Drug Accumulation Assays. azadC was rapidly dissolved in normal growth medium, aliquoted, and frozen in liquid nitrogen before storage at -70#{176}C.Cells in logarithmic growth were seeded at S X I0 cells/mb, and azadC was added at a final concentration of 1 or 2 p.M. After 2-3 days, cells were counted and RNA and DNA harvested. Following treatment with azadC, cells were washed Research. on July 14, 2017. © 1997 American Association for Cancer clincancerres.aacrjournals.org Downloaded from