Methylation status of breast cancer resistance protein detected by methylation-specific polymerase chain reaction analysis is correlated inversely with its expression in drug-resistant lung cancer cells
Hirofumi Nakano MD,
Second Department of Internal Medicine, Nagasaki University School of Medicine, Nagasaki, Japan
Breast cancer resistance protein (BCRP) functions as a drug efflux transporter that mediates drug resistance. Topoisomerase I inhibitors, including 7-ethyl-10-hydroxycamptothecin (SN-38), are substrates effluxed by BCRP. However, it remains unclear whether the overexpression of BCRP induces drug resistance during chemotherapy. The objectives of the current study were to examine a correlation of altered promoter methylation of BCRP with BCRP expression and to investigate the correlation between methylation status according to methylation-specific polymerase chain reaction (MSP) analysis and BCRP expression levels in several small cell and nonsmall cell lung cancer cells.
Non-BCRP-expressing PC-6 cells, which were sensitive to SN-38, were treated with DNA methyltransferase inhibitor to induce BCRP re-expression by means of reverse transcriptase-polymersae chain reaction, Western blot, and flow cytometric analyses. Subsequently, bisulfite sequencing analysis in both PC-6 cells and SN-38-resistant PC-6/SN2-5H, highly expressing BCRP cells was performed to identify the methylated region in the BCRP promoter. Finally, the authors established an MSP method on the basis of methylated and unmethylated DNA sequences.
DNA methyltransferase inhibitor treatment of PC-6 cells induced BCRP re-expression at the messenger RNA and protein levels. Bisulfite sequencing analysis revealed that both alleles at all CpG sites were methylated completely in PC-6 cells, whereas alleles at portions of CpG sites in PC-6/SN2-5H cells were unmethylated. There was an inverse correlation between promoter methylation of BCRP determined by MSP and BCRP expression in both small cell and nonsmall cell lung cancer cells.
Chemotherapy is the mainstay in treatment for various types of advanced cancer. However, it has been recognized that even highly sensitive cancers, such as leukemia, lymphoma, and small cell lung cancer, frequently exhibit drug resistance during the course of chemotherapy. The development of resistance to a single anticancer drug is accompanied by cross-resistance to several anticancer drugs and reduces the effectiveness of chemotherapy. This phenomenon is referred to as multidrug resistance (MDR).1 Thus, it is very important to overcome drug resistance and to elucidate its mechanisms, thus leading to better treatment for various cancers. One of the well known molecular mechanisms of drug resistance is to enhance extrusion of drugs and their metabolites by overexpression of energy-dependent pumps, such as adenosine triphosphate-binding cassette (ABC) transporters.2–4 ABC transporters are classified into 7 subfamilies (A through G), and it is well known that several transporters are correlated with drug resistance: P-glycoprotein (P-gp), the product of the MDR 1 gene (MDR1),5, 6 the MDR-associated proteins (MRPs),7, 8 and breast cancer resistance protein (BCRP).9, 10
BCRP, a novel member of the G (white) subfamily of ABC transporters, was newly identified as an over expressing molecule in MCF-7/doxorubicin-verapamil cells.9 BCRP is a 655 amino-acid polypeptide at 72 kD and is localized at the plasma membrane.11, 12 The anticancer drugs that are effluxed by BCRP are mitoxantrone, anthracyclines, and topotecan-derived and topoisomerase I inhibitors, including 7-ethyl-10-hydroxycamptothecin (SN-38) (an active metabolite of irinotecan) and methotrexate.10, 13–17 Moreover, because it is expressed physiologically in the placenta, bile canaliculi, colon, small intestine, and brain microvessel endothelium, BCRP may play a role in protecting the organism from potentially harmful xenobiotics in normal and cancer cells.9 In our previous study, BCRP was not expressed in human small cell lung cancer PC-6 cells but was overexpressed in the SN-38-resistant subline, PC-6/SN2-5H, which was selected from PC-6 cells by continuous exposure to SN-38.15 In addition, the antisense oligonucleotide significantly suppressed BCRP expression and enhanced the sensitivity to SN-38 in PC-6/SN2-5H cells, indicating that BCRP re-expression directly confers SN-38 resistance.15 However, the precise mechanisms underlying BCRP re-expression during acquisition of drug resistance remain unclear.
DNA methylation is a covalent modification of cytosine and is recognized as an epigenetic DNA modification in mammalian cells. It occurs at CpG dinucleotides by DNA methyltransferases. Approximately 70% to 80% of CpG sites are methylated with the notable exception of CpG-rich sequences, referred to as CpG islands, which generally are unmethylated and are located in the promoter region. It has been demonstrated that methylation of CpG islands in the promoter region of certain genes causes their transcriptional silencing in normal and malignant cells.18–21 Demethylation of CpG islands in the promoter region of MDR1 induces re-expression and overexpression of P-gp and demonstrates the MDR phenotype.22–26 These findings indicate that P-gp overexpression induced by demethylation of the MDR1 promoter is a mechanism of acquisition of drug resistance during chemotherapy.
From analogy to P-gp re-expression by demethylation of the MDR1 promoter, we hypothesize that promoter demethylation of BCRP may induce BCRP re-expression in drug-resistant lung cancer cells, leading to SN-38 resistance. Therefore, the objective of this study was to investigate whether altered methylation status of CpG sites in the promoter region of BCRP in PC-6 cells and PC-6/SN2-5H cells is correlated with BCRP re-expression at the messenger RNA (mRNA) and protein levels and whether methylation status analyzed by methylation-specific polymerase chain reaction (MSP) method also is correlated with BCRP mRNA expression levels in both small cell and nonsmall cell lung cancer cells.
MATERIALS AND METHODS
SN-38-resistant PC-6/SN2-5H cells were selected from parental PC-6 human small cell lung cancer cells by continuous exposure to SN-38, as reported previously.27 All other unselected human nonsmall cell lung cancer cell lines (NCI-H69, NCI-H358, NCI-H441, and NCI-H460) were obtained from American TypeCulture Collection (Manassas, Va). PC-6/SN2-5H cells over express BCRP and have highly functional efflux-pump activity, indicating high resistance to SN-38.15, 28 NCI-H358, NCI-H441, and NCI-H460 cells have moderate expression and functional pump activity of BCRP and demonstrate moderate to low resistance to SN-38.28 Conversely, PC-6 and NCI-H69 cells never express BCRP and have high sensitivity to SN-38.15, 28 PC-6/SN2-5H cells have no mutations of BCRP (for example, homozygous wild-type alleles at codons 376, 421, and 482) and no 944-949 deletions, which alter the substrate specificity of BCRP.29–32 Cells were cultured at 37 °C under a humidified atmosphere of 5% carbon dioxide in RPMI-1640 medium (Invitrogen, Carlsbad, Calif), 10% fetal bovine serum (Invitrogen), 200 mM glutamine (Invitrogen), and 250 μg/mL kanamycin (Meiji Seika, Ltd., Tokyo, Japan). All other reagents were purchased from Sigma Chemical Company (St. Louis, Mo).
DNA Methyltransferase Inhibitor Treatment of PC-6 Cells and BCRP Expression Analysis
PC-6 cells were incubated in RPMI-1640 with various concentrations of DNA methyltransferase inhibitor, 5-aza-2′-deoxycytidine (5-aza-dC) (Sigma Chemical Company). After culture for 24 hours, cells were treated with 0 μM, 0.1 μM, 0.5 μM, 1 μM, or 2 μM 5-aza-dC for 72 hours. Cells were harvested, and total RNA was isolated from each cell by using ISOGEN (Nippongene, Tokyo, Japan) according to the manufacturer's protocol. Reverse transcriptase (RT) reaction was performed on 3.5 μg of total RNA with random hexamer by using the Thermoscript RT-polymerase chain reaction (PCR) system (Invitrogen) according to the manufacturer's protocol. PCR primer sequences for BCRP were described previously.15 The human β-actin gene (ACTB) was used as an internal control to normalize expression of BCRP. The PCR primer sequences for ACTB were as follows: 5′-AAGATGACCCAGATCATGTTTGAG-3′ (sense primer) and 5′-AGGAGGAGCAATGATCTTGATCTT-3′ (antisense primer). PCR was performed in a 25-μL reaction mixture that contained aliquots of combinational DNA solution; 20 mM Tris-HCl, pH 8.4; 50 mM KCl; 1.5 mM MgCl2; 200 μM dinucleotide triphosphate (dNTP) (Promega, Madison, Wis); 0.6 μM each of the sense and antisense primers; and 1.25 U Taq DNA polymerase (Invitrogen) with a DNA thermal cycler (GeneAmp PCR System 9700; Applied Biosystems, Foster City, Calif) according to the following protocol: for BCRP, initial denaturation at 94 °C for 2 minutes; 30 cycles of denaturation at 94 °C for 90 seconds, annealing at 63 °C for 60 seconds, and extension at 72 °C for 60 seconds; and final extension at 72 °C for 5 minutes; and, for ACTB, initial denaturation at 94 °C for 2 minutes; 27 cycles of denaturation at 94 °C for 30 seconds, annealing at 57 °C for 30 seconds, and extension at 72 °C for 30 seconds; and final extension at 72 °C for 5 minutes. PCR products were electrophoresed on a 2% agarose-ME gel (Nacalai Tesque, Kyoto, Japan), observed with an ultraviolet transilluminator (Alpha Innotech, San Leandro, Calif) after ethidium bromide (Invitrogen) staining, and photographed. The signal intensity of each PCR product was estimated by using the Image Gauge 4.0 program (Fuji Photo Film Company, Ltd., Tokyo, Japan). To represent the relative BCRP mRNA expression level, we used the following formula: the signal intensity of BCRP divided by that of ACTB.
Western Blot Analysis of BCRP
Whole cell proteins from PC-6, PC-6/SN2-5H, and PC-6 cells with exposure to 1 μM 5-aza-dC for 72 hours were obtained, and 15 μg of protein from each lane were separated on a 7.5% sodium dodecyl sulfate-polyacrylamide electrophoresis gel, then electrotransferred to nitrocellulose membranes as described previously.28 By using BXP-21 (Kamiya, Seattle, Wash) of antihuman BCRP antibody (1:500) and the ECL detection system (Amersham, Bucks, U.K.), Western blot analysis was performed as described previously.28
Flow cytometric analysis was performed as described previously.28, 33 Because topotecan is one of the substrates that is effluxed by BCRP and has intrinsic fluorescent properties that are detectable by using a conventional flow cytometer,28, 33 cultured cells (approximately 2 × 106) were exposed to 30 μM topotecan for 15 minutes at 37 °C with or without 300 μM novobiocin, which is a good substrate for BCRP and inhibits its function competitively (topotecan transport by BCRP was inhibited by 75% under 160 μM novobiocin),33 then washed twice in ice-cold saline. Fluorescence of topotecan was measured through a 488-nanometer (nm) band-pass filter at an excitation wavelength of 585 nm by using a FACscan flow cytometer (Becton Dickinson, Mountain View, Calif) equipped with a 15 megawatt argon laser, because the detection of differences in retention has been associated with various levels of BCRP expression.28 Topotecan accumulation after incubation for 15 minutes was expressed in fluorescence units. In all fluorescence assays, cells without exposure to topotecan were used as controls.
Bisulfite Modification of DNA From PC-6 and PC-6/SN2-5H Cells
Genomic DNA was obtained from PC-6 and PC-6/SN2-5H cells by using the QuickGene DNA Tissue Kit S (Fujifilm, Tokyo, Japan) with QuickGene-800 (Fujifilm) according to the manufacturer's protocol. DNA was processed with the CpGenome DNA Modification Kit (Intergen Company, Purchase, NY) according to the manufacturer's protocol. The promoter region ranging from nucleotide (nt) −1381 to nt +261 with respect to the transcription initiation site of BCRP (GenBank Accession no. AF151530)34 was divided into 5 segments (A through E). Primer pairs for both PCR and direct DNA sequencing were designed in each region (Table 1). PCR was performed in a 25-μL reaction mixture that contained 50 ng of bisulfite-modified DNA or unmodified DNA; 20 mM Tris-HCl, pH 8.4; 50 mM KCl; 1.5 mM MgCl2; 240 μM dNTP; 0.4 μM each of the sense and antisense primers; and 1.5 U platinum Taq DNA polymerase (Invitrogen). Amplification was performed for bisulfite-modified DNA with initial denaturation at 94 °C for 2 minutes; 35 cycles of denaturation at 94 °C for 30 seconds, annealing for 60 seconds at each temperature for the selected primer pair, and extension at 72 °C for 30 seconds; and final extension at 72 °C for 5 minutes; and amplification was performed for unmodified DNA with initial denaturation at 94°C for 2 minutes; 35 cycles of denaturation at 94 °C for 30 seconds, annealing for 30 seconds at each temperature for the selected primer pair, and extension for at 72 °C for 30 seconds; and final extension at 72 °C for 5 minutes. A portion of PCR products was electrophoresed on a 2% agarose-ME gel, and the remaining products were used as a template for subsequent direct DNA sequencing.
Table 1. Primer Sequences of Polymerase Chain Reaction and Direct DNA Sequencing for Bisulfite-modified and Unmodified DNA
The promoter and exon1 regions ranging from nucleotide −1381 to nucleotide +261 with respect to the transcription initiation site of the breast cancer resistance protein gene (BCRP) was divided into 5 regions (A through E).
Direct DNA Sequencing Analysis
PCR products were treated with ExoSAP-IT (Amersham Pharmacia Biotech, Piscataway, NJ) to remove primers and excess nucleotides and were sequenced by using the Big Dye Terminator Cycle Sequencing FS Ready Reaction Kit (Applied Biosystems). The cycle sequencing was hot-started at 96 °C for 30 seconds; followed by 25 cycles at 10 seconds at 96 °C, 5 seconds at 50 °C, and 4 minutes at 72 °C; and 4 minutes at 60 °C. After sequencing reaction solutions were purified by using Sephadex G-50 superfine columns (Amersham Pharmacia Biotech), samples were dried and sequenced with the ABI PRIAM 310 or 3100 Genetic Analyzer (Applied Biosystems).
MSP for BCRP
To perform qualitative analysis of methylation of the BCRP promoter by PCR, we designed primer pairs for MSP35 on the basis of methylated and unmethylated DNA sequences in the promoter region of BCRP as follows: methylation-specific sense primer (M-BCRPf-s1): 5′-gcgtttcggttagtgacggc-3′ (nt −337 to −318) and antisense primer (M-BCRPr-s1): 5′-cccgcctccgaaatcgaacg-3′ (nt −213 to −232); and unmethylation-specific sense primer (U-BCRPf-s2): 5′-tagttttgttggtggtttagtgt-3′ (nt −147 to −125) and antisense primer (U-BCRPr-s2): 5′-aaccccaactaccaaaccaca-3′ (nt −60 to −80). PCR was performed in a 25-μL reaction mixture that contained 50 ng of bisulfite-modified DNA from PC-6, PC-6/SN2-5H, NCI-H69, NCI-H358, NCI-H441, or NCI-H460 cells; 20 mM Tris-HCl, pH 8.4; 50 mM KCl; 1.5 mM MgCl2; 240 μM dNTP; 0.4 μM M-BCRPf-s1/r-s1 or U-BCRPf-s2/r-s2; and 1.5 U platinum Taq DNA polymerase under the following conditions: for M-BCRPf-s1/r-s1, initial denaturation at 95 °C for 2 minutes; 35 cycles of denaturation at 94 °C for 30 seconds, annealing at 58 °C for 30 seconds, and extension at 72 °C for 30 seconds; and final extension at 72°C for 5 minutes; and, for U-BCRPf-s2/r-s2, initial denaturation at 94 °C for 2 minutes; 35 cycles of denaturation at 94 °C for 30 seconds, annealing at 64 °C for 30 seconds, and extension at 72 °C for 30 seconds; and final extension at 72 °C for 5 minutes. PCR products were electrophoresed on a 6% polyacrylamide electrophoresis gel (Nacalai Tesque). To evaluate the MSP method for BCRP properly, we examined the correlation of promoter methylation analyzed by MSP with BCRP mRNA expression in small cell lung cancer cells (PC-6 and PC-6/SN2-5H cells) as well as other nonsmall cell lung cancer cells (NCI-H69, NCI-H358, NCI-H441, and NCI-H460 cells).
A simple linear regression analysis was performed with the relative BCRP mRNA expression as a dependent variable and with the concentration of 5-aza-dC as an independent variable by using Prism 4 (GraphPad Software Inc., San Diego, Calif). A P value < .05 was considered statistically significant.
BCRP mRNA Expression in PC-6 Cells Treated With 5-Aza-Dc
After non-BCRP-expressing PC-6 cells were treated with 0.1 μM, 0.5 μM, 1 μM, or 2 μM of 5-aza-dC for 72 hours, the relative BCRP mRNA expression (the BCRP/ACTB ratio) was increased to 0.0573, 0.0695, 0.0845, and 0.0980, respectively, in duplicate experiments (Fig. 1). These results indicate that 5-aza-dC treatment induced BCRP re-expression in a dose-dependent manner (r2 = 0.9550; P = .0228) and that an alteration in the methylation status of the BCRP promoter is correlated with the activation of BCRP expression in small cell lung cancer cells.
BCRP Protein Expression and Efflux-pump Activity in PC-6 Cell Panels
Treatment with 1 μM 5-aza-dC for 72 hours induced PC-6 cells to express BCRP protein (Fig. 2). The expression of BCRP protein in PC-6 cells that were treated with 1 μM 5-aza-dC (Fig. 2, lane 2) relative to the expression of BCRP protein in PC-6/SN2-5H cells (Fig. 2, lane 1) was 0.19. However, in the parental PC-6 cells (Fig. 2, lane 3), BCRP expression was not observed. Moreover, the levels of BCRP protein expression nearly coincided with the levels of BCRP mRNA expression in PC-6 cells that were treated with 5-aza-dC (Figs. 1, 2).
Subsequently, the intracellular topotecan accumulation (Fig. 3, green line) in PC-6/SN2-5H cells was decreased remarkably compared with the accumulation in PC-6 cells (Fig. 3). Novobiocin, a competitive inhibitor of BCRP, restored the intracellular topotecan accumulation (Fig. 3, red line) in PC-6/SN2-5H cells but had no effect on PC-6 cells (Fig. 3). In PC-6 cells that were treated with 5-aza-dC, the intracellular topotecan accumulation was decreased to the levels between those in PC-6/SN2-5H cells and in PC-6 cells (green line), and novobiocin also restored the intracellular topotecan accumulation (Fig. 3, red line). These results indicate that the functional efflux-pump activity of BCRP almost corresponded to the level of BCRP protein expression in PC-6 cell panels and in PC-6 cells that were treated with 1 μM 5-aza-dC.
Fine Mapping of Methylation Sites in the BCRP Promoter in PC-6 Cell Panels
Comparison between bisulfite-modified and unmodified DNA sequences (Fig. 4) revealed that both alleles at all 89 CpG sites (segments A through E) in the promoter region of BCRP were methylated completely in PC-6 cells (Fig. 5). Conversely, in SN-38-resistant PC-6/SN2-5H cells, 79 CpG sites in the limited region (segments C through E; nt −720 to +261) were unmethylated at both alleles, but 10 CpG sites in segments A and B (nt −1381 to −721) were methylated (Fig. 5). These results suggest that unmethylation in the BCRP promoter is necessary for BCRP expression in SN-38-resistant small cell lung cancer cells.
Establishment of MSP for BCRP in Small Cell Lung Cancer Cells
We have established an MSP method that readily can identify methylated or unmethylated alleles in the examined cells. A PCR product size of 125 base pairs (bp) was detected in SN-38-sensitive PC-6 cells by using only a methylation-specific PCR primer pair. In SN-38-resistant PC-6/SN2-5H cells, a PCR product size of 88 bp was yielded by using only an unmethylation-specific PCR primer pair (Fig. 6). These MSP results were coincident with the results from the bisulfite sequencing analysis and the BCRP expression analysis.
Inverse Correlation Between Methylation of the BCRP Promoter and BCRP Expression in Nonsmall Cell Lung Cancer Cells
The relative expression of BCRP mRNA (BCRP/ACTB ratio) was 1.00, 0.60, 0.42, and 0.22 in PC-6/SN2-5H, NCI-H460, NCI-H441, and NCI-H358 cells, respectively (Fig. 6). In PC-6/SN2-5H cells in which BCRP was highly expressed, both alleles were unmethylated. In moderately BCRP-expressing cells (NCI-H460, NCI-H441, and NCI-H358 cells), 1 allele was methylated, but another allele was unmethylated. In non-BCRP-expressing PC-6 and NCI-H69 cells, both alleles were methylated (Fig. 6). These results indicate that the unmethylation of 1 allele in the promoter region of BCRP induces BCRP mRNA expression and that the MSP method we developed is useful for detecting methylation status of the BCRP promoter. Furthermore, there was an inverse correlation between methylation of the BCRP promoter and BCRP mRNA expression in both small cell and nonsmall cell lung cancer cells.
In the current study, we have demonstrated that demethylation of the BCRP promoter induced BCRP re-expression at the mRNA and protein levels in the SN-38-resistant lung cancer cell line PC-6/SN2-5H. This evidence was confirmed by bisulfite sequencing analysis and the treatment of non-BCRP-expressing PC-6 cells with 5-aza-dC, which acts as a DNA methyltransferase inhibitor only after its incorporation into DNA and leads to irreversible binding of methyltransferases to analog bases and to their depletion.25, 36 Moreover, BCRP was re-expressed by 5-aza-dC in a dose-dependent manner. These findings suggest that promoter methylation is responsible for transcriptional silencing of BCRP in PC-6 cells and that demethylation of at least 1 allele is necessary for BCRP re-expression and for BCRP-mediated resistance to SN-38 in PC-6/SN2-5H cells. Thus, it is likely that demethylation of the BCRP promoter may be one of mechanisms of BCRP expression in both small cell and nonsmall cell lung cancer cells and in multiple myeloma37 and renal carcinoma.38
However, the degree of BCRP expression at the mRNA and protein levels in PC-6 cells that were treated with 5-aza-dC was far below the degree of expression in PC-6/SN2-5H cells, as illustrated in Figures 1 and 2. Although the 5-aza-dC treatment may not induce demethylation of the BCRP promoter generally and completely in all PC-6 cells that have the hypermethylation, our results revealed that demethylation of the CpG promoter motif may be necessary in part but is not a sufficient condition for gene transcription.37
BCRP is transcribed by a TATA boxless promoter with multiple Sp1 sites and a CCAAT box.34, 39 We examined Sp1 mRNA expression by using semiquantitative RT-PCR in PC-6 and PC-6/SN2-5H cells. The results indicated that the expression of Sp1 was increased more in PC-6/SN2-5H cells than in PC-6 cells (data not shown), supporting the belief that Sp1 also contributes to BCRP re-expression.34 Thus, both promoter demethylation of BCRP and overexpression of transcription factors, including Sp1, may play a role in the induction of BCRP re-expression in SN-38-resistant lung cancer cells, although the former may have a minor and partial role, and the latter may contribute mainly and predominantly to BCRP expression. It remains to be elucidated whether promoter demethylation of BCRP and activation of transcription factors contribute to the up-regulation of BCRP expression in other cancer cell lines and tissues.
Gene silencing through hypermethylation of promoters has been studied with the MSP method in various cancers.40–44 In those studies, the high sensitivity and specificity of MSP successfully allowed accurate detection of the inactivation of tumor suppressor genes. We also have developed the MSP method to detect promoter methylation of BCRP. Moreover, because promoter methylation is correlated inversely with BCRP expression in both small cell and nonsmall cell lung cancer cells, it is easy to predict BCRP expression by MSP for BCRP using bisulfite-modified DNA. In addition, occasionally, it is difficult to extract RNA from transbronchial lung biopsy and microdissected specimens (because of the small amount of tissues) and also to maintain RNA quality, which depends on the length of time frozen tissues are analyzed after biopsy. Conversely, DNA facilitates the handling of extraction and the maintenance of quality compared with RNA. Furthermore, because BCRP expression confers resistance to several anticancer drugs, including irinotecan, topotecan, mitoxantrone, methotrexate, doxorubicin, and daunorubicin,10, 13–17 which are substrates of BCRP, assessment of BCRP expression is important for the choice of these drugs before chemotherapy. Therefore, the presence or absence of methylation of the BCRP promoter by MSP using DNA samples, and not RNA samples, before chemotherapy may be useful as a new biomarker in predicting drug resistance and response to anticancer drugs that are effluxed by BCRP to achieve an optimal treatment for each individual cancer patient.
We thank to Professor Norio Niikawa for his support.