Microarray-Based Detection of Multidrug Resistance in Human Tumor Cells by Expression Profiling of ATP-Binding Cassette Transporter Genes

Home , ABCA12, ABCA2, ABCA3, ABCA4, ABCA7, ABCA8

[CANCER RESEARCH 64, 8987–8993, December 15, 2004] Microarray-based Detection of Multidrug Resistance in Human Tumor Cells by Expression Profiling of ATP-binding Cassette Transporter Genes

Jean-Pierre Gillet,1 Thomas Efferth,2 Daniel Steinbach,3 Jacques Hamels,4 Franc¸oise de Longueville,5 Vincent Bertholet,5 and Jose´Remacle1 1Research Unit of Cellular Biology, University of Namur, Namur, Belgium; 2Center for Molecular Biology of the University of Heidelberg, Heidelberg, Germany; 3University Children’s Hospital Jena, Department of Pediatrics, Jena, Germany; 4Genetic and Pathology Institute, Department of Pathology and Molecular Onco-Hematology, Loverval, Belgium; and 5Eppendorf Array Technologies, Namur, Belgium

ABSTRACT transporter genes, e.g., ABCB1 (MDR1), ABCC1 (MRP1), and ABCG2 (MXR/BCRP; refs. 5–7). The parental sensitive cell lines are used as Different mechanisms of drug resistance, including ATP-binding cas- reference and their corresponding drug-resistant subline as test. sette (ABC) transporters, are responsible for treatment failure of tumors. We developed a low-density DNA microarray which contains 38 genes of the ABC transporter gene family. This tool has been validated with three MATERIALS AND METHODS different multidrug-resistant sublines (CEM/ADR5000, HL60/AR, and MCF7/CH1000) known to overexpress either the ABCB1 (MDR1), ABCC1 Cell Lines. Human T-lymphoblastoid leukemic ABCB1 (MDR1) express- (MRP1), or ABCG2 (MXR and BCRP) genes. When compared with their ing CCRF/ADR5000 cells selected with doxorubicin and parental ABCB1 drug-sensitive parental lines, we observed not only the overexpression of (MDR1)-negative CCRF-CEM cells were obtained from Dr. Axel Sauerbrey these genes in the multidrug-resistant cell lines but also of other ABC (Department for Pediatrics, University of Jena, Jena, Germany). These cells transporter genes pointing to their possible role in multidrug resistance. were cultured as described previously (5). These results were corroborated by quantitative real-time reverse tran- Promyelocytic ABCC1 (MRP1)-overexpressing HL60/AR leukemia cells scription-PCR. As the microarray allows the determination of the expres- and parental ABCC1 (MRP1)-negative HL60 drug-sensitive cells were ob- sion profile of many ABC transporters in a single hybridization experi- tained from Dr. Sauerbrey and seeded in RPMI 1640 supplemented with 10% ment, it may be useful as a diagnostic tool to detect drug resistance in fetal calf serum and 100 mmol/L daunorubicin for HL60/AR. Parental HL-60 clinical samples. cells were maintained under the same conditions without daunorubicin exposure (6). INTRODUCTION Human breast carcinoma parental MCF7 cell line and the multidrug-resistant MCF7/CH1000 subline were kindly provided by Dr. Douglas D. Ross Multidrug resistance (MDR) in tumor treatment is characterized by (University of Maryland Greenebaum Cancer Center, Department of Medicine, resistance to a broad spectrum of structurally unrelated cytotoxic University of Maryland School of Medicine, and the Baltimore Veterans drugs and is of prognostic relevance for treatment outcome (1). Drug Affairs Medical Center; Baltimore, MD; ref. 7). resistance may develop after prolonged exposure to a drug and is The panel of 60 human tumor cell lines of the Developmental Therapeutics mediated by different cellular mechanisms, including the expression Programme of the National Cancer Institute (NCI; Bethesda, MD) consisted of of ATP-binding cassette (ABC) transporters (2). The ABCB1 (MDR1), leukemia, melanoma, non–small cell lung cancer, colon cancer, renal cancer, ABCC1 (MRP1), and ABCG2 (MXR and BCRP) genes are the best ovarian cancer cell lines, cell lines of tumors of the central nervous system, known genes associated with MDR, and at least seven other ABC prostate carcinoma, and breast cancer. Their origin and processing have been described previously (8). transporters are also capable of transporting drugs (2). The presence of ϩ Isolation of mRNA. PolyA RNA was isolated with the FastTrack 2.0 other ABC transporters such as ABCC7 and ABCA4 are linked to other mRNA isolation kit (Invitrogen, Merelbeke, Belgium) with the manufacturer’s diseases such as cystic fibrosis or Stargardt’s disease (3). However, protocol for isolating mRNA starting from 4 ϫ 107 cells. RNA integrity was several ABC transporters have still unknown or undefined functions. verified by capillary electrophoresis on the Agilent 2100 Bioanalyzer (Agilent In recent years, MDR has been diagnosed in tumors by various types Technologies, Palo Alto, CA). of assays. There is actually no technique for obtaining a full picture of Synthesis of Labeled cDNA. Labeled cDNA were prepared with 1 ␮gof the different MDR parameters but mostly determination of individual mRNA. Three synthetic polyAϩ-tailed RNA samples were spiked at three genes or proteins of the ABC family. High-density DNA microarray different amounts (10, 1, and 0.1 ng per reaction) into the purified mRNA as analysis have been done to analyze differentially expressed genes in internal standard to assist in quantification and estimation of experimental tumor cells after drug treatment (4). Although the use of high-density variation introduced during labeling and analysis. The detailed procedure was already reported previously (9). DNA microarray is very useful for the screening of new markers, the Design of Low-Density Microarrays for 38 ABC Transporter Genes. quantitation of gene expression is rather poor. Low-density DNA The genes on the low-density microarray (termed DualChip human ABC) are microarrays are more suitable for routine applications because of their presented in Table 1. It contained two arrays per slide with a range of 41 simplicity, good reproducibility, easy data management, and low transporter genes composed of 38 ABC transporters, 1 cationic transporter, and costs. The aim of the present study was to test a novel low-density 2 ATP-sensitive potassium channels. To evaluate the reliability of the exper- microarray for the simultaneous expression analysis of 38 ABC trans- imental data, several positive and negative hybridization and detection controls porter genes in multidrug-resistant tumor cells. To investigate the are included on the microarray. Three internal standard controls and eight usefulness of such a new tool, we investigated three multidrug- housekeeping genes are arrayed on the slides for the normalization. DualChip resistant sublines that are known to express three different ABC human ABC is composed of single-strand DNA probes attached to the glass support by a covalent link. Each DNA probe is present in triplicates (Fig. 1A). The length of the DNA probes has been optimized, and the design of the probes Received 6/4/04; revised 9/16/04; accepted 10/13/04. has been done as described previously. Recent update and the absence of 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 several clones, 11 ABC transporters are absent on the array. The homology 18 U.S.C. Section 1734 solely to indicate this fact. between the different genes of this superfamily is very high. It is 35 to 40% in Note: J-P. Gillet and T. Efferth contributed equally to this work. average and for certain genes Ͼ60 to 70%. For this reason, in five cases, the Requests for reprints: Jean-Pierre Gillet, Research Unit of Cellular Biology, University capture-probe was complementary of two closely related genes (ABCA2/3, of Namur, Rue de Bruxelles 61, 5000 Namur, Belgium. Phone: 32-081-72-57-11; Fax: 32-081-72-41-35; E-mail: [email protected]. ABCB1/4, ABCC6/8/9, and Kir 6.1/6.2). The specificity of the capture-probe ©2004 American Association for Cancer Research. was checked by testing the binding of PCR-amplified ABC transporter clones. 8987

Table 1 Human ABC genes: their representation on the DualChip human ABC, chromosomal loci, expression and function Family Members On ABCChips Location Expression Function ABC A ABC A1 Yes 9q31.1 Ubiquitous Cholesterol efflux onto HDL ABC A2 Yes 9q34.3 Brain Drug resistance ABC A3 Yes 16p13.3 Lung Surfactant secretion ABC A4 Yes 1p21.3 Rod photoreceptors N-Retinylidiene-PE efflux ABC A5 Yes 17q24.3 Muscle, heart, testes ABC A6 Yes 17q24.3 Liver ABC A7 Yes 19p13.3 Spleen, thymus ABC A8 Yes 17q24.3 Ovary ABC A9 No 17q24.3 Heart ABC A10 No 17q24.3 Muscle, heart ABC A12 No 2q34 Stomach ABC A13 No 7p12.3 Low in all tissues ABC B ABC B1 Yes 7q21.12 Adrenal, kidney, brain Multidrug resistance ABC B2 Yes 6p21.3 All cells Peptide transport ABC B3 Yes 6p21.3 All cells Peptide transport ABC B4 Yes 7q21.12 Liver Phosphatidylcholine transport ABC B5 No 7p21.1 Ubiquitous ABC B6 Yes 2q35 Mitochondria Iron transport ABC B7 Yes Xq21–q22 Mitochondria Fe/S cluster transport ABC B8 Yes 7q36.1 Mitochondria ABC B9 Yes 12q24.31 Heart, brain ABC B10 Yes 1q42.13 Mitochondria ABC B11 Yes 2q24.3 Liver Bile salt transport ABC C ABC C1 Yes 16p13.12 Lung, testes Multidrug resistance ABC C2 Yes 10q24.2 Liver Organic anion efflux ABC C3 Yes 17q21.33 Lung, intestine, liver Drug resistance ABC C4 Yes 13q32.1 Prostate Nucleoside transport ABC C5 Yes 3q27.1 Ubiquitous Nucleoside transport ABC C6 Yes 16p13.12 Kidney, liver ABC C7 Yes 7q31.31 Exocrine tissues Chloride ion channel ABC C8 Yes 11p15.1 Pancreas Sulfonylurea receptor ABC C9A ϩ Kir 6.1 Yes 12p12.1 Muscle, heart K(ATP) channel regulation ABC C9B ϩ Kir 6.2 Yes K(ATP) channel regulation ABC C10 Yes 6p21.1 Low in all tissues ABC C11 No 16q12.1 Low in all tissues ABC C12 No 16q12.1 Low in all tissues ABC C13 No 21q11.2 Many tissues ABC D ABC D1 Yes Xq28 Peroxisomes VLCFA transport regulation ABC D2 Yes 12q11 Peroxisomes ABC D3 Yes 1p22.1 Peroxisomes ABC D4 Yes 14q24.3 Peroxisomes ABC E ABC E1 Yes 4q31.31 Ovary, testes, spleen Oligoadenylate binding protein ABC F ABC F1 Yes 6p21.1 Ubiquitous ABC F2 Yes 7q36.1 Ubiquitous ABC F3 Yes 3q27.1 Ubiquitous ABC G ABC G1 Yes 21q22.3 Ubiquitous Cholesterol transport? ABC G2 Yes 4q22 Placenta, intestine Multidrug resistance ABC G4 No 11q23 Liver ABC G5 No 2p21 Liver, intestine Sterol transport ABC G8 No 2p21 Liver, intestine Sterol transport

Genes On ABCChips Location Expression Function

Kir 6.1 Yes 12p11.23 Large variety of tissues K(ATP) channel regulation Kir 6.2 Yes 11p15.1 Large variety of tissues K(ATP) channel regulation IMPT Yes 11p15.5 Colon, breast, neck Chloroquine & quinidine transport NOTE. Details of the nomenclature can be found at http://www.gene.ucl.ac.uk/nomenclature/genefamily/abc.html. All human ABC genes have standard nomenclature developed by the Human Genome Organization. Three other transporters are spotted on the array. It concerns Kir 6.1, Kir 6.2, and IMPT genes.

Hybridization and Analysis. The DualChip human ABC hybridization Real-time reverse transcription-PCR (RT-PCR) was done on 16 genes, and the statistical analysis was carried out according to de Longueville et al. namely ABCA2, ABCA3, ABCA4, ABCA7, ABCB1/4, ABCB2, ABCC1, (9, 10). Briefly, the detection was done by using a Cy3-conjugated anti-biotin ABCC3, ABCC4, ABCC5, ABCC6, ABCC8, ABCC9, ABCG1, ABCG2, and IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) on biotinylated ␣-tubulin (housekeeping gene). mRNA of sensitive and multidrug-resistant cDNA. cell lines were used in the real-time RT-PCR (n ϭ 2), and each reaction was The statistical analysis consisted in the calculation of the statistically sig- done in triplicate. nificant change in the ratio of the drug-resistant compared with their parental The detailed procedure for PCR reaction mixtures is reported elsewhere drug-sensitive cells. (10). Validation of Relative Gene Expression by Real-Time PCR. The cDNA Fluorescence emission was detected for each PCR cycle, and the threshold ␮ were synthesized from 0.5 g of mRNA according the RNA-labeling protocol cycle (CP) values were determined. The CP value was defined as the actual described in ref. 9 with the following minor modifications: (a) a DNase PCR cycle when the fluorescence signal increased above the background treatment of mRNA was done before cDNA synthesis; (b) the deoxynucleoside threshold. Average CP values from duplicate PCR reactions were normalized triphosphate mixture contained dGTP, dATP, dTTP, and dCTP each at 500 to average CP values for housekeeping genes from the same cDNA prepara- ␮mol/L but no biotinylated dCTP; and (c) the second addition of reverse tions. The relative expression ratio of a target gene in resistant cells is transcriptase was omitted. calculated based on E (efficience) and the CP deviation of a resistant sample Gene-specific primers corresponded to the gene sequences present on the versus a sensitive sample and expressed in comparison to a reference gene DualChip human ABC (Eppendorf Array Technologies, Namur, Belgium). (housekeeping gene): Forward and reverse primers for real-time PCR amplification were designed ϭ ␦CP target (resistant-sensitive) ␦CP reference (resistant-sensitive) with the Primer Express Software (PE Applied Biosystems, Foster City, CA). ratio (Etarget) /(Ereference) . 8988

Fig. 1. A, design of DualChip human ABC. The array included 49 genes (including 8 housekeeping genes). Each capture-probe is spotted in triplicates. Three complete subarrays are schematically drawn. Six different internal standards are placed in different area for the normalization. B, fluorescence image of the DualChip human ABC hybridized with cDNA obtained from mRNA isolated from the drug-resistant ABCB1 (MDR1)-positive CCRF/ADR5000 subline, compared with the parental drug-sensitive ABCB1 (MDR1)-negative CCRF-CEM cell line. The red arrow shows ABCB1 saturated in the drug-resistant subline.

The detailed procedure is reported elsewhere (11). Values were reported as intensity ratios for genes, which were statistically significant for each average of triplicate analysis. subline, are presented in Fig. 2. COMPARE Analysis. The sulforhodamine B assay for the determination The CCRF/ADR5000 drug-resistant subline is known to be ABCB1 of drug sensitivity in these cell lines has been reported previously (12). The (MDR1) positive (5). Indeed, we found that the ratio of ABCB1 gene inhibition concentration 50% (IC ) values for drugs are included in the 50 expression in this resistant subline was Ͼ100-fold higher compared Standard Agents Database of the Developmental Therapeutics Programme of with the drug-sensitive parental line by the microarray technique. The the NCI.6 The mRNA expression values of 60 cell lines of 31 ABC transporter genes (represented by 68 different clones with individual GenBank accession expression of the genes in the parental line was very low, making the numbers) were selected from the NCI’s database. The other 18 members of the ratio calculation to be qualitative. The multidrug-resistant HL60/AR ABC transporter gene family were not represented in the database. The mRNA subline overexpresses the ABCC1 (MRP1) gene (6). This gene was expression has been determined by microarray analysis as reported previously indeed found to be overexpressed by a factor of 10.46 compared with (13, 14). COMPARE analysis were done to produce rank-ordered lists of genes the parental line with the DualChip human ABC. expressed in the 60 NCI cell lines. The methodology has been previously The multidrug-resistant subline MCF7/CH1000 has been reported described in detail (15). Briefly, every standard drug of the NCI’s database is to overexpress the ABCG2 (MXR/BCRP) gene (7), which was indeed ranked for similarity of its IC values to the mRNA expression for a given 50 observed with an overexpression of 10 times. ABC transporter. To derive COMPARE rankings, a scale index of correlations Besides the expected overexpression of the specific ABC trans- coefficients (R values) is created. In the standard COMPARE approach, porter genes, the three cell lines also showed several other overex- greater mRNA expression in cell lines correlate with greater IC50 values, e.g., drug resistance. pressed genes. In the CCRF/ADR5000 subline, six other ABC genes were overexpressed in addition to ABCB1. The expression pattern for the promyelocytic resistant cell line HL60/AR showed seven overex- RESULTS pressed ABC transporter genes. Several ABC transporter genes in- Characterization of ABC Gene Expression Pattern of Three volved in drug resistance were found to be overexpressed: ABCA2, Different Parental Drug-sensitive Cell Lines and Their Drug- which is known to be involved in drug resistance (16), ABCC1, and resistant Sublines. Fig. 1B shows the data of one representative ABCC4. In the MCF7/CH1000 drug-resistant cell line, three important experiment from the drug-resistant CCRF/ADR5000 subline com- ABC transporter genes included in drug resistance are overexpressed: pared with the parental sensitive-drug cell line CCRF-CEM. The ABCC3, ABCC5, and ABCG2, although ABCA7 was down-regulated. reliability and the reproducibility between assays were assessed by We have also observed an overexpression of ABCB6, ABCF3, and repeating the experiments three times for each cell line (n ϭ 3). The ABCG1. variability observed in triplicate experiments from the drug-resistant Validation of Relative Gene Expression by Real-Time RT-PCR. CCRF-ADR5000 subline compared with the parental sensitive drug The expression of ABC transporter genes with expression ratios Ͼ 2 CCRF-CEM cell line was 11.8% (calculated as the mean of variation and one gene with an expression ratio near one selected at random in coefficient for every quantitative significant gene) ranging from 6.3 to the multidrug-resistant cell lines were corroborated by quantitative 25.9%. In two others experiments, the drug-resistant HL60/AR sub- real-time RT-PCR as independent test method. Three genes were line compared with the parental sensitive drug HL60 cell line and the quantified on the parental drug-sensitive CCRF-CEM cell line and on multidrug-resistant MCF7/CH1000 subline compared with the paren- its drug-resistant CCRF/ADR5000 subline: ABCA7, ABCB1/4, and tal sensitive drug MCF7 line were, respectively, 16.1% ranging from ABCF2 (Fig. 3A). On the microarray, the overexpression of the 9.2 to 31.9 and 16.4% ranging from 6.2 to 26.7%. The fluorescence ABCB1/4 was found to be qualitative and estimated Ͼ100. The real-time RT-PCR gave a ratio of 799. The overexpression was found 6 Internet address: http://dtp.nci.nih.gov. for ABCA7 to be 2.15 on the microarray and 2.7 by real-time RT-PCR. 8989

Fig. 2. Ratios of differentially expressed genes in three multidrug-resistant sublines compared with their parental drug-sensitive cell lines. The results are given as the mean of the ratios and the SD of three different experiments. Spots framed in white show overexpressed genes, and the spot framed in gray is the down-regulated one.

For the ABCF2, the overexpression was 1.28 on the microarray and coefficients of r Ͼ 0.4 are listed in Table 2. This approach was applied 1.64 in real-time RT-PCR. to explore which of the members of the ABC transporter gene family The data obtained for the parental promyelocytic HL60 cell line and could be involved in drug transport processes and resistance. The its drug-resistant subline HL60/AR are summarized in Fig. 3B. The results indicate that 17 of 31 ABC transporters investigated might act overexpression of the 10 genes tested in multidrug-resistant HL60/AR as drug transporters (ABCA1, ABCA2, ABCA3, ABCA5, ABCA12, cells was corroborated by real-time RT-PCR. ABCA4, ABCA7, ABCB1, ABCB4, ABCB6, ABCB7, ABCB11, ABCC1, ABCC3, ABCB2, ABCC1, and ABCC4 were also found overexpressed in real- ABCC4, ABCC6, ABCF2, ABCF3, and ABCG2) because com- time RT-PCR with ratios slightly higher than on the microarray. We pounds of the Standard Agents Database whose IC50 values correlated have dissected ABCC6/8/9 and ABCA2/3 capture-probes. ABCC8 with the mRNA expression of ABC transporters could be identified. and ABCC9 were undetected by real-time RT-PCR in the drug- Vice versa, no candidate substrates were assigned to 14 other ABC resistant cells, whereas the ABCC6 was highly overexpressed. The transporters (ABCA4, ABCA6, ABCA8, ABCB8, ABCB10, ABCC2, chips gives only a qualitative result with an overexpression Ͼ 3 for ABCC5, ABCC7, ABCC8, ABCD2, ABCD3, ABCE1, ABCF1, and the sum of the three genes. We obtained a ratio for ABCA2 of 12.56, ABCG1; Table 2). although ABCA3 was down-regulated with a ratio of 0.022 by real- time RT-PCR. If we sum up the two results, we obtain a ratio for DISCUSSION ABCA2ϩA3 in real-time RT-PCR of 0.59. The ratio of the capture- probe ABCA2/3 was 1.8 on the array. Many studies have been done in tumors based on various methods Fig. 3C summarizes the data obtained for the parental MCF7 cell to detect the expression of single members of the ABC transporter line and its multidrug-resistant MCF7/CH1000 subline. Four overex- gene family, e.g., the MDR-conferring ABCB1 (MDR1), ABCC1 pressed genes were confirmed with real-time RT-PCR with very close (MRP1), and ABCG2 (BCRP) genes (17). The family of ABC trans- ratios obtained in the two methods and even an identical ratio of 2 for porters is still increasing, and the role of many members of the MDR the ABCC5 gene. is still unknown or not fully understood. To explore the MDR phe-

COMPARE Analysis. Finally, we correlated the IC50 values for notype in its entirety, it may be helpful to analyze the expression of compounds included into the NCI’s Standard Agent Database with the the ABC transporters simultaneously. DNA microarray technology is baseline mRNA expression level of 31 human ABC transporters one of the technologies that allows gene expression profiling and that (represented by 68 different clones with individual GenBank acces- has been applied in tumor cells associated with the treatment respon- sion numbers) of the 60 NCI cell lines with the COMPARE algorithm. siveness (18).

Drugs whose IC50 values correlated with microarray-based mRNA In the present investigation, we described the use of a low-density expression of ABC transporter genes with COMPARE correlation DNA microarray for the analysis of 38 ABC transporter genes and 8990

Table 2 Possible chemotherapy substrates of ABC transporters as determined by COMPARE analysis of microarray-based mRNA expression of 31 ABC transporters (represented by 68 clones with individual Genbank accession numbers) and the IC50 values for compounds of the Standard Agent database of the NCI (http://dtp.nci.nih.gov) Symbol Genbank Drug ABCA1 AJ012376 5-Hydroxypicolinaldehyde thiosemicarbazone AI344681 Geldanamycin, Tiazofurin, 6-Thioguanine, L-Cysteine ethyl ester methylcarbamate ABCA2 AB028985 Ifosfamide ABCA3 H51436 Tetrocarcin A, vinblastine, maytansin U78735 None ABCA4 H87722 None AF000148 None ABCA5 H26264 L-Buthionine sulfoximine U66672 NSC 366140 ABCA6 AI651024 None ABCA8 AB020629 None ABCA12 AL080207 Vinblastine, maytansin ABCB1 M14758 Dactinomycin, bruceantin, didemnin B, 4Ј- deoxydoxorubicin, mithramycin, NSC355644, echinomycin A, bisantrene, bactobolin, phyllanthoside, acodazole, doxorubicin, daunorubicin benzoylhydrazone, paclitaxel, vinblastine, tetrocarcin A, geldanamycin ABCB4 M23234 Echinomycin A ABCB6 C20962 None AF070598 6-Thioguanine, caracemide, NSC 291643, NSC 284751, 6-mercaptopurine, inosine dialdehyde, rifamycin ABCB7 AA056272 O6-Methylguanine AB005289 None ABCB8 U66688 None ABCB10 N58275 None U18237 None ABCB11 AF091582 Echinomycin A ABCC1 W46896 Echinomycin A, L-cysteine ethyl ester methylcarbamate, caracemide L05628 Caracemide, inosine dialdehyde, L-Cysteine ethyl ester methylcarbamate, tamoxifen X78338 Carademide, spirogermanium, L-cysteine ethyl ester methylcarbamate, inosine dialdehyde ABCC2 R91503 None R96618 None W89005 None U49248 NSC 354646 ABCC3 U83659 NSC 291643, inosine dialdehyde, L-cysteine ethyl ester methylcarbamate AF085692 NSC 291643, L-cysteine ethyl ester methylcarbamate, inosine dialdehyde ABCC4 U83660 Melphalan, chlorambucil, pancratistatin Fig. 3. Expression of ABC transporter genes in multidrug-resistant cell lines versus AF071202 Pancratistatin, NSC 354646, guanylhydrazone drug-sensitive parental cell lines. The values, presented in logarithmic scale, are the ABCC5 H93519 None fluorescence intensity ratios obtained on the DualChip human ABC and the ratios U83661 None measured by real-time RT-PCR for genes with ratio Ͼ 2 and a ratio near one selected at AF104942 None random in three cell line experiments (A) CEM/ADR5000 versus CCRF-CEM, (B) ABCC6 U66689 Vinblastine, maytansin, paclitaxel, tetrocarcin A, HL60/AR versus HL60, and (C) MCF7/CH1000 versus MCF7. qRT-PCR, quantitative dactinomycin, bactobolin, bruceantin, zinostatin, RT-PCR. mithramycin, geldanamyin, echinomycin A, NSC 354646 X95715 None R99091 Chlorozotocin, vinblastine, tetrocarcin A, maytansin, NSC three other transporters. IMPT1 (alias SLC22A1L solute carrier fam- 354646, didemnin B, paclitaxel, dactinomycin ily 22) was added as an organic cation transporter for chloroquine and ABCC7 M28668 None quinidine-related compounds. Kir 6.1 and 6.2 are also present because ABCC8 L78207 None AI288757 None they form channels in association with the sulfonylurea receptor SUR ABCD2 AJ000327 None (such ABCC8 and ABCC9). DualChip human ABC has been elabo- ABCD3 AA013086 None ABCE1 W31619 None rated as a first attempt to determine the cell resistance pattern. Other X76388 None transporter genes could be added later on because the gene family will ABCF1 W90495 None grow. AF027302 None ABCF2 W67806 Caracemide, spiromustine, maytansin, chlorambucil, 4- The DualChip human ABC was developed to give quantitative ipomeanol, hydroxyurea, NSC 293015, NSC, 267213, results. The variability from one experiment to the other showed that NSC 602668, carmustine, asalex, amsacrine, melphalan, the results were rather well reproducible with mean coefficient of teniposide, anguidine, uracil nitrogen mustard, doxorubicin, menogaril, etoposide, daunorubicin variation between 11.8 and 16.4% from one experiment to the other. benzoylhydrazone, deoxyspergualin, NSC 102627, We observed that the high variability generally occurs for low- aminoglutethimide, asparaginase, hepsulfam AA022488 Fludarabine expressed genes. This explains partly the spread of variability from AJ005016 NSC 357704 one gene to the other. ABCF3 AA045255 None The results presented in this article clearly showed an increased U66685 Caracemide, geldanamycin ABCG1 H68928 None expression of several times of the expected ABC transporter genes for X91249 None the three resistant cell lines analyzed. In addition, we also observed ABCG2 N59214 None significant overexpression of several other genes (Fig. 2). The values AF093771 Pancratistatin 8991

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2004 American Association for Cancer Research. DETECTION OF ABC TRANSPORTERS BY MICROARRAYS for these overexpressions were Ͼ60%, except for the ABCF2 in the expression in non–small cell lung cancer (26). The MRP3 and MRP5 CEM cells. Some of these genes are known to confer drug resistance expression levels in normal lung tissues and in tumors from patients such as ABCA2, ABCB1, ABCC1, ABCC3, ABCC4, ABCC5, and exposed to platinum drugs were significantly higher than those in ABCG2. Such overexpression could be seen as typical of the adapta- tissues from nonexposed patients (27, 28). In lung cancer cell lines, tion of the cells after prolonged incubation in the presence of drugs. Young et al. (29) could demonstrate a good correlation of mRNA and Three of these genes have been extensively studied. ABCB1 protein levels of ABCC2 and ABCC3 in response to cytostatic drugs. (MDR1) was the first human ABC transporter cloned and character- Tada et al. (30) showed that the expression of MDR1, MRP1, MRP2, ized through its ability to confer a MDR phenotype to cancer cells and MRP3 in recurrent and residual bladder tumors after chemother- (19). Analysis of multidrug-resistant cells not expressing ABCB1 led apeutic treatment was higher than that in untreated primary tumors. to the discovery of the ABCC1 (MRP1) protein, which plays a role in The role of other ABC transporters for drug resistance remains to protecting cells from chemical toxicity and oxidative stress and in be shown, i.e., ABCA7, whose substrates and function are still un- mediation of inflammatory responses involving cysteinyl leukotrienes known. We have detected ABCA2 in our panel of cell lines (16). (20). Finally, ABCG2 (MXR/BCRP) was identified as a drug trans- Interestingly, antisense treatment against ABCA2 of drug-resistant porter in multidrug-resistant cells, which do not express ABCB1 and ovarian carcinoma cells increased their drug sensitivity, suggesting ABCC1 (21). that ABCA2 may also function as a drug efflux pump. The present As a strategy to explore which ABC transporters might function as investigation also points to ABC transporter genes that were found drug transporters, we performed COMPARE analysis with com- overexpressed and have not thus far been recognized as associated pounds included in the NCI’s Standard Agent database6 and 31 ABC with MDR (ABCA4, ABCA7, ABCB2, ABCB3, ABCB6, ABCB8, transporters whose mRNA expression in 60 NCI cell lines has been ABCB9, ABCC6, ABCF2, ABCF3, and ABCG1). Developing a low- determined by microarrays (13, 14). The COMPARE computation density microarray enabled us to explore the potential implication of provided a list of drugs that could be considered as substrates for ABC many ABC transporter genes currently unknown to be associated with transporters. Although such correlation analysis does not provide clear MDR. Future studies have to address the substrates and function for evidence for a compound as being a true ABC transporter substrate, drug resistance of these novel ABC transporter genes. this strategy can be used to generate testable hypothesis. In many In conclusion, we have shown that gene expression profiles of ABC cases, the COMPARE analysis point to previously validated sub- transporter genes in multidrug-resistant cell lines can be obtained by strates, i.e., doxorubicin, vinblastine, paclitaxel, dactinomycin, and so quantitative analysis with low-density microarrays. The overexpres- forth for the ABCB1 (MDR1) gene. We only considered drugs whose sion of all of the genes observed with the arrays were confirmed with COMPARE correlation coefficient was r Ͼ 0.4, and the list of real-time PCR analysis. The determination of the expression profiles candidate substrates is considerably longer if correlations of r Ͻ 0.3 of ABC transporter genes in multidrug-resistant cells lines may open were also taken into account. Surprisingly, well-known substrates of new avenues for the diagnosis of MDR in the clinic and for monitor- some ABC transporters did not show up by COMPARE analysis with ing expression profiles in clinical biopsies and their correlation to r Ͼ 0.4, i.e., doxorubicin, etoposide, vincristine, or methotrexcate as clinical treatment. substrates for ABCC1 (MRP1) or doxorubicin, mitoxantrone, or to- potecan as substrates for ABCG2 (BCRP). This raises the possibility ACKNOWLEDGMENTS that other still unknown compounds might be better substrates for these ABC transporters. This is in accord with recently published We thank the Institut of Pathology and Genetic of Loverval (Loverval, results on the ABC transporter expression detected by real-time RT- Belgium) for its support and the members of Eppendorf Array Technologies PCR in the same NCI cell line panel as used here (22). Our aim was, Company of Namur (Namur, Belgium) for their collaboration. We also thank however, not to provide a complete list of possible substrates for ABC Drs. Axel Sauerbrey and Douglas D. Ross (University of Maryland Greene- transporters but to obtain information that ABC transporters could be baum Cancer Center, Department of Medicine, University of Maryland School considered as candidate drug transporters. In this respect, the COM- of Medicine, and the Baltimore Veterans Affairs Medical Center; Baltimore, PARE analysis point to 17 of 31 ABC transporters analyzed. The MD), respectively, for the generous provision of CCRF-CEM, HL60, and MCF7 drug-sensitive and drug-resistant cell lines. results of the COMPARE analysis reinforce the use of the DualChip human ABC as a tool to detect ABC transporter-associated drug resistance. REFERENCES As recently reviewed, 11 ABC transporters (ABCA2, ABCB1, 1. Ross DD. Novel mechanisms of drug resistance in leukemia. Leukemia (Baltimore) ABCB4, ABCB11, ABCC1–6, and ABCG2) have been associated 2000;14:467–73. with drug resistance and drug transport as of yet (2). Importantly, the 2. Gottesman MM, Fojo T, Bates SE. Multidrug resistance in cancer: role of ATP- dependent transporters. Nat Rev Cancer 2002;2:48–58. role of ABC transporters, apart from the well-known ABCB1 3. Efferth T. The human ATP-binding cassette transporter genes: from the bench to the (MDR1), ABCC1 (MRP1), and ABCG2 (BCRP) for clinical drug bedside. Curr Mol Med 2001;1:45–65. resistance, treatment outcome, and patient survival is increasingly 4. Watts GS, Futscher BW, Isett R, Gleason-Guzman M, Kunkel MW, Salmon SE. cDNA microarray analysis of multidrug resistance: doxorubicin selection produces recognized. A recent study of Steinbach et al. (23) provides first multiple defects in apoptosis signaling pathways. J Pharmacol Exp Ther 2001;299: evidence on the association between the clinical response to chemo- 434–41. 5. Kimmig A, Gekeler V, Neumann M, et al. Susceptibility of multidrug-resistant human therapy and the expression of ABCC2, ABCC3, ABCC4, and ABCC5 leukemia cell lines to human interleukin 2-activated killer cells. Cancer Res 1990; in childhood acute myeloid leukemia. This study strongly suggest that 50:6793–9. ABCC3 is involved in drug resistance in this disease and represents an 6. Brugger D, Herbart H, Gekeler V, et al. Functional analysis of P-glycoprotein and multidrug resistance associated protein related multidrug resistance in AML-blasts. interesting marker for risk-adapted therapy and a possible target for Leuk Res 1999;23:467–75. the development of specific drugs to overcome MDR. Moreover, 7. Doyle LA, Yang W, Abruzzo LV, et al. A multidrug resistance transporter from patients who expressed high levels of ABCC2 and ABCC3 had a human MCF-7 breast cancer cells. Proc Natl Acad Sci USA 1998;95:15665–70. 8. Alley MC, Monks A, Hursey ML, et al. Feasibility of drug screening with panels of particularly poor prognosis (23). The expression of ABCC3 is also human tumor cell lines using a microculture tetrazolium assay. Cancer Res 1988;48: associated with a poor outcome in childhood acute lymphoblastic 589–601. 9. de Longueville F, Meneses-Lorente G, Bertholet V, et al. Gene expression profiling leukemia (24). MRP3 expression correlates with unfavorable clinical of drug metabolism and toxicology markers using a low-density DNA microarray. outcome in ovarian carcinoma (25). The same applies for MRP2 Biochem Pharmacol 2002;64:137–49. 8992

10. de Longueville F, Marcq L, Dufrane S, et al. Use of a low-density microarray for 21. Allikmets R, Hutchinson A, Romano-Spica V, Dean M. A human placenta-specific studying gene expression patterns induced by hepatotoxicants on primary cultures of ATP-binding cassette gene (ABCP) on chromosome 4q22 that is involved in multi- rat hepatocytes. Toxicol Sci 2003;75:378–92. drug resistance. Cancer Res 1998;58:5337–9. 11. Pfaffl MW. A new mathematical model for relative quantification in real-time 22. Szakacs G, Lababidi S, Shankavaram U, et al. Predicting drug sensitivity and RT-PCR. Nucleic Acids Res 2001;29:e45. resistance: profiling ABC transporter genes in cancer cells. Cancer Cell 2004;6: 12. Rubinstein LV, Paull KD, Simon RM, et al. Comparison of in vitro anticancer-drug- 129–37. screening data generated with a tetrazolium assay versus a protein assay against a 23. Steinbach D, Voigt A, Hermann J, Zintl F, Sauerbrey A. Response to chemotherapy diverse panel of human tumor cell lines. J Natl Cancer Inst (Bethesda) 1990;82: and expression of the genes encoding the multidrug resistance-associated proteins 1113–8. MRP2, MRP3, MRP4, MRP5, and SMRP in childhood acute myeloid leukemia. Clin 13. Scherf U, Waltham M, Smith LH, et al. A gene expression database for the molecular Cancer Res 2003;9:1083–6. pharmacology of cancer. Nat Genet 2000;24:236–244. 24. Steinbach D, Cario G, Viehmann S, et al. The multidrug resistance-associated protein 14. Staunton JE, Coller HA, Tamayo P, et al. Chemosensitivity prediction by transcrip- 3 (MRP3) is associated with a poor outcome in childhood ALL and may account for tional profiling. Proc Natl Acad Sci USA 2001;98:10787–92. the worse prognosis in male patients and T-cell immunophenotype. Blood 2003;102: 15. Wosikowski K, Johnson K, Paull KD, Myers TG, Weinstein JN, Bates SE. Identifi- 4493–8. cation of epidermal growth factor receptor and c-erbB2 pathway inhibitors by 25. Ohishi Y, Uchiumi T, Kobayashi H, et al. ATP-binding cassette superfamily trans- correlation with gene expression patterns. J Natl Cancer Inst (Bethesda) 1997;89: porter gene expression in human primary ovarian carcinoma. Clin Cancer Res 1505–15. 2002;8:3767–75. 16. Laing NM, Kruh GD, Bell DW, et al. Amplification of the ATP-binding cassette 2 26. Yoh K, Yokose T, Minegishi Y, et al. Breast cancer resistance protein impacts clinical transporter gene is functionally linked with enhanced efflux of estramustine in ovarian outcome in platinum-based chemotherapy for advanced non-small cell lung cancer. carcinoma cells. Cancer Res 1998;58:1332–7. Clin Cancer Res 2004;10:1691–7. 17. Olesen LH, Norgaard JM, Pallisgaard N, Bukh A, Hokland P. Validation and clinical 27. Oguri T, Suzuki T, Nishio K, Fujiwara Y, Katoh O, Yamakido M. Increased implication of a quantitative real-time PCR determination of MDR1 gene expression: expression of the MRP5 gene is associated with exposure to platinum drugs in lung comparison with semi-quantitative PCR in 101 patients with acute myeloid leukemia. cancer. Int J Cancer 2000;86:95–100. Eur J Haematol 2003;70:296–303. 28. Oguri T, Fujitaka K, Ishikawa N, Kohno N. Association between expression of the 18. Kudoh K, Ramanna M, Ravatn R, et al. Monitoring the expression profiles of MRP3 gene and exposure to platinum drugs in lung cancer. Int J Cancer 2001;93: doxorubicin-induced and doxorubicin-resistant cancer cells by cDNA microarray. 584–9. Cancer Res 2000;60:4161–6. 29. Young LC, Cole SP, Deeley RG, Gerlach JH. Multidrug resistance proteins MRP3, 19. Riordan JR, Deuchars K, Kartner N, Alon N, Trent J, Ling V. Amplification of MRP1, and MRP2 in lung cancer: correlation of protein levels with drug response and P-glycoprotein genes in multidrug-resistant mammalian cell lines. Nature (Lond.) messenger RNA levels. Clin Cancer Res 2001;7:1798–804. 1985;316:817–9. 30. Tada Y, Migita T, Nagayama J, et al. Increased expression of multidrug resistance- 20. Cole SP, Deeley RG. Multidrug resistance mediated by the ATP-binding cassette associated proteins in bladder cancer during clinical course and drug resistance to transporter protein MRP. Bioessays 1998;20:931–40. doxorubicin. Int J Cancer 2002;98:630–5.

8993

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2004 American Association for Cancer Research. Microarray-based Detection of Multidrug Resistance in Human Tumor Cells by Expression Profiling of ATP-binding Cassette Transporter Genes

Jean-Pierre Gillet, Thomas Efferth, Daniel Steinbach, et al.

Cancer Res 2004;64:8987-8993.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/64/24/8987

Cited articles This article cites 29 articles, 13 of which you can access for free at: http://cancerres.aacrjournals.org/content/64/24/8987.full#ref-list-1

Citing articles This article has been cited by 18 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/64/24/8987.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/64/24/8987. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.